Pages and panels showed inconsistent loading states: the Backup center
and several admin pages (System, SSH, admin Overview), the Overview
Migrate/Peers panels, the per-app updater section and the backup engine
details modal rendered a bare 'Loading…' text line (updater-empty /
backup-empty-state) with no spinner, while Services/Config/Tasks used a
boxed card + spinner.
Add one shared loader — window.lpLoadingBox(message) + .lp-loading CSS in
the core/loading subsystem (the boxed card + accent spinner the good tabs
already use) — and route those bare-text loaders through it. The system
metric graph keeps its absolute overlay but gains the same spinner.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
'Checking for improvements… this usually takes a minute.' implied an active
scan finishing imminently. The scan is periodic and runs in the background
(CFG_UPDATER_SCAN_INTERVAL, default 30m), so reword to reflect that. Kept the
cadence generic since the interval is admin-configurable and not exposed here.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
'No installed apps to track.' was terse and gave no next step. Point the user
to the App Center so the empty Updates tab is actionable.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The custom-select enhancer only matched select.form-control /
select.theme-selector, so dropdowns rendered by the per-app config form
(config-form.js uses form-select / form-input / config-input), the app
tools form, the port-manager grid (port-* classes) and the instance
domain picker stayed as plain native OS dropdowns. The LibrePortal app's
Theme option is one of these.
Add those classes to ENHANCE_CLASSES, give the classless instance domain
select a form-control class, and add a compact button override so the
themed dropdown matches the dense port-manager input metrics.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
'No hotfix data yet — the automatic scan fetches the signed improvements index
within a couple of minutes.' leaked jargon (hotfix data, signed index) onto the
interface. Replace with a short, plain 'Checking for improvements…' line.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Empty-state 'no data yet' messages rendered at rgba(text, 0.55) (and one at
text-muted), which is hard to read on the dark recessed panels. Switch them to
the theme's --text-secondary muted token so they stay de-emphasized but legible.
Covers .updater-empty, .updater-detail-empty, .sys-detail-empty, .eo-modal-empty.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The last task tile's 10px bottom margin stacked on the list's 16px bottom
padding, leaving a larger gap below the last row than above the first.
Zero the last tile's bottom margin so the dark panel reads symmetric.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Clicking Reclaim space fired two notifications: routeAction → executeTask
already shows the rich 'Reclaim Space task started!' toast (icon + bold
LibrePortal + task link), then _reclaim added a second, plain, iconless
'Reclaiming space…' info toast on top of it. The image-removal path doesn't
double-notify like this. Drop the redundant one — the start toast plus the
completion toast give clean feedback on their own.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
.sys-detail-loading and .sys-detail-empty are absolutely-positioned overlays
that JS shows/hides via the `hidden` attribute. Their .sys-detail-* rule sets
`display: flex`, an author declaration that overrides the UA
`[hidden] { display: none }` — so `el.hidden = true` never actually hid them.
Both the 'Loading history…' and 'No samples in this range yet' overlays stayed
painted on top of a fully-populated chart (overlapping into garbled text).
Add `.sys-detail-loading[hidden], .sys-detail-empty[hidden] { display: none }`
(higher specificity than the bare class) so the hidden attribute wins.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The donut legend collided into the stat cards and the image-row details
were squished on tablet widths — the 220px sidebar leaves the content
cramped while the old breakpoints assumed full-viewport width.
- Headline collapses to a single column at <=1024px (was 800px), the two
stat cards reflow side-by-side, and on phones the donut stacks above its
legend with one stat per row. Legend labels now ellipsis instead of
overflowing into the stats.
- Image rows group the name+pill and the size/shared/age metadata so the
metadata drops onto its own line under the name at <=1024px instead of
competing for width; on phones the Delete button collapses to an icon.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The Images list on /admin/system/storage now has an All / In use / Unused
segmented filter (with live per-group counts), and the default All view
sorts in-use images to the top — the ones you can't reclaim lead, the
reclaimable ones follow. Select all / Clear All act on the visible rows,
so they honour the active filter.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
App-less system tasks (verify, regen, …) resolved to an empty displayName
and null icon in _taskNotificationDescriptor, so their completion toast
rendered an empty <strong></strong><br> — a blank bold line that showed as
a random gap above the message — and had no icon, unlike every other
notification. Treat any task with no app slug as a system task so it gets
the 'LibrePortal' subject and libreportal.svg icon.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The .sys-tasklist panel spaces rows with a flex gap, but each .task-item
also carries margin-bottom: 10px from the shared tasks styles. That margin
only stacks on the last row, so the list had ~8px above the first row and
~18px below the last row, looking lopsided. Zero the row margin inside the
list so spacing is symmetric.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Describe only the current useradd behaviour; drop the narration of the old
silent-failure bug (per the repo's no-tombstone-comments convention).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Drop the GitHub-release version comparison entirely. We install slirp4netns
from apt regardless, so comparing against the GitHub-latest tag only produced
a perpetual 'outdated' loop and a no-op re-install. apt-get install -y is
already idempotent, so run it unconditionally and report the resulting
version.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
slirp4netns --version prints multiple lines (version, commit, libslirp,
SLIRP_CONFIG_VERSION_MAX). The old 'awk {print $2}' ran on every line and
also picked the literal word 'version' from line 1, producing a multi-line
blob that leaked into the 'is outdated' notice. Read only the first line and
take field 3 (the actual number), strip the leading v from the GitHub tag so
the comparison is meaningful, and skip the check if the tag fetch fails.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The Services tab's loading placeholder (.services-loading) was a bare
centered row with no container chrome, unlike the Config and Tasks tab
loading cards. Give it the same boxed look (semi-opaque black fill,
hairline border, rounded corners, margin and min-height) so it reads as
a deliberate loading panel.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
showAppDetail() derived the target tab from the legacy ?tab= query
(searchParams.get('tab')), but the app is path-based now
(/app/<name>/<tab>), so that read was always null and defaulted to
'config'. Since loadTabContent() calls showAppDetail() on every switch,
clicking any non-config tab (services/backups/updater/tasks) immediately
rewrote the URL back to /app/<name> and rendered config.
Read the current main tab off the path via appPartsFromPath, honouring
it only when already on this app; cross-app/cold nav still starts at
config. The legacy ?tab= shape is already normalised to the path by the
SPA's handleAppDetail before this runs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
autoExpandTask (the monitorTask path) opened its row directly without
collapsing the others and never set highlightedTaskId — unlike every
other opener (toggleTaskDetails, selectTask), which enforce a single
open row. So a burst of monitored task creations, e.g. a multi-app
first install, stacked every panel open at once.
Wait for the row to render, then delegate to selectTask, which collapses
any other open panel, sets highlightedTaskId, attaches the right log
view (live stream vs snapshot) and scrolls into view. Setting
highlightedTaskId also makes monitorTask's own guard trip after the
first task, so the running-task auto-follow takes over from there.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Initial rootless setup ran 'systemctl --user start docker' immediately
after install, but the rootless net/port-driver override.conf (and the
daemon-reload that loads it) aren't written until further down. So the
first start always failed — 'Job for docker.service failed' plus a
spurious '✗ Error Setting up Rootless' in the error report — even though
the later 'systemctl --user restart docker' brought the daemon up fine
once the override was in place.
Drop the premature start from the install step (keep install + enable);
the restart after the override is written is now the first real start.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Folder creation output interleaved the system/containers/backups roots
and their children. Regroup the array so each root is immediately
followed by its own children (alphabetical), keeping parents before
children since the mkdir has no -p.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The rootless WebUI container reads its bind-mount sources (configs/webui/*)
through the container-owner GROUP since a2376e2 switched those files from
world-readable to 0640 group=container-owner. But the WebUI credential
randomizer rewrites webui_logins via `sed -i` as the non-root manager, which
recreates the file with the manager's own group — dropping the container-owner
group. The installer then started the container immediately, so node hit
EACCES on /app/webui_logins at require-time (parseConfigFile) and exited 1;
nothing listened on the WebUI port. `libreportal webui login reset` had the
same latent bug (rewrite → restart). Under the old world-readable model a
post-sed file stayed o+r so the container could still read it, which is why
this only surfaced on fresh rootless installs after a2376e2.
Fix: make reconcileWebuiDirOwnership the single "ready the WebUI for its
container" pass — it now also restores the configs/webui bind access (new
`webui-bind` ownership action) on top of the container-dir chown. Reorder the
installer so the credential randomizer runs BEFORE the before-start permission
pass, making that pass the last ownership touch before the container starts;
and call reconcileWebuiDirOwnership before the restart in login reset.
Live box recovered via `libreportal-ownership reconcile`; WebUI 200.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
_webui_bind_access granted o+r to every file in configs/webui so the
rootless container could read its bind-mount sources — but that also made
secrets like webui_logins world-readable to any local user. Under rootless
the container's gid 0 maps to the container owner's gid, so group access is
sufficient: chown the webui dir + files to MANAGER:container-owner, dir
0751 (traverse, not list), files 0640. Container reads via group; other
local users get nothing; the manager (owner) still rewrites them.
Verified live: container READ ok, world READ denied, manager rw, WebUI
login still 200. Live helper updated in lockstep with this source.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Design note for a per-app Files tab scoped to LibrePortal-managed files
(not system files): four file buckets (hidden/view-only/editable/lever),
the advanced/dev mode as the single escalation lever (not per-file flags),
and the hard rule that the flag is UX-only while the locked-down task CLI
stays the security boundary (jail + secret allowlist).
Includes the live UID-access spike: the manager owns and can write the
config tree (/libreportal-system/configs) directly, but the container tree
(/libreportal-containers/<app>) is dockerinstall-owned — readable, not
writable — so config edits need no helper while compose-class edits do.
webui_logins is manager-readable, so secret-hiding must live in the CLI
allowlist, not in perms.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
/app/<name>/config/<sub> URLs (e.g. .../config/ports) are generated by the app
itself and shown in the address bar, but a refresh or deep-link always reset to
the first config category. Cause: showAppDetail() rebuilt the URL via
appPath(appName, targetTab) with NO sub argument and pushState'd it BEFORE
renderAppDetail() read the sub back off the path — so the /<sub> segment was
already gone and preferredCategory stayed null.
Preserve the sub when the URL already points at this app's config (matched by
appPartsFromPath().app === appName), so cold-load/refresh lands on the encoded
sub-tab. Cross-app switches still start at the first category. The sibling
showAppDetailWithConfig() (the grid 'manage' button) is intentionally left to
land on the first category.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The per-service rich detail panel (.service-rich: limits, healthcheck,
networks, mounts) and the live mem chip both did `const fmt =
window.SystemFmt` and bailed when it was absent. SystemFmt is defined by
the lazy admin System page module, so on the app Services tab it's usually
undefined — `_renderRichDetail` returned '' and the panel was never in the
DOM, so flipping Advanced revealed nothing.
Give the Services component its own SVC_FMT formatter (mirrors SystemFmt)
and use `window.SystemFmt || SVC_FMT` everywhere, dropping the hard bail.
The rich panel now renders whether or not /admin/system was ever visited.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The port manager's 'Show advanced fields' control was a raw <input type=
checkbox> + <span>, out of step with the modern pill switch the Services tab
uses for its Advanced toggle. Swap the markup to the shared .lp-ui-advanced-
toggle (track + thumb) structure — defined in services.css, which is loaded
globally — keeping .port-manager-show-advanced on the input for the JS hook.
Drop the now-dead .port-manager-advanced-toggle CSS.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
App-config fields rendered as a loose vertical stack with no grouping. Give
each .app-config .form-field the .admin-card glass recipe (var(--card-bg) fill,
faint rgba(text,0.10) hairline border, 12px radius, 16px padding) so each
option reads as a discrete container. Spacing now comes from the .panel-fields
flex gap, so the per-field margin-bottom is dropped to avoid doubling.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The pinned 'Overview' sidebar entry used a rounded, inset pill (margin +
border-radius:8px) so its hover/active highlight floated in the middle of the
sidebar, unlike the full-width app-category rows below it. Drop the margin and
radius, adopt the .category padding (15px 20px), border-bottom separator and
var(--surface-hover) hover, so the highlight spans the full sidebar width with
square corners and matches the categories. Keeps font-weight 600 + the
page-updater active tint as its only distinguishing marks.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The recessed task-list box had flex:1, so its background filled the full
height and ran well past the last task. Move the scroll onto .tasks-terminal
and let .tasks-list size to its content, so the box ends at the last task
(and still scrolls when the list overflows). Scrollbar styling follows to
the new scroll container.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
On /overview Backups the card surface lived on .main, which wraps both the
body and the flipped footer — so the background overhung past the action
buttons. Move the card surface onto .backup-page-body (rounded bottom,
joined to the tab strip) and let the footer sit transparent below it,
matching the app Config tab's .tabs-content + .config-actions split.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The fleet Overview area (Overview/Updates/Improvements/Backups/Migrate) now
lives at /apps/overview* instead of /overview*, reflecting that it belongs to
App Center. The Backups tab is therefore /apps/overview/backups, and the old
standalone Backup Center page is removed entirely:
- apps feature owns /apps/overview* (covered by the existing /apps* route); its
mount() dispatches /apps/overview -> fleet Overview before the grid check.
- _legacyRedirect() rewrites old short URLs so bookmarks/links keep working and
the address bar shows the canonical path:
/overview[/tab] -> /apps/overview[/tab]
/backup[/sub] -> /apps/overview/backups[/sub]
/updater, /peers redirects retargeted to /apps/overview*
- Removed the standalone backup feature: components/backup/{index.js,feature.json},
its manifest entry, the /backup route registrations and the dead handleBackup().
The BackupPage classes stay — the Overview Backups tab embeds them.
- Repointed every backups/overview link: the admin dashboard's 'Open backup
center', the app-card 'Open backup center' button + snapshot-overflow link,
the sidebar Overview entry, the improvements deep-link, and the Migrate
'go to locations' deep-link.
Also drop the redundant inline Check button from the Security empty state
(same rationale as Improvements: the host auto-scan repopulates it and the
header carries a manual Check).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Apply the same recessed-panel treatment as the Dashboard/Tasks to the
System page: gauges, Trends, Storage, Host and Per-app now sit inside one
dark rounded box under the header divider. Generalise the Dashboard's
.admin-card-grid-wrap into a shared .admin-panel class so both admin pages
use one source of truth.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Match the fleet Overview's .ov-tab-body treatment: the /admin/dashboard
card grid (under the header divider) and the /tasks list now sit inside a
recessed dark rounded box instead of floating directly on the page
gradient.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The Improvements tab's empty state ('No hotfix data yet …') rendered an
inline 'Check now' button. It was redundant: the embedding Overview header
already carries a manual Check, and the host-side auto-scan repopulates the
signed improvements index on its own within a couple of minutes (the empty
message already says so). Remove the button so the empty state is just the
self-explanatory, automation-backed message.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The config reconcile pass printed one 'Reconciled config: <name> (backup:
.<name>.bak)' line per changed file. Drop the per-file message entirely:
the intro notice and the two per-section '...completed.' confirmations are
enough, and the backup mention added noise. The hidden .<file>.bak sibling
is still written for safety — it's just no longer announced.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Holding the singleton flock at startup proves no other processor is alive
to heartbeat or complete anything, so every task still marked running is
a corpse from a killed predecessor. Recover them all before the first
dispatch (recoverOrphans now takes an 'all' mode) instead of waiting out
the 60s heartbeat-staleness window — which used to leave a dead task
showing 'running' alongside the genuinely-running next task for a minute
whenever the service was restarted mid-task (e.g. by the deploy chain
during initial setup). The idle-loop pass keeps the stale-only gate.
refactor(dashboard): slim the storage card back to chart + percentage
The disk card was only ever meant to be the donut and the % figure; drop
the Apps/Docker/Other/Free legend rows and signal the deeper view with a
corner expand glyph instead (the System page's chart-expand icon) — the
card already opens /admin/system/storage on click.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Services tab restart button POSTed to a backend endpoint that (a)
checked the app's compose path from INSIDE the webui container, where
the host's containers root isn't mounted — so every restart failed with
'Compose file not found' — and (b) queued a raw 'docker compose restart'
that the host task processor would run as the manager user, which can't
talk to the rootless daemon anyway. Errors surfaced via a bare alert().
Per-service restart now follows the exact shape of the whole-app verbs:
- CLI: 'libreportal app restart <app> [service]' — the optional service
arg makes dockerRestartApp restart just that compose service, via
dockerCommandRun (right user in rootless mode) from the app dir on the
host, where the compose file actually lives. Service names validated
against compose-legal characters before touching a shell line.
- WebUI: the button dispatches a 'service_restart' task action through
the task router (mutations-via-tasks), runs in the background with the
standard task toast + link — no page switch — and failures use the
notification system instead of alert(). Because the task runs host-
side, restarting the WebUI's own libreportal-service now works too.
- Backend: the mutating restart endpoint and its now-unused helpers are
removed; service-routes.js is read-only surface (status + log tails).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Reconcile backups now land as .<file>.bak instead of <file>.bak, so they
no longer clutter the configs folder. The .bak suffix is kept, so every
existing walker/sourcing exclusion still applies.
Also exclude dotfiles and *.bak from findConfigFileForOption: it walked
the configs tree with no backup exclusion, so depending on directory
order a 'config update' could resolve a key to the backup file and write
the user's change there — silently lost.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move the WebUI-updater settings out of general_terminal into their own
advanced webui-category file (webui_logs precedent): new
configs/webui/webui_updater holds CFG_UPDATER_SCAN_INTERVAL and the
migrated CFG_HOTFIX_AUTO, listed in webui/.category.
The move only reaches existing installs if the config convergence
machinery works, and three pieces of it silently didn't:
- checkConfigFilesMissingFiles walked a stale hardcoded category list
('general features network' — features doesn't exist; webui/backup/
security never healed). Derive the categories from the template tree
instead, and heal .category metadata too: copy it when absent and
merge missing SUBCATEGORY_ORDER entries when present, so healed files
actually appear in the WebUI Config editor. core_categories removed.
- Option reconciliation never touched ANY nested config file: configs_dir
carries a trailing slash, so rel stripping missed ('configs//'), the
template lookup failed, and reconcileConfigFile early-returned for
every file. Strip the slash before matching.
- reconcileConfigFile's AUTO_DELETE=false branch read a never-populated
live_line array, losing the dropped keys it promised to keep. Populate
it alongside live_value.
Also exclude *.bak from config sourcing (reconciliation writes <file>.bak
next to live configs — now that it runs, sourcing backups would resurrect
deleted keys), and add 'libreportal config check' as a non-interactive
front door to the converge pass (was only reachable via install flows and
the interactive menu).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the stat-tile grid with a needs-attention board: a health hero on
top (one big status circle — green all-clear / amber something-to-do /
neutral pre-scan — matching the admin dashboard's dot language) over one
status row per area (updates, security, improvements, backups). Rows that
want a decision are tinted with their area hue and carry their action
buttons inline (Review / Update all / Open Backups); healthy areas
collapse to a quiet neutral one-liner. An Everything / Needs action chip
pair filters the board down to just the actionable rows.
Board rows deep-link with intent: Security lands on the Updates tab
pre-filtered to the affected apps via a new data-filter hop on the goto
action. backupSummary() now splits never-backed-up from week-stale apps
so the backup row can say which it is.
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the click-to-scan-only flow with a self-throttled auto-scan that
rides the existing task-processor idle poll (the same shape as the
network-drift check — no new daemon, unit, or endpoint):
- 'libreportal updater check auto' gates on the age of the generated
updates.json vs CFG_UPDATER_SCAN_INTERVAL (minutes, default 30,
0 disables); a fresh file makes the 60s tick a single stat() + return.
Manual checks and post-update rescans reset the clock for free, and a
missing file means the first scan runs ~a minute after install.
- Eligible signed hotfixes keep flowing through artifactApplyAuto, which
only enqueues ordinary tasks — mutations stay on the task path.
- Open updater surfaces (standalone /updater and the fleet Overview's
headless UpdaterPage) follow along with a 60s static-JSON re-read that
repaints only when a generated_at stamp changed; timer released via
dispose() on unmount, ticks skipped while hidden.
- Empty states now say the first scan happens automatically; Check now
stays as the immediate manual override.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The dashboard carried a persistent, full-width "LibrePortal vX.Y.Z" /
up-to-date strip — passive info occupying prime real estate above the
user's actual content, and out of place next to genuinely actionable
cards. Version/update state is chrome, so it now lives in one persistent
pill in the global topbar (every page), with detail + actions behind the
existing modal.
The pill is calm by default and escalates only when warranted:
* up to date -> subtle "✓ Up to date"
* local/dev -> neutral version chip ("v0.2.0")
* checking -> spinning "Checking…"
* update waiting -> accented, pulsing-dot "⟳ Update"
Clicking opens the details modal (unchanged) for the full readout and the
Update now / Check for updates actions.
The dashboard update banner is removed. The network-notifier's banner
stays — it's attention-only (shown solely on a real, actionable network
conflict), which is exactly when a dashboard banner earns its place; its
topbar badge now anchors after the pill, and the dashboard data-loader
re-asserts that banner on mount.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Backups tab's embedded BackupPage repeated the active section as a
big page header (icon + 'Dashboard' + subtitle) right under the nested
strip that already names it. Embedded-scoped CSS now hides the title
block and flips the header below the body (flex order), so its actions
(Refresh + per-section primary) become a bottom-left footer row — the
same place app-detail tabs keep theirs. The export dropdown flips to
open upward from the footer. The standalone /backup page is untouched.
The Migrate ▸ Peers sub-tab drops its page header (breadcrumb + title
+ blurb) the same way: the peer list/empty state now sit in the shared
recessed .ov-tab-body container with the four actions in a bottom-left
.peers-actions footer.
Signed-off-by: librelad <librelad@digitalangels.vip>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Every fleet Overview tab now follows the per-app detail tab idiom the
rest of the app uses: title + description on the left, action buttons
on the right, a divider underneath, and the body inside the recessed
dark container (.tasks-container recipe).
- renderHeader() gains an action slot; Check/Check now/Update all move
out of in-body toolbars into the header (Updates keeps its filter
chips in the body; the Apps-tracked stat card drops its duplicate
Check button; UpdaterPage.renderImprovements can skip its toolbar).
- String tabs wrap their body in .ov-tab-body — margin/padding 16px,
rgba(bg,.2) panel — mirroring backup/tasks/updater containers.
- The Backups tab's embedded nested strip (Dashboard/Backups/Locations/
Configuration) now sits on the same surface as every other tab strip:
added to the nebula sidebar-bg anchor rule (it was stuck on the
lighter --hover-bg) and its buttons use .main-tab-button type.
Signed-off-by: librelad <librelad@digitalangels.vip>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The global tasks page still deep-linked a single task with a ?task=<id>
query while the rest of the SPA moved to path-based permalinks
(/app/<name>, /admin/<…>). Bring it in line: the task is now a path
segment, /tasks/<category>/<id>.
Task ids are guaranteed `task_<digits>_<base36>` (isValidTaskId), so the
redundant `task_` prefix is dropped in the URL and restored on read via a
new window.taskPath / window.taskPartsFromPath helper pair (mirrors
appPath/appPartsFromPath). The parser still accepts the legacy ?task=
query and the full-prefixed id, so old links, bookmarks and notifications
keep resolving.
Updated every builder (tasks-manager updateURL + notification url,
task-id link, task-actions, admin config-form, setup-wizard handoff with
its &from=setup flag) and the notification navigation handler / button
text to recognise the path form.
Signed-off-by: librelad <librelad@digitalangels.vip>
Landing on /tasks (directly, via a deep link, or from the setup-wizard
handoff) now opens the row the visit is actually about:
- init() re-reads the URL on every SPA (re)mount, so ?task= deep links
work after the first visit instead of using constructor-stale state.
- applyInitialSelection() opens the deep-linked task, or — for setup
handoffs whose first task the queue has already moved past, and for
plain visits with no deep link — the currently running task (else the
next queued one).
- The selection then follows the queue: when a new task starts running
the open panel moves with it, until the user manually toggles a row
or switches category (their choice then wins for the visit).
- selectTask() is the shared programmatic open: exclusive expand, live
log stream for active tasks, smooth scroll into view.
Signed-off-by: librelad <librelad@digitalangels.vip>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The app-detail header is a flex row, so the switcher rendered inline to the
right of the service buttons. Wrap the detail header and make the bar span
100% so it sits under the title/service buttons, above the tab strip, as
intended.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replaces the per-instance grid cards + per-card "New instance" action with a
single card per app type and an in-page family switcher — the UX you asked
for (swap between instances via LibrePortal navigation; manage on the page).
- Grid: hide any app declaring INSTANCE_OF (loadApps filter), so there's one
card per type. A subtle "N instances" chip replaces the old card button.
- App detail: a "family switcher" bar under the title for multi-instance
types and their instances — a pill per member (base + each instance) that
path-navigates to that slug's detail (current tab kept), plus "+ Add"
(existing create modal). When viewing an instance, a "Remove instance"
action sits in the bar.
- Remove flow: instance_remove task verb (-> existing `libreportal instance
remove`) with an on-brand confirm modal; lands back on the base type page.
Frontend-only; the instance create/remove backend and CLI are unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Lets a *multi-instance-capable* app run as several fully isolated instances
on one box (e.g. two Bookstack/WordPress sites, or a "family" + "work"
Nextcloud) — distinct data, DB, subdomain, backups and update cadence.
Design: an instance is just another app. It gets its own slug (<type>_<id>),
its own CFG_<SLUG>_* namespace, deployed dir, DB row, IP/port allocation and
host, so the entire existing pipeline (scan, install, services, routing,
updater, backups) treats it like any app with zero changes. All
instance-specific rewriting is confined to a clone of the type's template;
the shipped template and the core engine are untouched.
Gating: opt-in per app via CFG_<TYPE>_MULTI_INSTANCE=true. Only Bookstack
carries it for now (the validated reference). The other 31 apps are
unaffected — the feature is invisible unless the flag is present.
- scripts/instance/instance_create.sh — clone + re-namespace config, rewrite
compose identity (container_name / Traefik routers / backup labels) and
per-app tools, set a hostname-safe subdomain (PORT field 10), then hand off
to dockerInstallApp. Plus instanceList / instanceRemove.
- libreportal instance create|remove|list — new CLI category; mutations route
through the task system (no new mutating API endpoint).
- WebUI: "instance of <type>" badge + a "New instance" card action on capable
apps, and a create modal (name + domain# + subdomain, live host preview)
that dispatches the standard task. Capability/instance-of read straight off
the already-exposed app config.
Known follow-ups (documented): flip the flag on more apps after a compose
identity check (Nextcloud next); per-app tools are best-effort isolated.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The per-app Updates tab rendered a bare version/badge bar + detail with no
title, no dividers and no recessed container — unlike the Config / Backups /
Tasks tabs.
- Add a .updater-title block (⬆️ Updates + description, Check/Update actions,
bottom-border divider) mirroring .backup-title.
- Wrap the body in a .updater-detail-container recessed dark panel (same recipe
as .backup-snapshots-container / .tasks-container).
- Separate the Version/Security/Recovery/History sections with divider lines
(scoped to the container; fleet row-details keep their gap-only spacing).
- renderAppDetail() gains an opt-in Version section so the version/badge reads
as a section in the panel; fleet rows omit it (the row head shows it already).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Migrate and Backups detached their sub-tab strip from the content (16px gap)
and Backups had no top-level header and a transparent (card-less) body — so
both read as different from the per-app Config sub-tabs.
- Migrate: nest .tabs-content inside the same .tabs-wrapper as .tabs-list (the
canonical app-detail structure) and drop the gap, so the strip joins the
content card with no space and clean corners.
- Backups: inject the shared "Backups" .config-title header above the embedded
BackupPage; drop the sub-tab strip's bottom gap; give the content (.main) the
connected .tabs-content surface (card bg + rounded bottom) and round the
strip's outer top corners — so it reads as one tabs-within-tabs unit.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Drop the floating fleet header; every Overview tab now renders its heading
inside the pane, under the tab strip, in the per-app detail .config-title
format (emoji + title + description) — matching the Migrate tab.
Introduces a small modular system so the area can't drift:
- OV_TAB_META is the single source of truth for each tab's icon/title/blurb.
- renderHeader(id) is the only thing that turns it into markup; renderTab()
prepends it for the string tabs and mountMigrate() injects it once for the
static Migrate pane. Body renderers now only ever produce the body.
Retires the now-dead floating-header plumbing: updateHeader(), the
ov-backups-active/ov-migrate-active hide toggles + CSS, and the .overview-header
rules. The tabbed interface owns the top padding the header used to provide.
Backups is the documented exception — it embeds the full BackupPage, which
supplies its own header.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Give the fleet Overview › Migrate tab the same in-content header + sub-tab
styling the per-app detail tabs use:
- Add a .config-title header (icon + title + description) inside the Migrate
pane, above the Restore/Peers sub-tabs, and hide the generic floating fleet
header for Migrate (mirrors how Backups already supplies its own heading).
- Make the .tab-button icon↔label gap explicit (flex + 6px gap) so sub-tabs
render identically whether or not the markup has whitespace between the
emoji and name spans; align the Backups sub-tab strip gap to match.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
installDebianUbuntu ran apt (bare on line 14, via sudo on 17/20) during
the startPreInstall pass. Under the hardened de-sudo model the runtime is
the manager (libreportal, non-root) and the LP_SYSTEM sudoers allowlist
scopes systemctl/ufw/sysctl/loginctl/service but NOT apt — so every apt
call failed (exit 100, 'Updating System Operating system.').
Detect privilege once: run apt directly when root (the install-time path,
which also bootstraps sudo on a bare box), and skip cleanly with a notice
when we're the unprivileged manager. OS/security updates are a host /
install-time concern there, deliberately kept out of the manager's reach.
Also routes the trailing sysctl mkdir/touch through the same prefix.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The Beginner/Advanced experience cards are tap targets, but had no
user-select guard. Rapid clicking — notably the Advanced card's 10-tap
dev-mode unlock — selected the card title/description text as a side
effect. Add user-select: none to .setup-level-card.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Surfaces network_status.json in the WebUI, attention-only: a rose badge
+ dashboard banner appear ONLY when conflicts_found, with a details panel
listing the stranded apps and a 'Heal now' button that runs the heal
through the task pipeline (libreportal system network heal). Re-reads on
task completion via taskRefresh.
Cloned from update-notifier; the badge anchors just after the update
badge so the two coexist in a stable order. New --page-network hue.
Wired into system-loader scripts[], topbar onTopbarReady, and index.html.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Closes the gap behind the vpn-recreate bug: when the shared network is
recreated with a different /24, every app's stored static IP is left
outside it and adoptDockerSubnet only realigns CFG, not the apps.
- networkScanConflicts (network_conflicts.sh): read-only scan diffing each
active network_resources IP against docker's real subnet (via ipInSubnet).
Per-service routing-aware — skips gateway-routed services whose ipv4 is
commented out in the deployed compose, so gluetun apps don't false-positive.
Distinguishes 'daemon down' (benign) from 'network missing' (real).
- webuiSystemNetworkCheck (webui_system_network.sh): self-throttled generator
that writes frontend/data/system/network_status.json (modelled on
verify_status.json). Wired into webuiSystemUpdate AND run unconditionally
every ~60s from the task-processor poll (regen webui is mtime-gated and
would never fire on drift, which touches no source file).
- networkHealConflicts (network_heal.sh) + 'libreportal system network
check|heal [app]': the heal adopts docker's subnet in-process, then re-IPs
stranded apps with reset_network=ip (ports preserved), gluetun first.
Mutating path runs only through the task system (dual-mode, like update
apply); read-only check runs inline.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Foundations for network-drift healing:
- ipInSubnet(ip, cidr): prefix-aware CIDR membership (pure bash), so
stored IPs can be checked against docker's real subnet. Honours the
actual prefix, so a healthy /16-subnet + /24-ip-range install is not
mistaken for drift.
- dockerInstallApp now accepts reset_network="ip": re-roll the static IP
from the current subnet but PRESERVE published host ports (clears only
IP rows; LIBREPORTAL_RESET_IP_ONLY keeps port_allocate reusing existing
ports). This is the heal path — a subnet move strands the IP, not the
port, so we don't churn bookmarks/forwards/proxy upstreams. reset="true"
still re-rolls both.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two latent issues uncovered while designing network-drift detection:
- adoptDockerSubnet's comment claimed apps' IPs stay inside docker's
subnet after adoption. False: IPs are pinned to the old subnet's first
three octets, so adopting a different /24 base strands every app IP
out-of-subnet. Document the real behaviour + the heal paths.
- ipAllocation fell through from the existing-row branch to the
unconditional INSERT, which would violate UNIQUE(app,type,service).
Unreachable on today's reset path (rows are deleted first) but a hazard
for any direct caller; add an explicit return after reuse/reset.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
installDockerNetwork errored with 'network with name <x> already exists'
on re-runs: the requirement check sets DOCKER_NETWORK_SETUP_NEEDED=true
whenever 'docker network inspect' returns non-zero, but that also happens
when the rootless daemon socket isn't reachable yet — indistinguishable
from the network being genuinely absent. A prior install also leaves the
network behind, so the flag fires on every re-install.
Re-check existence right before creating and converge: if the network is
already there, leave it in place and adopt its real subnet into CFG rather
than erroring. This also stops the spurious subnet randomization (and the
resulting CFG drift) that ran before the doomed create.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The 'Installing System Requirements' step ran apt-get install with no
output until checkSuccess reported afterwards, so it looked frozen
while packages were being fetched. Print a notice up front.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Restore empty-state 'Open Locations' now deep-links to the backup center's
Locations sub-tab: the embedded center honors /overview/backups/<sub> and
switchTab()s to it after mount (was landing on Dashboard).
- PeersPage.notify + MigratePage.notify use the real window.notificationSystem
(were calling a never-defined window.showNotification → console-only).
- Remove the now-dead Admin config-manager peers branch (Peers left Admin).
- Trim the dead migrate.json/peers fetch + hostnameToPeerName from BackupPage
(no consumer after the migrate removal).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Remove the Admin sidebar Peers entry; /peers and /admin/tools/peers now
redirect to /overview/migrate/peers (its new home next to cross-host Restore).
- Re-skin the embedded Backups center's sub-tab strip from pills to the per-app
Config .tabs-list/.tab-button segmented look (full-width bar, accent underline)
so every nested sub-tab row is consistent.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The embedded backup center's lazy bundle still listed backup-migrate.js, which
was just removed — its 404 failed the whole load chain. Drop it.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Migrate now lives at Overview › Migrate › Restore (standalone MigratePage). Strip
it out of BackupPage: drop the migrate sidebar item/panel/modal from the
fragment, the 'migrate' tab from the allowed set / titleFor / subtitleFor /
iconFor, the renderMigrate() call, and the migrate-host/app/confirm click
handlers; delete the now-orphaned backup-migrate.js. The backup center is now
Dashboard/Backups/Locations/Configuration.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New 5th Overview tab 'Migrate' with a nested segmented sub-tab row reusing the
per-app Config-tab .tabs-list/.tab-button design:
- Restore: a standalone MigratePage (cross-host migrate moved out of BackupPage
into its own controller + fragment + modal; own data fetch + task dispatch).
- Peers: reuses the existing PeersPage (container-parameterized) + its template.
Both lazy-loaded on first open and disposed on apps-feature unmount. Additive —
migrate is still in the backup center and Peers still in Admin until the next
commits remove the duplicates.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- BackupPage task-refresh run-guard is instance-relative now (window.backupPage
OR window.overviewBackupPage), so the embedded center's own auto-refresh works
instead of relying on OverviewManager's overlapping coverage.
- _ensureBackupAssets no longer memoizes a rejected promise — a transient
script-load failure no longer bricks the backup center until reload; the next
open retries.
- spaClean.loadScript removes the failed <script> element on error so the
getElementById dedupe can't make a retry resolve without re-fetching.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The 'already mounted?' guard checked #backup-section, but the per-app Backups
tab also defines that id — detect the prior mount via the pane's own
.backup-layout instead so the check is correct, not coincidental.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Instead of a glance + 'Open backup center' button, the Backups tab now mounts
the real BackupPage (dashboard/snapshots/locations/migrate/configuration) inline,
with its sidebar restyled as a nested tab strip and its own header taking over.
- BackupPage gains an embedded mode (opts.embedded): no /backup URL coupling, so
sub-tabs switch in-page under /overview/backups. Backward compatible.
- OverviewManager lazy-loads the backup bundle + fragment on first open, news a
BackupPage({embedded:true}), and disposes it on apps-feature unmount. Colliding
ids (#sidebar/#mobile-overlay) are stripped on inject.
- Revert the Admin backup-config surface — the embedded center (incl.
Configuration) is now the single home for backup settings.
- The updater needs no equivalent: its sections were already unpacked into the
Overview/Updates/Improvements tabs + the per-row expander.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- HIGH: renderAppDetail gated the Roll back button on last_snapshot* fields no
generator emits, so it never rendered. Derive recoverability from update
history (updater_apply always snapshots first) so the affordance is reachable.
- MED: per-app Updates tab now repaints on update/rollback/check/hotfix task
completion (mirrors the backups card) instead of going stale until re-click.
- MED: in-page tab switches now sync spaClean.currentRoute, so the sidebar
Overview entry no longer no-ops after switching tabs.
- LOW: keyboard activation (Enter/Space) for the role=button expander heads,
backup tiles, and sidebar Overview entry.
- LOW: preserve ALL expanded Updates rows across a background repaint, not just
the single ?app= deep-link (split toggle into _openDetail/_closeDetail).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The fleet Overview area (Overview · Updates · Improvements · Backups) and the
backup center now live under App Center, so the standalone top-nav items are
redundant. Top nav is now Dashboard · App Center · Admin · Tasks.
- Remove the Backups and Updates anchors from topbar.html.
- Remove the nav{} blocks from updater/backup feature.json + manifest (so they
don't resurface when the nav kernel lands).
- Highlight App Center for /overview and /backup; drop the dead /updater branch.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- _legacyRedirect: /updater[/tab] -> /overview[/tab] (security/recovery/history
fold into the Updates expander -> /overview/updates). /backup is intentionally
NOT redirected — it stays the operational backup center (locations/migrate/
snapshots), reached from Overview › Backups.
- Re-point the per-app hotfix chip to /overview/improvements.
- Unhide the existing backup config category in the Admin sidebar so
engine/schedule/retention config lives under Admin (same generated category
the backup center binds, so edits stay in sync).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New 'Updates' tab in the app detail page, beside Backups. Reuses the headless
UpdaterPage + renderAppDetail() scoped to the single app, so the per-app and
fleet views share one data/render path. UpdaterPage is added to the apps script
bundle so it's available on app pages; the tab is disabled while a task runs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Open the per-app row named by ?app=<name> on load/repaint and write it back on
toggle, so an expanded Updates row is a shareable URL — mirrors the Tasks page's
?task=<id> pattern.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Introduce a per-fleet Overview area inside the apps shell, reachable from a new
'Overview' entry pinned above the apps sidebar search. Selecting it renders a
top-tabbed view in the main pane — Overview · Updates · Improvements · Backups —
mirroring the per-app tabbed layout, with the apps sidebar persistent.
- TabController: generic root-scoped show/hide tab host (core/ui-state).
- OverviewManager: drives the 4 tabs. Reuses a headless UpdaterPage for all
update/CVE/improvement data + rendering (its renderX() are pure HTML
producers) and reads backup/dashboard.json directly for backup health.
- Overview tab: combined update + backup health cards.
- Updates tab: per-app expander table (CVEs/recovery/history inline via the new
UpdaterPage.renderAppDetail) + All/Updates/Security filter chips.
- Improvements tab: reuses the updater's signed-hotfix renderer.
- Backups tab: fleet backup-health tiles; actions deep-link per app.
- Additive only: /overview* routes on the apps feature; old /updater and
/backup pages untouched. Cleanup (redirects, nav, Admin config move) is next.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Update §8.7 + the banner + §1 TODOs to reflect that Phases 2–5 shipped today
(apply/revert pipeline, severity-split auto-apply, the WebUI Improvements stream
+ per-app chip, and make_hotfix.sh). Only the registry/marketplace stays
deferred (demand-gated by design).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The maintainer-side tool that turns a small hotfix SPEC into the two signed
artifacts the install verifies + applies (completes the hotfix product):
dist/<channel>/payloads/<id>.json(.minisig) the bounded declarative op list
dist/<channel>/index.json(.minisig) the catalog (entry upserted, serial++)
laid out exactly like get.libreportal.org serves it (local-serve testable).
- Reads a spec (envelope fields + an embedded ops array); inlines any
op `content_file` to content_b64 for convenience.
- Validates id charset + every op name against the applier's CLOSED vocabulary,
so a typo can't ship an artifact that fails-closed on every box.
- Builds the payload (sha256), the envelope (payload ref {kind,url,sha256,sig}),
and upserts it into index.json — bumping index_serial, refreshing valid_until
(LP_HOTFIX_VALID_DAYS, default 30), and recording the publisher in the
publishers map with role + the footprint public key.
- minisign-signs the payload + index when LP_MINISIGN_SECKEY is set (the offline
key, kept on the release machine, same as make_release.sh); unsigned otherwise
for local testing — `libreportal artifact apply` refuses to apply unsigned.
Verified end-to-end (unsigned mode): produces a valid index.json + payload.json
matching the §8.1 envelope that lpFetchIndex / artifactApply consume.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Surfaces the hotfix channel in the WebUI. Primary home is the Updates &
Improvements page (the updater component) — its own "Improvements" tab — with a
secondary chip on the App detail page (fork 3 locality = both).
Updater component (components/updater):
- New "Improvements" sidebar tab + panel; renderImprovements() reads the host-
generated artifacts_available.json (severity badge, scope chip, applied/auto/
not-applicable badges, plain-English why). Apply/Revert buttons dispatch
artifact_apply / artifact_revert through the TASK system (services.tasks.route)
— no mutating API. Apply is disabled when the index is UNSIGNED.
- Overview gains an "Improvements" stat card; task-refresh now also repaints on
artifact_* task completion; URL tab routing + dispose teardown extended.
Task plumbing (core/tasks): artifactApply/artifactRevert action methods (id is
charset-guarded before it enters the command string) + artifact_apply/
artifact_revert routeAction cases. Task list/format gain icons + friendly labels.
Apps component: an amber "⚡ N improvements" chip on an installed app's detail
header (populated async from artifacts_available.json filtered by app, applicable
& not-applied), linking to /updater/improvements. Best-effort, never throws.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- CFG_HOTFIX_AUTO (security-breakage|all|off, default security-breakage) seeded in
general_terminal; reaches existing installs via the add-only config reconciler.
- webui_artifact_scan.sh (webuiArtifactScan): fetch+verify the signed index, write
artifacts_available.json ATOMICALLY (build in temp → jq-validate → one write;
keep the prior file on any failure — never emits broken JSON). Annotates each
artifact with applied (a per-id record exists) + applicable (target installed).
- artifactApplyAuto + `libreportal artifact apply-auto`: enqueue apply tasks for
the eligible signed hotfixes — only when the index is VERIFIED-signed, only
auto==true + in the severity policy + applicable + not already applied. Each
apply is its own task (visible in the log + History), never applied inline.
- `updater check` now also refreshes the index (webuiArtifactScan) and runs
artifactApplyAuto — one front door, no second phone-home.
Unit-tested 13/13: policy filtering (security-breakage / off / all), auto:false
exclusion, already-applied skip, non-installed-app skip, unsigned-index fail-closed,
and the scan transform's signed/applied/applicable fields.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
A 4-lens adversarial security review of the Phase 2 applier raised 19 issues
and confirmed 17 after per-finding verification. All are trust-boundary (they
require the signing key), but several break the explicit "no code-exec, always
reversible, nothing-silent" contract, so all 17 are fixed:
Trust path — fail CLOSED, never misreport:
- lpFetchIndex now surfaces the real signature state (LP_INDEX_SIGSTATE);
artifactApply REFUSES to mutate unless the index is actually verified, and
_artifactFetchPayload refuses an unsigned payload. The read path still
tolerates dev/unsigned but now says "UNSIGNED" instead of "Signed + verified".
- valid_until and index_serial are now MANDATORY + numeric in lpFetchIndex
(missing = refuse) — closes the anti-withholding / anti-rollback fail-opens.
Injection / code-exec (defense in depth even for a signed payload):
- runFileWrite rootless branch no longer builds a `bash -c` shell string with the
destination interpolated — it uses the argv form (like runFileOp), so a path
with a quote can't inject a command as the install user. (shared-helper fix)
- op paths must match a safe-filename charset (no quotes/$/backtick/;/newline);
set-config-key values and set-compose-image refs are charset-guarded too.
- content_b64 is validated as real base64 at precheck.
Reversibility / honest failure:
- dockerComposeUp now returns the real compose exit status (it always returned 0,
so the updater's rollback gate AND the apply's start-failure detection were
fail-open). (shared-helper fix)
- set-config-key undo captures the WHOLE config file (lossless) instead of a
lossy re-parsed scalar; edit-only (rejects an absent key).
- _artifactReplayUndoFile returns non-zero if any inverse op fails; auto-rollback
and revert now record "rollback-incomplete"/"revert-incomplete" + isError
instead of falsely claiming success, and revert keeps the record for retry.
- applied-record write failure is checked — apply rolls back rather than leave an
un-revertable change. System-scope regen failure is no longer swallowed.
- Writes are path-aware (configs/ -> runInstallWrite, container tree ->
runFileWrite) so system-scope hotfixes write/restore correctly.
- Checked lazy-sourcing surfaces a clear error instead of a bare exit 127.
Unit-tested 35/35 (adds: command-sub value rejection, bad image-ref, invalid
base64, quote/metachar path-injection rejection, replay-failure reporting).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The mutating side of the unified distribution primitive (spec §8.3). Hotfixes
can now be applied and reverted, first-party, through the task system.
New scripts/cli/commands/artifact/cli_artifact_apply.sh:
- artifactApply <id>: resolve+gate (applies_when / min_lp / max_lp /
max_footprint / publishers-map role) → fetch+verify payload (sha256 pinned by
the signed index + minisig) → dry-precheck ALL ops (all-or-nothing) → best-
effort snapshot → apply each op recording a precise inverse → bring app up →
auto-rollback (replay undo LIFO, snapshot fallback) → applied-record + History.
- artifactRevert <id>: replay the applied-record's undo log (LIFO).
- Bounded, CLOSED op vocabulary (no run-script/exec, ever): set-config-key,
set-compose-image, patch-file-if-checksum-matches, set-data-file. An
unsupported op rejects the whole artifact at precheck (fail-closed).
- Write-target firewall: scope:app → containers/<app>/ only; scope:system →
configs/ only; the install tree (our code) is off-limits to hotfixes (fork 1).
Drift guards (expect_current / checksum) skip cleanly rather than clobber.
- Two-tier trust: index minisig-verified vs the footprint key (lpFetchIndex)
covers the envelope; payload sha256-pinned + minisig-verified; publishers-map
role gate (a non-official publisher can't claim official). Community per-
artifact-key sigs are gated off until that tier is enabled.
cli_artifact_commands.sh: apply/revert via the task system (artifact_apply /
artifact_revert types — no allowlist needed), + read-only `applied` list.
cli_updater_commands.sh:
- FIX verified safety bug: updaterApplyApp/RollbackApp called `libreportal backup
app "$app"` and `... restore latest`, which parse the app name as the ACTION,
hit the dispatcher's `*)` default (exits 0) — so updates ran with NO snapshot
and rollback was a silent no-op. Call backupAppStart / restoreAppStart directly.
- FIX updaterRecordHistory jq-silent-skip: was `command -v jq || return 0`
(silently dropped the audit entry). Now fail-closed with a brace-agnostic
bash-native prepend fallback; extended with artifact_id/serial/undo_id.
fetch.sh: add _lpJsonEsc (shared JSON-escape for the jq-free fallbacks).
Regenerated source arrays + lazy-load manifest for the new file/functions.
Unit-tested 31/31: every op apply+precheck+undo round-trip, the path-allowlist
firewall (incl. .. traversal + install-tree + cross-app rejection), all-or-
nothing abort, unsupported-op rejection, and the History bash-native fallback
(records + preserves prior entries without jq). A full signed-apply e2e needs
minisign + the signing key (Phase 5 make_hotfix.sh).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The four-lens design panel finished (marketplace-first ranked top) and
confirmed the format; graft in the strongest refinements it surfaced so
the spec is genuinely "done":
- Publishers MAP trust anchor: `publisher` is now a key into an index-root
`publishers` map ({display, role, key}) the team-signed index vouches
for, not an inline {name,trust}. An artifact's claimed trust is honored
only if the publisher's role permits AND its sig verifies against that
key — so a community key can never self-certify as official. This is the
load-bearing trust mechanism for the marketplace seam.
- Two-tier reversibility: a per-op `undo` array (precise revert) plus the
snapshot (dirty-op fallback).
- All-or-nothing dry-precheck-all before any snapshot; unknown op rejects
the whole artifact at validation.
- Canonical-bytes signing rule (sign the exact artifact bytes, never
re-serialize on the box) + warrant-canary countersigning index_serial.
- Op vocabulary grown to the full set (set-data-file as the bridge to
bundles; set/unset-compose-env; ensure-compose-up/restart-service).
- Envelope gains version/supersedes/reversible + richer applies_when
(image_match/requires/conflicts).
- CFG_HOTFIX_AUTO + staged rollout / randomized delay / recall-via-supersedes.
- Flag the VERIFIED existing bug: updaterRecordHistory silently skips the
audit entry when jq is absent (cli_updater_commands.sh:154-168) — Phase 2
must make it fail-closed; "nothing silent" depends on it.
- Phases re-sequenced (P2 heart, P3 auto-apply, P4 WebUI, P5 make_hotfix.sh,
deferred registry).
Spec-only change — no code; the Phase 1 read primitive is unaffected (it's
a generic verified fetch; publisher/envelope internals are Phase 2).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
curl's raw "(6) Could not resolve host" / 404 noise leaked through on
the index.json download while the .minisig fetch was already silenced —
inconsistent and confusing. The caller's clean isError covers the
failure, so route the index download's stderr to /dev/null too.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Build the read side of the unified distribution primitive from
docs/roadmap/updates-and-distribution.md: one team-signed catalog
(index.json) on the same channel as latest.json, listing type-tagged
artifact envelopes. A hotfix is the first artifact type; apps/themes/
components are future envelope rows through the SAME pipe — the
marketplace seam is just the `type` + `payload.kind` fields.
Phase 1 is fetch + verify + parse only (NO mutation; the snapshot →
ops → rollback → History apply verb is Phase 2):
- Factor `lpVerifyMinisig` out of `lpFetchRelease` (scripts/source/
fetch.sh) — one trust anchor (the root-owned footprint key) now
shared by releases and the index; refactor `lpFetchRelease` to use
it (behaviour-preserving, still fail-closed).
- scripts/source/artifacts.sh: `lpFetchIndex` — download →
verify-before-parse → `valid_until` freshness (anti-withholding) →
`index_serial` monotonic high-water (anti-rollback, TUF-lite) → emit
verified JSON. Trust core is jq-free; parsing accessors prefer jq
with a grep fallback.
- `libreportal artifact index` (scripts/cli/commands/artifact/) —
read-only front door that fetches, verifies and lists. Runs directly
like `updater check` (no task; no mutation).
- Regenerate the source arrays + lazy-load function manifest for the
new files.
Doc: promote the format from vision to spec (§8) — 3 layers
(INDEX/ENVELOPE/PIPELINE), the bounded declarative op vocabulary (no
run-script, ever), the apply pipeline mapped onto existing functions,
the marketplace seam, and resolutions for all five open forks.
Self-tested 12/12: trust core fails closed (real key + no minisign →
refuse), happy path, stale-refused, rollback-refused, signature-refused,
jq + grep parsing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The teardown audit found the backup-stacking leak class across 4 more feature
modules (12 confirmed leaks); unmount() left document/window listeners, intervals,
and SSE subscriptions firing on stale controllers after navigation:
- admin: overview/ssh/peers/system each leaked a document click listener ->
AbortController + dispose() per page; admin unmount() aborts each.
- dashboard: the 1 Hz update-countdown interval + the LiveSystem view sub ->
stopUpdateCountdown()/detachDashboardLive(), registered via ctx.sub().
- tasks: constructor-started global live-log poller (discarded handle) -> stored
+ idempotent + cleared on unmount + re-armed on mount; per-task monitorTask
window listeners + interval -> tracked in a map, released on unmount.
- apps: app-tabbed reconcile setTimeout loop + watchdog window/document listeners
+ popstate -> per-instance AbortController + dispose() that clears the timer,
resets the guards, and unloads the active tab's Services intervals + log SSE.
All mirror the kernel's MountContext teardown discipline. 12 files, all pass
node --check. Backup (fixed earlier) re-confirmed clean by the audit.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The original report: clicking a backup sidebar tab loaded content on top of
the old content. Root cause (flagged in the unmount comment as deferred):
BackupPage.bindEvents() attaches document-level click/input/change listeners
guarded only by the instance-level this.eventBound, and unmount() nulled
window.backupPage WITHOUT removing them. Each revisit added another full set of
listeners bound to a stale BackupPage, all firing on every click and mutating
the live DOM (double tab-switches, double modal opens, stale-instance renders).
Fix (mirrors the kernel's MountContext pattern): give BackupPage an
AbortController, bind the three document listeners to its signal, add dispose()
that aborts them (+ drops the task-refresh reg + clears the timer), and call it
from the feature module's unmount(). Revisits now start clean — one live
instance, one set of listeners.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The honest-checkSuccess + masking fixes immediately surfaced a real masked
failure in error_report.log: updateDockerSudoPassword (run every system scan
from start_scan.sh) does 'sudo passwd $sudo_user_name', but Model A's scoped
sudoers grants only LP_HELPERS/LP_SYSTEM + run-as-install-user — not passwd.
So at runtime (manager, non-root) it failed exit 1 every scan, masked until now.
The password is set at install (root, chpasswd) and admin login is key-based,
so the runtime re-sync is legacy + impossible under de-sudo: guard it to skip
unless EUID 0. (Validates the surfacing mechanism working as intended.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
'local result=$(cmd)' resets $? to 0 (the local builtin's own exit), so the
following checkSuccess always saw success regardless of cmd's real exit — the
mechanism that masked the de-sudo write failures. Split declaration from
assignment ('local result; result=$(cmd)') across all 235 active-code sites
(84 files) so the command's exit reaches checkSuccess. No behaviour change
beyond $? now being accurate (no set -e in runtime code; multi-line
assignments transform safely).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
checkSuccess silently reported '✓ Success' for failed commands, which is how
the de-sudo write gaps (throttle stamp, passwords, updater) hid. Rework it:
- Capture the real exit code up front; success path unchanged.
- On failure, ALWAYS append to a greppable $logs_dir/error_report.log tagged
with the caller's script:line + exit code — a failure can't hide behind a
green check anymore.
- New CFG_REQUIREMENT_CONTINUE_ON_ERROR (default true): log + continue so one
failure doesn't abort the run and we surface EVERY issue in a single pass.
Flip it off later for strict abort/prompt (the prior behaviour, preserved).
Documents the 'local VAR=$(cmd); checkSuccess' footgun (local resets $?), which
the next commit fixes across the tree.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
updater_check/apply/apply_all/rollback tasks fell through every per-type
branch of the Tasks panel, so they showed the generic custom gear icon, a
raw/truncated command title, and (for the app:'updater' sentinel) a broken
hidden app icon. Wired them in like every other task type:
- tasks-format.js formatCommandForUser PATTERNS: added the 'libreportal updater'
command rows (Apps - Check for Updates / Update All / <App> - Update /
<App> - Roll Back) — only the *self*-update 'libreportal update' was mapped.
- tasks-format.js formatActionTitle: added the updater_* short labels.
- tasks-list-render.js getTaskTypeIcon: updater_check 🔍 / apply ⬆️ /
apply_all ⬆️ / rollback ↩️ (reusing existing verify/update/restore classes).
- tasks-list-render.js renderTaskIcons: treat app:'updater' as a sentinel like
app:'system' so updater_check/apply_all fall back to the LibrePortal logo
instead of a 404'd /core/icons/apps/updater.svg (apply/rollback keep their
real app icon).
node --check clean.
Signed-off-by: librelad <librelad@digitalangels.vip>
Two more cases of the manager writing directly into the container-owned
/libreportal-containers tree (same class as the regen-poll stamp), both masked
by a '✓ Success' that printed anyway:
- Password replacers (config/password/*): used 'runInstallOp sed -i' (manager)
on app configs copied into the container tree, so sed -i EACCES'd its temp
file and the substitution silently failed — the adguard.config 'couldn't open
temporary file', leaving the literal RANDOMIZEDPASSWORD placeholder. Added
runCfgOp (picks runFileOp vs runInstallOp by the target file's location) and
routed every $file grep/sed/awk through it: password, username, hex, vapid,
appkey, and bcrypt.
- Updater generator (webui_updater_scan): 'runFileOp cp <manager-tmp>' can't
read the manager's 0600 mktemp as the container user, so it fell through to a
manager 'cp' that EACCES'd on the container-owned out_dir. Switched the three
writes to 'runFileWrite < tmp' (manager shell reads the tmp; container user
tees the write).
Both deploy via the normal quick path (relocatable scripts) — no footprint bump,
no reinstall.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
A self-referential array — files_source.sh enumerates the arrays/ files — only
picks up a newly-created arrays/ file on the next regen pass. The task-folder
move created files_task.sh; this pass adds it to source_scripts so the committed
arrays match a fresh regen (and make_release's stale-array guard stays happy).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
maybeRegenPoll truncates $REGEN_POLL_STAMP (.regen_poll_at) to throttle the
self-heal 'regen webui' poll, but the stamp lives in the docker-install-owned
TASK_DIR — the manager-run processor can't write there, so the truncate
EACCES'd every poll (swallowed by || true). The stamp never updated, so the
throttle read last=0 forever and 'regen webui' ran on every idle tick (and
spammed the journal ~16x/min).
Fix: pre-create the stamp world-writable in setupTaskDir, exactly like the
lock file and FIFO already are (runFileOp install -m 666). Truncate then
lands, the mtime advances, and the poll throttles to REGEN_POLL_INTERVAL.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The task processor is a systemd-service daemon, not a cron job — move it out
of the misleadingly-named scripts/crontab/task/ to scripts/task/.
To stop the systemd unit from baking the processor's in-tree path (the footprint
coupling that forces a reinstall on every reorg), the unit now ExecStarts the
stable wrapper: /usr/local/bin/libreportal __task-processor. start.sh intercepts
that early (after paths.sh, before the heavy load), exports install_scripts_dir,
and exec's the processor with start_script. Future moves/renames need only the
one hand-off updated + a regen — no footprint bump.
- git mv scripts/crontab/task -> scripts/task (filenames kept; cron-watchdog grep
+ function names unchanged)
- libreportal-svc: ExecStart -> stable wrapper launcher
- start.sh: __task-processor internal launcher (export install_scripts_dir; exec)
- crontab_task_processor.sh: fix self-location ../.. -> .. for the new 1-level
depth (latent bug the move would otherwise have introduced)
- regen files_*/function_manifest; add task_scripts to the app/cli aggregates
- footprint_version 3 -> 4 (root-owned svc unit changed -> needs a root reinstall)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Verified-dead assets from the feng-shui audit, zero consumers:
- core/icons/categories/utils.svg — no 'utils' app category exists (the only
'utils' refs are unrelated system health-check names); category icons are
requested as /core/icons/categories/<id>.svg and no id is 'utils'.
- core/icons/apps/portainer.svg — Portainer was retired to
scripts/unused/OLD_CONTAINERS/; no live containers/portainer/, and apps.json
is generated only from live containers, so the icon is never requested.
Both git-recoverable if a portainer app / utils category is ever (re)added.
Signed-off-by: librelad <librelad@digitalangels.vip>
The shipped frontend carried ~600 muted '// console.…' debug statements (and
their multi-line commented continuation lines) left over from development —
clutter across 30 files. Removed them with a guarded pass that ONLY ever deletes
lines starting with // (so it can never alter behaviour), consuming each
commented console opener plus its continuation comment lines until the
string-stripped parens balance.
665 lines removed, 30 files; 0 insertions. Verified every deleted line is a //
comment (no code touched), real prose comments preserved, full node --check
sweep clean.
Signed-off-by: librelad <librelad@digitalangels.vip>
From the feng-shui audit naming findings:
- admin/overview/js/admin-overview.js -> overview-page.js (class AdminOverview ->
OverviewPage, window globals + the 'admin-overview' task-refresh id ->
overview-page, lazy-load path + typeof/new in config-manager.js).
- admin/system/js/admin-system.js -> system-page.js (class AdminSystem ->
SystemPage; now sits beside its -page sub-views system-metric-page.js /
system-storage-page.js).
- tasks/js/tasks-logs-modal.js -> tasks-log-modal.js (singular 'log' to match its
sibling tasks-log-stream.js; single path ref in system-loader.js).
These were the only page controllers breaking the dominant <thing>-page.js /
<Thing>Page convention (ssh-page/peers-page/backup-page/updater-page/
system-metric-page/system-storage-page). Pure renames; node --check clean.
Signed-off-by: librelad <librelad@digitalangels.vip>
- docs: remove the docs/README.md index and docs/CONTRIBUTING.md pointer
(duplicate filenames); the canonical contributing guide stays at
docs/contributing/contributing.md. Clean tree, no name collisions.
- scripts/system/*: 6 helper headers + host_access.sh said the helpers
install to /usr/local/sbin, but init.sh installs all of them to
/usr/local/lib/libreportal/ (verified via initRootHelpers + the sudoers
Cmnd_Alias). Corrected. The only remaining /usr/local/sbin is the legit
PATH export in the task processor.
- frontend kernel: drop migration-era comments that are now false post-
modularization (feature-registry 'passive/phase 0/unused', lifecycle
'ctx.services lands with Phase 2', manifest 'scan generator lands') —
describe current behaviour instead.
Comment-only edits to scripts/system/* — no footprint_version bump (no
behavioural change; bumping would force needless reinstalls).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Audited every doc against the code. Three fixes:
- system-footprint.md: add the libreportal-crowdsec root helper row
(init.sh installs 8 helpers; the table listed 7). appcfg row clarified
to 'CrowdSec-bouncer' config since the new helper does the host install.
- .gitattributes: add 'site export-ignore' — development.md documents the
website as never-shipping, but the rule was missing, so site/ was landing
in release tarballs. No runtime refs to site/; hosting lives in the Infra repo.
- promise.md: fix LICENSE link (../../LICENSE) after the docs/ reshuffle.
Everything else (install-and-use, development, contributing) verified current:
all install/uninstall/update flags, release scripts, fetch fns, footprint_version,
service name, and config keys check out against the code.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Sort docs/ into guide/ contributing/ architecture/ roadmap/ and rename
to consistent kebab-case (USER->guide/install-and-use, FOOTPRINT->
architecture/system-footprint, frontend-modularization->architecture/
webui-architecture, etc.). Add a docs/README.md index and a docs/
CONTRIBUTING.md pointer so the forge still surfaces the contributing
guide. Fix every reference (README, init.sh comments, frontend code
comments, and the USER<->DEVELOPMENT cross-links). History preserved
via git mv. Root stays README.md + CLAUDE.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Comment-only tidy from the feng-shui audit — no code behavior changes. The
features/ directory was renamed to components/ during modularization, but
several header banners and inline comments still named the old path:
- 6 component module headers (admin/tasks/backup/dashboard/updater/index.js +
updater/js/updater-page.js) now name their real components/<id>/… path
- core/kernel/js/spa.js + core/tasks/js/task-router.js comments
- backend/routes/features.js doc-banner (drop a components/<id>/ folder …)
- core/update-notifier/css/update-notifier.css header (js/update-notifier.js)
Guarded the rewrite so the LIVE /api/features/list endpoint name (feature-
registry.js sources + backend route) is untouched — only stale 'features/<path>'
directory references were updated.
Signed-off-by: librelad <librelad@digitalangels.vip>
The modularization shipped (2026-05-30), so the design doc was stale and
internally contradictory: it described a features/ tree (real tree is
components/), a shell-generator/Node route that were never built, and a
'partially implemented' status. Replace the 59KB design exploration with
a short, accurate description of the component-module system as it exists
(components/<id> pages, core/ subsystems, the kernel: feature-registry/
services/lifecycle/spa, static manifest discovery, mount/unmount contract,
eager global CSS). Fix one stale features/ path in a spa.js comment.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
From the feng-shui audit (all adversarially verified):
- BUG (high): apps-grid.js category tiles used onerror fallback
/core/icons/categories/default.svg, which doesn't exist (the dir has
misc.svg as its generic icon, which data-loader.js already uses). Any
category missing its named SVG showed a broken-image glyph. Repointed to
/core/icons/categories/misc.svg.
- TIDY: core/forms was the lone depth-3 nesting — JS at forms/controls/js/
while its CSS sat at forms/css/ and every other core subsystem uses
<name>/js/. 'controls/' grouped nothing (just the 2 custom-* widgets), so
flattened to core/forms/js/ (+ updated index.html). forms is now symmetric.
- CONSISTENCY: components/manifest.dev.json entries carried nav.order but not
the top-level 'order' that each feature.json has; added it so the API-down
fallback matches the live /api/features/list scan.
Signed-off-by: librelad <librelad@digitalangels.vip>
Captures the brainstorm on hotfixes, the updater reframe to
'Updates & Improvements', and registry-not-marketplace distribution:
one signed/declarative/reversible primitive behind hotfixes, app
installs, and themes. Vision/TODO doc with open forks, not a spec.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The config-category icons sat at admin/config/icons/CONFIG/ — the inner config/
duplicates the subsystem name; they belong in the icons root. Moved all 6
(backup, features, general, network, security, webui) up to
components/admin/config/icons/ and updated the two consumers (config-manager.js
header icon, config-sidebar.js category icons).
Also fixed the backup-engine logos: scripts/backup/engines/{restic,kopia,borg}
.json pointed 'logo' at /icons/config/backup.svg — a path that 404'd on two
counts (missing the components/admin/config prefix AND the now-removed config/
nesting), so the engine-details modal logo silently hid. Repointed to the real
served path /components/admin/config/icons/backup.svg.
(Left the meaningful icon groupings alone — admin/system/icons/{cpu,os} and
apps/core/icons/vpn are vendor/OS/provider logo sets, not redundant nesting.
The backup engines borrowing an admin-config icon is a minor smell; a dedicated
backup-engine icon could replace it later if wanted.)
Signed-off-by: librelad <librelad@digitalangels.vip>
Brings core/ in line with components/ — each subsystem now sorts its files into
js/ css/ html/ subfolders (and the nested auth/ + controls/ groups keep theirs):
core/topbar/{js/{topbar,mobile-menu}.js, css/{topbar,sidebar}.css, html/topbar.html}
core/theme/{js/theme-registry.js, css/{tokens,themes,base,aurora-background}.css}
core/forms/{css/{forms,config}.css, controls/js/{custom-number,custom-select}.js}
core/boot/{js/{system-loader,system-orchestrator}.js, auth/{js/auth-manager.js,css/login.css}}
core/{config,tasks,kernel}/js/… core/overlays/{js,css}/… core/setup/{js,css}/…
core/{app-meta,backup-card,data-loader,dom,live,notifications,ui-mode,ui-state}/js/…
core/{loading,update-notifier}/{js,css}/…
50 files relocated (pure git mv). All path literals rewritten from a generated
old→new map across index.html, system-loader.js bundles, topbar.js's internal
fetch (/core/topbar/html/topbar.html), and the three backup-app-card loaders. No
OLD path contained a js/css/html segment, so no double-prefixing was possible.
core/icons/ left as-is (shared asset tree). All 50 /core asset refs verified to
resolve; full node --check sweep clean.
Signed-off-by: librelad <librelad@digitalangels.vip>
The frontend modularization moved icons to frontend/core/icons/ and updated the
frontend JS, but the host-side generators were never updated — they wrote the
apps.json/categories 'icon' field as /icons/apps/<app>.svg and /icons/categories/
<cat>.svg, and webui_app_icons.sh / webui_config.sh synced icon files into the
non-existent frontend/icons/apps. Those served paths 404 (text/html catch-all),
so every app card fell back to default.svg (the generic box) instead of its real
logo.
Repointed to /core/icons/... (where the SVGs actually live and serve as
image/svg+xml):
- webui_config.sh: icon dir + emitted apps.json icon path
- webui_app_icons.sh: icon sync dir + comment
- webui_container_setup.sh: comment
- webui_create_app_categories.sh: 11 category icon paths
Source fix only — the live apps.json refreshes on the next host-side regen
(lpRegen). NOT touched: scripts/backup/engines/*.json '/icons/config/backup.svg'
(that SVG lives at the oddly-nested components/admin/config/icons/config/ and
serves at neither path — needs a placement decision, flagged separately).
Signed-off-by: librelad <librelad@digitalangels.vip>
The generic core/css/{base,components,screens} buckets are gone; every shared
stylesheet now lives beside the subsystem that owns it:
base/tokens.css, base/themes.css, components/aurora-background.css -> core/theme/
base/style.css -> core/theme/base.css (carve deferred)
components/modal.css -> core/overlays/
components/topbar.css, components/sidebar.css -> core/topbar/
components/forms.css, components/config.css -> core/forms/ (config.css under forms)
components/update-notifier.css -> core/update-notifier/
screens/login.css -> core/boot/auth/
screens/loading-screen.css -> core/loading/
screens/setup-wizard.css -> core/setup/
href-only rewrites in index.html; the <link> line ORDER is unchanged, so the
cascade is preserved (no @import anywhere). All 13 /core css hrefs verified to
resolve. (The JS for overlays/topbar/forms/update-notifier/loading/setup/auth
co-locates in the next phase.)
Signed-off-by: librelad <librelad@digitalangels.vip>
The Phase-2 rename put DataLoader in core/data/, but update.sh's deploy rsync
uses --exclude 'data/' (to protect the runtime frontend/data/ dir the backend
serves auth-gated under /data). rsync's pattern matches a 'data' dir ANYWHERE in
the tree, so core/data/ was silently excluded from the served copy — the file
404'd and the dashboard showed 0 apps / Loading… while every sibling subsystem
deployed fine. Renamed the folder to core/data-loader/ (segment 'data-loader' ≠
'data') so it ships. No code/class change.
Signed-off-by: librelad <librelad@digitalangels.vip>
The generic core/lib/ wrapper (and its task/config/util sub-buckets) is gone.
Each child is now a named core subsystem describing what it IS:
core/lib/task/ -> core/tasks/ (task kernel: bus, refresh,
manager, router, actions,
commands, parameter-preserve)
core/lib/config/ -> core/config/ (config-shared.js→field-factory.js,
config-options.js→options.js;
options-before-factory order kept)
core/lib/util/system-live -> core/live/live-system.js
core/lib/util/lp-ui -> core/ui-mode/lp-ui.js (stays FIRST eager — no FOUC)
core/lib/util/data-loader -> core/data/data-loader.js
core/lib/util/dom-helpers -> core/dom/dom-helpers.js
core/lib/util/ui-helpers -> core/app-meta/app-helpers.js (getAppIcon survivor)
core/lib/util/dismissible -> core/ui-state/dismissible.js (generic+eager, stays
core — NOT a backup widget)
core/boot/theme-registry -> core/theme/theme-registry.js (theming, not bootstrap)
Path-only moves (git mv) + literal rewrites in index.html, system-loader.js
(config/task/apps bundles) and apps-manager ensureTaskScripts. Class/global
names unchanged (ConfigShared/ConfigOptions/LiveSystem/getAppIcon) so consumers
are untouched. All 16 referenced paths verified to resolve; full node --check
sweep clean.
Signed-off-by: librelad <librelad@digitalangels.vip>
Verified-dead removals (zero consumers, confirmed by adversarial dependency
audit):
- core/lib/util/router.js — legacy class Router superseded by kernel/spa.js;
self-instantiated, never exposed, and added a SECOND competing popstate
listener. Dropped the file + its eager index.html tag.
- core/lib/task/task-global-functions.js — wired window.installApp/uninstallApp/
etc. that nothing calls (real calls go through class methods / the task
router). Dropped the file + its task-system scripts[] entry + the
setupTaskGlobalFunctions() block in system-loader.js.
- TopbarComponent.createNavigationHighlighting + clearAllNavigationHighlighting —
dead statics; window.topbarNavigationHighlighting was never set.
- ui-helpers.js: getAppStatus/formatAppName/getAppShortName (dead), the stale
setupMobileMenu/closeMobileMenu (superseded by core/ui/mobile-menu.js's
#mobile-drawer impl), setupActiveNavigation + the safe* helpers (verbatim dups
of dom-helpers). Only getAppIcon remains. dom-helpers loses dead
setupActiveNavigation + waitForElement; it is now the sole safe* source.
Bug fixes surfaced during the audit:
- system-orchestrator.js called this._wireLogout() which is defined nowhere —
threw on the 'Continue Anyway' boot path. Removed the dangling call (logout is
wired in topbar.setupLogout()).
- active-nav highlighting never updated on SPA navigation (it depended on the
never-set global). spa.js now calls the live window.topbar?.setActiveNav?.()
after each route handler.
No structural moves yet; full node --check sweep clean.
Signed-off-by: librelad <librelad@digitalangels.vip>
core/css held 13 flat stylesheets. Grouped by role, matching the JS sub-system
split:
core/css/base/ foundation (3): tokens, style, themes
core/css/components/ chrome + widgets (7): topbar, sidebar, modal, forms,
config, update-notifier, aurora-background
core/css/screens/ full-page screens (3): login, loading-screen,
setup-wizard
Path-only move (git mv) + href rewrite in index.html. No @import anywhere and
the <link> line order is unchanged, so the cascade is preserved; all 13 paths
verified to resolve.
Signed-off-by: librelad <librelad@digitalangels.vip>
backup/ held 15 loose .js plus css/html at the component root. Split into
per-tab sub-systems mirroring the admin/apps layout:
backup/core/{js,css,html}/ schema, base BackupPage, fetch-client,
cron-schedule, backup.css, backup-content.html
backup/dashboard/js/ backup-dashboard
backup/snapshots/js/ backup-snapshots, backup-snapshot-actions
backup/locations/js/ backup-locations, backup-location-fields,
backup-location-modal, backup-ssh-key
backup/migrate/js/ backup-migrate
backup/configuration/js/ backup-configuration, backup-retention-presets,
backup-engine-details
Updated the scripts[] array + loadFragment in index.js, the backup.css href in
index.html, and the handleBackup() fallback paths in spa.js. No behaviour
change — files moved verbatim (prototype-augment clusters), all 17 referenced
paths verified to resolve, node --check clean.
Signed-off-by: librelad <librelad@digitalangels.vip>
admin had sub-system folders (config/overview/system/ssh/peers) AND a bare
root css/ (admin.css) — inconsistent. A component's shared base belongs in a
core/ sub-folder once it has sub-systems (matching apps/core/), so admin.css
moves to admin/core/css/admin.css. Audit confirms all components are now
consistent: sub-system components (apps, admin) use core/ + sub-systems;
single-page components (backup/dashboard/tasks/updater) stay flat js/css/html
(no sub-systems, so no core/ needed).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
config-redirect/, peers/, ssh/ were redirect-only shim components (just a
feature.json each) whose handlers re-entered navigate() and were silently
no-op'd by the isLoading guard — i.e. the legacy /config /peers /ssh URLs were
broken, and nothing in the UI links to them. Replace all three with a single
_legacyRedirect() at the top of navigate() (before the guard, so it actually
works) that rewrites them to the canonical /admin/* path. Removed the 3
folders, their manifest entries, the 3 dead spa.js handlers, and their
setupRoutes() lines. components/ is now exactly the 6 real pages; the real
SSH/Peers pages live (as before) under admin/ssh and admin/peers.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
app-detail was a separate sibling component but is really the apps feature's
detail view (shares the apps controllers + apps-unified-layout). Merge it in:
the apps feature now owns /apps*, /app, /app* and its mount() dispatches grid
vs detail by path (checks /apps first for wildcard precedence). Removed the
components/app-detail/ folder + its manifest entry. (The 1-level feature scan
means a feature must be a direct components/<id>/ child — folding the routes
into apps is the correct way to express the relationship.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Brace-aware split of apps-manager.js (4154->2015 base) into apps-grid,
config-form, port-manager-integration, gluetun-vpn, config-dirty-guard,
install-console, service-buttons-sidebar — augmenting AppsManager.prototype.
Removed the dead duplicate initializeSimpleTabs (kept the live later def).
Used a self-checking extractor (node --check per cluster, auto-revert on
failure): install-dispatch contains a regex literal (/^\d{16}$/) that trips
brace-bounding, so its methods were safely LEFT in the base rather than risk a
bad split. ServiceButtons class + expandServiceLinks + bootstrap also remain
inline. Verified: all 117 methods preserved (none lost), all files
node --check clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Faithful brace-aware split of tasks-manager.js (2664->491 line base) into
list-render, data-load, log-stream, row-expand, actions, modals, logs-modal,
format — augmenting TasksManager.prototype. First removed 4 provably-dead
duplicate defs (the earlier init/setupAutoRefresh/startLogStreaming/loadTaskLogs
that the later defs already overrode — behavior-preserving). Methods relocated
verbatim via a brace-aware Node extractor (handles strings/templates/comments/
regex, fixing the line-heuristic over-capture). Verified: all 60 (deduped)
methods present exactly once, no dups, all 9 files node --check clean. Wired
after the base in the task-system loader (async=false-ordered).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Faithful prototype-augment split of backup-page.js (2353->753 line base) into
fetch-client, dashboard, snapshots, locations, location-fields, ssh-key,
retention-presets, configuration, engine-details, location-modal,
snapshot-actions, migrate (+ the earlier cron-schedule). Methods relocated
verbatim (mechanical sed/awk extraction, no logic change); all augment
BackupPage.prototype and load after the base via the ordered kernel loader.
Verified: all 99 original methods present exactly once across base+clusters,
no duplicates, all 14 files node --check clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
First god-file decomposition slice: lift the standalone cron-next utility
(nextCronFireTime/_cronFieldSet/formatRelativeFuture/formatScheduleClock) out
of backup-page.js into backup-cron-schedule.js, augmenting BackupPage.prototype.
Extracted verbatim via sed (no logic change); loaded after backup-page.js in
the feature's ordered scripts array. backup-page.js 2470 -> ~2353 lines.
Proves the faithful prototype-split pattern on the verify-confirmed-safe loader.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
De-clutter each component into sub-system folders (apps: core/ port-manager/
services/ tools/ routing/; admin: config/ overview/ system/ ssh/ peers/) with
the standard js/ css/ html/ icons/ layout inside; single-page components
(backup/dashboard/tasks/updater) get js/ css/ html/. Single-feature icon sets
moved into their sub-system (vpn -> apps/core/icons, config/cpu/os ->
admin/{config,system}/icons); shared app + category icons stay in core/icons.
feature.json + index.js stay at each component root (the scanned descriptor +
entry). Every controller/CSS/fragment/icon path reference rewritten; verified
no stale refs, all JS valid.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
system-loader's loadScript appended scripts without async=false, so a
component's scripts[] executed in download-finish order, not array order.
That's a latent nondeterminism today (duplicate-method 'last wins' depended
on it) and a hard blocker for splitting a class across files (a base file
must run before prototype-augment files). Forcing async=false makes the
boot loader honour array order, matching the kernel's ctx.loadScripts.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Restructure the updater page to mirror Backups exactly:
- updater-content.html: shared .sidebar/.category (with .category-icon SVGs)
outside the page card, .config-section + .page-header + a padded
.updater-page-body — fixes the missing content padding.
- updater.css: layout copied from .backup-* (flex, padding-bottom:48px,
body padding:22px); dropped the custom sidebar/header styles (now using the
shared chrome); kept the updater-specific content widgets.
- updater-page.js: delegate clicks on .updater-layout (sidebar is now a
sibling of the card) so clicking a sidebar entry opens that tab; fill the
shared page-header icon slot.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- scripts/webui/data/generators/updater/webui_updater_scan.sh (webuiUpdaterScan):
writes frontend/data/updater/generated/{updates,cves,history}.json from the
installed-apps DB (current image per app from compose). Available-version +
CVE-scanner are clearly-marked pluggable hooks; always emits valid JSON.
- scripts/cli/commands/updater/{cli_updater_commands.sh,cli_updater_header.sh}:
auto-dispatched as 'libreportal updater <sub>' (check/apply/apply-all/rollback).
apply does disaster-recovery FIRST — snapshots the app via the backup engine,
then pulls + recreates (real dockerComposeUp/compose-pull helpers), records
history, and auto-rolls-back on failure. Standard LIBREPORTAL_TASK_EXEC
enqueue/exec split so WebUI + CLI share locking + audit trail.
New .sh files: the array/function-manifest regen self-heals on deploy; the
check path also sources its generator on demand to cover the gap.
NOTE: host-side bash — written to the repo's conventions but not runnable in
this env; this is the surface to test (the WebUI feature is lp-shot-verified).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add updater_check/apply/apply_all/rollback cases to task-router routeAction
and the matching task-actions methods, each running the locked-down
'libreportal updater …' CLI as a task (apply/rollback snapshot-before-update
host-side for disaster recovery). Verified: /updater + all tabs render, the
new Updates nav button shows globally, and the feature auto-discovers via
/api/features/list.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
New self-contained feature in features/updater/ (mirrors the backup feature):
- index.js + feature.json (auto-discovered; routes /updater + sub-tabs).
- updater-page.js: 5 tabs — Overview (update/CVE/recovery counts), Updates
(per-app current->available + Update/Update-all), Security (CVEs by severity,
links to NVD), Recovery (per-app rollback points; snapshot-before-update),
History. Reads /data/updater/generated/{updates,cves,history}.json; falls
back to the installed-apps list so it's useful before the first scan. All
actions route through services.tasks (updater_check/apply/apply_all/rollback)
— no new mutating API.
- updater.css (self-contained, teal --page-updater hue) + updater-content.html.
- New topbar 'Updates' nav button (nav-updater) + active-highlighting in
topbar.js + spa.js. Kernel: setupRoutesFromManifest now allows module-only
features (no legacy handler) — this is the first such feature.
Backend generator + 'libreportal updater' task land next.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- features/admin/: the 10 admin-owned config controllers, the 5 admin pages
(overview/system/charts/metric/storage), ssh-page.js, peers-page.js, plus
admin.css/ip-whitelist.css/ssh.css (eager). config-manager.js kept last in
the load order (it news the sub-managers).
- shared/js/: config-shared.js + config-options.js (ConfigShared/ConfigOptions
globals consumed cross-feature by backup/apps/tasks).
- shared/css/: forms.css + config.css (generic form + config-form primitives
borrowed by apps/backup/admin).
- Updated all path strings in system-loader.js (config component) and
config-manager.js (lazyLoad of admin/ssh/peers controllers); index.html CSS
hrefs. No /js/components/{config,admin,ssh,peers}/ refs remain.
js/components/ now holds only shared UI (topbar, notifications, eo-modal,
update-notifier, mobile-menu, confirmation-dialog).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- features/backup/: backup-schema.js, backup-page.js, backup.css (eager).
- shared/js/backup-app-card.js: cross-feature (used by backup AND the
app-detail Backups tab), so it goes to shared/, not buried in backup.
- Updated paths: feature scripts array, spa.js handleBackup fallback,
system-loader apps-manager component (app-card), index.html backup.css href.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move the 6 app controllers (apps-manager, app-tabbed-manager, port-manager,
routing-manager, services-manager, tools-manager) and their 7 stylesheets
(apps, apps-layout, tools, services, service-buttons, routing, port-manager)
into features/apps/. JS paths only existed in system-loader.js (the
apps-manager + app-tabbed-manager components); CSS hrefs in index.html (kept
eager). The cross-feature task-manager (/shared/task/) + backup-app-card
entries in the apps-manager component are untouched here. Globals unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- features/tasks/: tasks-manager.js (the /tasks page controller) + tasks.css.
- shared/task/: the 6 cross-cutting task-kernel files (event-bus, commands,
actions, router, global-functions, manager) + task-refresh-coordinator.js —
used by tasks AND apps/app-detail/backup, so they go to shared/, not a
feature. task-parameter-preserve.js stays at js/ (shared root).
- Updated all path strings: system-loader.js task-system + apps-manager
components, apps-manager loadTaskSystem(), index.html (refresh-coordinator +
tasks.css). Globals (taskEventBus/tasksManager/TaskManager/...) unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move dashboard.js + dashboard.css into the feature folder (kept eager-linked
from the new paths in index.html — dashboard.js stays eager because
system-loader calls its bare globals at boot). No reference sites outside
index.html (the feature's index.js uses globals, not the controller path).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Both are orphans with zero inbound references (verified by grep across the
whole frontend): app-manager.js only self-assigns window.appManager and is
in no loader/scripts array; config-router.js only self-defines ConfigRouter
and is referenced nowhere (config-manager owns rendering directly). First,
safest step of the feature reorganization.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- kernel/lifecycle.js: ctx.loadScripts now loads in array order (sequential),
so a feature can list a base file before files that augment it. Strictly
safer than the previous parallel load.
- Extract the module-level schema/retention data (the BACKUP_* maps + the
retention-preset detector, 83 lines) out of backup-page.js into a new
backup-schema.js, loaded first. Verbatim move — no logic change. First slice
of the backup decomposition (god-file: 2553 -> 2470 lines).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add an Implementation status section: kernel + manifest routing, all pages
migrated to feature modules, folder auto-discovery via /api/features/list
(supersedes the §3 shell-regen generator), DI seam, and the shared token
layer are shipped + verified. God-file decomposition and the base.css/CSS
split remain as the large internal refactors still to do.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Introduces kernel/services.js (window.LP.services): an additive, typed,
LAZY view onto the existing cross-cutting singletons — tasks{bus,refresh,
route}, live, auth, data, notify, theme, modal, router. It constructs
nothing (pure getters onto the live globals), so there's no double-init and
the globals stay authoritative. MountContext now injects it as ctx.services.
Slot names/globals were verified against the real code (workflow map): the
design doc's §4 list was wrong in several places — no window.taskManager
(client slot dropped), tasks.route lives on tasksManager.router, auth has no
status(), DataLoader isn't a window prop (lexical fallback), modal/router are
split surfaces (grouped/bound objects).
Migrated the 4 cross-cutting refs in the feature modules onto ctx.services
(admin: router.adminCategoryFromPath + tasks.refresh; backup: tasks.refresh;
app-detail: router.appPath). Page-owned controllers stay feature-globals.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Themes are already modular via folder discovery (GET /api/themes/list scans
themes/<name>/). This brings the SAME model to pages:
- backend/routes/features.js: public GET /api/features/list scans
frontend/features/<id>/feature.json and returns the page manifest. The
Node process reads its own bind-mounted /app/frontend — no runFileOp /
regen / source-array plumbing needed (sidesteps the shell-generator gotchas).
- features/<id>/feature.json: each page now self-describes (id, routes,
module, handler, navId, nav, order). 6 real features + 3 redirect-only
(config/peers/ssh) so behaviour is preserved exactly.
- kernel loadManifest() prefers /api/features/list, falls back to the static
features/manifest.dev.json when the endpoint isn't up yet.
Result: dropping a features/<id>/ folder registers a page; deleting it
removes it — zero central edits, exactly like dropping a theme folder.
(Backend route needs a Node restart to activate; the static-manifest
fallback keeps everything working until then.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The kernel now loads each feature's self-registering index.js from the
manifest (new 'module' field) before building routes, so index.html no
longer hardcodes a per-feature <script> list — one of the four registries
the modularization targets is now gone. Adding a page = drop a
features/<id>/ folder + a manifest entry; no index.html edit.
loadScript is idempotent and non-fatal: a module that 404s or fails to
register leaves its route on the legacy handler. Manifest-load failure
still falls back to the built-in route table.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
AppTabbedManager.initialize() re-runs on every /app navigation (its
'initialized' flag is never set true), and setupURLMonitoring() /
setupTaskEventListeners() add window listeners (popstate, taskCreated/
Completed/Updated) unguarded — so each app-detail visit stacked another
listener set. Bind them once via a _listenersWired flag (mirrors the
existing _watchdogStarted guard). Pre-existing leak surfaced by the
feature-migration review.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
features/admin/index.js owns all /admin* sub-routes (overview, config/<cat>,
system[/sub], tools/ssh-access, tools/peers). mount() parses the category and
delegates to the system-loader configManager singleton's renderConfig();
unmount() stops AdminSystem's live SSE sub + 30s interval, drops the
admin-overview task-refresh registration, and nulls the per-visit
sub-controllers (adminOverview/adminSystem/sshPage/peersPage) while leaving
configManager intact. Legacy /config /ssh /peers redirect handlers unchanged.
With this, every WebUI page now routes through the feature-module kernel;
the legacy handleX() methods remain only as fallbacks.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- features/apps/index.js (/apps*) and features/app-detail/index.js (/app*)
as two features; /apps* registered first so wildcard precedence holds.
Both drive the system-loader-pre-initialized singletons (appsManager /
appTabbedManager) via .initialize(), mirroring the legacy handlers exactly
(incl. app-detail's legacy ?app=/?=name parsing + ?tab=/?config= rewrite).
- kernel/lifecycle.js: ctx.nav(path, addToHistory) so app-detail's empty-name
redirect matches navigate('/apps', false) exactly.
unmount is a no-op for both (shared singletons); the app-tabbed-manager
listener-rebind leak is pre-existing and handled in a follow-up.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Two more pages on the feature-module contract (specs produced + adversarially
verified by workflow):
- features/dashboard/index.js: landing page; mount() folds in the data-reload
that used to be a navigate() special-case (now deleted from spa.js, so it
fires exactly once). No controller class — uses the eager dashboard.js globals.
- features/tasks/index.js: re-inits the system-loader tasksManager singleton;
unmount() clears the 30s auto-refresh interval + open log streams WITHOUT
stopping the shared SSE bus or nulling the singleton.
Verifier fixes applied: deleted the duplicate dashboard reload in navigate();
dropped a dead detachDashboardLive() call; fixed an invalid try{}while(false)
in the tasks unmount.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Introduces the kernel lifecycle and migrates the first real page to the
feature-module contract:
- kernel/lifecycle.js: MountContext (loadScripts/loadFragment/setContent
+ an AbortController/unsub teardown ledger so mounts can't leak
listeners or live streams).
- features/backup/index.js: Backup Center as a self-contained module
(LP.features.register with mount/unmount); heavy backup-page.js stays
lazy-loaded on first mount.
- spa.js: routes whose feature has a registered mount() are driven
through the kernel; everything else still uses its legacy handleX().
navigate() unmounts the current feature first. Both fall back to the
legacy handler if a module is missing or mount throws.
Strangler step: /backup now flows manifest -> registry -> mount/unmount.
The other pages are untouched. handleBackup remains as the fallback.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
First slice of the per-module CSS strategy: introduce
shared/css/tokens.css as the always-present, theme-agnostic base token
layer, loaded before all other stylesheets.
- Defines --font-mono (ui-monospace stack). Several feature sheets used
var(--font-mono, monospace) with no definition, falling back to bare
monospace; this unifies them with the richer stack used elsewhere.
- Hoists the --page-* identity hues out of css/admin.css so they belong
to the base layer rather than the admin stylesheet. Values unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
LibrePortalSPAClean now builds its route table from window.LP.features
(features/manifest.dev.json) instead of the hardcoded setupRoutes() Map.
Manifest order is preserved so findRouteHandler()'s wildcard precedence
(/apps* before /app*) is unchanged. All-or-nothing fallback to the
built-in table if the manifest is missing/empty or names an unknown
handler, so routing is never left half-wired.
Rendering is unchanged — handlers still do the work; only the routing
source moved.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds the foundation of the feature-module architecture
(docs/frontend-modularization.md) as inert, additive code:
- kernel/feature-registry.js: window.LP.features — runtime register(),
manifest loader, route-table + nav builders.
- features/manifest.dev.json: hand-committed manifest mirroring spa.js
setupRoutes() exactly (route -> handler + navId).
- index.html loads the kernel before spa.js.
Zero behaviour change: nothing consults the registry yet. Phase 0b flips
routing to be registry-driven with the spa.js Map as fallback.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Synthesized architecture for turning the no-build vanilla-JS WebUI into a
scan-and-manifest feature system mirroring the backend container scan:
self-contained features/<id>/ folders, a navigation kernel, uniform
mount/unmount lifecycle, DI service context replacing ~80 window globals,
per-feature CSS, god-file decomposition, and a strangler migration roadmap.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The empty-folder reaper only ever fired on folders with no real data
(empty, or only a regenerable .config and/or migrate.txt marker), yet
prompted 'THIS WILL WIPE ALL DATA' before each removal — a question
about data that didn't exist. Collapse the four duplicated branches into
one reason-string path, clean these leftovers automatically, and fix the
stale $app_name used in the DB-delete (it deleted the wrong row when
looping over $folder_name).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Directive for agents working on the repo: after a user-visible WebUI change,
screenshot the route with the lp-shot helper and review the PNG before marking
done, rather than relying on syntax checks alone. Box-specific setup (port
discovery, auth, install) is kept out of the repo.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Remove the "Paste a public key…" line (the section description already
explains it) and stretch the Login / Add-a-key cards to equal height
(.ssh-cols align-items: start -> stretch).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The `libreportal verify` task showed its raw command and no icon. Add its
formatCommandForUser pattern ("LibrePortal - Verify System"), a 🛡️ type icon,
a formatActionTitle entry, and include it in isLibrePortalSystemTask so it shows
the LibrePortal logo like other system tasks.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The System header used a cpu icon while the sidebar (and the Dashboard's System
card) use the activity-pulse icon. Swap the header to the same activity pulse so
System reads consistently everywhere.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Wrap both the OS and CPU logos in a 36x36 rounded icon tile with a subtle
background — same treatment as the backup app-list tiles — at a uniform
size, with the logo centred inside.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Renames the Admin landing to "Dashboard" in both the page title and the sidebar
label, and adds a leading header icon (the grid icon, matching the Backup
dashboard) via the shared .page-header-icon-slot. System gets a cpu icon in its
header too. The slot styles come from the globally-loaded backup.css, so no new
CSS is needed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Both sections are light on content, so a two-column grid uses the space better
than full-width stacking. Wrapped them in .ssh-cols (1fr 1fr, stacks under
640px); Authorized keys stays full width below.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The Intel/AMD logos are wide wordmarks that sat in a square 24x24 box, so
the actual mark rendered only a few px tall and looked invisible. Crop each
viewBox to the wordmark and size the CPU icon by height with auto width so
it shows at a legible scale. Bump the OS distro icon from 18 to 22px.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Reverts the solid fills (too heavy) back to the translucent hue tint but keeps
white text — white on a tint-over-dark-card reads cleanly on nebula, which the
old coloured text didn't. Hues restored to the brighter originals. Verify now
gets its own green token (--page-verify) per the usual verify=green convention,
and Backups takes the blue.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the hand-drawn distro marks with the official artwork (Simple
Icons, brand-coloured, bundled locally — no external calls) for
Debian/Ubuntu/Fedora/Arch + a generic Linux fallback. Add Intel/AMD logos
under /icons/cpu/ and show the vendor logo beside the CPU, with the model
string stripped of trademark noise (Intel(R) Core(TM), ®/™, "CPU") since
the logo conveys the vendor — e.g. "Core i5-8250U @ 1.60GHz".
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The translucent tinted buttons washed out against nebula's glassy aurora
background. Switch .admin-action-btn to a solid fill with white text (the bg can
no longer bleed through) and deepen the --page-* hues enough for white-text
contrast. Hover uses brightness() so it's theme-agnostic.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Bundle a small set of distro marks (Debian/Ubuntu/Fedora/Arch) plus a
generic Linux/Tux fallback under /icons/os/, and show the icon next to the
OS value, keyed off the cleaned distro name. Locally bundled — no external
calls — and unknown distros fall back to the generic glyph.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Introduces per-area identity hues as reusable tokens (--page-updates/backups/
ssh/system + -rgb companions) and a generic .admin-action-btn that takes its
colour from --page set on the card. The Overview buttons now read in their
area's hue — Updates blue, Backups emerald, SSH violet, System amber — with the
icon following via currentColor; "Update now" is the filled (primary) variant.
The tokens are the foundation to extend each area's identity (page headers,
accents) going forward, not just these buttons.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The admin ops/health board now lives at /admin/dashboard (adminPath
('overview') emits it; the topbar Admin link points there). Bare /admin
still resolves to the same board — no redirect, both paths render it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Update/Verify (refresh, shield-check), Manage backups (archive), Manage SSH
(key), View system stats (activity). Icons inherit the button's text colour —
no per-button colour, so the footer stays calm; the icon plus the card's status
dot do the distinguishing.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
System is live stats, not configuration, so it shouldn't live under
/admin/config. adminPath('system') now emits /admin/system; the path
parser locates 'system' positionally; all nav targets, breadcrumbs and the
dashboard disk-card link point at /admin/system{,/storage,/metric/<k>}.
adminCategoryFromPath already resolves /admin/<x> to that category, so
ConfigManager still mounts AdminSystem unchanged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The "View/Open storage breakdown" links in the System storage summary were
plain text links; restyle them as the same accent pill CTA (icon + label)
used on the Disk metric page.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
"Live · binary ring backed" meant nothing to a user; show a per-metric
sentence describing what the chart is (e.g. "Root-filesystem usage over
time.") instead.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move the Disk page's "View storage breakdown" button beneath the time-range
picker (actions stack in a column, right-aligned) instead of beside it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The plain text link read as unstyled; give it the same pill shape as the
Reclaim Space button (icon + label) in the accent colour, so it looks
intentional without borrowing Reclaim's caution-orange.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add a "View storage breakdown →" button on the Disk metric page (disk
only) that opens /admin/config/system/storage — keeps the trend page and
the breakdown page focused but a click apart, instead of merging them.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The "LibrePortal X GB" sublabel under the disk percentage read as clutter;
remove it. The % and the coloured LibrePortal portion of the ring stay.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The System-page storage ring is an overview, so showing individual app
(compose-project) names was wrong — it now shows a single "Applications"
slice totalling all apps, alongside Images and Build cache. The per-app
breakdown stays on the full Storage page.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the disk gauge's concentric inner ring with a single ring whose
leading portion is coloured to mark the LibrePortal share of the disk
(one ring: total disk used overall, LibrePortal highlighted within it).
On the full-screen Disk metric page, add a flat reference line marking
LibrePortal's current share alongside the disk-usage trend. The gauge
gains a `segment` option; the chart line is a "now" value (no historical
LP series yet), so it's flat across the range.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- Disk gauge (System page) gains an inner ring for the LibrePortal slice of
the disk, so it shows total disk used AND how much of that is us.
- System-page storage summary now shows the full LibrePortal breakdown
(apps + images + build cache), not just the Docker engine categories.
- Fix the chart colours: Images use the Reclaim orange, build cache the
deeper red (warm = reclaimable overhead), apps a cool palette.
- Images list: dark container, Clear All / Select all moved into the section
head (count text dropped), each image shows its app's icon.
- Storage by app restyled to the same Tasks-style list (app icons,
expandable folders), minus the selection controls.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Post-task UI refresh was scattered: every page added its own taskCompleted
listener and hard-coded which actions it cared about, so it was easy to add a
task and forget the refresh (stale UI), with no single place to see the wiring.
Adds TaskRefreshCoordinator (window.taskRefresh): one listener, with dedupe
(the SSE bus + synthetic fallbacks double-fire) and opt-in debounce (bursts
coalesce; per-task handlers run every time). Components now register a refresh
entry; window.taskRefresh.table() is the introspectable "what reloads when" map.
Migrated onto it: apps (install/uninstall/tool/config_update lifecycle +
restore/update/rebuild state), backups (backup/restore/delete), the update
badge, and the admin overview integrity badge. Gaps closed: restore/update/
rebuild now repaint app+service data. (start/stop/restart intentionally omitted —
no live status surface to refresh today; revisit if a running/stopped badge is
added. Storage reclaim/image-rm keep their own in-page refresh.)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
config_update re-deploys apps (ports, subdomains, Open URLs, routing) but the
WebUI only unlocked the nav on completion — leaving stale URLs/routing until a
manual reload. apps-manager now refreshes apps + services and repaints the
current view when a config_update task completes, via a reusable
refreshAppsAndView() helper (also the basis for the upcoming task-refresh
registry).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the frontpage liquid-fill disk circle with a real donut split into
Apps · Docker · Other · Free, keeping disk % used in the centre. Apps come
from the per-app generator (root-device bytes), Docker from system df; both
are clamped within "used" so skew can't overflow. Live disk ticks redraw it,
and the card now clicks through to the full Storage breakdown.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The trends grid lost its disk chart; add it back as a 6th card plotting
root-filesystem % over time, alongside CPU/Memory/Network/Load/Swap. The
'disk' history series and metric-detail entry already existed, so the
expand-to-detail flow works unchanged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replaces the read-only "Largest images" top-10 table with a Tasks-style list of
ALL Docker images, with select-one / select-multiple / clear-all removal that
mirrors the Tasks page UX (row checkboxes, master select-all, a button that
morphs Clear All ↔ Delete Selected (N), an eo confirm modal).
Deletion routes through the task system, NOT a new web API: a new
`libreportal system image rm [--force] <ids>` CLI subcommand (validates each
ref, loops runFileOp docker image rm, reports a tally) is invoked via the
system_image_rm task action — same pattern as Reclaim. The web backend change
is read-only (uncap the existing /storage image list). In-use images are
skipped by default with an opt-in "force-remove" toggle (warned). The page
stays put, toasts, and refreshes on the task's completion event.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Put everything back in a single donut instead of an app-only one: a slice
per app plus Docker's images and build cache, with one legend listing them
all. Colours run continuously so app and Docker slices stay distinct, and
the Docker cards below reuse the same colours. The per-folder app
drill-down table stays. This is the "all the data, with app usage added"
view that was asked for.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The Storage page now leads with on-disk usage by app: the headline donut
is split by app (each a coloured slice, total app data in the centre),
and the "Storage by app" table is its legend — swatch + bar colours match
the slices, rows expand to the per-folder breakdown. Docker's engine
figures (images + build cache) drop to a secondary section below. This is
the integration that was asked for; the donut is your data, not Docker's
overhead.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Like named volumes, a container's writable layer is a near-zero scratch
number for LibrePortal (app data lives in bind mounts, shown per-app), so
sitting it next to per-app storage just confused things. Remove the
"Containers" slice/card and its backend summation, and reframe the Docker
breakdown as "Docker engine" overhead (images + build cache) — clearly
separate from your app data.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Extend the app-storage generator to record every bind mount's size and
in-container path, grouped into a per-app folder list. The "Storage by
app" rows are now expandable: click an app to see where its space goes
(e.g. /var/lib/mysql vs /data), with external-drive folders flagged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
LibrePortal apps keep data in bind mounts, so Docker named-volume
accounting is always ~empty and just reads as a confusing "0 B". Now that
per-app on-disk usage covers the real "what's filling my disk" question,
remove volumes end to end: the donut slice, category card, "Largest
volumes" table and the System-page count, plus the backend's volume
summation and top_volumes payload. Reclaim copy no longer references
volumes (it reassures about app data instead).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The old comment implied the key was still a REPLACE_ME placeholder. Reword to
describe current behaviour (signature required for release installs) plus how to
rotate the key.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replaces the REPLACE_ME placeholder public key in libreportal.pub and install.sh
with the real LibrePortal release-signing public key (id BC92526B3ECA7F41). The
secret half is held offline by the maintainer.
This activates the signature-required path everywhere it was wired but inert:
install.sh now REQUIRES a valid tarball signature on release installs, the
updater (fetch.sh) requires it on update, and the integrity check (verify.sh)
will report a real "Verified" state once a signed release is installed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Docker only tracks where an app's data lives (its bind mounts), not how
big a bind-mounted host dir is — so named-volume accounting reads ~0 for
LibrePortal, whose app data lives in bind mounts. Add a generator that
reads each app's mount map from `docker inspect` and `du`s the directories
(via runFileOp, so it runs as the data-owning user and isn't blocked by
rootless UID mapping). `du -x` keeps each measurement on its own
filesystem, so data on a separate disk is reported as a distinct
"external" total. The generator self-throttles to ~10 min since du is
heavier than the per-minute metrics. Surfaced as a "Storage by app"
section on the Storage page.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds per-file integrity attestation on top of the existing signed-tarball
release flow. make_release now generates a SHA256SUMS manifest over the shipped
tree and (when a key is configured) signs it, riding both inside the release
tarball so they land in the install tree with no extra download.
lpVerifyInstall (scripts/source/verify.sh) re-hashes the install tree against
that manifest and verifies the manifest's minisign signature against the
root-owned footprint pubkey, yielding states: verified / modified / tampered /
unsigned / unverifiable / development. webuiSystemVerify writes verify_status.json
(throttled daily, force on demand, also after each update apply), surfaced as an
Integrity line + "Verify now" button on the Admin → Overview Updates card and a
row in the update details panel. `libreportal verify` exposes the same check on
the CLI.
Honest framing: this is a self-check (run by the software it verifies), so red
fires only for genuine modified/tampered states; the badge tooltip points to
out-of-band `minisign -Vm` for an independent guarantee.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Drop the OS row and compact Memory/Uptime so the System card matches
the density of its sibling cards. Full detail stays on the deep system
stats page.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Three fixes from testing the storage page:
- Placement: the "Reclaim space" button moves into the page header,
top-right (matching the metric page), instead of sitting in the body.
- It now actually reclaims: build cache needs -a to drop (docker reports
0 B "reclaimable" without it, but it's pure cache — safe to clear), so
the CLI uses `docker builder prune -af`. Previously the safe scope
freed ~nothing on a box whose reclaimable was mostly cache.
- Honest "Reclaimable" number: /api/system/storage was counting the
whole build cache AND unused tagged images, overstating what the safe
prune frees (e.g. 340 MB shown, ~96 MB per docker, button cleared 0).
Reclaimable now = dangling images + build cache only; stopped
containers and volumes are never counted (the safe prune never touches
them). Headline now matches the button's effect.
Also simplify the CLI output (drop the jargony scope notice and the
reclaimed-total greps) and re-enable the now-persistent header button
after the post-reclaim refreshes.
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds a `libreportal system reclaim` CLI command and an orange "Reclaim
space" button on /admin/config/system/storage (the v2 prune control the
page always hinted at).
Scope is deliberately SAFE: build cache + dangling (untagged) images
only (docker builder prune -f + docker image prune -f via the
rootless-aware runFileOp). It never touches volumes (app data) or
tagged/in-use images, so nothing an app relies on is removed.
Wiring mirrors system_update: a systemReclaim() action + system_reclaim
route case run the command verbatim through the task processor. The
button confirms via showConfirmation, shows a spinner, and re-reads
storage usage as the prune lands. Button styled with --status-warning to
match the Reclaimable stat it sits under, with a note clarifying scope.
Signed-off-by: librelad <librelad@digitalangels.vip>
The expanded snapshot detail reused the shared .task-meta/.meta-item
layout, which forces each field onto one nowrap line and clips long
values (the full date string, repo paths) mid-string. Give the backup
snapshot its own scoped label-over-value grid plus full-width Tags/Paths
blocks that wrap, surface app=/host=/engine= tags as their own fields,
and show a readable date (full timestamp on hover). Applied to both the
global Snapshots tab and the per-app Backups card so they match.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The page has no live feed, so a completed backup/restore wasn't reflected
until a manual Refresh or re-navigation. Subscribe to the TaskEventBus
'taskCompleted' event and repaint on backup/restore completions. Debounced
to coalesce the burst when several per-app tasks finish together; only the
mounted instance reacts and a stale listener removes itself. The Refresh
button stays as a manual pull.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The primary header button defaulted to "Backup all apps" on every tab that
wasn't Locations/Configuration, so it showed on Dashboard and Migrate where
it isn't wanted (Dashboard backs up via the status-grid tiles; Migrate is
about moving hosts, not backing up). Keep it on the Backups tab and hide the
header action entirely on Dashboard and Migrate.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Promote a compact Storage summary (breakdown donut + per-category legend
+ reclaimable) onto the System index, replacing the thin Docker strip and
its easily-missed "Open breakdown" link; it links through to the full
breakdown page. Drop the Disk usage trend chart, which duplicated the
Disk gauge's root-mount %.
Extract the donut + segment builders onto SystemStoragePage so the index
summary and the full page share one renderer. This also fixes a donut
stacking bug: the SVG used the final cumulative fraction for every
slice's dashoffset instead of each slice's own running offset, so the
ring only partially filled. It now fills proportionally.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The metric detail page showed the empty overlay ("No samples in this
range yet — check back in a minute") whenever the initial history fetch
returned zero points, but nothing ever hid it again. Live ticks then
pushed samples in, drew the chart and filled the now/peak/avg/min stats
— while the overlay stayed up, contradicting the visible data.
Make _renderChart the single authority for the overlay: hidden whenever
there are points, shown only when there are none. Live data clears it as
soon as the first sample lands; switching to a genuinely empty range
brings it back.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Load ring was driven by load1_percent = min(100, load1/cores*100)
and coloured by the generic gauge thresholds (red at >=90%). On a
low-core box that pinned it red whenever load merely approached the core
count — which is normal "fully used" territory, not a problem.
Drive the ring from raw load1 with max = cores*2 (so load == cores sits
mid-gauge) and colour by load-per-core: green below capacity, orange
around capacity (>=1.0x), red only once load clearly exceeds it (>=1.7x,
tasks genuinely queuing). cpu.cores rides the live SSE payload, so the
colour is correct on live ticks too.
Signed-off-by: librelad <librelad@digitalangels.vip>
The OS/Kernel/Uptime/CPU/Swap strip was the only section on the System
page rendered without a .sys-section-head, so it had no title and butted
directly against the charts above it (no top spacing). Add a "Host" head
— matching the Docker / Per-app pattern — which supplies both the label
and the section's 26px top margin. "Host" rather than "System" since the
page H1 is already "System".
Signed-off-by: librelad <librelad@digitalangels.vip>
The admin overview Manage backups action called window.librePortalSPA,
a global that is never assigned, so the optional-chaining call silently
no-op'd. The router is window.spaClean; point the call at it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
System-config backups (libreportal backup system) carry no app slug, so the
notification descriptor resolved a blank subject + no icon, and a system-only
pick collapsed to `backup all` when no apps were installed. Give them the
LibrePortal icon + a "Configs" subject, add backup-system to the system-task
logo detection, and guard the whole-fleet collapse on having >=1 app. Rename
the visible subject from "System config" to "Configs" throughout the backup UI.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The shared .task-header has no gap, so a full-width chip row (status /
cpu / mem / ports / ip) left the last live chip butting against the
Restart button. Add a 12px gap on .service-item rows only (matching
.task-info's internal gap) so the resource chip and Restart no longer
touch, without affecting the Tasks page that reuses .task-header.
Signed-off-by: librelad <librelad@digitalangels.vip>
Was centred in the open details panel; move it to the left edge so it
lines up with the rest of the panel content.
Signed-off-by: librelad <librelad@digitalangels.vip>
Bump the dev strip's horizontal padding 14px -> 18px to match
.setup-level-card's content inset, so the strip icon/text sit on the
same left edge as the card titles above it instead of ~4px inboard.
Padding sides (not an icon margin) keeps the whole row aligned and
leaves the top/bottom entrance animation untouched.
Signed-off-by: librelad <librelad@digitalangels.vip>
Swap the strip's shared 🛠️ emoji for the inline "tool" SVG used by the
topbar Developer-mode banner — a real, dedicated icon that ties the two
dev-mode surfaces together and no longer doubles the Advanced card's
glyph.
Enrich the entrance: the box grows in and settles, a one-shot accent
glow pulses for the "unlocked" beat, a subtle shine sweeps across, the
icon pops with a slight overshoot/wiggle, and the text slides in just
behind it. All gated behind prefers-reduced-motion.
Signed-off-by: librelad <librelad@digitalangels.vip>
Tap the Advanced card 10 times and a full-width "Dev mode activated"
strip slides in beneath the two cards — the same 10-tap pattern as the
topbar logo and services-manager unlocks, now at install time. The
choice rides the setup payload (dev_mode) so setup_apply.sh persists
CFG_DEV_MODE=true, and it's mirrored in-process via LpUi.dev so the
next surface already reflects it. 10 more taps toggles it back off.
Counting the Advanced radio's click (not the label's) sidesteps the
label->input double-fire; the radio is pointer-events:none, so each tap
reaches it exactly once. The strip is [hidden] by default (no phantom
gap in the flex column) and replays its entrance keyframes each reveal.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Beginner/Advanced cards on the first setup step had three-sentence
descriptions that read as a wall of operator detail — the opposite of
the friendly first impression the step is meant to give. Trim each to a
single game-intro-style line; the reversibility note and the Advanced
toggle still cover the details for anyone who wants them.
Signed-off-by: librelad <librelad@digitalangels.vip>
The uninstall branch of webuiUpdateAppLog removed the per-app WebUI log
with a bare `rm`. The log lives in the container data plane and is owned
by the container user, often without a write bit. A bare rm (run as root
via `sudo init.sh uninstall`) prompts interactively for write-protected
files — which hangs an otherwise-unattended deploy: the uninstall phase
of a `full` redeploy stopped dead at "rm: remove write-protected regular
file '.../frontend/logs/apps/<app>.log'?".
Route it through runFileOp rm -f (as the container-data owner, force) to
match the neighbouring uninstall_app.sh and the install branch's
owner-aware createTouch/runFileWrite helpers. No prompt, correct owner.
Signed-off-by: librelad <librelad@digitalangels.vip>
The "<X> needs to be installed" feature cards (Enable Whitelist, Authelia
Integration, Headscale Integration, …) were rendering with broken
proportions inside narrow form-grid columns: the body squashed into a
~30-char column and the install button stretched vertically as the only
flex item with room to grow.
Switch to a 2-row CSS grid:
┌────────┬──────────────────────────┐
│ icon │ title │ row 1: who is this for
│ │ reason │
├────────┴──────────────────────────┤
│ [ Install <Service> ] │ row 2: full-width fix-it
└───────────────────────────────────┘
icon grid-row 1, col 1
body grid-row 1, col 2
action grid-row 2, col 1 / -1, width 100%
Reads top-to-bottom regardless of how narrow the host column is, so the
Features tab's 3-column grid stops looking broken. The old @media
(max-width: 560px) responsive override is gone — the grid layout works
at every width, no breakpoint needed.
Signed-off-by: librelad <librelad@digitalangels.vip>
The /backup → Backups tab was the last surface still rendering snapshots
as a plain HTML table — every other backup-related list had moved to the
.task-item card pattern shared with Services. Cohesion-only refactor:
both surfaces now look identical, with the global view adding the
fields the per-app view doesn't need.
HTML: drops <table class="backup-snapshot-table"> + its <tbody>,
replaces with a single <div id="backup-snapshot-list"
class="backup-snapshot-rows"> that the same .backup-snapshot-flash
deep-link highlight already targets.
renderSnapshots() now emits .task-item cards via the new
_renderSnapshotRow() helper. Each card carries:
app icon · "12h ago" title · app-name chip (linked) · location pill
· timestamp chip · short-ID chip Restore · Delete · Details
Extras vs the per-app card:
- App-name chip — global list isn't scoped to one app, so each row
needs to name the app it belongs to. The chip is the deep-link to
/app/<name>/backups?snapshot=<id> (replaces the dashed-underline
"link" treatment on the old App / ID table cells).
- Delete button alongside Restore — destructive cleanup lives on the
global view, not on the per-app card.
- "System config" rows (snapshots without an app=<slug> tag) get the
LibrePortal icon and no app-link (no per-app page to open).
Detail panel (expanded via header / Details button) shows App, Backup
ID, Location, full timestamp, Host, Tags, Paths — the same shape as
the per-app version, plus Host (relevant on the global multi-host view).
Click delegation:
- [data-action="toggle-snapshot-row"] on the header + Details button
toggles .task-details-open
- Restore / Delete buttons now stopPropagation so clicking them
doesn't also toggle the panel
- Existing [data-deep-link] handler is reused by the app-name chip
Signed-off-by: librelad <librelad@digitalangels.vip>
The Backup status card sat with just a heading + tooltip on the right;
the Locations card on the same row already had a hint pill ("Active
destinations"). Mirror that pattern: show the next scheduled backup
time pushed to the right of the heading, so the user can see at a
glance when the daily run will fire without digging into Configuration.
Derived purely client-side from CFG_BACKUP_CRONTAB_APP (read off the
already-loaded window.systemConfigs map) — no backend surface needed:
- nextCronFireTime(expr) parses a 5-field crontab (minute hour dom
month dow) supporting *, N, lists (N,M,O), ranges (N-M), and
steps (* /N, N-M/S). Walks one minute at a time from now+1, honours
the POSIX OR rule for DOM+DOW, caps at 366 days so an unmatchable
expression doesn't loop forever, returns null on bad syntax so the
UI falls back gracefully.
- formatRelativeFuture(when) — formatRelative's future-tense sibling:
"in 6h", "tomorrow", "in 3d".
- formatScheduleClock(when) — "at 05:00" today, "Mon 05:00" otherwise.
Hint slot rendered in #backup-next-run. Three states:
- parseable + computable "Next backup tomorrow · at 05:00"
+ title with absolute time + schedule
- unparseable schedule "Schedule: <raw>" with title hint
- empty CFG_BACKUP_CRONTAB_APP "No schedule set" with title hint
Smoke-tested the cron parser against "0 5 * * *", "*/15 * * * *",
"30 23 * * 0", "0 0 1 * *", "", "garbage", and "0 5 * *" (4 fields).
Signed-off-by: librelad <librelad@digitalangels.vip>
The Open-backup-center button was rendering as the raw .btn-secondary
fallback (muted grey) because the amber-tinted theme override was
scoped only to .config-actions and .console-actions — .backup-title-
actions wasn't in the selector list. Result: same shape as Back-to-Apps,
totally different colour, looked off.
Add .backup-title-actions .btn-secondary to both override blocks (the
nebula-theme rule and the default themes.css fallback) so Open backup
center now matches Back-to-Apps and the Config Reset button: solid
amber in default themes, translucent amber under nebula.
Comment also reframed — these aren't "Back to Apps"-specific anymore;
they're "step away from this page" secondary actions as a family.
Signed-off-by: librelad <librelad@digitalangels.vip>
Two fixes to the .lp-ui-advanced-toggle on the Services tab header:
1. The thumb flipped from --text-primary (white-ish) to --text-on-accent
(a dark navy on the default theme) when toggled on, which read as a
"black circle" inside the accent track. Other toggles in the project
(.eo-toggle in modal.css, .routing-toggle in routing.css) keep the
thumb white in both states — only the track changes colour. Dropping
the checked-state thumb fill brings this toggle in line.
2. The toggle was floating bare in the header row next to nothing,
which looked out of place compared to the contained button-style
controls in the same slot on Backups (Backup now / Open backup
center). Wrapped it in a chip: neutral rgba(text, 0.06) bg + 0.15
border + 6×12 padding, hover bumps both alphas. Same recipe a
.task-btn uses for its resting state, so the toggle visually reads
as a control sitting in line with the rest of the row's actions.
Signed-off-by: librelad <librelad@digitalangels.vip>
"Backup now" and "Open backup center" looked off compared to the rest of
the app page — the secondary link sat underlined with a trailing arrow
glyph instead of a real button, and neither carried an icon. Re-skins
both to use the .btn .btn-primary / .btn .btn-secondary pattern the
config Save / Reset buttons use, so the three action surfaces on an
app page read as one family.
Backup now .btn .btn-primary + upload-cloud SVG (16x16)
Open backup center .btn .btn-secondary + external-link SVG (16x16)
The "Open backup center" link is now SPA-routed (preventDefault + call
navigateToRoute) so clicking it doesn't trigger a full page reload —
same behaviour as the deep-link cells in the global Snapshots table.
href is still /backup so cmd/ctrl-click and right-click → open-in-new-tab
still work the natural way.
Applied to both apps-unified-layout.html and the legacy app-content.html
since the existing app-page surface lives in both templates.
Signed-off-by: librelad <librelad@digitalangels.vip>
Fixed-width tracks + cap formula kept the box pinned to "N cards at
328px" outer regardless of viewport size, so zooming out left a
massive empty band between the box's right edge and the layout edge.
The box was no longer "dynamic" in any real sense — it scaled with
the card count, not with the available content.
Switching grid-template-columns to repeat(auto-fit, minmax(--app-min,
1fr)) lets cards stretch to fill the row, and auto-fit collapses
trailing empty tracks so a 2-card row in a 3-track-wide viewport
doesn't leave a 328px hole at the end. Zoom in/out now just widens
or narrows the cards; the box always reaches the layout edge.
This drops the cross-category card-width uniformity that the earlier
fixed-width pass introduced — a 2-card category now lays out as two
wide cards while a 3-card category gets three narrower ones. That's
mutually exclusive with "box always full width" without leaving
holes, and the user has shifted priority to full-width-always.
JS cleanup: dropped updateAppsCount + its window-resize listener +
its callsites in renderApps/filterAppsByQuery — no more --app-count
or column-count measurement needed when the grid handles everything
natively.
Signed-off-by: librelad <librelad@digitalangels.vip>
Each tool row's Run button gains a small play-triangle SVG to the left
of the label, matching the iconography pattern the Services tab uses
for its Restart and Open buttons. Same green colour (currentColor), so
the icon inherits the success/destructive variants without extra CSS.
Button becomes a flex container with a 6px gap so icon + label stay
nicely centred regardless of label width.
Signed-off-by: librelad <librelad@digitalangels.vip>
Tool descriptions were leaking internal vocabulary (Django superuser,
Postgres bcrypt update, htpasswd in protectionauth.yml, gitea admin
user change-password CLI, trusted_domains list, …) and repeating the
label as a full sentence. Beginners don't care, and even experienced
users don't need the CLI name to know what a button does.
Rewrites every tool description to a single short sentence plain
enough that a first-time installer can read it without context.
Conventions applied across the board:
- One sentence, sentence-case
- Plain English: "Set a new password", "Add a new user",
"Permanently remove a user", "List every user"
- "Leave blank to generate one" only where it's actually useful
(password fields), and matches the field placeholder text
- No CLI names, no schema field names, no internal file paths
- Destructive actions stop saying "permanently" twice (the action
label + the confirm modal already cover that)
- Field placeholders harmonised: "Leave blank for random" /
"Leave blank to generate" → consistently "Leave blank to generate"
Touched files (descriptions only — no logic, no fields removed):
containers/adguard/tools/adguard.tools.json
containers/bookstack/tools/bookstack.tools.json
containers/dashy/tools/dashy.tools.json
containers/focalboard/tools/focalboard.tools.json
containers/gitea/tools/gitea.tools.json
containers/gluetun/tools/gluetun.tools.json
containers/invidious/tools/invidious.tools.json
containers/linkding/tools/linkding.tools.json
containers/nextcloud/tools/nextcloud.tools.json
containers/pihole/tools/pihole.tools.json
containers/traefik/tools/traefik.tools.json
Signed-off-by: librelad <librelad@digitalangels.vip>
User's empirical fix: on a 1280-class viewport (sidebar 220, content
~1010), --app-min 300 made the grid pick 3 cols because
floor((1010+20)/(300+20)) = 3, which left a 4-card category landing
as 3+1 with the orphan-row gap that's been the running visual
complaint. Bumping --app-min to 328 changes the floor to
floor((1010+20)/(328+20)) = 2, so the same 4-card category becomes
2+2 with no orphan.
Wider monitors are unaffected — a 1056px content area still fits 3
tracks of 328 (3*328 + 2*20 = 1024 ≤ 1056), and 1700px+ content
still fits 4. The cards-per-row count only drops on the narrow band
where 300 would otherwise have squeezed a third just-too-tight
column in.
Signed-off-by: librelad <librelad@digitalangels.vip>
Apply the same .services-rows / .tasks-container pocket pattern to the
Tools tab so the three app-page tabs (Services, Tools, Tasks/Backups)
share one visual language: rows live inside a sunken dark panel that
reads as a contained area inside the tab-pane's glass surface.
.tools-rows gains
background: rgba(var(--bg-rgb), 0.2)
border-radius: 8px
margin: 16px
padding: 16px (was 1rem 1.25rem 2rem)
For the multi-category tabs case the pocket sits flush below the tab
bar — drop the top margin via the .tools-tab-bar ~ .tools-cat-pane
sibling selector so the tabs and pocket read as one element.
Signed-off-by: librelad <librelad@digitalangels.vip>
The orphan-1 rebalance (4-on-3 → 2x2) cost too much: it dropped a
4-card category from 3 cards per row to 2 across the board, and would
do the same for any N where N % maxCols == 1. User feedback: 3
densely-packed cards with a small orphan-row gap reads better than
2 wider cards in a 2x2 layout — denser rows feel more compact and
let the eye scan more apps at once.
Back to the post-d4b7731 state: fixed-width tracks (auto-fill,
--app-min) so card widths line up across categories, plus the
sentinel that sets --app-count to 99 when visible cards meet the
natural full-width column count so the box reaches the layout edge.
The 4-on-3 case is now 3+1 again — the lone card on row 2 has empty
cells to its right, accepted as the lesser of two evils.
If the orphan ever becomes a real visual issue, the next move would
be JS-rendered last row (own sub-grid sized to its item count) rather
than reducing the column count globally.
Signed-off-by: librelad <librelad@digitalangels.vip>
Screenshot showed a 4-card category laying out as 3+1 (three cards on
row 1, Wireguard Easy alone on row 2 with two card-shaped empty cells
on its right). Fixed-width tracks + auto-fill kept the cards aligned
across categories but couldn't avoid the orphan — pure CSS grid has
no way to collapse partial-row trailing cells when the column above
them is filled.
apps-manager.js now picks --app-cols deliberately: the natural
column count for the viewport, reduced by one when the last row
would otherwise be exactly one orphan card. 4 cards on a 3-col
viewport becomes 2x2; 5 cards stays at 3+2; 6 stays at 3+3+0; 7
drops to 2-col so the last row gets a partner (still has one orphan
at the very end since 7 is prime, but never below 2 cols — a single
column stack reads worse than an orphan).
CSS swap: grid-template-columns now consumes the new --app-cols
custom property and uses minmax(--app-min, 1fr) so cards stretch
within their tracks (the orphan-prevention dance means widths can
vary across categories now — tradeoff for never having internal
gaps). 1-card view still shrinks the box via the existing formula
so a lone card isn't stretched across the full row.
Signed-off-by: librelad <librelad@digitalangels.vip>
The fixed-width tracks change kept card widths uniform across
categories but reintroduced the "large gap on the right" outside the
glass box — with cards locked at 300px and the cap formula tracking
exactly N cards, the box stopped wherever the visible cards ended,
leaving up to 150px+ of empty parent space to its right on a wide
viewport.
Bringing back the natural-columns sentinel from the earlier pass.
updateAppsCount measures the parent's inner width (minus the
section's 90px of margin/padding/border), computes the column count
the auto-fill grid would pick at full width, and passes a huge
sentinel (99) as --app-count whenever visible cards >= that count.
The formula then overshoots the 100%-44px parent cap and the box
runs edge-to-edge. Cards themselves still come out at --app-min
because the grid template is repeat(auto-fill, var(--app-min)) — no
1fr stretching — so the cross-category uniformity from the previous
fix is preserved.
Net effect: 1-2 cards on a 3-col viewport still shrink (no card-shaped
hole), 3+ cards reach the right edge of the layout, every card lines
up across categories regardless of which branch fires.
Signed-off-by: librelad <librelad@digitalangels.vip>
The per-app Backups tab was the odd one out: snapshots and the
"Backup now" / "Open backup center" buttons all sat inside a single
flat .backup-app-card with no styling parity to Services or Tasks.
The Services tab uses .services-title (20px header + bottom border)
on top of a recessed .services-rows panel; Tasks uses the same recipe
with .tasks-title + .tasks-container. Backups now matches.
.backup-title is the header row — h3 + subtitle on the left,
Backup-now (primary) and Open-backup-center (secondary) buttons
pinned to the right so they stay reachable regardless of how long
the snapshot list grows. No pagination needed: the renderer already
soft-caps the displayed list at 50 with an "Open backup center"
overflow link, and per-app snapshot counts almost never exceed that.
.backup-snapshots-container is the dark panel (rgba bg 0.2, radius 8,
padding/margin 16) wrapping the existing status line + snapshot rows.
JS untouched — it still writes to #backup-app-card-status and
#backup-app-card-snapshots; only the outer shell changed.
Signed-off-by: librelad <librelad@digitalangels.vip>
Two bundled UI fixes.
1. Tools page Run / destructive buttons — the base recipe in tools.css
(rgba green/red 0.12 bg + 0.30 border + full-saturation text) reads
muddy against Nebula's cosmic gradient, same readability problem
.install-btn / .uninstall-btn had before the nebula overrides bumped
them to 0.35/0.65 with --text-primary text. .tool-run-btn and its
.destructive variant now ride those same overrides so Run pops as
green-tint and the dangerous variant pops as red-tint, both with
neutral text against the gradient.
2. Services tab row — the "Logs" button now reads "Details" because
that's what it actually toggles (meta + rich detail + log toggle).
The data-action moves from toggle-logs to toggle-details, and the
expanded panel no longer auto-opens a log stream. A small footer
"Show logs" / "Hide logs" toggle at the bottom of the open panel
explicitly opts in to tailing, kicking off the existing SSE stream
on click (auto-updates while shown). Closing the parent details
panel also resets the log block back to its hidden state so the
next reopen starts clean. app-tabbed-manager's task-running button
disable was taught about the new actions so they stay clickable
while a long task is running.
Signed-off-by: librelad <librelad@digitalangels.vip>
"Snapshot" is restic's term and leaks the tool's vocabulary into the
WebUI. Users think in "backups" — the on-page label even says "Backups"
already; only the secondary copy still said "snapshot". Renames the
remaining user-visible mentions while leaving code identifiers, API
keys, data attributes, CSS class names, and the ?snapshot= deep-link
param untouched (those are internal contracts and changing them would
churn for no user-visible win).
Renamed surfaces:
- Per-app Backups tab header:
"Snapshots for <app>" → "Backups for <app>"
"across all configured repositories" → "across all configured locations"
- BackupAppCard:
"No snapshots yet" → "No backups yet"
"No snapshots found" → "No backups found"
"Showing the most recent 50 of N snapshots" → "...of N backups"
ID-chip tooltip "Snapshot ID" → "Backup ID"
Detail panel "Snapshot ID:" → "Backup ID:"
- Backup retention preset descriptions (KEEP_LAST/DAILY/WEEKLY/MONTHLY/
YEARLY) — "snapshot per day/week/..." → "backup per day/week/..."
- Personal preset hint: "6 monthly snapshots" → "6 monthly backups"
- Restore confirmation modal hint: "snapshot restored in place" →
"backup restored in place"
- Config-warning banner copy adjusted so it doesn't introduce
"snapshots" as a noun
- Retention "Keep last" input suffix: "snapshots" → "backups"
- Cross-host migrate tooltip: "snapshot" → "backup"
Signed-off-by: librelad <librelad@digitalangels.vip>
The auto-fill minmax(300px, 1fr) template stretched cards to fill the
glass box, so a 2-card category landed at ~301px each (the box
shrunk-and-stretched to a hair over 2*300) while a 3-card category
(box now full-width) landed at ~323px each. Cards visibly didn't
align between categories — the user spotted the 22px difference.
Switching the grid template to fixed-width tracks
(repeat(auto-fill, var(--app-min))) means cards are always exactly
--app-min (300px / 280px at ≤1024) regardless of how many are
visible. Card positions and widths line up across every category.
The natural-columns sentinel from the previous pass is no longer
load-bearing — with fixed-width cards, "full width" at high N gives
no extra card-width benefit, only trailing space inside the box.
updateAppsCount drops the measurement step and just sets the visible
count, letting the formula shrink the box around the cards.
Signed-off-by: librelad <librelad@digitalangels.vip>
Previous cap shrank the box to "exactly N cards at min width", which
made 3 cards sit a few pixels short of the layout edge on a 3-column
viewport while 4 cards (which wraps internally) ran edge-to-edge —
visually inconsistent and the user flagged the gap.
updateAppsCount now measures the parent's available inner width
(minus the section's own 90px overhead: 22 margin + 22 padding + 1
border, doubled) and computes the natural column count the auto-fill
grid would pick at full width. If visible cards >= that count, the
function passes a sentinel (99) as --app-count so the formula
overshoots the 100%-44px parent cap and yields the layout-edge box.
Otherwise the cap still kicks in to hide card-shaped holes for 1-2
cards.
Also wired a window resize listener in the constructor so dragging
the window, snapping it, or opening devtools re-evaluates the
decision — the natural column count is viewport-dependent.
Signed-off-by: librelad <librelad@digitalangels.vip>
Two follow-ups to the dynamic-width change:
1. The box was centred (margin: 22px auto), which moved cards out of
their original left position whenever the cap kicked in. Revert to
margin: 22px so the cards keep their left X — the box just shortens
on the right when there are few visible cards.
2. The formula assumed content-box, but style.css:4 sets
* { box-sizing: border-box } globally. With border-box max-width is
the outer width, so a 2-card cap of 664px gave content = 664 - 44
(padding) - 2 (border) = 618, just under the 620 needed to keep
2 columns at minmax(300px, 1fr) with gap 20 — grid silently dropped
to 1 column and the cards stacked. Formula now adds 46px (padding
+ border) plus 2px of sub-pixel buffer, so 2 cards have 622px of
content and reliably stay on one row.
Signed-off-by: librelad <librelad@digitalangels.vip>
Restores the per-app snapshot list (regressed during the backup-system
revamp) and rebuilds it on the same .task-item visual the Services tab
uses, so the two app-page tabs read as a matched pair. Wires the three-
level navigation the user asked for end-to-end:
/backup global dashboard + snapshot table
└─ click app tile → /app/<name>/backups
└─ click any snapshot row expands to detail in place
└─ click App / ID cell → /app/<name>/backups?snapshot=<id>
(auto-expands + scrolls + flashes)
Per-app Backups tab (BackupAppCard):
- Snapshots render as task-item rows: app icon, "12h ago" title,
location pill, full timestamp chip, short-ID monospace chip,
Restore + Details actions.
- Click the row header (or "Details") to toggle a .task-details panel
showing snapshot ID, location, full timestamp, host, tags, and the
paths the snapshot covers.
- Shows up to the 50 most recent; >50 surfaces a hint to the global
backup center for the full list.
- flattenSnapshots() now carries hostname/tags/paths through so the
detail panel has real content.
Cross-page navigation:
- Dashboard app-tile click navigates to /app/<name>/backups instead of
opening the pick-now modal. The pick-now action is preserved as an
explicit "Back up" pill that appears top-right on hover/focus.
System tile keeps the old modal click (no dedicated page yet).
- Global Snapshots table — the App and ID cells are now SPA-routed
links to /app/<name>/backups?snapshot=<id>. Snapshots without an
app=<slug> tag (system backups) stay plain text. Routed via
navigateToRoute so the SPA mounts in place instead of a full reload.
Deep-link mechanism:
- BackupAppCard._honorSnapshotDeepLink reads ?snapshot=<id> on render,
finds the matching .backup-snapshot-item, opens its details, scrolls
it into view, and applies a brief .backup-snapshot-flash (animated
box-shadow pulse) so the user's eye lands on it after the SPA jump.
CSS:
- backup.css gains .backup-snapshot-rows, the location pill, the
monospace ID chip, the tag chips, the deep-link flash keyframes,
the tile "Back up" pill (.backup-app-tile-action — only visible on
hover/focus to keep the dashboard calm at rest), and the dashed
underline link style for the snapshot-table deep-link cells.
Signed-off-by: librelad <librelad@digitalangels.vip>
The inline "Delete location" action was the last spot on the Backup
page still using the native browser confirm() — the snapshot delete
already uses the styled backup-modal, so the location delete sat out
as the odd one. Adds a new #backup-delete-location-modal matching the
existing modal shell (header / body / backup-danger-btn footer),
swaps deleteInlineLocation() to open it instead of confirm(), and
wires the confirm button to a new confirmDeleteLocation() that does
the actual `libreportal backup location remove <idx>` task.
Behaviour is the same — confirm body text moves into the modal as a
muted hint paragraph using backup-card-hint, location name bolded
for scannability. expandedLocs cleanup also moves into the confirm
handler so the row collapses only when the user actually deletes.
Signed-off-by: librelad <librelad@digitalangels.vip>
The existing --skip-docker-images flag keeps a lot more than just images:
the docker-install user, the rootless dockerd, the rootless sysctl
drop-ins, AND the image/build cache. So a reinstall after using it
already skips the slow `dockerd-rootless-setuptool.sh install` step —
which is the meat of why anyone reaches for this flag on a local dev
loop. The name "--skip-docker-images" undersells what it actually does
and "skip the rootless install" is the user-facing intent.
Adds --skip-rootless as an alias of --skip-docker-images (same flag
variable, no behaviour change). Both spellings continue to work — anything
scripting the old name keeps working — but the help text, examples, and
the uninstall printf now use the clearer --skip-rootless. Same name
shift in scripts/update.sh: SKIP_ROOTLESS=1 is the new env-var spelling,
SKIP_DOCKER_IMAGES=1 is the back-compat alias.
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds the install-time Beginner/Advanced choice the user described, with
the linked dev-mode escape hatch and global body-class machinery that
any surface can hang advanced/dev-only DOM off.
Three-tier mental model, two flags in the data model:
Beginner default. nothing extra shown.
Advanced .lp-advanced DOM revealed; advanced wizard steps shown
Adv+Dev .lp-dev DOM also revealed; dev-only fields visible
Linking rule (enforced inside LpUi):
- enabling dev auto-enables advanced (dev w/o advanced is incoherent)
- disabling advanced auto-disables dev
Wire shape:
CFG_INSTALL_LEVEL beginner | advanced (general_basic)
CFG_DEV_MODE existing, unchanged behaviour
window.LpUi.{advanced,dev} {get(), set(), apply()}
localStorage keys lp.ui.advanced, lp.ui.dev, lp.ui.seeded
body classes lp-ui--advanced, lp-ui--dev
events lp-ui-advanced-changed, lp-ui-dev-changed
global CSS gates body:not(.lp-ui--advanced) .lp-advanced { hide }
body:not(.lp-ui--dev) .lp-dev { hide }
Setup wizard:
- New step 1 "Choose your experience" with Beginner/Advanced cards.
Beginner is preselected so race-through gets the safe default.
- Picking a level updates totalSteps live (4 for beginner, 5 for
advanced) so the progress bar reflects the choice.
- Metrics step (Prometheus + Grafana) is gated to Advanced — beginner
never sees it, never gets asked, never installs them by accident.
- Submit payload now carries install_level; setup-routes.js validates
it against the enum (beginner|advanced).
- scripts/setup/setup_apply.sh writes it to CFG_INSTALL_LEVEL via
updateConfigOption.
- On submit, LpUi.advanced.set is called immediately so the next
surface (running-tasks page) is already in the right mode — no
refresh needed.
WebUI bootstrap:
- js/utils/lp-ui.js loads first thing in index.html (before any other
bootstrap) so body.lp-ui--advanced is applied pre-paint — no FOUC
of advanced content on a fresh tab.
- On first run, seeds lp.ui.advanced from CFG_INSTALL_LEVEL.
Subsequent loads honour the user's per-browser override.
- Mirrors CFG_DEV_MODE → lp.ui.dev on the seed pass.
Dev-mode unlock:
- Existing 10-click LibrePortal-logo easter egg unchanged.
- NEW: same 10-click unlock on the Advanced toggle (in services-manager).
Reuses the countdown-toast pattern; on the 10th click delegates to
the topbar's _setDevMode so there's one canonical setter and the
config_update task path stays singular.
- TopbarComponent now exposes its instance as window.topbar so the
toggle's tap handler can reach _setDevMode.
- topbar._setDevMode also calls LpUi.dev.set(enabled) so the body
class flips immediately (no reload needed to see dev-only DOM).
Convention rolled out:
- Services tab's .service-rich panel was already gated on
body.lp-ui--advanced.
- .lp-advanced / .lp-dev are now first-class hide classes any
component can tag DOM with — see style.css globals.
Signed-off-by: librelad <librelad@digitalangels.vip>
Two reasons the back button was unreliable:
1. The very first history entry (the URL the user landed on) had
state: null because handleInitialRoute() called navigate(path,
false), and the pushState branch only ran when addToHistory=true.
When the user later pushState'd forward and then hit back, the
popstate handler's guard "e.state && e.state.route" was false on
the initial entry, so it silently did nothing — back appeared
broken. Now navigate() replaceState's the current entry whenever
addToHistory=false, so the initial entry (and any back-compat
URL rewrite) always carries its route. The popstate handler also
now falls back to window.location when state.route is missing,
so third-party history manipulation can't break us.
2. Open SSE streams (LiveSystem, taskEventBus, services-manager log
tails) block the browser's back-forward cache. Without BFCache,
back has to fully re-mount the page instead of restoring it
instantly the way Amazon/GitHub feel. Now pagehide closes every
live bus we own, and pageshow(persisted=true) reopens them when
the page is restored from BFCache. Log tails aren't auto-resumed
— Resume overlay handles that if the user comes back to a
services tab.
Public surface added: LiveSystem.pause()/resume() and
ServicesManager.pauseStreams(). TaskEventBus already had stop()/
start(). The legacy-URL rewrite in handleAppDetail also now
replaceState's with { route: canonical } instead of {} so the
stamp is consistent across all internal history updates.
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds a foundational global UI-mode flag — Beginner (default) vs Advanced —
gated by a single toggle on the Services tab title row. First foothold
of a project-wide pattern: any surface that grows extra-technical detail
(mounts, limits, internals, raw IDs, …) will gate it on the same flag,
so a newcomer doesn't see a wall of operator information while a power
user gets everything site-wide with one flip.
How it's wired:
window.LpUi.advanced — { get(), set(on), apply() }
localStorage key — lp.ui.advanced ('0' | '1')
body class — lp-ui--advanced
event — window 'lp-ui-advanced-changed' { advanced }
Surfaces gate their advanced-only DOM via CSS:
body:not(.lp-ui--advanced) .service-rich { display: none; }
So flipping the toggle is instant and DOM-free — no re-render needed.
The Services tab's rich container panel (limits, image, healthcheck,
networks, mounts) is the first thing behind the flag; live CPU%/memory
chips in each row stay visible always because they read just as easily
as a status colour and are useful to everyone.
Title row gets a small slider toggle styled in the project's accent —
unobtrusive, labelled "Advanced". Default OFF (Beginner).
The same _renderRow reorders the log block above the rich-detail block
inside .task-details, so when Advanced is on AND a row is expanded, the
live log appears right where the "Logs" click landed rather than below
a wall of metadata. Helps with the old simple-click feel even when the
extra detail is showing.
Plumbed deliberately to be project-wide so the upcoming first-install
"Beginner vs Advanced" wizard step can seed the flag (planned:
CFG_INSTALL_LEVEL in general config → emit body class server-side at
template render time → no FOUC on a fresh load).
Signed-off-by: librelad <librelad@digitalangels.vip>
The Admin → System area was growing a parallel per-container surface
(/admin/config/system/app/<name>) alongside the existing per-app Services
tab on the app page. Two pages onto the same thing is the kind of
duplication that rots fast — they drift, users have to remember which
one to use, and the next person adding a feature has to decide twice.
This commit consolidates onto the existing Services tab (which already
has compose-service awareness, docker socket access, restart actions via
the task system, and live log streaming) and decommissions the parallel
admin sub-page:
- Delete system-app-page.js and its lazyLoad entry. The dispatch in
admin-system.js for the 'app' view now redirects to the app page's
Services tab so old bookmarks still resolve cleanly.
- System index per-app rows navigate to /app/<name>/services (not
/admin/config/system/app/<name>) and the row hint copy is updated
to match.
- Services tab gains the rich container detail the old admin page
rendered, fed by /api/system/containers + /containers/:id +
/containers/:id/stats:
* Inline live chips in each service header: CPU% and memory
(with limit + percent if a limit is set). Memory chip flips
amber at 80% and red at 95% of the configured limit.
* New "service-rich" panel inside the existing expandable
details section (above the log block, so the existing Logs
toggle reveals both):
- Image + image-id + uptime + restart count
- Memory / CPU / PIDs limits + restart policy
- Healthcheck pill + last 3 probes (collapsible per-probe)
- Networks table (name, IP, gateway, MAC)
- Mounts table with type badges (volume/bind/tmpfs)
* Live stats refresh every 5 s; existing status refresh stays
on 10 s. Both gated on the Services tab being active.
- Backups for the app already live on the existing /app/<name>/backups
tab (loadAppBackups → BackupAppCard.render), so the navigational
promise of "one place per-app" is already met — System index just
needed to route there.
- CSS: services.css picks up .service-live-chip (with warn/danger
colour cues) and the full .service-rich block (grid, tables, mount
badges, healthcheck pills).
Signed-off-by: librelad <librelad@digitalangels.vip>
Promotes the admin → System area from a single index page with a transient
overlay into a real router with four addressable sub-pages, plus a docker-
api-backed read surface to drive them.
URLs:
/admin/config/system index (gauges + trends + per-app table)
/admin/config/system/metric/<key> single-metric deep-dive
/admin/config/system/app/<name> per-container app deep-dive
/admin/config/system/storage docker disk-usage breakdown
The path resolves to category=`system` in adminCategoryFromPath, so the
existing SPA dispatch still drops you into AdminSystem; AdminSystem then
reads the rest of the path and mounts the right sub-renderer into
config-section. Each sub-page owns its own DOM + lifecycle and is disposed
when the orchestrator re-mounts on the next navigation. Browser back, page
reload, and shareable URLs all work — no modal, no overlay state, no
fragile open/close lifecycle. Esc on the metric page navigates back to the
index.
Backend (containers/libreportal/backend):
- utils/docker.js — shared client for the bind-mounted Docker socket
(extracted from service-routes.js' inline copy). dockerRequest,
dockerStream, and a multiplex-log decoder for /containers/:id/logs.
- routes/docker-info-routes.js mounted at /api/system, contributes:
GET /containers full list, plus grouped-by-app shape
GET /containers/:id inspect projection (limits, mounts,
networks, ports, health, restart count)
GET /containers/:id/stats one-shot CPU% / memory / network /
blkio / pids (derived from precpu/cpu
deltas, like `docker stats`)
GET /containers/:id/logs last N lines, multiplex-decoded
GET /storage `docker system df` rolled up per
category, plus top-10 images +
top-10 volumes by size
Frontend (containers/libreportal/frontend/js/components/admin):
- admin-system.js — refactored into orchestrator + index view. _parsePath
drives dispatch; sub-views are window.SystemMetricPage /
SystemAppPage / SystemStoragePage classes mounted into config-section.
The per-app table is now keyboard-focusable rows that navigate to the
per-container page; the Docker strip grows a "Storage" tile that
navigates to the storage page.
- system-metric-page.js (renamed from system-detail.js, rewritten as an
in-flow page renderer). Same chart visuals as the old overlay — grid,
axis, area gradient, peak/min/now markers, hover crosshair + tooltip
scrubbing, per-metric accent theming — but rendered into the page
instead of a fixed-position panel. Range picker reflects to ?range=
so refresh preserves the selection. 1 Hz SSE feed splices into the
chart tail in real time.
- system-app-page.js — for each container in the app stack: status,
image, image-id, uptime; live stats card (cpu / mem with limit-pct /
rx / tx / blkio r-w / pids, polled every 2s with warn+danger colour
cues at 80% and 95% of memory limit); limits panel (memory, cpu,
pids, restart policy, restart count, started-ago); healthcheck
status + last 3 probes; networks table (name, IP, gateway, MAC);
published ports; mounts table with type badges; collapsible log tail
with refresh.
- system-storage-page.js — donut chart (cumulative-arc, hand-rolled
SVG) splits total in-use disk by images / volumes / containers /
build cache; per-category cards with size + reclaimable; top-10
images and top-10 volumes tables with "unused" / "orphan" badges.
CSS (containers/libreportal/frontend/css/admin.css):
Overlay-specific rules (.sys-detail wrapper, backdrop, panel, close
button, body lock) removed. Inner chart rules (stats grid, svg, grid,
axes, peak/min/now, crosshair, tooltip, foot) retained and reused by
the metric page. New blocks for .sys-metric-page, .sys-app-page (with
stat warn/danger colour states, health pills, mount-type badges, log
pre styling), .sys-storage-page (donut + legend + headline + per-
category cards + orphan/unused badges), .sys-app-row (clickable
rows with arrow + accent hover), .sys-stat-link (clickable Docker
strip tile).
Signed-off-by: librelad <librelad@digitalangels.vip>
The task-event bus translates the backend's task.upsert SSE events into
window-level taskCreated / taskUpdated / taskCompleted CustomEvents. It
fired taskCompleted whenever a task's current status was terminal AND
the previously-known status was not — including the case where the bus
had never seen the task before at all (prevStatus undefined → wasTerminal
false → "transition" detected).
Why this misfired: the backend re-broadcasts the full task object on any
inode change to the task file, not just on logical status changes. The
periodic ownership/permission repair sweep (crontab_check_processor.sh)
chowns the entire tasks directory, which bumps ctime on every task file
and trips fs.watch, which broadcasts task.upsert for each one. If the
page was loaded after a task had already finished, the bus saw that
task for the first time as already terminal and fired a "task completed"
toast — for tasks that completed minutes or hours earlier.
Fix: when an upsert is for a task the bus has never seen AND that task
is already terminal, bootstrap silently. We have no evidence the task
transitioned now — it might have transitioned hours ago. The real
running→terminal transition (bus knew about the task while it was
running, then receives a terminal upsert) still notifies, which is what
users actually want to know about.
Signed-off-by: librelad <librelad@digitalangels.vip>
Replaces the JSON history file behind /api/system/history with a fixed-size
binary ring buffer on disk and adds a second, downsampled tier so the chart
can now span seven days, not just twenty-four hours.
Two on-disk rings under frontend/data/system/:
metrics_ring_1m.bin 1440 pts @ 1 min ( 24 h)
metrics_ring_5m.bin 2016 pts @ 5 min ( 7 d)
Each point is 32 bytes (uint32 timestamp + 7 float32 metrics — cpu / mem /
swap / disk / load1 / net_rx / net_tx); files carry a 32-byte header with
magic, version, capacity, head, count, bucket seconds, and last bucket time
so they're self-describing and torn-write recoverable.
A persistent 1-minute ticker inside the backend (independent of whether
anyone's subscribed to /api/system/stream) composes points from /proc plus
the bash generator's latest snapshots and appends to the 1m ring; every
five minutes it averages the last five 1m points into the 5m ring. On
first run, the writer backfills the 1m ring from the legacy
metrics_history.json so first paint already has 24 h.
/api/system/history?range=N auto-selects the tier (≤1440 → 1m, else 5m),
keeps the existing { points, updated } shape, and additionally returns
`tier` for clients that care. Falls back to the legacy JSON on cold start.
Admin → System: 7d added to the range picker (now 1h / 6h / 24h / 7d),
swap + load1 promoted to their own trend cards, and every gauge / chart
card grows an Expand affordance that opens a fullscreen single-metric
deep-dive overlay:
- Big themed chart with grid, gradient area, peak/min/now markers, and
a live-pulsing "now" dot
- Hover crosshair + tooltip scrubs the series with formatted time +
value
- now / peak / avg / min stat strip with deltas
- Range picker (1h / 6h / 24h / 7d) re-fetches and re-themes per metric
- 1 Hz live SSE feed updates the overlay's now-stat in real time
- Escape / backdrop / close button all dismiss
- Per-metric accent colour (cpu=accent, mem=info, disk/swap=warning,
net_rx=success, net_tx=accent, load=accent) flows through gradient,
border, dot, and stats card
Zero new dependencies — hand-rolled SVG and pointer events throughout.
Signed-off-by: librelad <librelad@digitalangels.vip>
The glass box was a CSS Grid with auto-fill columns of minmax(300px,
1fr), so it always painted across the full content area. With only 2
apps on a wide row the third/fourth column slots remained inside the
border as empty space — visually a card-shaped hole.
Drive the box's max-width off a --app-count CSS variable, capped at
(100% - 44px) so it can't escape the layout's symmetric 22px gutter.
margin: 22px auto keeps the horizontal padding symmetric in both the
capped (auto-centers the smaller box) and full-width (auto collapses
to 22+22) cases. --app-min (300/280 at the ≤1024 breakpoint) feeds
both the grid template and the cap formula so the responsive column
width stays a single source of truth.
apps-manager.js sets --app-count to the count of visible .app-card
elements after every render and after the sidebar search filter, so
filtering down to 2 hits also collapses the box. Floor of 1 keeps the
empty state usable.
Mobile (≤768) overrides max-width to none — single column already
fills, and the 10px gutter shouldn't be auto-centered.
Signed-off-by: librelad <librelad@digitalangels.vip>
The previous commit added body.has-dev-banner shifts for .sidebar and
.apps-layout assuming they were position:fixed top:60 like the topbar.
They aren't — on desktop both sit in flex flow (.sidebar is
position:relative, .apps-layout is just a flex container), so
top:96px pushed the sidebar 96px down from its natural slot, leaving
a big visible gap above the category list.
Scope the sidebar nudge to the mobile media query where it actually
becomes fixed (also covers .sidebar-container, the unified apps
layout's mobile drawer). Replace the wrong .apps-layout top rule with
a height tweak — it sizes itself off (100vh - 60px) and was overflowing
the viewport by 36px when the banner was on; calc(100vh - 96px)
accounts for the banner.
Topbar shift (top:0 → 36) stays unchanged; that one was correct.
Signed-off-by: librelad <librelad@digitalangels.vip>
Banner was fixed at top: 60px (just below the 60px-tall topbar) at
z-index 999 — same vertical band as the sidebar (top: 60px, z-index
100) and the apps-layout subnav, so it covered the top 36px of both
when dev mode was on.
Moved to top: 0, z-index 1001 (above the topbar). When the banner is
visible, body.has-dev-banner now also shifts every other fixed-
positioned chrome element down by the banner's 36px:
.topbar 0 → 36
.sidebar 60 → 96
.apps-layout 60 → 96
.mobile-drawer 60 → 96 (already had this override)
Body padding-top stays at 96px (banner + topbar) — content offset is
unchanged. Standard environment-banner placement (Stripe test-mode,
GitHub staff-mode) and makes "you're in dev mode" actually visible
above your nav.
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds /api/system/stream — a Server-Sent Events feed driven by a single
per-process ticker that reads /proc directly and splices in the latest
host-side metrics.json each second. Subscribers share the connection so
N open tabs cost one ticker, and the ticker pauses entirely when nobody
is listening.
Frontend gets a singleton LiveSystem EventSource manager with auto-
reconnect, Page-Visibility integration (closes on tab hide), and last-
sample replay for late subscribers. Admin -> System gauges and the
dashboard memory + disk tile now tick at 1 Hz; trend charts and the
per-app table keep their 30 s poll because the underlying files only
regenerate once a minute.
Also adds /api/system/history as a thin range-query wrapper over the
existing 24 h JSON ring buffer — the binary ring backend will slot in
behind it in the next phase without changing the response shape.
Signed-off-by: librelad <librelad@digitalangels.vip>
Old: "Backup status — system config + every app — at a glance."
New: "Check what's protected — and when it last ran."
The em-dash chain was filler and "at a glance" was redundant on the
dashboard tab (which is the at-a-glance view). New copy leads with
what the admin is here to do.
Signed-off-by: librelad <librelad@digitalangels.vip>
The custom-drawn green box + white tick was reading too utilitarian
against the row's other buttons (and the tick itself had defaulted
black against the dim green fill, hard to spot). Switches both
.task-select-box (per-row) and the master Select-all to the same
chrome the setup wizard uses for its app-pick cards:
- accent gradient fill on :checked (was status-success)
- 12px white SVG checkmark (inline data: URL, same one as
.setup-app input[type=checkbox]:checked::after)
- subtle inset border at rest, accent glow on hover/focus
- 0.22s setupCheckPop / taskCheckPop pop-in on tick
- indeterminate state on the master shows a horizontal dash,
drawn from a second inline SVG (still white on accent)
Sized to 18px so the row checkbox sits clean alongside the 22px-tall
.task-btn buttons. The master in the action bar reuses the same box
spec (no separate variant), matching the wizard's "one checkbox style,
many places" pattern.
Signed-off-by: librelad <librelad@digitalangels.vip>
Clicking the LibrePortal logo 6→9 times spawned four separate
"X clicks away from being a developer" notifications stacked on top of
each other — visual noise for a delightful-bonus interaction.
Now the easter egg keeps a single reference to its current toast and
mutates the `.notification-message` text in place on each subsequent
click. When the toast's 10s auto-remove timer expires mid-sequence
(slow clicker) the next click opens a fresh one — same fallback for
the idle-reset path that clears the count after 3s.
`_devToast` now returns the notification element so the easter-egg
handler can grab it; previously it returned undefined, fine for the
one-shot toasts but no longer enough for the rolling-update pattern.
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds per-row checkboxes (right of the Delete button, per request), a
master "Select all" toggle in the action bar, and morphs Clear All into
"Delete Selected (N)" the moment 1+ rows are ticked. Both paths go
through the same _showClearAllModal redesigned in 1ccc4bb — same UX,
same "Cancel running too" toggle, same logic; only the title + eyebrow
shift to reflect which mode the user came in through:
all → "Delete all N tasks?" eyebrow "Delete Tasks"
selected → "Delete N selected tasks?" eyebrow "Delete Selected"
State lives in this.selectedTaskIds (Set<string>). The row checkboxes
fire toggleTaskSelection(id, checked); the master fires toggleSelectAll
which ticks/unticks every visible row's checkbox in one pass (visible,
not all-of-this.tasks — so category filters DTRT).
_updateSelectionUI() reconciles three things on every change:
- the Clear All button label + title attr
- the master checkbox's checked/indeterminate state (some-but-not-all
visible → indeterminate dash, all → checked, none → unchecked)
- hooked into renderTasks() so category-switches don't leave stale
UI
performClearAll(opts) now accepts opts.targets — the subset to operate
on. clearAllTasks() passes either the selection or this.tasks depending
on mode. The active-task cancel-or-skip logic (cancelRunning toggle) is
unchanged — runs identically over the smaller set.
CSS:
.task-select — 22×22 framed checkbox matching the .task-btn
buttons it sits next to (border, hover green,
focus outline)
.task-select-box — custom box with check + indeterminate dash
drawn via ::after, no SVG dependency
.task-select-all — text-style toggle in the action bar with the
same custom box
No new globals. Hooked up via the existing window.tasksManager.
Signed-off-by: librelad <librelad@digitalangels.vip>
Two small uninstall-output tweaks.
1. dockerComposeDownRemove now ALWAYS calls dockerRemoveApp (the
`docker ps -aqf name=…` → stop + rm sweep) as a fallback, even when
the compose-down step is skipped because the app dir is missing.
Before, a partial prior uninstall (compose file gone but containers
still running) produced "App directory not found. Skipping container
shutdown." and then proceeded as if the uninstall were complete —
leaving the actual containers running. The name-based sweep also
runs after a successful compose-down to catch anything compose
wouldn't pick up (renamed services, orphans from earlier failures).
While here: the OS_TYPE gate (only Ubuntu/Debian) is gone too —
`docker compose down` works on any OS with docker, and gating it
meant Arch/etc. users got NO compose teardown at all.
2. The step-2 header "Keeping Docker images (pass --delete-images to
remove)" trimmed to just "Keeping Docker images". The `isNotice`
line below already explains the reuse-on-reinstall behaviour; the
CLI-flag hint reads as noise in the WebUI task log where users
can't act on it anyway. CLI users can still pass --delete-images
(cli_app_commands.sh wires it as before) or tick the WebUI's
"Also delete docker image" checkbox.
Signed-off-by: librelad <librelad@digitalangels.vip>
.task-command was still using var(--status-success) (#28a745) which reads
muddy olive against the nebula gradient — the same dimming the status
pills and apps-installed pill already work around with #86efac. The
empty-state row ("$ No tasks found …") was the most visible offender.
Switches .task-command to the same bright mint already used elsewhere.
Same edit, while I was there: the empty-state copy interpolated
categoryName.toLowerCase() as `No ${cat} tasks found`, so the "All Tasks"
category produced "No all tasks tasks found". Special-cases the all
bucket and strips the trailing word when the category name already
includes it ("Running Tasks" → "No running tasks found", not "running
tasks tasks").
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerDeleteData (uninstall) and the wipe-before-restore step in
restoreAppStart both did `runFileOp rm -rf $containers_dir$app_name`,
which runs as $CFG_DOCKER_INSTALL_USER (dockerinstall, uid 1002 on
rootless). That user owns app-template files but CANNOT remove
container sub-UID dirs created by the daemon's userns mapping —
postgres data at uid 232070, nextcloud html at uid 33, etc. The rm
therefore silently failed with
rm: cannot remove '/libreportal-containers/invidious/postgresdata':
Permission denied
while still reporting "<app> successfully uninstalled" — leaving the
sub-UID directory tree on disk to confuse the next install and leak
storage.
Fix: route the wipe through a new `app-data-remove` action in the
root-owned libreportal-ownership helper. Root can rm sub-UID files
unconditionally. The helper validates the app name (alphanumeric +
. _ -, no traversal), refuses the WebUI's own slot (libreportal), and
is idempotent when the dir is already gone.
Two callers updated:
- scripts/docker/app/uninstall/delete_data.sh
- scripts/restore/restore_app_start.sh
The helper itself ships root-owned at /usr/local/lib/libreportal/, so a
fresh install or release upgrade is needed to pick up the new action.
Bumped init.sh footprint_version 2 → 3 so the runtime updater
prompts a root re-install on the next release.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Clear All confirmation was the last destructive task action still
running through window.showConfirmation (the legacy dialog system) —
visually inconsistent with the rest of the tasks page (single-row delete,
Uninstall, etc., which all use openEoModal). Switches it to the same
eo-modal shape used by _showDeleteTaskModal so the destructive-confirm
family looks unified.
While here, adds a "Cancel running tasks too" toggle inside the modal,
off by default. Backed by the existing .eo-toggle.eo-toggle-card style
(modal.css). Drives a new opts.cancelRunning in performClearAll:
Off → skip any running/queued/pending tasks; only terminal rows
are deleted. The success toast reports the split.
On → cancel each active task first (POST /cancel), wait for the
terminal status via SSE, then delete (with the 409→force
fallback the single-row deleteTask already uses).
Body composition mirrors the per-task delete modal:
- danger empty-state ("This cannot be undone")
- badge row with Total / Running / Terminal counts
- the toggle (only shown when runningCount > 0 — no need otherwise)
The action button's label live-updates as the toggle changes:
toggle off + running rows → "Delete N (skip M running)"
toggle on / no runners → "Delete N Tasks"
So the user sees exactly what they're about to do before clicking.
Cancel / backdrop / X all resolve to no-op (same contract as
_showDeleteTaskModal). Modal returns {confirmed, cancelRunning} so the
caller knows which path to take.
Sets up the multi-select work next: the modal already accepts an
arbitrary tasks array; the upcoming "Delete selected" is a one-line
call into the same _showClearAllModal with a filtered list.
Signed-off-by: librelad <librelad@digitalangels.vip>
checkApplicationsConfigFilesMissingVariables does a `find $containers_dir
-maxdepth 2 -type f -name '*.config'` to enumerate every app's live
config. runFileOp drops privileges to $CFG_DOCKER_INSTALL_USER
(dockerinstall), which is intentionally the *manager* for the rootless
data plane — but doesn't own the per-app container sub-UID dirs (e.g.
invidious/postgresdata uid 232070, nextcloud/html uid 33).
At maxdepth 2 find doesn't actually need to descend into those dirs to
satisfy the name filter, but it tries to anyway and emits chatter like
find: '.../invidious/postgresdata': Permission denied
every time the function runs (config-reconcile path on install / app
start / restart). Cosmetic only — the actual .config files are at the
right depth and ARE found — but it shows up in the live CLI output
during installs.
2>/dev/null on the find. The function's purpose is purely to enumerate
LibrePortal-managed .config files; sub-UID data dirs are by design
unreachable to the manager and there's no signal in that error.
Signed-off-by: librelad <librelad@digitalangels.vip>
"Task deleted successfully" was a plain single-line toast while every
other task notification (started, completed, failed, cancelled) renders
as <App>-in-bold on the first line + "<Action> task <verb>" on the
second, with the app icon on the left and the per-action emoji as
custom-icon. Inconsistent.
Now reads e.g.
[🗑️] Ipinfo
Install task deleted.
with the ipinfo logo as the row icon, matching the install/completion
toast format.
Also factored the three duplicate "task → identity (display name + app
icon + friendly action title + emoji)" blocks (taskCompleted listener,
delete-modal title, delete notification) into one helper —
_taskNotificationDescriptor(task) — so the four surfaces (started,
completed/failed/cancelled, delete modal, delete notification) always
agree on what to call a task. Net -20 lines.
Signed-off-by: librelad <librelad@digitalangels.vip>
load_sources.sh calls checkConfigFilesMissingFiles() after init.sh +
variables.sh but BEFORE initilize_files.sh sources the function
manifest. checkConfigFilesMissingFiles uses runInstallOp (in
docker/command/run_privileged.sh) to copy missing config templates —
under LP_LAZY=1 that's an autoload stub that only exists once the
manifest is sourced. So when any template is genuinely missing, the
copy call hits "runInstallOp: command not found" and the file silently
never gets copied.
Symptom on a fresh CLI invocation (foreground or processor subprocess
inheriting LIBREPORTAL_TASK_EXEC=1) where a new config category was
added:
config_check_missing.sh: line 33: runInstallOp: command not found
✓ Success 1 config files were missing and have been added to the
configs folder. ← false success: the count incremented but the
copy itself didn't happen
Fix: source run_privileged.sh directly in load_sources.sh just before
the missing-files check. The file is pure function definitions (runAsManager
/ runFileOp / runFileWrite / runInstallOp / runInstallWrite), no side
effects, ~150 lines — safe to source unconditionally and idempotent
with the eager/lazy load that happens later. Adds <1ms to every CLI
invocation; saves silent failures on the rare path that calls it.
Signed-off-by: librelad <librelad@digitalangels.vip>
WebUI-created tasks emit camelCase initial fields (createdAt, startedAt,
completedAt, heartbeatAt, exitCode, errorMessage) per
tasks-manager.js / task-manager.js conventions, with createdAt in
ISO-UTC-with-ms (`2026-05-27T13:01:26.345Z`). The processor then layers
snake_case status fields (started_at, heartbeat_at, …) on top as the
task runs.
The CLI's cliTaskRun was writing snake_case only — `created_at` with
local-tz offset. The task panel's renderer reads `task.createdAt`
directly (no alias), so CLI-queued tasks showed blank Created/Started
columns until the processor wrote its own snake_case overlay
(which doesn't include createdAt at all). Visible symptom: dates
"broken" on CLI-queued tasks.
Now the initial JSON cliTaskRun writes matches what the WebUI's
"Install" button writes:
{
id, command, status: queued,
createdAt: "<ISO-UTC-with-ms>",
startedAt: null, completedAt: null, heartbeatAt: null,
exitCode: null, errorMessage: null,
type, app
}
Processor side is unchanged (still adds snake_case overlay on
status transitions — that's how WebUI tasks already work). No JSON
shape change for in-flight tasks.
ALSO (out-of-repo): /home/user/Documents/Scripts/update.sh now restarts
the systemd `libreportal.service` task processor after the docker
`libreportal-service` container restart. Same reason — both pre-load
code at startup, both need a restart to pick up changes. Without this,
deploys silently kept a stale processor running old code while the
disk reflected the new code; the install task-routing recursion I just
saw was a direct consequence.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Delete Task confirmation modal was rendering the raw command
("libreportal app install ipinfo") as its title with no app icon, while
the rest of the WebUI (task notifications, task rows) shows
"Install Ipinfo" and the ipinfo logo. Inconsistent and slightly
intimidating for a confirmation step.
Now mirrors the completion-notification flow:
- Title: `${formatActionTitle(task.type)} ${getAppDisplayName(task.app)}`
→ "Install Ipinfo", "Backup Nextcloud", etc.
- Icon: /icons/apps/<slug>.svg (or libreportal.svg for system tasks)
- Tool tasks borrow the same tool-catalog-lookup the completion toast
uses so a tool deletion reads as "Manage Shortcuts" rather than the
raw tool id.
Reuses the existing TasksManager.formatActionTitle() helper so any
future task type added to that map flows through here automatically.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Features section was a grab-bag of ~27 toggles, most of which are
either category-specific (firewall, SSL, Docker network, SSH hardening)
or install-time choices that brick the box if flipped on a live
install (the WebUI / config / CLI / Docker requirements). One page
made auditing easier but flattened the risk hierarchy.
Reorganised so each toggle lives where it conceptually belongs, and
the dangerous install-time set is double-gated:
network_docker (Advanced) DOCKER_NETWORK, DOCKER_NETWORK_PRUNE,
DOCKER_SWITCHER
network_firewall (Advanced) UFW, UFWD, WHITELIST_PORT_UPDATER [new]
network_domains (field-Adv) SSLCERTS
security_ssh (Advanced) SSHKEY_DOWNLOADER, SSH_DISABLE_PASSWORDS,
BCRYPT_SAVE, GLUETUN_FOR_ALL [new]
general_terminal (Advanced) CRONTAB, CONFIGS_CHECK,
CONFIGS_AUTO_UPDATE, CONFIGS_AUTO_DELETE,
MISSING_IPS, CONTINUE_PROMPT,
SUGGEST_INSTALLS, SUGGEST_METRICS
general_install (Adv+DEV) CONFIG, COMMAND, WEBUI, WEBUI_SERVICE,
DATABASE, PASSWORDS, DOCKER_CE,
DOCKER_COMPOSE
The install-time eight are marked **ADVANCED** **DEV** — invisible
unless Developer Mode is on AND "Show Advanced Options" is expanded.
Each field's description was updated to note "Disabling on an existing
install will brick the system" / "install-time choice only" so a user
who does get to the toggle understands the gun before pulling the
trigger.
Other cleanup that fell out:
- Removed `configs/features/` directory entirely.
- Added the two new subcategories to SUBCATEGORY_ORDER in
network/.category and security/.category.
- Dropped the `category === 'features'` Danger Zone header special-case
in config-manager.js and its .danger-zone-section--header-only CSS
variant (sole user).
- Trimmed an obsolete "Edit the features config" notice in
check_requirements.sh.
Signed-off-by: librelad <librelad@digitalangels.vip>
Extends the install-routing spike (e5273a4) to every long-running CLI
command, so CLI and WebUI now share one execution path everywhere:
app install ← already done
app uninstall
app start / stop / restart / up / down / reload
app backup
app restore
update apply
backup app create (matches `app backup` — same end target)
Each handler now has the same shape:
if [[ "$LIBREPORTAL_TASK_EXEC" == "1" ]]; then
<inline call> # processor's recursive invocation
else
cliTaskRun "<cmd>" <type> <app> # user invocation: enqueue + follow
fi
Processor change — crontab_task_processor.sh:
Adds `export LIBREPORTAL_TASK_EXEC=1` next to LIBREPORTAL_NONINTERACTIVE.
Universal bypass: every task command the processor runs (CLI-queued OR
pre-existing WebUI-queued like `libreportal app install adguard`)
inherits the env var, so the inline branch fires and we never
re-enqueue. This also lets us drop the env-var prefix the install spike
was baking into the command string (e5273a4) — cleaner task files +
one place to think about the bypass.
`backup app schedule` (the cron-driven path that already enqueues via
createTaskFile in backup_app_schedule.sh) is left alone — different
entry point, different runtime context, already correctly task-routed.
Why route the fast ones too (start/stop/restart/up/down):
Consistency beats the ~1s task-roundtrip latency for a CLI button.
Locking now serialises a CLI `app stop foo` against a WebUI restart of
the same app; the audit trail covers every state change. Cheap to
revert any individually if the latency turns out to bother someone.
Validated live earlier with `libreportal app install dashy` — task file
written, processor dispatched, follower streamed install live, exit 0
propagated. Same machinery now powers the other 9 handlers.
Signed-off-by: librelad <librelad@digitalangels.vip>
Spike — closes the gap where the CLI install bypassed the very task system
the WebUI uses. Now both surfaces hit the same path:
user types `libreportal app install dashy`
→ CLI enqueues a task file in $TASK_DIR (identical shape to the
WebUI's createTaskFile)
→ pokes $TASK_DIR/.queue.fifo so the processor dispatches in <100ms
instead of waiting up to IDLE_POLL_SECS
→ CLI tails the task log + polls .status, exits with the task's
exit_code on terminal state
→ Ctrl-C detaches the follower without killing the task — the
WebUI's tasks panel keeps showing it
Bypass: the recursive command in the task file is prefixed
`LIBREPORTAL_TASK_EXEC=1 libreportal app install <name>`. The install
branch in cli_app_commands.sh honours that env var by running inline,
which is what the processor's eval invocation hits. No processor
changes — the bypass travels with the task.
Wins:
- one log file per install, shared by CLI + WebUI (audit trail + replay)
- locking serialises CLI + WebUI installs (no more two-frontend race)
- WebUI's "current task" indicator now reflects CLI work too
- free `--detach` for fire-and-forget queueing
New: scripts/cli/task/cli_task_run.sh
cliTaskRun <cmd> [type] [app] [--detach]
Enqueues + follows; --detach prints the task id and exits 0.
cliTaskFollow <task_id>
`tail -F` the log + jq-poll the status; returns the task's exit_code.
Designed to be reused for `libreportal task log <id>` reattach later.
Trade-off: ~200-500ms latency before the first byte (write task file,
processor wakes, opens log, follower starts tailing). Negligible for
install/update/backup — fast commands (list/status/config get) still
run inline. The current branch only changes `app install`; uninstall +
update + backup can be moved on the same pattern once this lands clean.
Signed-off-by: librelad <librelad@digitalangels.vip>
Two unrelated UI polish items.
1. Task notification titles. The "Config_update task completed!" toast
was leaking the literal task-type id because the friendly-name map
in the taskCompleted listener didn't list `config_update`,
`update_config`, or `system_update`, and the fallback only
capitalized the first letter. Same fallback was duplicated in
task-actions.js's started-toast path.
Extracted into `TasksManager.formatActionTitle(action)`:
- Adds the missing entries (`config_update`/`update_config` → "Update
Config", `system_update` → "Update System").
- Smarter fallback: snake/kebab → Title Case, so an unmapped future
type renders as e.g. "Foo Bar" instead of "Foo_bar".
- Both the started (task-actions.js) and completed/failed/cancelled
(tasks-manager.js) notification paths now route through it, so the
started/done pair always reads the same.
2. Dev-mode strip styling. Earlier amber-on-amber recipe read as a
warning state; the strip is just informational. Switched to a
neutral glass surface (rgba(text,0.04) + rgba(text,0.12) border),
text-primary copy, and the icon alone in var(--accent). Label is
now centered with the close button absolute-positioned on the
right (24px / 56px right padding so the centered text never sits
under the X).
Signed-off-by: librelad <librelad@digitalangels.vip>
Two small dev-mode UX changes.
1. Banner. When CFG_DEV_MODE is on, a 36px amber-tinted strip sits flush
under the topbar with "You are currently running in Developer mode"
and a dismiss X. Dismissal is remembered in localStorage and cleared
whenever dev mode is toggled back on, so re-enabling the mode brings
the banner back. Body picks up `.has-dev-banner` while visible to
bump padding-top by the strip's height (also adjusts the mobile
drawer's top/height).
2. Toast. The auto-enable message dropped the trailing
"Click the LibrePortal logo 10× to disable." — too noisy on every
git/local page load; the easter egg is still discoverable. New
message is just "Developer mode auto-enabled — you're on a <mode>
install."
Signed-off-by: librelad <librelad@digitalangels.vip>
Finishes the installApp refactor started in d941f59 (Wave A). Every app
whose <app>.sh was either pure boilerplate (Wave B) or boilerplate +
small custom logic (Wave C) now routes through the generic driver in
scripts/app/install/app_install.sh; bespoke logic moved to declarative
hooks in containers/<app>/scripts/<app>_install_hooks.sh.
Net: ~4,000 lines of duplicated 10-step sequence gone. From 31 per-app
.sh files (pre-Wave-A) down to 2 intentional keepers.
DELETED outright (pure boilerplate — driver replaces them identically):
jellyfin, mastodon, focalboard, ipinfo, speedtest, dashy, invidious,
nextcloud, ollama, vaultwarden, pihole
DELETED + hook-extracted (small bespoke step preserved in a hook):
bookstack, moneyapp, owncloud, trilium, searxng, gitea, headscale,
unbound, prometheus, grafana, gluetun, wireguard, jitsimeet, authelia,
traefik, adguard, onlyoffice
KEPT (intentional special cases):
crowdsec — host-app pattern (no docker compose, runs as apt+
systemd via installCrowdsecHost; uninstall/stop/
restart hooks already live in this file and are
invoked by dockerUninstall/Stop/RestartApp directly).
libreportal — WebUI bootstrap. Pre-compose image build + post-install
webuiLibrePortalUpdate + bootstrap-time suppression of
menuShowFinalMessages don't fit the generic flow.
Driver change — scripts/app/install/app_install.sh:
Moved monitoringToggleAppConfig "$app_name" "docker-compose.yml" from
the post-start integrations block into the install body at post-compose
(right after dockerComposeSetupFile, before docker-compose up). The
toggle edits the compose file on disk — running it after start meant
the container had already been brought up with the unmodified compose,
so the metrics endpoint wouldn't reflect CFG_<APP>_MONITORING until
the next restart. Matches the original ordering in every per-app .sh
that used to call it inline.
Hook surface (declare-f-gated, silent no-op when absent):
<slug>_install_pre before any install work
<slug>_install_post_setup after dockerConfigSetupToContainer
<slug>_install_post_compose after dockerComposeSetupFile (+ the
shared monitoring toggle on the compose)
<slug>_install_post_start after dockerComposeUpdateAndStartApp
<slug>_install_message_data echoes extra argv for menuShowFinalMessages
<slug>_install_post very last thing, after the final message
+ the existing _uninstall_pre/_post, _stop_post, _restart_post
Notable extractions:
bookstack — _install_post_start: probe :PORT_1/login until 200/302,
then `bookstack:create-admin` inside the container with
CFG_BOOKSTACK_ADMIN_{EMAIL,PASSWORD}; falls back to the
seeded admin@admin.com on timeout.
adguard — _install_post_start drives the wizard's HTTP API
(POST /control/install/configure) so the admin doesn't
click through five pages, then pins the admin bind back
to 0.0.0.0:3000 (matches the compose mapping) and health
checks. _install_message_data echoes user/password to
menuShowFinalMessages.
authelia — _install_pre requirements; _install_post_compose copies
configuration.yml + users_database.yml, substitutes
theme/domain/host, generates JWT/session/storage secrets,
toggles monitoring on configuration.yml; _install_post_start
argon2-hashes the admin password via the container, writes
users_database.yml, restarts; _install_post echoes creds.
traefik — _install_pre prompts for the LE email if CFG_TRAEFIK_EMAIL
is unset; _install_post_compose copies static + dynamic
configs, wires CFG_TRAEFIK_DASHBOARD_ACCESS (local-only /
domain-only / public), toggles monitoring on traefik.yml,
then traefikUpdateWhitelist + traefikSetupLoginCredentials.
wireguard — _install_pre host-conflict guard (/etc/wireguard/params);
_install_post_compose persists CFG_WIREGUARD_SUBNET,
resolves WG_HOST (domain+traefik → host_setup, else IP),
runs runAppCfg wireguard-ip-forward; _install_post_start
restarts after wg-easy installs its iptables rules.
jitsimeet — _install_post_setup downloads the tagged release zip from
GitHub; _install_post_compose mass-edits the .env and runs
gen-passwords.sh; _install_post_start rewrites nginx
default site to usedport1/2 + restart.
prometheus — _install_post_compose seeds prometheus.yml under
$containers_dir/prometheus/prometheus/; _install_post_start
sets 0777 on storage dirs so the container TSDB can write
regardless of host UID mapping.
grafana — _install_pre requirements; _install_post_start 0777 on
grafana_storage.
gluetun — _install_post_start refreshes the provider snapshot,
reattaches every routed app (the netns container ID is
stale after gluetun gets recreated), then prompts to
onboard any existing apps.
+ the smaller bookstack-shape extractions for owncloud (version scrape),
trilium / searxng (wait-for-first-boot-config), gitea (Prometheus
bearer token sync), headscale / unbound (config copy), moneyapp
(Auth.js AUTH_URL), onlyoffice (compose-resolved user/pass into the
final message).
Manifest + arrays regenerated. Verified end-to-end:
- bash -n on every hook file + the driver: clean
- Each hook file sources cleanly in a subshell, exposes only the
intended functions, flagged lazy-loadable (not eager)
- Smoke-stubbed install run for jellyfin (pure), nextcloud (pure),
bookstack (hooked), crowdsec (kept): correct dispatch in all cases —
deleted apps route to installApp, kept apps still hit their real
function
Signed-off-by: librelad <librelad@digitalangels.vip>
routing-manager.js read CFG_<APP>_HOST_NAME for its preview URL, but that
key was retired by the per-port subdomain refactor (2e4f420, 2026-05-22)
and no .config defines it anymore. The lookup always returned undefined,
so even with a configured domain the preview fell through to the
`<your-domain>` placeholder instead of showing the real host.
Now derives the preview from the port's own subdomain (parts[10] of
the 12-col PORT row), matching the canonical host_setup rule in
scripts/network/variables/variables_init_app.sh:
@ / root -> apex (`https://<domain>`)
set -> `https://<sub>.<domain>`
empty -> `https://<app>.<domain>`
Also adds `subdomain` to the port object emitted by _collectPorts so
this and any future per-row consumer can read it.
Signed-off-by: librelad <librelad@digitalangels.vip>
The 31 containers/<app>/<app>.sh files each defined install<App>() with
the SAME 10-step sequence — ~4,000 lines of duplicated boilerplate.
Replaces all that with one generic driver + hook surface.
scripts/app/install/app_install.sh:
installApp <slug> [config_variables]
— Dispatches on $<slug> (c/u/s/r/i) the same way the per-app .sh
files did. Same convention; dockerInstallApp's existing
`declare $app=i` callsite needs no change.
— Runs the standard sequence: dockerConfigSetupToContainer →
dockerComposeSetupFile → optional .env copy → fixPermissions →
dockerComposeUpdateAndStartApp → standard post-install steps
(appUpdateSpecifics, setupHeadscale, databaseInstallApp,
webuiContainerSetup, monitoring registration) → final message.
— Hooks (all declare-f-gated, silent no-op when absent):
<slug>_install_pre / _post_setup / _post_compose / _post_start
<slug>_install_message_data (echoes extra args for menu)
<slug>_install_post
<slug>_uninstall_pre / _post
<slug>_stop_post
<slug>_restart_post
Hooks live in containers/<app>/tools/<app>_tools.sh (auto-sourced
per the modular-per-app-tools convention).
function_install_app.sh:
When no install<App>() function exists, fall through to
`installApp <app_name>` instead of erroring. So an app with no .sh
at all becomes a zero-byte addition — drop in <app>.config +
docker-compose.yml + <app>.svg, done.
containers/linkding/linkding.sh:
Deleted (canary). Linkding's body was 100% standard sequence;
fallback handles it identically. Smoke-tested with stubbed helpers
— dispatcher fires, generic runs full flow, monitoring integration
+ final-message hook plumbing all intact.
Wave B (next): delete the .sh for every other 'pure-boilerplate' app
(~15 candidates per the survey). Wave C: extract custom logic from
the 7 fat apps into hooks before deleting their .sh.
Signed-off-by: librelad <librelad@digitalangels.vip>
The app-detail page was the last corner of the SPA still using query
parameters for navigation state. Two related complaints surfaced it:
- `/app/adguard?tab=tasks` should mirror admin (`/admin/tools/peers`,
`/admin/config/network`) and be `/app/adguard/tasks`.
- The config sub-tab (general / advanced / features / network / …)
had no URL representation at all — `showTab` was a pure visual
swap with no history push, so refreshing a deep config sub-tab
sent the user back to the default first category.
New URL shape:
/app/<name> → config tab, default sub-tab
/app/<name>/<tab> → non-config main tab (tasks, backups, …)
/app/<name>/config/<category> → config tab + specific sub-tab
…?task=<id> → optional deep-link to a single task
Mirrors `adminPath` / `adminCategoryFromPath`. Two new helpers in
spa.js carry the convention:
window.appPath(name, tab, sub, taskId) → URL
window.appPartsFromPath(pathname) → { app, tab, sub }
Every URL constructor in the WebUI was replaced with `window.appPath`:
spa.js — handleAppDetail back-compat redirect
app-tabbed-manager.js — getTabFromURL + new getConfigSubFromURL
(path first, ?tab= fallback for legacy)
updateURL + updateApp use appPath
the inline task-deep-link constructor
apps-manager.js — showAppDetail + showAppDetailWithConfig
showTab now pushes /app/<n>/config/<sub>
renderAppDetail picks the sub-tab out of
the URL on first load
4 fallback task-URL constructors
tasks-manager.js — completion-notification URL
task-actions.js — start-notification URL
notifications.js — 2 task deep-link URLs
Back-compat: handleAppDetail detects legacy `?tab=` / `?config=` /
`?task=` queries and replaceState()s the URL to the canonical path
shape BEFORE anything else reads URL state — old bookmarks land on
the right page and end up with a clean URL.
Verified by running every appPath / appPartsFromPath case (including
the `logs` → `tasks` legacy alias) and confirming the round-trip is
identity. JS syntax checks clean on all six files. No remaining
hardcoded `/app/<x>?tab=` strings outside the back-compat comment.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The advanced field was carrying a four-clause explanation walking through
"quiesce", "live dump", "stop fallback", and the live-option eligibility
rules — useful background for engineers, overwhelming as a hover tooltip
for the typical user picking from the dropdown.
New tooltip:
"How the app is prepared before each backup. Automatic is recommended;
pick Live only if you need zero downtime."
Two lines, plain language, points at the right default. The dropdown's
own labels (Automatic (recommended) / Stop→snapshot→start / Pause→snapshot
→unpause / Live — no downtime) carry the technical specifics; the
tooltip no longer has to.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
gluetun_providers.sh writes its working files ($raw, $headers) next to
$output_file, which lives at
containers_dir/libreportal/frontend/data/apps/generated/gluetun-providers.json
— dockerinstall-owned in rootless. The five rm -f calls on those paths
were unwrapped, so the manager running the script (e.g. from the
task processor) would get Permission denied — same class as the
updateConfigOption sed -i bug that was just fixed.
$tmp comes from mktemp (/tmp), so the rm -f for it stays unwrapped.
Audit context: this was the only remaining raw filesystem op against
container-tree paths in any containers/*/*.sh. The rest of the
container .sh files are clean — every sed -i / chmod / chown / cp / mv
is already routed through runFileOp or runFileWrite, and the
per-app install bodies delegate fs work to high-level helpers
(dockerConfigSetupToContainer, copyResource, dockerComposeSetupFile)
which themselves use the wrappers.
Hooks (<app>_migrate_pre/_post, restoreAppRunHook pre/post) are
present in the framework but unused by any app today — that's by
design (opt-in per-app). If a future app needs federation-key rotation
post-migrate, or a hostname rewrite that the generic URL-rewrite
layer doesn't cover, those slots are ready.
Signed-off-by: librelad <librelad@digitalangels.vip>
Started, completed, and failed task toasts were rendered by three
different code paths producing three different layouts:
• task-actions.js executeTask "App: AdGuard\n…task started!" (with type emoji)
• task-actions.js executeTaskMonitoring "App: AdGuard\n…task started!" (without type emoji) — dead code
• tasks-manager.js createAndExecuteTask "Task created: install adguard" (raw shape) — dead code
• tasks-manager.js complete/fail notif "App: AdGuard\n…task completed!" (with type emoji)
…plus the system-task path was reading the literal `'system'` slug into
the toast: "App: System / Config_update task started!" with a 404'd
/icons/apps/system.svg (the same bug renderTaskIcons had on the row
itself, fixed in 59ee92b).
Three changes:
1. Drop the "App: " / "System: " label prefix on every toast. The bold
line is now just the subject name (the row's title still carries the
semantic with its leading App-or-LibrePortal icon). Three tasks of
the same app no longer read like a column heading repeated.
2. Treat `appName === 'system'` as the LibrePortal sentinel everywhere
the toast renders — displayName resolves to "LibrePortal" and the
app-icon slot loads /icons/libreportal.svg. Mirrors the row-icon
fix in 59ee92b. The completion-path `isSystemTask` check now also
accepts `appName === 'system'` in addition to `setup-*` types.
3. Delete the dead code that produced the inconsistent shapes:
- executeTaskMonitoring in task-actions.js (no callers anywhere)
- window.createAndExecuteTask in tasks-manager.js (no callers; only
surviving reference was a stale comment in app-tabbed-manager.js,
updated to point at executeTask instead)
Net: every task toast in the WebUI now follows the same three-slot
layout — [type emoji] [app/LibrePortal logo] <strong>Name</strong> +
"Action task started/completed/failed/cancelled!".
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Symptom (reported on adguard install):
sed: couldn't open temporary file /libreportal-containers/adguard/sedZaGX2H:
Permission denied
✗ Error Updated CFG_ADGUARD_BACKUP to true
! Notice Non-interactive mode: aborting on error.
Root cause: updateConfigOption ran a raw `sed -i` regardless of which tree
the config file sits in. Works fine for /libreportal-system/configs/*
(manager-owned) but breaks on /libreportal-containers/<app>/<app>.config
(dockerinstall-owned in rootless mode) — sed -i writes its temp file next
to the target, inheriting the directory's perms, and the manager can't
write inside dockerinstall dirs. EVERY app installer that mutates a
CFG_<APP>_* value (the autogenerated random password, BACKUP toggle,
PORT override, etc.) goes through this function, so this was a latent
ticking bomb across all containers.
Fix: pick the helper based on path —
- under $containers_dir → runFileOp (escalates to
dockerinstall in rootless, runs as manager in rooted)
- otherwise → runInstallOp (always manager)
Read paths (grep / source) stay unwrapped — both dirs are world-readable;
only the write needs the privilege swap.
Net: no more 'Permission denied' on app installs; the de-sudo pattern is
now respected end-to-end for CFG writes.
Signed-off-by: librelad <librelad@digitalangels.vip>
Reshape the dashboard's Backup status grid into a click-to-pick UI:
- Removed the inline Back-up / Restore buttons from the System config
tile. Same shape as an app tile now; LibrePortal app icon instead of
the server-stack glyph.
- Grid is 2 columns (was auto-fill min 220px). Tiles are wider, read
better, and the System tile no longer needs to span a full row to fit
inline buttons.
- Click any tile (System or app) → opens a new "Back up" modal:
* System config first (key=__system__, LibrePortal icon)
* Every installed app, alphabetical
* Checkbox per row + 'Select all' / 'Clear' shortcuts
* The tile clicked is pre-ticked
- Confirm queues backup tasks:
* Everything ticked → single `libreportal backup all` (which also
runs `backup system`) — one task instead of N
* Subset → one task per ticked item (`backup system`
and/or `backup app create <slug>`)
Restore for System config used to live on the dashboard's inline
'Restore' button. It's now reachable via the Backups tab — system
snapshots appear in the snapshot list with the standard per-row
Restore action — same path apps already use. No new UI required;
just one fewer dashboard button.
Signed-off-by: librelad <librelad@digitalangels.vip>
The user Dashboard carried a small chevron link "Admin overview →" just
above the installed-apps grid. The topbar already has a top-level "Admin"
nav-item (topbar.html:34) that goes to the same /admin route. The
dashboard link was a redundant second entry point with no extra value;
removing it tightens the dashboard layout without losing navigation.
Drops:
- dashboard-content.html: the <a class="dashboard-admin-link"> block
- admin.css: the .dashboard-admin-link rule + :hover (now orphaned)
The /admin route, the topbar Admin nav-item, and the AdminOverview JS
component all stay as-is — only the dashboard-side entry point goes.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Was: 'What's saved. Save System config first — if anything breaks, you
need it to get everything else back.' — read a bit kindergarten.
Now: 'Latest backup per app + System config. Back up System first —
it's needed to restore the rest.' — same info, tighter, still
reads at a glance.
Signed-off-by: librelad <librelad@digitalangels.vip>
The two-line hint under 'Backup status' was redundant — the System
config tile speaks for itself once it's there. Replaced with an ℹ️
tooltip on the heading (same pattern as 'Cross-host migrate' on the
Migrate tab).
Tooltip text deliberately plain: 'What's saved. Save System config
first — if anything breaks, you need it to get everything else back.'
No 'bare-metal restore' jargon, no 'snapshot' — the kind of sentence
that lands for someone who's never heard of either.
Signed-off-by: librelad <librelad@digitalangels.vip>
The dashboard had two parallel sections — 'Per-app status' (every app's
latest backup) and a standalone 'System config' card below it. Folded
them into one grid: a single 'Backup status' card with the System config
tile rendered FIRST, then every app tile.
Why first: a bare-metal restore needs the system config (CFG_* +
backup-location credentials) — without it the backups exist but the
keys to reach them don't. Putting it at eye-level above the app tiles
makes the dependency visible.
System tile reuses the .backup-app-tile shape: server-stack icon,
'System config' as the name, status dot + 'Last backed up X ago' /
'No backup yet'. Plus two compact inline action buttons (Back up /
Restore) on the right that wire into the same data-action handlers
the old standalone card used — no behaviour change, just the visual
container.
grid-column: 1 / -1 on the system tile makes it span the row so the
two action buttons fit alongside the meta text without crushing the
app-tile grid template.
Section header: 'Per-app status' → 'Backup status' + hint 'System
config and every installed app's latest backup. System config always
first — a bare-metal restore needs it.' Dashboard subtitle updated
to match.
Signed-off-by: librelad <librelad@digitalangels.vip>
`libreportal config update` and `libreportal system update` tasks are
submitted with `task.app === 'system'` (see task-actions.js' configUpdate
+ systemUpdate). renderTaskIcons hit its first branch on any truthy
task.app and built an <img src="/icons/apps/system.svg"…> which 404s
(there is no per-app icon called "system"). The onerror handler then
hid the broken image, so those task rows showed only the 🛠️ type emoji
and no LibrePortal logo — visually inconsistent with sibling system-level
rows like "LibrePortal - Finalize Setup" (which happens to carry
`app: 'libreportal'`, matching a real icon, and renders correctly via
the same branch).
Treat `app: 'system'` as a category sentinel rather than a real slug:
skip the per-app icon path, fall through to the system-task branch that
loads /icons/libreportal.svg directly. That icon is already shipped + the
data shape stays intact ('system' is the meaningful category, not a lie
about the app identity).
Net: "LibrePortal - Apply Configuration" and "LibrePortal - System Update"
now show the LibrePortal logo alongside their type emoji, matching the
Setup / Update / Backup-All rows.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
formatCommandForUser was a 90-line if/else chain that grew one branch at
a time as new task shapes appeared. Three sites had escaped coverage and
fell through to the raw-command fallback:
libreportal config update '<changes>' → shown as the raw 50-char clip
libreportal peer add <name> <kind> … → same
libreportal regen webui --force → same
Restructure as a declarative `PATTERNS` array of `{match, title}` rows.
Each row is one regex + one title (string OR function for per-app rows
that extract the app slug). The matcher iterates once; first match wins.
Adding a new task shape is now a one-line append — no new code branch,
no copy-paste of the `if/match/return` boilerplate.
Behaviour-equivalent for every previously-formatted command (verified
by running 15 sample command strings through the new function against
the old expected titles); the three previously-broken ones now resolve:
libreportal config update CFG_DEV_MODE=true → "LibrePortal - Apply Configuration"
libreportal peer add Alice host … → "LibrePortal - Add Peer"
libreportal regen webui --force → "LibrePortal - Regenerate WebUI Data"
Plus a couple I noticed while in there:
libreportal backup system → "LibrePortal - Backup System Config"
libreportal peer remove / peer pair → friendly equivalents
The two non-table fall-throughs (the toolsCatalog-aware `app tool` lookup,
and the generic `libreportal app <action> <app>` map) stay inline since
they need richer logic than the table can carry — but everything else
lives in the one scannable list.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The 'No backups from other hosts visible…' empty state was rendering as
centred text inside the outer card, which read as floating prose rather
than a defined block. Wrapped it in a bordered callout (matches the
visual weight of the per-app task cards): rounded border, surface-2
background, padding, plus a centred location-pin glyph above the
message and the existing 'Open Locations' button as the CTA.
Inline styles so it works against the existing theme vars without
needing a new CSS rule.
Signed-off-by: librelad <librelad@digitalangels.vip>
`webuiSystemUpdate` was calling `webuiSystemMetrics` and getting
"command not found": the lazy-load manifest was missing it (and
`webuiSystemApps`), even though both are defined in
webui_system_metrics.sh.
Cause: the manifest scanner's awk-based depth tracker is line-based.
The first function in that file, `_metricsReadCpu`, contains an
embedded awk script:
_metricsReadCpu() {
awk '/^cpu /{ <-- the open brace runs to end-of-line, depth += 1
...
}' /proc/stat <-- the close brace is mid-line, NOT counted
} <-- depth-- but we started one too high
The increment rule fired for the `/^cpu /{` line (open brace at EOL),
the matching close on the `}' /proc/stat` line was not counted because
the existing depth>0 branch only decremented when stripped was EXACTLY
`}`. Result: depth ended at 1 instead of 0 after _metricsReadCpu, and
every subsequent funcname() header was treated as "still inside a
function" and silently dropped. Same pattern across the ~6 files the
lazy-loader memory called out as "false-positive eager".
Fix: in the depth>0 branch, count open AND close braces per line
symmetrically and apply the net delta. Drops the redundant "exactly
`}`" special-case rule, clamps depth at 0 as a sanity backstop. False
positives from braces inside string literals could under/over-count
locally, but the clamp + the next legitimate `funcname()` header at
depth-0 re-anchors the tracker.
Result on a full regen:
- 858 → 876 functions indexed (+18 previously-stranded)
- webuiSystemMetrics + webuiSystemApps now correctly autoloaded
- eager-file count: 9 → 11 (two files that genuinely have both
function defs and top-level side effects are now correctly seen
both as eager AND get their functions indexed — net win on
every axis)
Verified live: the WebUI updater that was failing with "command not
found" now prints "Updated system information..." cleanly.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Symptom: after any commit / deploy on this box, the WebUI would log
users out ~60 seconds after they logged back in. Looked like a
short session timeout; was actually the auth file being deleted.
Cause: my recent update.sh change added --delete to the frontend
rsync so source-tree file removals propagate to the live install.
Excludes only protected data/. .auth.json sits at the top of
frontend/ (never in the source repo — it's the persisted credentials
+ JWT secret), so --delete nuked it on every deploy. The next
container start regenerated it with a fresh secret; all existing
cookies (signed with the old secret) became invalid. The dashboard's
60-second auto-refresh hits /data/system/*.json which is auth-gated,
gets 401, and the global 401 interceptor in auth-manager.js shows
the re-login overlay. Hence 'logged out after 60 seconds'.
Fix: extend the rsync exclude list with:
--exclude '.*' (any top-level dotfile — covers .auth.json
and future runtime state of the same shape)
--exclude '*.lock' (lockfiles like setup.lock if any ever land
outside data/)
--exclude '*.bak' (backup files from manual edits)
data/ exclude kept. JWT lifetime stays at 30 days as designed.
Also: feat(webui): icon on the 'Open Locations' button in the
backup → Migrate tab's empty state. Matches the location-pin icon
used by the sidebar's Locations entry so the visual carries over
when the user clicks through.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Migrate tab carried two walls of explanation text — a 3-line hint
under the h2 ("Pulls a snapshot taken on another host…") and an even
longer empty-state paragraph ("Either no other LibrePortal has backed
up to a location this host can see, or this is the only host using its
locations…"). Both spelled out diagnosis the user can infer from the
empty list itself, and the tone didn't match the rest of the backup
page (cards elsewhere have a short title + a 4-6 word hint, with any
long explanation as a hover title attribute).
Three changes:
1. h2 down to "Cross-host migrate" with a small ℹ️ carrying the full
explanation as a title= tooltip — matches the existing tooltip
pattern in the Locations form (BACKUP_RETENTION_PRESET_META).
The short subtitle "Restore an app from another LibrePortal" stays
as backup-card-hint, mirroring "Per-app status / Latest backup per
app on this host" elsewhere on the page.
2. The empty state is now the standard `<div class="backup-empty-state">`
container (same shape Locations + Snapshots use), one trimmed line
("No backups from other hosts visible in any enabled location.
Add a shared backup location on both hosts to enable cross-host
migrate.") instead of two paragraphs.
3. Added an "Open Locations" CTA button inside the empty state — the
#1 next-step for a user staring at this empty list is to add a
shared location, which lives one tab over. New data-action
"go-to-locations" wired through the existing event-delegation
handler in backup-page.js calling switchTab('locations').
The renderMigrate JS still toggles #backup-migrate-empty.hidden — the
wrapper id is unchanged, only its inner markup tightened. No
behavioural change beyond the CTA + tab switch.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Existing installs were locked out of template-side description /
options / marker changes because reconciliation kept the user's whole
`CFG_KEY=value # comment` line verbatim. So new metadata like the
**DEV** marker I'm adding for the developer-mode feature wouldn't take
effect on already-deployed boxes — only fresh installs.
Updated reconcileConfigFile to split each line into value-part and
comment-part. User value is still sacred; the comment (title,
description, [options], **ADVANCED**/**DEV** markers) now comes from
the template. Field renames, label tweaks, marker additions/removals
shipped in a release reach existing installs on the next CLI
invocation (which runs the reconciler).
Specifically unblocks: the developer-mode WebUI feature (CFG_DEV_MODE
field gets added by the existing add-only path; CFG_INSTALL_MODE and
CFG_RELEASE_CHANNEL now pick up their new **DEV** markers and the
'Release - Stable' / 'Bleeding Edge' labels).
Signed-off-by: librelad <librelad@digitalangels.vip>
What this delivers (Stage 1+2 of the dev-mode feature):
1. New `**DEV**` marker for config fields. Mirrors the existing
`**ADVANCED**` pattern: stays in the description string, frontend
strips it for display, presence flips a 'hide unless dev mode is on'
behaviour. Implemented in ConfigUtils.cleanDescription /
isDevField / isDevModeOn and in ConfigShared._filterDevKeys, which
the two generateFieldsForCategory* helpers now call before rendering.
2. New CFG_DEV_MODE field in configs/general/general_install. Visible
under Advanced; defaults to false. The canonical place to toggle
dev mode (the WebUI easter egg writes to it, the auto-detector
writes to it, and users can flip it directly here too).
3. Marked CFG_INSTALL_MODE and CFG_RELEASE_CHANNEL with `**DEV**`.
Normal users no longer see either field — they install Release-
Stable and that's the whole story. Devs see both with the
user-facing labels you asked for:
CFG_INSTALL_MODE Release - Stable | Git clone | Local folder
CFG_RELEASE_CHANNEL Release - Stable | Release - Bleeding Edge
(CFG_INSTALL_MODE label for the release option also renamed to match.)
4. 10-click LibrePortal-logo easter egg in topbar.js:
- Counter on any .libreportal-logo click; idle-reset after 3 s
- Toast countdown from click 6 ('4 clicks away from being a developer…')
- At 10: toggles CFG_DEV_MODE via the standard config_update task
(same path the Config form uses); shows '🛠️ Developer mode
unlocked. Reload to see the extra options.'
- Re-using the same logo when dev mode is on toggles it back off
('… away from disabling developer mode') — symmetric, no separate UI
5. Auto-detect: on every WebUI load, if CFG_INSTALL_MODE is git or
local AND CFG_DEV_MODE is off, auto-flip to on with a one-time
toast 'Developer mode auto-enabled — you're on a git install.
Click the LibrePortal logo 10× to disable.' Stops dev-install
users getting locked out of the very options they need to manage
their install. Idempotent — runs once per page load; no-op if
already on or on release.
Disable surfaces: (a) CFG_DEV_MODE in Advanced on the Config form is
the canonical toggle; (b) 10 more logo clicks. A 3rd surface (a System
page banner) is deferred — those two cover the practical cases.
Signed-off-by: librelad <librelad@digitalangels.vip>
`libreportal app generate <name>` (and the menu's "g. Generate App" entry)
was broken three independent ways and incompatible with the per-app
architecture the project actually uses now:
1. Copies from $install_containers_dir/template/ which doesn't exist —
the only template/ in the tree was in scripts/unused/OLD_CONTAINERS/
and was never installed into the live tree. cp -r would just fail.
2. Every sed call used BSD/macOS syntax `sed -i '' -e …`. On Linux
(every distro this targets) the empty '' becomes a positional file
argument, so the substitutions never ran. 8 calls, all broken.
3. Even if it had run, the produced skeleton would have been a
pre-modular-tools / pre-per-port-subdomain app shape: no tools/,
no scripts/ subdir, HOST_NAME=test in the .config. Every active
containers/<app>/ today carries the modular layout the rest of the
framework expects.
Plus the recent cleanups (the prompt loop fix in 9ffc8e4, the per-port
subdomain refactor in 2e4f420) had been peeling pieces off it without
the root question — does the function still belong? — getting asked.
Delete the whole surface:
- scripts/app/app_generate.sh (157 lines, the function body)
- scripts/unused/OLD_CONTAINERS/template/ (the never-installed source
files appGenerate would have copied — stale enough to still carry
HOST_NAME=test, CFG_<X>_HOST_NAME, and 248 lines of compose template)
- menu entry "g. Generate App" + its dispatch in menu_main.sh
- "generate" case branch in cli_app_commands.sh
- `libreportal app generate` line in cli_app_header.sh
- The corresponding entries auto-drop from files_app.sh +
function_manifest.sh via regen.
New apps are added the way the catalog already grew — by hand-crafting
containers/<app>/{<app>.sh, <app>.config, docker-compose.yml,
tools/<app>.tools.json, scripts/<app>_*.sh}. Copying an existing app's
folder + renaming is the closest thing to a "generator" and it's a one-
command operation.
Net: -556 lines, no behaviour lost (the function never worked).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
initilize_files.sh's load-comment said "Phase 6 will tackle this with a
precompiled cache file" — but Phase 6 deliberately shipped a different
(simpler) win, and Phase 7 was reconsidered and rejected (the ~30 ms
saving doesn't justify the invalidation complexity across every CFG
write site). Rewrite the comment to describe current behaviour: the two
config scans run because bash can't lazy-load variables, and the
precompile was looked at and dropped.
app_generate.sh had a `read -p "" host_name` inside a `while true` with
no break — anyone who actually ran `libreportal app generate` would
have hung at the hostname prompt forever. The value was then fed into
a `sed 's/HOST_NAME=test/HOST_NAME='"$host_name"'/g'` against the new
app's .config file, but the post-2e4f420 template no longer carries
HOST_NAME=test (per-port subdomains in the PORT row drive routing now).
So the prompt was infinite-looping to gather a value that fed a no-op
sed. Drop both — the function loses no real capability since the
subdomain field is set per-port via the standard PORT row.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The installer (rootless_docker.sh:123) already defaulted CFG_ROOTLESS_NET
to pasta when unset — but the bundled configs/network/network_rootless
shipped CFG_ROOTLESS_NET=slirp4netns with a description warning about
the AppArmor caveat. That made the WebUI Config page surface slirp4netns
as the selected option even though the install script preferred pasta if
unset, and the warning told users they'd have to hand-relax the AppArmor
profile if they switched.
Both are now obsolete:
- CFG_ROOTLESS_NET=pasta is now the explicit default in the bundled
config (matches the installer's implicit default).
- Description drops the AppArmor manual-fix warning since the
installer applies the local override automatically
(installRootlessApparmorForPasta, shipped in the previous commit).
Dropdown order swapped too — pasta now top of the list as the
recommended option, slirp4netns kept as 'legacy fallback'.
The live install on this box already runs pasta (manually flipped
during debugging); the CFG file was synced to match so a future
rootless reinstall doesn't revert.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Debian-shipped passt AppArmor profile (/etc/apparmor.d/usr.bin.passt)
denies the accesses pasta needs to plumb rootlesskit's netns:
- ptrace_read on the rootlesskit child to enter its user namespace
- read /run/user/<uid>/dockerd-rootless/netns (the netns file)
- read /proc/<pid>/net/{tcp,tcp6,udp,udp6} for implicit port forwarding
Without these the rootless docker daemon fails with:
pasta failed with exit code 1:
Couldn't open user namespace /proc/<pid>/ns/user: Permission denied
scripts/docker/install/rootless/rootless_apparmor.sh:
New installRootlessApparmorForPasta() — idempotent fixup.
1. Adds `include if exists <local/usr.bin.passt>` to the main profile
(one line; re-adding is a no-op via grep).
2. Writes /etc/apparmor.d/local/usr.bin.passt with the four rules
pasta needs. The /local/ pattern is the standard Debian AppArmor
hook for site-managed overrides — survives `apt upgrade passt`
because it's outside the package's managed paths.
3. Reloads via apparmor_parser -r.
Called from installDockerRootless after the override.conf write, gated
on $rootless_net == pasta. slirp4netns installs skip it.
This box was already manually patched while debugging the pasta swap —
the installer-side change makes it idempotent across reinstalls and
applies the same fix on any other host that installs rootless docker
with pasta as the net driver.
Signed-off-by: librelad <librelad@digitalangels.vip>
The 1h max-age set in Phase A caused a cache-vs-deploy mismatch when
Phase B refactored config-manager.js to lazy-load admin-overview.js et
al. The new index.html no longer eager-loads those scripts, but
browsers with the cached (pre-Phase-B) config-manager.js didn't do the
lazy-load either — so AdminOverview / AdminSystem / etc. were
undefined and the admin tools rendered 'failed to load' errors.
60s is the right balance: rapid in-session clicks skip the network
round-trip, but a deploy is visible within a minute. ETag-based 304s
still keep the per-request cost tiny when nothing changed.
Signed-off-by: librelad <librelad@digitalangels.vip>
Two webui-data generators wrote to a temp file via bare `cat > "$temp_file"`
then `runFileOp mv` to the final path. The temp file's path sits inside
$containers_dir/libreportal/frontend/data/<x>/generated/ — owned by the
dockerinstall user (the data plane). The generators run as the manager,
who can't open paths under that tree for write, so every WebUI update
hit:
webui_backup_migrate.sh: line 125: …/migrate.json.tmp.<pid>: Permission denied
mv: cannot stat '…/migrate.json.tmp.<pid>': No such file or directory
webui_peers.sh: line 23: …/peers.json.tmp.<pid>: Permission denied
mv: cannot stat '…/peers.json.tmp.<pid>': No such file or directory
Pipe the heredoc through `runFileWrite "$output_file"` instead — same
shape the 5 sibling generators in this dir (backup_app_status,
backup_locations, backup_passwords, backup_snapshots, backup_dashboard)
already use. runFileWrite routes the write via the install user that
owns the data tree, so the file lands on disk in one step (no temp +
mv dance needed). The unused `local temp_file=...` declarations dropped
out cleanly.
The trailing `runFileOp chmod 644 "$output_file"` stays — it's the only
guarantee the WebUI container (which reads these files RO) sees them as
world-readable regardless of dockerinstall's umask.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
initUpdateConfigOption (init.sh) and commandUpdateConfigOption (the CLI
wrapper heredoc) both rewrite CFG_<NAME>= lines with a sed s-command
using `|` as the delimiter. The escaping covered only `/` and `&` in
$escaped_value, and $comment_part wasn't escaped at all — so any line
whose comment contains a literal `|` blew up the substitution:
sed: -e expression #1, char 167: unknown option to `s'
The trigger in the install log is the CFG_INSTALL_MODE comment:
# Installation Mode - ... [release:Release tarball (recommended)|git:Git clone (dev)|local:Local folder (dev)]
Two sed errors in install-20260526-223006.log, both same line — once
from initUpdateConfigOption during the initial-values pass, once from
the CFG_INSTALL_MODE re-set later. The substitution silently failed
(line not rewritten) and the install continued.
Switch the delimiter to SOH (\x01). Text-based config values + comments
never contain that byte, so the delimiter never needs to be escaped.
Only `&` (whole-match insertion in the replacement) and `\` (escape
char) remain hazardous in the replacement field, and BOTH are now
neutralised in $escaped_value AND in $comment_part.
Verified against the actual offending line: the old form reproduces
`sed: unknown option to 's'` at char 165; the new form rewrites cleanly
with every `|` in the comment preserved.
Same fix applied to both functions — initUpdateConfigOption lives at
install-time, commandUpdateConfigOption is baked into the CLI wrapper
at /usr/local/lib/libreportal/libreportal; new installs pick up both
from this commit.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The wall-of-dashes "✗ Error This PERMANENTLY removes EVERYTHING" listing
made the most consequential prompt in the project look like a routine
error log: same icon as a failed command, unaligned columns, no visual
grouping. Replaced with a structured block:
- Single red ⚠ "PERMANENT — there is no undo" callout (instead of the
✗ "Error" prefix, which semantically means a thing failed — this is
a pre-action warning).
- Four bold section headings (Filesystem / Users / System integration
/ Containers + binaries) so the reader can scan by category.
- Aligned %-34s path column with dim trailing descriptors — the eye
can sweep the left edge without re-anchoring per line.
- Green "Left in place:" reassurance lands at the end (same content as
before, just promoted from two isNotice lines into one styled line).
Pure-presentation change — no behavioural difference, same destroy list,
same DELETE LIBREPORTAL prompt. Verified the printf format renders
cleanly with the colour vars from variables.sh.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
7 page-specific controllers were eager-loaded in index.html on every cold
visit, even when the user lands on /dashboard and never opens /backup,
/admin, etc. Moved them to lazy-load via spa.js's existing loadScript()
helper, fired from each route's handler on first navigation:
/js/components/backup/backup-page.js — handleBackup()
/js/components/backup/backup-app-card.js — handleBackup()
/js/components/ssh/ssh-page.js — config-manager ssh-access
/js/components/peers/peers-page.js — config-manager peers
/js/components/admin/admin-overview.js — config-manager overview
/js/components/admin/charts.js — config-manager overview
/js/components/admin/admin-system.js — config-manager system
config-manager.js gets a tiny `lazyLoad` helper that delegates to
window.spaClean.loadScript with a graceful fallback when the SPA hasn't
booted (legacy paths). loadScript is idempotent — subsequent visits to
the same route are no-ops, so we don't re-fetch after the first nav.
Cold-load impact on /dashboard (the most common landing):
Before: 25 sync <script> tags loading ~1.7 MB raw / ~430 KB gzipped
After: 18 sync <script> tags loading ~1.5 MB raw / ~380 KB gzipped
+ corresponding parse-cost reduction on the client (no longer parsing
backup-page.js + apps-related JS just to render the dashboard)
Page-specific JS still loads cleanly when the user navigates there — a
single extra network round-trip per route on first visit, then cached
for 1h (per Phase A's cache headers). Compression (Phase A) means the
deferred JS is ~75 % smaller on the wire than it would have been
pre-Phase-A.
Sister update to .../Scripts/update.sh: rsync now uses --delete so
file removals in the source tree (this commit deletes 7 script tags;
earlier commits deleted config-manager-old.js) propagate to the live
install. Excludes still protect frontend/data/.
Signed-off-by: librelad <librelad@digitalangels.vip>
Three WebUI cold-load wins:
1. DELETED containers/libreportal/frontend/js/components/config/config-manager-old.js
66 KB / 68189 bytes. Zero references anywhere in source or deployed
tree (confirmed via grep across containers/libreportal/). Pure dead
code from a previous refactor — removed.
2. ADDED `compression` middleware (defensive require)
Gzip-compresses JS/CSS/HTML/JSON responses. Typical ~70 % wire-size
reduction → the 1.7 MB cold-load drops to ~500 KB. New package.json
dependency; container's node_modules is baked into the image so the
require is wrapped in try/catch to degrade silently until the image
is next rebuilt (libreportal app install libreportal, or a full
deploy). Once active: free wire-size win on every response.
3. ADDED static cache headers via staticOptions on express.static
- JS/CSS/icons: Cache-Control: max-age=3600 + ETag
(1h browser cache, cheap 304 revalidation after)
- HTML files: Cache-Control: no-cache + ETag
(always revalidates so SPA shell updates land
immediately after a deploy; 304 if unchanged)
Repeat navigation in the same browser session skips ~25 script-tag
round-trips entirely.
Net effect once compression deploys:
- Cold load: 1.7 MB → ~500 KB on the wire (~70 % shrink)
- Warm load: 25 conditional requests → 0 (served from cache for 1h)
- Deploy lands: HTML revalidates immediately, JS/CSS picks up after 1h
or hard refresh
Phase B (defer non-critical scripts via SPA loadScript) and Phase C
(rebuild image / split the bind-mount story for node_modules) come
next; this commit is the safe Phase A foundation.
Signed-off-by: librelad <librelad@digitalangels.vip>
CrowdSec's host-side install (the agent + nftables bouncer the LibrePortal
Traefik plugin talks to) had stayed on blanket sudo throughout the rootless +
de-sudo hardening: `sudo apt-get install crowdsec`, `curl | sudo bash`,
`sudo sed -i /etc/crowdsec/config.yaml`, `sudo touch + sudo chmod /var/log/
crowdsec*.log`, `echo $key | sudo tee /etc/crowdsec/traefik_bouncer.key`,
plus `sudo cscli capi register / console enroll / bouncers add`. None of
those are in the scoped LP_HELPERS / LP_SYSTEM sudoers grant the manager
now holds, so any user who enabled crowdsec would have hit hard sudo
failures on every privileged step.
Follow the libreportal-appcfg / libreportal-bininstall pattern: one new
root-owned helper at /usr/local/lib/libreportal/libreportal-crowdsec
that does every privileged op behind a fixed action vocabulary with strict
argument validation. The manager calls in via runCrowdsec — the scoped
sudoers grants exactly one binary, the same trust boundary the other
helpers rely on.
Actions:
install apt repo + agent + firewall-bouncer + enable +
crowdsecurity/{linux,sshd} collections + reload
(idempotent — skips parts already in place)
services <verb> enable | disable | restart
capi <verb> register | unregister | status
console <verb> enroll <token> | disenroll | status
token format strictly validated
bouncer-traefik-init cscli register + write the manager-owned key file
atomically (returns EXISTS or GENERATED:<key>)
bouncer-priority bouncer yaml nftables priority → -100
(moved from libreportal-appcfg; one helper for
every crowdsec root op)
bind-lapi flip listen_uri to 0.0.0.0:8080 in config.yaml
prometheus <on…|off> flip the prometheus block (validated addr/port)
touch-host-logs create + chmod 0644 /var/log/crowdsec*.log so the
libreportal container can tail them
Wired in via:
- new sudoers Cmnd_Alias entry for the helper in LP_HELPERS
- new helper baked alongside the others by initRootHelpers
(replaces __SYSTEM_DIR__ / __CONTAINERS_DIR__ / __MANAGER__ at
install, with safe runtime fallbacks if unbaked)
- new runCrowdsec dispatch in scripts/docker/command/run_privileged.sh
containers/crowdsec/scripts/crowdsec_install_host.sh now drives the whole
flow through runCrowdsec — every `sudo …` is gone, the compose-toggle sed
uses runFileOp, and the security_crowdsec CFG mirror uses runInstallOp
(configs/ is manager-owned). Net: install script shrinks ~80 lines while
gaining a single auditable trust boundary. crowdsec_fix_priority.sh swung
over to runCrowdsec bouncer-priority too — the appcfg crowdsec_priority
action drops out cleanly.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
runBackupOp dropped privileges to $docker_install_user with `sudo -E`,
which preserves the CALLER's environment — including HOME. The caller is
the manager (libreportal), so restic-running-as-dockerinstall ended up
with HOME=/home/libreportal and tried to mkdir
`/home/libreportal/.cache/restic` for its cache. dockerinstall can't
write into libreportal's home, so every backup ran with:
unable to open cache: mkdir /home/libreportal/.cache/restic: permission denied
twice (once in backup, once in the verify-via-scratch-restore step), with
restic falling back to a no-cache run that's a few × slower than it
should be.
Add `-H` (sudo's "reset HOME to target user's home"). Now restic sees
HOME=/home/dockerinstall, creates ~/.cache/restic there (dockerinstall
owns its own home, no help needed), and the warning is gone. Confirmed
live: a `backup app create linkding` round-trip is silent on cache, and
the dir lands at /home/dockerinstall/.cache/restic, mode 0700, correctly
owned.
All restic/borg/kopia calls funnel through runBackupOp, so this single
character fix covers every backup-tool invocation.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Three small bugs in the legacy git-update flow that all hung off the
same never-set variable:
1. backup_install_dir was referenced in 4 files (reset_git_backup,
install_git_backup, use_git_backup, config_git_check) but DEFINED
nowhere — never has been, in any branch or tag. Resolved to "", so
"$backup_install_dir/$backupFolder" became "/backup_<ts>" (filesystem
root, perm denied). Add it to libreportalDerivePaths beside the other
roots, point it at $backup_dir/install (a sibling of restic's per-
location subdirs at $backup_dir/<idx>), and add it to initFolders so
it exists on first install.
2. gitCleanInstallBackups' find expression was
find ... -mindepth 1 -type f ! -name '*.zip' -o -type d ! -name '*.zip' -exec rm -rf {} +
`-o` binds looser than the implicit -a, so the -exec only applied to
the second clause. That meant: every non-.zip DIR anywhere under the
tree got deleted; every non-.zip FILE got matched and ignored. Even
once $backup_install_dir resolved correctly the cleanup would've
wrecked unrelated dirs.
Collapsed to `-mindepth 1 -maxdepth 1 ! -name '*.zip' -exec rm -rf {} +`
— direct children of $backup_install_dir, kill everything that isn't
a zip, let -rf take care of the dirs. Synthetic-tree smoke test
confirms only the .zip files survive.
3. use_git_backup.sh had a typo'd doubled var:
copyFolders "$backup_install_dir$backup_install_dir/$backup_file_without_zip/" ...
Reduced to the single $backup_install_dir/$backup_file_without_zip/.
All three only fire in the manual-update path (libreportal update apply
under git/local install mode); the install-blocking path is unaffected.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Tree-wide audit (working tree + deployed install + every local/remote ref
+ every reachable commit + unreachable objects via git fsck) found zero
external callers. Existed dead since v0.1.0 — never wired in.
The function set DOMAINSUBNAME, TIMEZONE, DOCKER_NETWORK (all duplicates
of fills that happen elsewhere) plus the two unique-to-it CONFIGS_DIR_TAG
+ CONTAINERS_DIR_TAG. Those two are already wired directly into the
standard tag-fill block in dockerConfigSetupFileWithData (commit 521f08b),
so dropping the source file leaves no behavioural gap.
Also tighten the comment that explained why we inlined the two tags —
don't reference the function we're deleting in the same change. Describe
the current behaviour, not the history (per repo convention).
Regenerated the auto arrays + function_manifest.sh: the 3 stale entries
referencing this function drop out cleanly. files_cli.sh / files_config.sh
/ files_source.sh also rebuilt — no net content change beyond dropping
this one path.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The location_loader's find ran through runFileOp (sudo -u libreportal),
which forks a sudo shell purely for the find call. configs/backup/
locations/ is already manager-owned (not in the dockerinstall-owned
containers tree), so the sudo step adds ~50 ms of process-creation
overhead for zero security benefit.
Plain `find` now, with a tiny [[ ! -r || ! -x ]] guard that falls
back to runFileOp if someone relocates the dir to a non-traversable
location. Same observable behaviour, ~50 ms saved per CLI invocation.
This is the simpler half of Phase 6 — the libreportal_configs scan
itself was already plain-find (only ~11 ms total for 22 files). The
remaining costly scan is app_configs against /libreportal-containers/,
which legitimately needs sudo because dockerinstall owns that tree.
Precompiling its content is possible but adds invalidation complexity
(updateConfigOption writes happen at runtime) — deferred for now;
better return on simpler interventions.
Signed-off-by: librelad <librelad@digitalangels.vip>
The early loop in sourceInitilize() sources every .sh under source/files/
recursively — including the new arrays/function_manifest.sh, which now
carries ~860 autoload stub definitions (~50 ms parse cost). Even in
eager mode where lazy infrastructure is never touched, every invocation
was paying that cost up front.
The manifest is only needed in lazy mode, where it's sourced explicitly
at the top of the lazy branch. Excluding it from the early loop:
- Eager mode: drops the ~50 ms regression introduced by Phase 5.
- Lazy mode: unchanged — the explicit source still runs.
This brings eager back to the pre-Phase-5 baseline and lets the lazy
container-stub gain (skipping sourceScanFiles containers, ~70 ms) show
through as a real saving.
Signed-off-by: librelad <librelad@digitalangels.vip>
Containers used to be eager-loaded via `sourceScanFiles "containers"`
even under LP_LAZY=1 — sourcing all ~160 installer functions up front.
Phase 5 brings them into the autoload-stub mechanism.
generate_function_manifest.sh now scans BOTH scripts/ AND containers/
(maxdepth 3, matching sourceScanFiles' existing prune), with a per-entry
root selector so stub emission uses the right base directory:
scripts/peer/peer_add.sh → source "${install_scripts_dir}peer/peer_add.sh"
containers/linkding/linkding.sh → source "${install_containers_dir}linkding/linkding.sh"
New manifest exports:
LP_FN_MAP funcname → relpath (existing)
LP_FN_ROOT funcname → scripts|containers NEW
LP_EAGER_FILES "<root>:<relpath>" entries NEW format
~860 autoload stubs (was ~700; +160 from containers)
Loader changes (initilize_files.sh):
- Parses LP_EAGER_FILES entries as `root:path`, dispatches to the
right install_*_dir. Pre-Phase-5 entries without a colon default to
scripts (backwards-compatible).
- sourceScanFiles "containers" is skipped when LP_LAZY=1 AND
LP_FN_MAP is loaded (manifest-driven autoload covers it).
Eager mode and lazy-with-missing-manifest both still run the scan.
Measurement target: ~70 ms saved on top of Phase 4. Verified separately
in the commit message of the next deploy.
Signed-off-by: librelad <librelad@digitalangels.vip>
`commandReloadConfigs` (baked into /usr/local/lib/libreportal/libreportal) and
`initCheckConfigs` both iterate every category dir's contents and `source` each
entry, with only a string-suffix exclusion for `.category` markers — no
`-f` test. That worked when `configs/<category>/` held only flat files.
The new backup system parks per-location configs at
`configs/backup/locations/<idx>/location.config`, so `configs/backup/locations/`
is now a SUBDIRECTORY inside the backup category. Sourcing it tripped:
source: /libreportal-system/configs/backup/locations: is a directory
…surfacing whenever something triggered a drift-driven config reload (e.g.
during a `regen --force` or a release-mode re-fetch). The nested location
configs already have their own dedicated loader (`sourceBackupLocations`)
that handles the depth-3 walk; the category-level loop just needs to leave
that subtree alone.
Collapse both loops to the cleaner guard `initReloadConfigs` and
`commandUpdateConfigOption` already use:
if [ -f "$config_file" ] && [[ ! "$config_file" =~ \.category$ ]]; then
…which both excludes directories (the bug) and the `.category` markers in
one shot, and drops a small pile of `should_load`/`filename` boilerplate
along the way. Verified live on dev-ai (CLI tool dispatch now works
through a drift-triggered reload without exiting non-zero).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
scripts/source/load_sources.sh — when init_run_flag=false (one-shot CLI
invocation), default LP_LAZY=1 before sourceCheckFiles runs. Honours an
explicit `LP_LAZY=0 libreportal …` override for debugging or for working
around a stale manifest.
Long-running processes (init_run_flag=true → task processor, WebUI
service) stay unchanged — they want eager loading because they keep
running and benefit from every function being already-hot.
Measured on this box, 3 runs each:
EAGER (default before this commit):
real 0.91s / 1.12s / 0.94s avg ~0.99s wall
user 0.40s / 0.41s / 0.42s avg ~0.41s CPU
LAZY (new default for CLI):
real 0.63s / 0.65s / 0.66s avg ~0.65s wall
user 0.26s / 0.27s / 0.26s avg ~0.26s CPU
Wall: ~340ms saved per invocation (34%).
User: ~150ms saved per invocation (37%).
Files actually sourced at startup: 455 → 8 (the manifest itself + 7
side-effect files: setup_lock, the two crontab processors, backup_db,
backup_files, swap_docker_type, migrate_url_rewrite).
Safety nets:
- Missing manifest auto-falls-back to eager loading (init.sh check
in the lazy branch sets LP_LAZY=0 if function_manifest.sh is absent).
- Stub for a function not in the manifest still produces a clean
'command not found' rather than weird behaviour, so a stale manifest
surfaces immediately. `libreportal regen arrays` regenerates both
files_*.sh and function_manifest.sh.
Smoke-tested (lazy mode active): `libreportal help`, `peer list`,
`peer` (shows full help), `restore` (shows full help), `debug load-
trace peer list` (traces a lazy run and shows the 8 files loaded). All
output identical to eager mode.
Signed-off-by: librelad <librelad@digitalangels.vip>
The previous run had 32 eager files; 24 of them were the auto-generated
files_*.sh arrays (only useful to the eager loader) + the manifest
itself (which the lazy loader sources explicitly). Eager-sourcing them
under lazy mode was pure overhead — ~55ms on the manifest alone (it was
being parsed twice, once via the explicit lazy-loader source and once
via the LP_EAGER_FILES loop).
Down to 8 eager files (the genuinely-side-effecting ones: setup_lock.sh,
the two crontab task processors, backup_db.sh, backup_files.sh,
docker swap_docker_type.sh, migrate_url_rewrite.sh, cli_debug_commands.sh).
The files_*.sh arrays are still sourced by the eager loader's existing
path — that's unchanged. Lazy mode just doesn't need them because it
never iterates files_libreportal_app[@] / files_libreportal_cli[@].
Signed-off-by: librelad <librelad@digitalangels.vip>
scripts/source/loading/initilize_files.sh gains an LP_LAZY=1 branch:
- Sources scripts/source/files/arrays/function_manifest.sh once. The
manifest defines LP_FN_MAP, LP_EAGER_FILES, AND ~700 autoload stubs
(precompiled by the generator — one parse cost vs evaluating 700
snippets at startup).
- Eager-sources every file listed in LP_EAGER_FILES (top-level side
effects: variable assignments, source calls, bare commands). These
can't safely be deferred — they'd skip the side effect, not just the
function definition.
- Skips the bulk loop that sources every files_to_source[@] entry.
Default behaviour (LP_LAZY unset or 0) is byte-identical to the previous
loader — every file gets eager-sourced up front. Long-running processes
(WebUI service, task processor) leave LP_LAZY unset because their first
call to anything wants the function already hot.
Each autoload stub looks like:
funcname() {
source "${install_scripts_dir}path/to/file.sh"
funcname "$@"
}
First call sources the real file, which redefines the function with the
real body; the stub's trailing `funcname "$@"` then calls the freshly-
defined real implementation. Sourcing the file also redefines stubs for
any sibling functions the same file declares, so they don't re-source.
Safety nets:
- Missing manifest → fall back to eager loading (`export LP_LAZY=0`).
No regression risk if someone enables LP_LAZY=1 on a stale install
whose regen never ran.
- LP_LOAD_TRACE=1 still works in lazy mode — it records the manifest
parse + each eager file (tagged LAZY-manifest / LAZY-EAGER) so Phase 4
can measure the actual saving.
No automatic flip yet — this commit only adds the path. Phase 4 will set
LP_LAZY=1 by default for the CLI entrypoint (and re-measure with the
trace tool from Phase 1).
Signed-off-by: librelad <librelad@digitalangels.vip>
scripts/source/files/generate_function_manifest.sh — scans every .sh in
scripts/ (skip-list matches generate_arrays.sh, plus excludes peer_shell.sh
which is a standalone forced-command target) and emits
scripts/source/files/arrays/function_manifest.sh:
declare -gA LP_FN_MAP=(
[acquireSingletonLock]="crontab/task/crontab_task_processor.sh"
[adoptDockerSubnet]="checks/requirements/check_docker_network.sh"
... # 698 entries
)
LP_EAGER_FILES=(
"backup/db/backup_db.sh"
"source/files/arrays/files_app.sh"
... # 32 entries (~7% of files)
)
The lazy loader (Phase 3) consumes LP_FN_MAP to install autoload stubs of
the form `name() { source "$LP_FN_MAP[name]"; name "$@"; }`. First call
sources the real file, which redefines the stub with the real body;
subsequent calls hit the real one. LP_EAGER_FILES enumerates files with
top-level side effects (variable assignments, source calls, bare commands
outside any function) — those MUST always source so the side effects fire.
Heuristic correctness, in order of importance:
- Function header detection requires EMPTY parens (`name()`), not just
`name(` — otherwise lines like `if (...)`, `for (...)`, `while (...)`
in embedded awk/perl get misread as bash function defs.
- Handles three function styles: `name() {` (same line), `name()\n{`
(LibrePortal convention), and one-liners `name() { body; }`.
- Tracks { } balance for inside-function depth, with the safe fallback that
ambiguous cases get marked eager (false positive = no behaviour change;
false negative would skip a needed source).
- Files containing embedded awk/perl with their own { } blocks (about 6 of
them: cli_debug_commands.sh, crontab_task_processor.sh, backup_db.sh,
backup_files.sh, etc.) get false-positive flagged eager — acceptable
because they just stay eager-loaded, matching today's behaviour.
- Collisions report to stderr (last-write wins, same as eager-load
semantics); no collisions found in the current tree.
Wiring:
- lpRegenArrays (`libreportal regen arrays`) now also runs the manifest
generator when the existing arrays need regen, keeping the two in sync.
- update.sh's quick-deploy regen step does the same after copying files
to the live install. Best-effort: failures don't abort because lazy
loading is opt-in (LP_LAZY=1) in Phase 3 and not the default yet.
Scanned: 454 files, indexed 698 function definitions, 32 eager (9 real,
23 auto-generated arrays + the manifest itself). 0 name collisions.
No behaviour change in this commit — the manifest is just data the loader
in Phase 3 will use. The default eager loading path is untouched.
Signed-off-by: librelad <librelad@digitalangels.vip>
First step toward an autoload-style lazy loader for the 499-file source
tree (current cold load ~1s wall / 340ms user-time per CLI invocation,
mostly spent sourcing files the command never calls). This commit only
measures — no behaviour change unless LP_LOAD_TRACE=1.
LP_LOAD_TRACE=1 instrumentation (scripts/source/loading/initilize_files.sh):
Wraps each in the main file-list loop with EPOCHREALTIME
before/after, writes `<elapsed_ms>\t<file_relpath>` to
$LP_LOAD_TRACE_FILE (default /tmp/libreportal-load-trace.<pid>.log).
Zero overhead when the env var is unset (one [[ test per file).
libreportal debug load-trace [cmd...]:
New `debug` CLI category. Spawns a child `libreportal <args>` (default
'help') with LP_LOAD_TRACE=1, then awk-aggregates the trace: wall vs
cumulative source time, file count, top-15 hottest files. The diff
between wall and cumulative-source = bash startup + dispatch + the
command's own work.
Used in the next phases to (a) validate that the lazy loader actually
delivers the speedup we expect and (b) flag any single file that hogs
disproportionate time (rare `heredoc | sed | base64` style work at
source time would show up here as a >10ms entry).
Signed-off-by: librelad <librelad@digitalangels.vip>
Linkding has shipped without any Tools-tab actions since v0.1.0 — the only
artifact was scripts/menu/tools/manage_linkding.sh, a dead legacy CLI menu
referencing an `appLinkdingSetupUser` function that was never defined.
Build the real thing, mirroring bookstack's pattern (manifest + thin tool
wrappers + auth_adapter that drives the app's native admin shell):
containers/linkding/tools/linkding.tools.json — manifest, 5 tools
containers/linkding/tools/linkding_<id>.sh — one wrapper per tool
containers/linkding/scripts/linkding_auth.sh — Django shell driver
Tools (all category=users, so the WebUI's custom user-list panel and its
row-level 🔑 / 👑 / 🗑 buttons light up):
reset_password — set_password on an existing user (random if blank)
create_account — create_user / create_superuser
list_users — emits EZ_USER\t<username>\t<username>\t<role> rows
(linkding is username-primary, so username goes into
both display slots — keeps the panel click-through
identifier consistent with the other tools' fields)
delete_user — delete by username (destructive, confirm gated)
set_admin — toggle is_superuser + is_staff
Implementation runs entirely inside the linkding-service container via
`runFileOp docker exec ... python manage.py shell -c "<code>"`, reading
inputs through `-e` env vars so quoting stays safe. Django's default
get_user_model() User is used directly — passwords hash exactly the way
the web UI does, admin flags map to the same fields the UI reads.
Also drop the dead legacy stub (scripts/menu/tools/manage_linkding.sh)
and regenerate files_menu.sh so the source-scan no longer pulls it in.
Nothing referenced linkdingToolsMenu — verified by tree-wide grep.
Verified live on dev-ai (Debian 12, linkding installed, Django 5 + sqlite):
$ libreportal app tool linkding create_account 'username=alice|password=…|admin=true'
✓ Linkding user created — Username: alice — Password: …
$ libreportal app tool linkding list_users ''
EZ_USER alice alice admin
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The previous `<img src="/icons/config/security.svg">` icon hardcoded
`stroke="#1e90ff"` (dodger blue) rather than `currentColor`, so on
themes where it should pick up the sidebar foreground colour it just
disappeared or visually clashed. The other Tools / admin sidebar items
(Overview, System, Peers) all use inline SVGs with `stroke=currentColor`
and follow the theme correctly.
Switched SSH Access to an inline key icon in the same style — circle
shackle bottom-left, shaft going up-right to a notched bit. Matches the
'what is this thing' framing: an SSH access page is fundamentally about
managing keys.
security.svg itself is left untouched (might be used elsewhere).
Signed-off-by: librelad <librelad@digitalangels.vip>
Two related UI tidies — both removing surface area from the topbar / Tools
group rather than adding new pages.
Peers → /admin/tools/peers
Was a top-level /peers route with its own topbar nav item, which doubled
the navigation surface for what's really an admin tool (same shape as
SSH Access). Now lives under the Admin sidebar's Tools group alongside
SSH Access. /peers is kept as a legacy redirect → /admin/tools/peers.
Plumbing:
- config-sidebar.js gains a Peers entry under the Tools label.
- config-manager.js gains a 'peers' branch that fetches
peers-content.html into config-section, then inits PeersPage.
- window.adminPath() learns 'peers' → /admin/tools/peers.
- spa.js handlePeers() is now a redirect (mirrors handleSsh).
- topbar.html drops the Peers nav item.
- peers-content.html slimmed to a config-section template (no
standalone page wrapper) so it embeds cleanly under the admin shell.
- PeersPage gains a rootId constructor arg for symmetry with SshPage
(queries still work globally — IDs are unique).
System lifted out of the Tools group
User feedback: 'overview/system are kinda like, the same thing'. Moved
System to sit right under Overview at the top of the sidebar, before
the 'Config' label. Both surfaces are admin-landing pages (Overview =
ops/health summary, System = live host + per-app stats) — distinct from
config form pages or the Tools utilities.
config-sidebar.js: System block moved to the top section (right after
Overview's click handler). Original Tools-group instance removed.
Signed-off-by: librelad <librelad@digitalangels.vip>
Polish pass for the migration system. Two concrete additions; the live-mirror
and full drift-verify ideas from the original plan are intentionally
deferred — both need real-world test data to land correctly, and the kernel
already exposes everything they'd need.
Per-app migrate hooks (scripts/migrate/migrate_hooks.sh):
Apps can declare two optional functions in their tools.sh (already
auto-sourced per [[libreportal-modular-app-tools]]):
<app>_migrate_pre() — runs before stop+wipe
<app>_migrate_post() — runs after restart, before the user sees it
Each receives:
$1 = source identifier (peer name or backup-tag hostname)
$2 = transport ("restic" | "direct-ssh")
migrateRunHook() is now called from both migration apply paths:
- migrate_apply.sh (restic-mediated, shared backup channel)
- peer_pull.sh (direct-SSH, peer-shell stream)
Use cases: rotate federation keys after a Mastodon move, regenerate
OIDC client secrets, drop SaaS-style locks, fix hostname-baked configs
the URL-rewrite layer doesn't cover.
Hooks are optional — apps without them inherit the standard flow.
Failed hooks emit a non-fatal notice (the rest of the migrate still
reaches 'done') so a single bad hook can't strand an otherwise-working
app in stopped state.
Peer friendly-name overlay (Migrate tab):
Was deferred from Phase 2 because it required Phase 3's UI to feel
cohesive. BackupPage.refreshAll() now also fetches peers.json and builds
a hostname → peer-name lookup. renderMigrate() shows
'homelab (host: homelab.lan)'
for any backup-channel peer that matches the source host, and falls back
to the bare hostname when no peer is defined. Same data, friendlier UI.
Skipped (genuinely deferred, not just out of time):
- Live mirror / warm-standby (continuous one-way sync). Needs a scheduler
+ drift-state to track. Right place for it is a separate feature on top
of the existing kernel rather than bolted onto migrate.
- Drift-verify ("what would change if I migrated?"). Cheap to write but
needs a real cross-host pair to validate against — adding it untested
would just be theatre.
Signed-off-by: librelad <librelad@digitalangels.vip>
End-to-end direct-ssh-direct: two LibrePortal instances exchange pairing
tokens, each authorizes the other to call a locked-down peer-shell dispatcher
via SSH forced-command, then either side can pull live app data from the
other without needing a shared backup repo.
Push and Connect-via-relay are deferred — push is symmetric to pull (same
forced-command, opposite verb), and the relay variant waits for Connect to
actually exist (config_json + kind enum already future-proofed in Phase 2).
Key generation (peer_key.sh):
One ed25519 keypair per install at ~<manager>/.ssh/libreportal-peer{,.pub}.
Generated lazily on the first peer-related call. Used as our outbound
SSH identity AND as the pubkey other instances authorize.
Forced-command dispatcher (peer_shell.sh):
Standalone script, deployed by peerInstallShell() to
~<manager>/.local/bin/peer-shell. authorized_keys entries look like:
command="~/.local/bin/peer-shell <peer-name>",no-pty,no-port-forwarding,
no-X11-forwarding,no-agent-forwarding,no-user-rc ssh-ed25519 AAAA… peer:<name>
sshd hands us $SSH_ORIGINAL_COMMAND; we parse, whitelist the verb, and
refuse anything else. Verbs:
ping Liveness probe (JSON ok:true).
list-apps JSON {peer, apps:[{slug, size_kb}]}.
stream-app tar of containers_dir/<slug> to stdout (slug strictly
validated — lowercase alnum+dash; rejects path traversal).
Audit log appended to ~/.local/state/libreportal/peer-shell.log. Excluded
from the generated source arrays (would crash any sourcing shell on empty
SSH_ORIGINAL_COMMAND); generate_arrays.sh skip-list extended.
Pairing token (peer_pairing.sh):
Format: lp-peer|v1|<name>|<user>|<host>|<port>|<base64-pubkey>|<fingerprint>
Pipe-delimited because the SHA256 fingerprint and base64 pubkey both
contain ':'. peerPairingParse decodes + re-derives the fingerprint from
the actual key, refusing tokens with mismatched fingerprints (catches
truncation / tampering). peerPairingAccept:
1. Installs peer-shell (peerInstallShell).
2. Appends to authorized_keys with the lockdown options above.
3. Inserts a peers row (kind=direct-ssh-direct, config carries host,
port, user, fingerprint).
Symmetric — user runs accept on BOTH sides with the other's token to
enable bidirectional calls.
Outbound SSH (peer_remote.sh):
peerExec <name> <verb> [args] — looks up the peer's connection config and
ssh's in with the right key, BatchMode + ConnectTimeout + accept-new for
the host key. peerPing wraps it and updates peers.status + last_seen.
Pull-an-app (peer_pull.sh):
peerPullApp <peer> <app> [--no-pre-backup] [--keep-urls]
1. peerPing (refuse if unreachable).
2. migratePreBackupDestination (reuses the Phase 0 safety wrapper —
same restic-tagged pre-migrate snapshot as the backup-channel flow).
3. Stop + wipe destination's app folder.
4. peerExec stream-app | tar -x (pipefail; bails on partial transfers).
5. migrateApplyUrlRewrite + dockerComposeUpdateAndStartApp install
(URL repointing, idempotent install path).
6. dockerComposeUp + post-restore hooks.
Identical Stage-2..6 to migrateApplyApp; only the data source differs
(tar-over-SSH instead of restic-restore).
CLI (cli_peer_commands.sh + header):
libreportal peer token — emit this host's pairing token
libreportal peer pair <token> [name] — accept a token (override name)
libreportal peer apps <peer> — live peer-shell list-apps
libreportal peer pull <peer> <app> [--no-pre-backup] [--keep-urls]
WebUI (/peers):
Header gains 'Show my token' and 'Pair with token' buttons (both open
modals around the matching CLI verbs). Token modal warns the user that
the token is credentials. Pair modal accepts a free-form override name.
Direct-SSH peer cards gain a 'List apps' button that opens an inline
drawer showing the peer's live app inventory (via peer apps) with per-
app 'Pull' buttons. Pull modal has the same two safety toggles as the
Migrate tab (pre-backup ON, URL rewrite ON by default).
Backup-channel manual-add modal kept; direct-SSH must use the token flow.
Smoke-tested:
- All 16 peer-subsystem functions register without crashing the shell.
- peer-shell ping ⇒ {ok:true}; unknown-verb refused; path-traversal slug
refused; valid-slug streams.
- Token emit→parse round-trip preserves every field; garbage rejected
with not-a-token; v99 rejected with unsupported-version.
Signed-off-by: librelad <librelad@digitalangels.vip>
The standalone WireGuard install used to flip net.ipv4.ip_forward by
appending+uncommenting `/etc/sysctl/99-custom.conf` via blanket sudo
(sudo tee, sudo sed, sudo sysctl -p). Two problems with that on a
de-sudoed manager:
- The path is non-standard. The conventional location is
/etc/sysctl.d/*.conf (drop-ins, loaded by sysctl --system) — the
old file may not even exist, leaving forwarding silently off.
- `sudo tee /etc` and `sudo sed -i /etc` are not in LP_SYSTEM. The
manager has lost the broad sudo it once had, so this would now
fail outright on every wireguard install.
Add a `wireguard-ip-forward` action to libreportal-appcfg that:
- writes /etc/sysctl.d/99-libreportal-wireguard.conf (a drop-in we
own and rewrite idempotently), and
- reloads via `sysctl --system` (with a `sysctl -p <dropin>` fallback).
containers/wireguard/wireguard.sh now calls `runAppCfg wireguard-ip-forward`
through the existing helper-dispatch path — the whole edit runs as root
in one validated step, no `sudo` in the per-app script.
Same de-sudo pattern as adguard-auth / crowdsec-priority / owncloud-config
already use.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Sweep of every containers/<app>/<app>.sh after the install-side fix that
went into config_file_setup_data.sh — these were the same class of bug:
bare `sudo sed -i` / `sudo docker exec` calls left over from when the
manager carried NOPASSWD:ALL. After the rootless+de-sudo hardening (Model
A, sudoers scoped to LP_HELPERS + LP_SYSTEM only) those calls fail at
runtime, so every per-app routine that uses one would refuse on install
or in its post-install tweak step.
Each call routes through the existing `runFileOp` shim, which picks the
right path per CFG_DOCKER_INSTALL_TYPE (dockerinstall in rootless, manager
in rootful) — same pattern setup_dns.sh / authelia.sh / config_file_setup_data.sh
already use.
Fixed:
gitea.sh:65 — sync GITEA_METRICS_TOKEN into prometheus-scrape.yml
owncloud.sh:88 — fill OWNCLOUD_SETUP_* in the setup-webform html
searxng.sh:87 — flip simple_style: auto → CFG_SEARXNG_THEME
trilium.sh:89 — rewrite trilium-data/config.ini port=
bookstack.sh:139 — bookstack:create-admin via `docker exec`
bookstack.sh:148 — admin@admin.com cleanup via `docker exec ... tinker`
`bash -n` clean on every touched file. Untested live (none of these apps
are installed on the verify VM) but mechanically equivalent to the
already-validated config_file_setup_data.sh fix.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
A peer is a named reference to another LibrePortal instance. Phase 2 only
implements kind=backup-channel (friendly label over a hostname that shows
up in a shared backup repo); direct-ssh-direct and direct-ssh-via-relay
(Connect's blind-relay) are reserved enum values for Phase 3.
DB schema (db_create_tables.sh):
CREATE TABLE peers (
id INTEGER PRIMARY KEY AUTOINCREMENT,
name TEXT UNIQUE NOT NULL,
kind TEXT NOT NULL DEFAULT 'backup-channel',
config_json TEXT NOT NULL DEFAULT '{}',
status TEXT DEFAULT 'unknown',
last_seen TEXT,
created_at TEXT DEFAULT CURRENT_TIMESTAMP
);
+ indexes on name and kind.
config_json is kind-specific so new transports don't need a schema
migration. For backup-channel it carries {"hostname":"","loc_idx":N}.
Bash module (scripts/peer/):
peer_helpers.sh _peerDb, peerSqlEscape, peerValidateName/Kind.
peer_add.sh peerAdd <name> <kind> [k=v ...] → INSERT, refresh
generator. Rejects unimplemented kinds early so users
don't create dead-end peer records.
peer_remove.sh peerRemove <name> → DELETE.
peer_list.sh peerList → JSON array; peerGet, peerNameForHostname
(reverse-lookup for the migrate-tab overlay).
peer_check.sh peerCheckReachable, peerCheckAll. For backup-channel
'reachable' = at least one snapshot from that hostname
visible in (preferred|any enabled) location. Updates
status + last_seen so UI dots render without re-probing.
CLI (scripts/cli/commands/peer/):
libreportal peer list
libreportal peer get <name>
libreportal peer add <name> backup-channel hostname=<host> [loc_idx=<n>]
libreportal peer remove <name>
libreportal peer check [name]
Auto-routed by cli_initialize.sh's category-discovery.
WebUI data generator (scripts/webui/data/generators/peers/webui_peers.sh):
Emits data/peers/generated/peers.json with the peerList output and a
generated_at envelope. Hooked into webuiLibrePortalUpdate alongside the
backup generators.
Frontend:
- New top-level /peers route in spa.js (PeersPage class, peers-content.html).
- 'Peers' nav item in the topbar between Backups and the right-side controls.
- Add-peer modal with friendly-name + kind + hostname + preferred-location
selector (populated from the existing backup-locations data).
- Per-peer card with status dot, last-checked time, Check + Remove buttons.
- Phase 3 kinds appear in the kind dropdown as disabled options so users
can see what's coming.
Source-array wiring:
- generate_arrays.sh auto-created files_peer.sh from the new peer/ dir.
- cli_files.sh + app_files.sh include ${peer_scripts[@]} alphabetically.
- files_webui.sh auto-picked-up the new peers/ generator subfolder.
The migrate-tab friendly-name overlay (use peer names in /backup/migrate
when a peer record exists for a hostname) is intentionally deferred — it's
a 5-line frontend lookup once peers.json is loaded; cleaner to add after
Phase 3 ships its peer-detail view.
Signed-off-by: librelad <librelad@digitalangels.vip>
configFileSetupData runs as the manager (libreportal user) during install,
but writes into /libreportal-containers/<app>/, which is owned by the
container user (dockerinstall) under the three-root layout. The six bare
`sed -i` calls in this function were missing the `runFileOp` wrapper that
every other in-tree sed-on-app-files call already uses (e.g. setup_dns.sh's
WG_DEFAULT_DNS edits), so on first run `sed -i` failed to create its temp
file in the live dir:
sed: couldn't open temporary file /libreportal-containers/linkding/sedaCaUNU: Permission denied
✗ Error Updated DOMAINSUBNAMEHERE with: bookmark.
! Notice Non-interactive mode: aborting on error.
…which aborted the install at step 3 of every per-app config setup.
Replace `result=$(sed -i ...)` with `result=$(runFileOp sed -i ...)` so each
substitution runs as the owner of the target file (via the bin-install
helper). All six call sites use the same pattern — done as a single
`replace_all` over the unique prefix.
Tags fixed: DOMAINSUBNAMEHERE, APPADDRESSHERE, DOMAINSUBNAME_DATA,
TIMEZONE_DATA, EMAILHERE, HOSTIPHERE.
Verified live on a fresh install: `libreportal app install linkding` now
completes cleanly through all 10 install steps and lands the container.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Phase 1 of the migration-system refresh. Surfaces Phase 0's kernel
(libreportal restore migrate ...) as a WebUI flow so users don't have
to drop to the CLI to pull an app from a peer's backups.
backend / data generator:
scripts/webui/data/generators/backup/webui_backup_migrate.sh
Walks every enabled backup location, lists every (other_host, app)
pair with snapshot count + latest id/date, and emits a single
destination summary block (installed apps, running apps, disk free)
so the frontend can compute collisions and warnings without per-row
API round-trips. Filters out our own hostname — we don't migrate to
ourselves. Output: data/backup/generated/migrate.json.
Hooked into the standard webuiLibrePortalUpdate refresh pipeline,
so 'libreportal regen webui' (and the periodic task-processor poll)
keep it fresh on their own.
frontend:
- New 'Migrate' sidebar tab on /backup, sits between Locations and
Configuration. Path-based URL: /backup/migrate.
- Per-source-host cards listing every available app, with snapshot
count + relative-time hint, collision dot when the app is already
installed here, and per-app + per-host migrate buttons.
- Confirm modal with two checkboxes matching the kernel's defaults:
[✓] Back up the destination's existing copy first (pre-migrate
backup; auto-disabled when there's nothing to back up)
[✓] Rewrite host-bound URLs to this host (URL rewrite
— uncheck only to keep source hostnames)
On confirm, runs 'libreportal restore migrate app/system …' via the
task system; opt-out checkboxes append --no-pre-backup / --keep-urls
only when the user un-ticks, matching the kernel's default-on flags.
- Empty state when no other hosts have visible backups, explaining
the shared-backup-location prerequisite.
The CLI dispatcher hooks (Phase 0) wire restore migrate app/system to
migrateApplyApp/migrateApplySystem, so the WebUI gets pre-backup safety,
URL rewrite, and structured progress (when --json-progress is set; not
needed here yet — the task system's log tail is enough for v1).
Signed-off-by: librelad <librelad@digitalangels.vip>
The libreportal compose template binds /libreportal-system/configs/webui/* and
/libreportal-containers into the WebUI container via #LIBREPORTAL|CONFIGS_DIR_TAG|
and CONTAINERS_DIR_TAG placeholders. The only code that knew how to substitute
those — tagsProcessorStandardReplacements — had NO callers anywhere (verified
by grep across the whole tree). Result: the deployed compose retained
CONFIGS_DIR_DATA / CONTAINERS_DIR_DATA placeholders and the safety check in
dockerComposeUp refused to start it.
Wire the two missing substitutions into the standard tag-fill block in
dockerConfigSetupFileWithData where TIMEZONE_TAG / CATEGORY_TAG / TITLE_TAG
already live — applies to every app's templating, idempotent, and unblocks the
fresh-install path. The orphaned tagsProcessorStandardReplacements function
duplicates 3 other tags that ARE filled elsewhere; left in place pending a
follow-up cleanup, but no longer the source-of-truth for these two.
Confirmed live: after running `libreportal app install libreportal` post-fix
the compose templated cleanly (no remaining *_DATA placeholders) and the WebUI
container came up — http://<host>:7270 returns HTTP/1.1 200 OK.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Triage of a broken fresh install:
1. init.sh → all root setup → completeInitMessage hands off to
`libreportal run install` as the manager.
2. start.sh sources load_sources.sh, which calls sourceCheckFiles "run".
3. sourceCheckFiles "run" calls checkUpdates — its only path to startLoad on
a non-local mode is via the git/release recovery branches.
4. git fails (the deployed install dir has no .git), lpFetchRelease fails (no
reachable release manifest), none of the recovery branches converge on
startLoad, and the install silently exits with WebUI + service unset.
Fix: completeInitMessage exports LIBREPORTAL_INITIAL_INSTALL=1, and the
sourceCheckFiles "run" branch calls startLoad directly when that's set — same
endpoint the local-mode branch hits. We just installed the latest code from
this tree; checking for updates on the first run was nonsensical and the
recovery gauntlet would only break things.
Confirmed by re-running uninstall + install: the install now reaches the
Pre-Installation / database / WebUI build / crontab / WebUI compose-up steps
and produces a working WebUI. (A separate compose-tag bug surfaced next —
fixed in the follow-up commit.)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Phase 0 of the migration-system refresh. Replaces the 77-line
scripts/migrate/ with a properly-shaped kernel that Phase 1 (WebUI) and
Phase 3 (direct peer SSH) can both build on.
New module layout (6 files):
migrate_progress.sh — migrateEmit JSON-per-line helper; opt-in via
MIGRATE_JSON_PROGRESS=1, writes to fd 3 if open
(clean WebUI streaming channel) else stdout.
migrate_discover.sh — migrateDiscoverHosts / migrateDiscoverApps /
migrateDiscoverAppDetail (JSON {snapshots, latest_*}).
Old migrateDiscoverAppsForHost kept as back-compat.
migrate_preflight.sh — migratePreflight emits one JSON object with
snapshot{id,date}, destination{installed,running,
disk_free_kb}, collision{occurs,default_action,
pre_backup_default}, url_rewrite{default_action,
per_app_opt_out}, warnings[], errors[].
Exit 0 on usable preflight, 1 on hard error.
migrate_url_rewrite.sh— Host-bound CFG_<APP>_* fields (URL/HOST/DOMAIN/
DOMAIN_PREFIX/HOSTNAME/PUBLIC_URL) get rewritten
from the destination's install-template after
restore — so a moved app stops claiming the
source's hostnames. Per-app opt-out via
CFG_<APP>_MIGRATE_URL_REWRITE=false. All other
fields (DB passwords, API keys, prefs) carry
over from the source unchanged.
migrate_pre_backup.sh — migratePreBackupDestination takes a snapshot of
the destination's existing <app> (tagged
pre-migrate=<UTC timestamp>) before the wipe.
Default ON; opt-out with --no-pre-backup. Safety
net for the always-replace collision policy.
migrate_apply.sh — migrateApplyApp / migrateApplySystem. Parses
--no-pre-backup / --keep-urls / --json-progress
opts, runs preflight → pre-backup → restoreAppStart
(existing flow) → URL rewrite → re-deploy compose.
migrateApp / migrateSystem kept as shims so the
old CLI surface still works.
CLI dispatcher (cli_restore_commands.sh + cli_restore_header.sh):
Existing 'restore migrate app/system/discover' calls all still work.
New verbs:
restore migrate list <host> [loc_idx]
restore migrate preflight <host> <app> [loc_idx] ← JSON, for the WebUI
Design choices baked in (per the spec):
- Always-replace collision (no multi-install of an app), safety net is the
on-by-default pre-migrate backup.
- URL rewrite by host-bound suffix list, not per-field allowlist — works
out-of-the-box for new apps without extra config.
- migrateEmit fd-3 contract is what Phase 1's WebUI will stream; falls
back to stdout in interactive CLI so dev/debug just works.
- Transport-agnostic: nothing in this kernel knows whether the backup
location is local/SSH/S3/Connect — engineSnapshotsJson + engineBackupApp
do that, so Connect (the future blind-relay) plugs in as 'just another
location kind' with zero kernel changes.
Smoke-tested: all 13 public functions register; JSON emit produces correct
escaping (quoted strings vs bare numerics) and respects MIGRATE_JSON_PROGRESS.
Signed-off-by: librelad <librelad@digitalangels.vip>
I skipped focalboard earlier citing "DB+files overlap" (sqlite lives in the same
dir as the file-capture target). But linkding / vaultwarden / headscale all have
that exact same shape and we just labeled them in 12b4d68. gitea has had it for
ages and it's proven — the DB dump excludes the raw .db from the snapshot, the
file-capture grabs the dir (incl. live sqlite), and restore replays the dump over
the captured tree. The torn live-sqlite copy is harmless bloat.
So focalboard gets the same treatment for consistency. Coverage now: 9 apps
(adguard set aside, jellyfin still pending DB declaration).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Now that uid auto-discover is in (d424473), adding more apps is just naming the
container + path — no uid guessing. Four more apps get complete backups of their
private data dirs (previously: incomplete on the file side because libreportal
can't read sub-UID-owned content from the host).
- linkding-service -> /etc/linkding/data (overlaps with sqlite db, same pattern as gitea — dump replays over the captured tree on restore, harmless)
- vaultwarden-service -> /data (same overlap pattern)
- headscale-service -> /var/lib/headscale (same overlap pattern)
- mastodon-service -> /mastodon/public/system (uploads; postgres handled separately by backup.db)
Coverage now: nextcloud, bookstack, gitea, owncloud, linkding, vaultwarden,
headscale, mastodon. Skipped jellyfin — it has multiple internal sqlite DBs and
no backup.db declared; adding just backup.files without backup.db / backup.live
wouldn't activate live capture, and adding backup.live blind could yield torn
sqlites. That one wants proper DB declaration first.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The migrate/ helpers were either uncallable or no-ops:
- migrateAppsToNewNetwork + updateComposeFileNetwork: never called from
anywhere. The intended sed-on-compose subnet rewrite would also have
fought the tag system / network_resources DB.
- checkAppNetworkCompatibility: called from updateDockerNetworkConfig as
a gate, but never explicitly returns, so it's effectively always-true
and both branches do the same work. Pure noise.
- getInstalledApps: only used by the above.
- updateDockerNetworkConfig: collapses to a 2-line 'CFG := docker's
reported subnet' adoption — inlined into check_docker_network.sh as
adoptDockerSubnet(), which is what it actually does.
The legitimate 'subnet changed, refresh apps' path is already covered by
the idempotent per-app reinstall (dockerInstallApp ... reset_network=true
→ clears DB allocations → installer re-runs → ipUpdateComposeTags picks
fresh IPs from the current CFG_NETWORK_SUBNET). Migration (infrastructure
regen) vs restore (data) stays clean: reinstall regenerates compose+IPs,
restore lays data on top. No new pathway needed.
Files dropped:
scripts/docker/network/migrate/migrate_apps_to_new_network.sh
scripts/docker/network/migrate/migrate_check_app_network_compatibility.sh
scripts/docker/network/migrate/migrate_get_installed_apps.sh
scripts/docker/network/migrate/migrate_update_compose_file_network.sh
scripts/docker/network/migrate/migrate_update_docker_network_config.sh
Plus the now-empty migrate/ subdir; files_docker.sh regenerated to drop
the references.
Signed-off-by: librelad <librelad@digitalangels.vip>
The hardcoded uid:gid in libreportal.backup.files labels was brittle: matched the
default PUID in the compose, but a PUID change (or new image version) would drift
silently and the next restore would chown to a stale owner. Make it impossible to
drift by letting the engine learn the uid at capture time.
backup_files.sh:
- After a successful tar capture, run `stat -c '%u:%g'` inside the container and
write the result to a <host_subdir>.lp-owner sidecar in staging. The sidecar
rides in the snapshot alongside the captured tree.
- Restore reads it back when the descriptor doesn't pin uid:gid; falls back to
0:0 with a clear notice if missing.
- The 5-field form (with explicit uid:gid) is still supported as an override; it
wins and skips the sidecar write entirely.
Update all 4 current labels to the new 3-field form
"<container>:<container_path>:<host_subdir>" (nextcloud, bookstack, gitea,
owncloud). Engine handles both formats during the transition.
Verified with stubs: 3-field capture writes the sidecar with the discovered
33:33; restore reads it back; 5-field override correctly skips the sidecar
write. backup_files.sh parses.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Three closeouts in one pass:
1. DEVELOPMENT.md — consolidated hook-conventions table covering all 8 per-app
hook types (tools / update-specifics / compose-tags / webui-refresh / the two
traefik markers / the two network-provider hooks). One place to look instead
of inferring from the codebase.
2. Nextcloud APCu wired alongside Redis: appUpdateSpecifics_nextcloud now sets
memcache.local=\OC\Memcache\APCu too (was deferred from the fpm switch). APCu
= cheap in-process cache; the fpm-alpine image ships the extension. CLI mode
may emit a harmless "no memory cache" notice on `occ` runs — Nextcloud is
graceful, the FPM worker still uses APCu fine.
3. Container-side file-capture rollout to 3 confident cases:
- bookstack: lscr.io/linuxserver/bookstack with PUID=1000 → /config (1000:1000)
- gitea: gitea/gitea with USER_UID=1000 → /data (1000:1000)
- owncloud: owncloud/server (Apache/PHP) → /mnt/data (33:33, www-data)
Snapshots are now complete for these (the dir's excluded from the raw restic
pass and captured live through the container as a tar → libreportal-owned
staging, same proven pattern as Nextcloud). Less-evidenced candidates left
for live verification: linkding, mastodon, jellyfin, trilium, focalboard,
invidious, vaultwarden, headscale-service — each needs its in-container uid
confirmed before labeling (wrong uid won't break backup, but restore would
chown to the wrong owner).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
On installs that went through the EasyDocker rename (or any case where the
manager user's uid was recycled), /home/<manager> stays owned by the old uid
(useradd doesn't take over an existing home dir). Files inside, including
restic's ~/.cache/restic, become unreadable by the new manager → restic logs
"mkdir: permission denied" every backup. Non-fatal but slows them.
Same recycled-uid pattern as the cron-spool cleanup right above this block:
chown -R only when the directory's owner uid != the manager's current uid.
Idempotent — a fresh install or one that's already correct is a no-op.
Closes the EasyDocker-artifact item from the live-backups memory.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
_focalboardSqlite called sqlite3 on /data/focalboard.db, but the compose mount
puts the DB at /opt/focalboard/data/focalboard.db (mount: ./data:/opt/focalboard/data).
/data inside the container isn't mounted, so every auth tool (set/reset password,
create/delete user, set admin) silently failed against a nonexistent file.
The memory flagged this as a "DB not persisted" bug, but the compose mount was
already corrected at some point; the auth adapter was the half that didn't get
the fix. Backup label was also already correct (data/focalboard.db relative to
the live app dir resolves to the same file via the mount).
One-line path correction.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Drop Apache+mod_php for the actual performance win — nginx + PHP-FPM — without
the LinuxServer image cascade (custom auto-install, /custom-cont-init.d, abc-vs-
www-data rewrites in the auth adapter + every tool, HTTPS-by-default quirks).
The official fpm-alpine image keeps env-var auto-install and the www-data user,
so the auth adapter, all tools, and the compose-tags hook keep working unchanged.
- Compose: nextcloud-service is now fpm-alpine (still container_name=nextcloud-
service so docker exec ... nextcloud-service php occ in the auth adapter is
untouched). New nextcloud-web nginx sidecar serves :80 over the shared ./html
volume, terminating FastCGI to nextcloud-service:9000. Traefik labels + PORTS_
TAG_1 move to nextcloud-web (the HTTP face); backup.files stays on -service
(the file-owning brain). nextcloud-db + nextcloud-redis unchanged.
- resources/nginx.conf: Nextcloud's recommended nginx config, trimmed for
behind-Traefik (no TLS), large-upload + caldav/carddav/.well-known redirects.
- scripts/nextcloud_update_specifics.sh: NEW post-install hook —
appUpdateSpecifics_nextcloud waits for first-boot occ install to complete
(config.php + occ status=installed), then wires Redis as memcache.distributed
+ memcache.locking via occ config:system:set. Idempotent.
Auto-install is unchanged (official image's NEXTCLOUD_ADMIN_USER + MYSQL_* env
flow). Redis caching now actually USED by Nextcloud (previously the container
was up but config.php had no memcache config). Container-side backup capture
still the right answer for the perm boundary — image change doesn't affect it.
Verified statically: yaml structure, hook parses + dispatches + has the right
graceful-timeout fallback when occ isn't reachable. Live verification (sync
performance + actual Redis hit rate + traefik proxy of FastCGI) needs a fresh
install on a throwaway box.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
If it's gluetun code, it lives with gluetun. Both functions in
scripts/config/tags/processors/tags_processor_network_mode.sh manipulate gluetun
markers / gluetun's compose, so move them into containers/gluetun/scripts/
gluetun_network.sh and rename to the per-app-hook convention:
tagsProcessorNetworkMode -> appNetworkApplyMode_gluetun
tagsProcessorGluetunForwardedPorts -> appNetworkRegisterPorts_gluetun
Central call sites are now provider-agnostic — no "gluetun" literal anywhere:
- docker_config_setup_data.sh: an app routing via CFG_<APP>_NETWORK=<provider>
triggers `appNetworkApplyMode_<provider>` + `appNetworkRegisterPorts_<provider>`
via declare -F, so any future gateway provider plugs in with no engine edits.
- uninstall_app.sh: loops every `appNetworkRegisterPorts_*` hook (each self-skips
when its provider isn't installed), so removing a routed app refreshes the
right provider with no provider name in central code.
Delete tags_processor_network_mode.sh; regenerate arrays. Verified with stubs:
default mode no-ops, gluetun-routed app fires both hooks, gluetun itself is
skipped, unknown provider is silently no-op, uninstall loop calls registerPorts.
Drive-by cleanup: 9 stale "${X_scripts[@]}" array references in app_files.sh /
cli_files.sh (gluetun + headscale from this session's moves, plus 7 pre-existing:
command/ssl/swapfile/ufw/ufwd/user — all from older refactors that left them
behind). Each expanded to nothing at runtime (harmless), but they're dead
misleading refs. Cleaned both files; every remaining array ref now points to a
real files_*.sh.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The 5 service arms in appInstallCheckRequirements (traefik/gluetun/authelia/
headscale/prometheus) were identical _appReqServiceInstalled calls. Collapse them
into one generic default: any requirement naming a real container is a service
prerequisite — so a new service requirement now needs NO code here, just list it
in the app's CFG_<APP>_REQUIRES. domain + mail stay as their own special types; a
requirement that isn't a known app is still treated as a typo and ignored (safety
net preserved). Flavor messages kept via a small optional reason map
(_appReqServiceMsg); unknown-to-the-map services get a clean generic message.
Stays central (it's the requirements engine, not per-app logic) but is now
extensible without edits. Verified with stubs: met→rc0, absent service→flavor or
generic msg, brand-new container service→generic (zero code), typo→ignored.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Caught in the final review — config-options.js referenced the pre-rename function
name. Comment-only fix.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerCheckAllowedInstall was a one-app `case` whose only active caller was the
wireguard app itself — so inline its check (abort if a host WireGuard exists at
/etc/wireguard/params, which would collide on the wg kernel module + UDP 51820)
directly into containers/wireguard/wireguard.sh and delete
scripts/docker/app/checks/allowed_install.sh.
The protection is unchanged; wireguard is now fully self-contained and the last
app name leaves central install code. Regenerated arrays. (The only remaining
dockerCheckAllowedInstall references are in scripts/unused/ — retired apps,
never sourced.)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Last app-specific bits out of central infra (from the per-app audit):
- traefik middleware: replace the hardcoded onlyoffice/owncloud exclude-list +
onlyoffice-headers special-case (in traefik_middlewares.sh AND
traefik_port_middlewares.sh) with two per-app hooks an app ships in
containers/<app>/scripts/<app>_traefik.sh:
appTraefikSkipsDefaultMiddleware_<app> (marker: opt out of default@file)
appTraefikExtraMiddlewares_<app> (echo extra middleware entries)
onlyoffice defines both; owncloud defines the skip marker. Two narrow hooks
(not one clever one) so behavior — incl. the different onlyoffice-headers
ordering between the two files — is preserved exactly. Verified with stubs:
identical middleware strings across normal/onlyoffice/owncloud × authelia/wl.
- moneyapp: add a placeholder icon (geometric banknote SVG, 512x512) so it no
longer falls back to default.svg in the WebUI.
Central traefik/compose code is now app-agnostic.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Mirror the restic system-config adapters for the other two engines, each in that
engine's own convention, so system backup/restore/status/retention work on any
location regardless of engine:
- kopia: BackupSystemToLocation (--tags system:config), SystemSnapshotsJson
(filter tag system:config), RestoreSystemLatest, ForgetSystem (per-source policy
on $configs_dir + maintenance).
- borg: BackupSystemToLocation (archive system-<host>-<ts>, comment system=config;
no app is named "system" so the namespace can't collide), SystemSnapshotsJson
(--glob-archives system-<host>-*), RestoreSystemLatest, ForgetSystem (prune the
system-<host>-* glob).
No dispatcher change needed — engineBackupSystem/SystemSnapshotsJson/
RestoreSystemLatest/ForgetSystem already resolve <engine><fn> per location. All
three engines now define the full set; syntax clean.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Final-review gaps in the system-config backup:
1. Scheduled (cron) backups skipped it — backupScheduleEnabledApps only queued
per-app backups, so the daily schedule never refreshed the system config (and
thus the backup-location creds could go stale). Now it queues a
`libreportal backup system` task (or runs inline on terminal-only installs),
and skips the reproducible libreportal app for consistency with backupAllApps.
2. No retention on system snapshots — they bypass backupAppStart's per-app forget,
so they accumulated unbounded. Add resticForgetSystem (tag system=config,
respects append-only + the same keep-* policy) + engineForgetSystem dispatcher;
backupSystemConfig now applies retention across all locations after snapshotting.
Verified with stubs: backupSystemConfig snapshots AND prunes on every location;
engineForgetSystem pairs with resticForgetSystem; scheduled createTaskFile call
matches the existing 3-arg signature.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- Add .backup-system-card { margin-top: 20px } — the card stands alone below the
two-column cards row (which has no bottom margin), so it was butting against it.
- Add a server-stack icon to the card header (matches the nebula stroke-icon style).
- DEVELOPMENT.md: document the dashboard "System config" card + its last-backup
status (tag system=config → `system` in the dashboard JSON), the CLI/auto paths,
and that the libreportal app is excluded from the per-app grid.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Make the system config a tracked backup, not just action buttons:
- engine: resticSystemSnapshotsJson (tag system=config) + engineSystemSnapshotsJson
dispatcher — query the system snapshots the way per-app status is queried.
- webui_backup_dashboard.sh: emit a "system": { latest_snapshot, latest_time }
object (latest system snapshot on the primary location), and exclude the
libreportal WebUI app from the per-app grid (it's intentionally not backed up, so
it no longer shows a perpetual "No backup yet" tile).
- backup dashboard card: a status line (dot + "Last backed up <relative>" / "No
backup yet"), populated in renderDashboard from d.system — mirrors the app tiles.
Verified: shell + JS parse; dashboard content assembles to valid JSON with the
system key; engine query defined + dispatched; frontend reads d.system into the
#backup-system-status element.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add a "System config" card to the backup dashboard with two actions wired through
the task processor (same path as "Backup all apps"):
- "Back up now" -> libreportal backup system
- "Restore…" -> libreportal restore system (confirm dialog explains it lands
in a staging folder and never overwrites live config)
Card copy explains why it matters (the backup-location creds otherwise live only on
the box). Click handlers + runBackupSystem/confirmRestoreSystem added; JS parses,
data-actions match handlers, commands match the CLI subcommands.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Expose the system-config backup on demand (not just within 'backup all'):
- `libreportal backup system` -> backupSystemConfig (snapshot the system
config — settings, WebUI creds, backup-location creds — to all enabled locations)
- `libreportal restore system [loc_idx]` -> backupRestoreSystemConfig (restore the
latest system snapshot into a staging dir; never overwrites live config)
Distinct from the existing 'restore migrate system' (which restores all *apps*
from another host). Help text updated for both. Routing verified with stubs.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
(a) Docs: reserve tools/ scripts/ resources/ as LibrePortal folder names (apps must
not bind-mount to them); document resources/ as the home for nest-able data AND for
.sh payloads that execute on load (vs scripts/ for sourced functions); document the
backup model (what's captured vs reproducible).
(b) System-config backup so a bare-metal restore is self-sufficient — this is why
the system root is its own tree. New scripts/backup/system/backup_system.sh:
- backupSystemConfig snapshots <system>/configs (global settings, WebUI creds, and
the BACKUP-LOCATION creds — otherwise the keys to reach your own backups live only
on the box) to every enabled location. Lightweight static-dir snapshot — it does
NOT go through backupAppStart (no containers to quiesce / DBs to dump).
- restic adapter resticBackupSystemToLocation (tag system=config) + dispatcher
engineBackupSystem; restore via resticRestoreSystemLatest / engineRestoreSystemLatest
+ backupRestoreSystemConfig (restores to a STAGING dir — never auto-overwrites
live config).
- backupAllApps runs it after the app loop.
WebUI exclusion: backupAllApps skips the 'libreportal' app — its frontend + generated
JSON regenerate, and its only state (the login) is in the system config now captured
above. Nothing in its data dir warrants a snapshot.
Verified with stubs: app loop skips libreportal + invokes the system backup; the
system backup dispatches to both locations; backup/restore function names pair with
the dispatcher. NOTE: restic-only (the sole live engine adapter); end-to-end repo
round-trip still needs a live box before being relied on.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move the whole central scripts/headscale/ tree into containers/headscale/, the
last app-specific dir living centrally:
- 11 sourced function files (incl. the former local/ remote/ subdirs) flattened
into containers/headscale/scripts/ — flat because the container scan is
maxdepth 3, so one subfolder level is the limit; basenames already encode the
local/remote distinction.
- tailscale.sh is a CONTAINER PAYLOAD (ends in a bare `install_tailscale` call,
runs apt/curl) — it must never be sourced into the manager, so it goes to
containers/headscale/resources/ (pruned by the scan), NOT scripts/. Verified
install_tailscale does not leak into the runtime after sourcing.
- Fix tailscaleInstallToContainer to copy the payload from its new resources/
path (it previously referenced ${install_scripts_dir}tailscale.sh, which never
matched the file's actual location) and drop the dead commented docker-cp line.
- Remove the now-moot headscale special-case from generate_arrays.sh; regenerate
(files_headscale.sh drops — headscale is fully container-scanned now).
All 11 functions source + define cleanly; callers resolve by name regardless of
location.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
docker_config_setup_data.sh's "App Specific" if/elif ladder (pihole, nextcloud,
searxng, speedtest, vaultwarden, wireguard, gluetun) becomes a generic hook
dispatch: an app needing computed (non-CFG) compose tags ships
containers/<app>/scripts/<app>_compose_tags.sh defining appSetupComposeTags_<app>
(live-sourced by the container scan, called with the compose path; reads
host_setup/public_ip_v4/CFG_* from scope). Same declare -F pattern as the tool /
update-specifics / webui-refresh hooks.
- 7 per-app hook files added; central ladder replaced by the dispatch.
- The generic gluetun network-mode block stays (any app may route through gluetun);
tagsProcessorGluetunForwardedPorts stays central (hook + network-mode both use it).
- Regenerate arrays (hooks live under containers/, not arrayed).
Verified with stubs: each hook emits exactly the tags the old branch did
(pihole REV_SERVER, nextcloud trusted-domains, gluetun VPN set + forwarded ports,
etc.); apps without a hook are a clean no-op.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Drop the appWebuiRefresh_gluetun -> webuiGenerateGluetunProviders wrapper; rename
the function itself to appWebuiRefresh_gluetun and point the installer + the
gluetun_refresh_providers tool at it. One name, no indirection.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move scripts/webui/data/generators/apps/webui_gluetun_providers.sh ->
containers/gluetun/scripts/gluetun_providers.sh and replace the gluetun-specific
gated call in webui_updater.sh with a generic per-app loop: an installed app may
define appWebuiRefresh_<app> (in its scripts/) for data it wants refreshed on
every WebUI update. gluetun provides appWebuiRefresh_gluetun (a thin wrapper over
webuiGenerateGluetunProviders).
- No gluetun-specific code remains in central WebUI code — it's a true drop-in.
- Install gate preserved + generalized: the loop iterates the manager-owned
install templates (listable) and tests each app's live compose directly (works
without list perm on the container-user data dir), so non-users never pay for it.
- webuiGenerateGluetunProviders keeps its name (still called by the installer and
the gluetun_refresh_providers tool); now sourced via the container scan.
- Regenerate arrays (generator drops out of files_webui).
Loop verified with stubs: only installed apps with a defined hook fire; apps
without a hook are skipped; nothing fires when nothing's installed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the central app-name if-ladder in app_update_specifics.sh with a generic
dispatcher: each app ships containers/<app>/scripts/<app>_update_specifics.sh
defining appUpdateSpecifics_<app> (live-sourced by the container scan, dispatched
by `declare -F` — same pattern as tools). A hook may set shouldrestart=true. Apps
with no specifics ship no hook.
- Move the adguard/pihole (DNS updater), dashy (conf refresh), focalboard (nobody
ownership + restart), and libreportal (webui regen) branches to per-app hooks.
- Move scripts/gluetun/gluetun_route_apps.sh -> containers/gluetun/scripts/
(scripts/gluetun/ removed).
- Move scripts/install/install_crowdsec.sh -> containers/crowdsec/scripts/
crowdsec_install_host.sh; fix the path note in crowdsec.sh.
- Regenerate arrays (moved files drop out; the per-app files are container-scanned,
not arrayed).
Dispatch verified with stubs: adguard/pihole/dashy/focalboard/libreportal behave
identically to the old ladder (incl. shouldrestart propagation), apps without a
hook are a clean no-op. The CLI itself had no per-app branches — app-specific CLI
is already the (now fully modular) tools system.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The rename was justified partly by an anticipated second `libreportal-regen`
unit — which we then decided not to create (the poll rides the existing task
processor). What's left is cosmetic, and it isn't worth a footprint_version bump
(which forces a root re-install on every existing box) plus the dual-name
migration cruft.
Reverting also means the rename was the ONLY footprint change in the regen work,
so the whole regen system now ships as a plain manager-owned code deploy — no
root re-install needed. footprint_version stays 2.
Kept only the accurate FOOTPRINT.md note that the service also drives the poll.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The single systemd unit is the task processor (and now also drives the periodic
regen poll), so name it for what it does instead of the ambiguous bare
"libreportal.service" — clearer now that the runtime has more than one concern.
- svc helper: SERVICE_NAME=libreportal-taskprocessor.service; _drop_legacy()
stops/removes the pre-rename unit on install (idempotent migration) so an
upgraded box never runs two processors.
- init.sh: read baked roots from the new unit (fall back to the old name);
uninstall removes both names; bump footprint_version 2 -> 3 (root-owned unit
changed, so a manager-run update flags "root re-install needed").
- check_webui_systemd: accept either name during the transition.
- docs/FOOTPRINT.md: new unit name + uninstall command.
No sudoers change — it allows /usr/bin/systemctl generically, not a named unit.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add `lpRegen` (scripts/webui/webui_regen.sh) — one entry point that rebuilds the
file-derived artifacts whose sources changed, so callers don't have to know which
generator owns what. Self-heal is a cheap `find -newer` mtime compare (no watcher
/ daemon): a stage runs only when a source is newer than its artifact, or --force.
- `libreportal regen [all|webui|arrays] [--force]` CLI command (new category).
- Task processor idle tick runs a throttled `regen webui` poll, so an app dropped
in out-of-band (drag-drop / marketplace) appears on its own — no manual command,
no inotify (works on the relocatable/external-drive roots where inotify can't).
- make_release.sh guards against shipping stale source arrays (regenerate; abort
if the committed tree was out of date), killing the "forgot generate_arrays" bug
class at the build boundary.
- Document the front door in DEVELOPMENT.md.
webui scope rebuilds from containers/<app>/{*.config,tools/*.tools.json}; arrays
scope from scripts/** (a dev/build concern — a no-op on a normal install). Gate
logic verified in a sandbox (clean/config-newer/tools-newer/force/missing).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Each app now carries everything under containers/<app>/: Tools-tab actions in
tools/ (declaration <app>.tools.json + function <app>_<tool_id>.sh) and logic
helpers in scripts/ (e.g. <app>_auth.sh). The container scan live-sources every
.sh under the app (maxdepth 3, prunes only resources/) and webui_tools.sh
auto-merges the .tools.json, so an app is a true drop-in — no central edit, no
array regen.
- Empty the central webui_tools.sh heredoc; all 34 tools across 11 apps now
come from per-app declarations (verified byte-identical to the old output).
- Retire the orphaned mattermost tool scripts to scripts/unused (there is no
containers/mattermost; its install fn already lived in unused).
- Update the dispatch comment/error path, the auth-adapter doc, and
DEVELOPMENT.md to the new convention.
- Regenerate static arrays (files_app.sh no longer lists app/containers/*).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Establish the self-contained tools convention and prove it on a core app:
- discovery now reads containers/<app>/tools/<app>.tools.json (the tools/ subfolder);
tool functions live at containers/<app>/tools/*.sh, auto-sourced by the container
scan (depth 3) — no scripts/app/ entry, no array regen.
- adguard migrated: its 2 Tools-tab actions (reset_password, apply_dns_updater) moved
to containers/adguard/tools/ + tools/adguard.tools.json, and dropped from the
central webui_tools.sh heredoc. adguard_auth.sh stays in scripts/app/ — it's a logic
helper, NOT a tool (the key distinction: only DECLARED tools move).
Central + per-app styles coexist (pihole etc. still central), so the remaining apps
can migrate one at a time with nothing breaking. Verified: heredoc valid sans adguard,
per-app merge re-adds adguard's 2 tools, scripts array dropped the moved fns.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The previous commit accidentally tracked site/dist/** and the generated
site/src/_data/*.json — leftover Eleventy build output from the legacy site/
location (the active site now lives in containers/weblibreportal). They are
generator artifacts, not source. Untrack them (kept on disk) and gitignore so
they can't be swept in again.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
webui_tools.sh now merges any containers/<app>/<app>.tools.json into apps-tools.json
(jq, sets .apps[<app>]) on top of the central heredoc. So a dropped-in app — e.g.
from LibrePortal-Infra — registers its own Tools-tab actions WITHOUT editing this
file. Combined with the container scan already sourcing containers/<app>/*.sh live,
an app can now be fully self-contained (install fn + tool fns in <app>.sh + tool
declarations in <app>.tools.json) → true copy-on-top deploy, no array regen, no
central edits. Core apps in the heredoc are unaffected; invalid tools files are
skipped with a notice. Verified the merge (drop-in registers, core preserved).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The SSH Access page was boxed to max-width 860px and centered, unlike the
Overview and System admin pages (.admin-page) which span the full content
width. Drop the cap and match .admin-page padding so /admin/tools/ssh-access
looks like the rest of the Admin area.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
getlibreportal (downloads host) + weblibreportal (website) — including the website
Eleventy source and the publish tool functions — now live in the separate
LibrePortal-Infra repo (Webstar/LibrePortal-Infra). They're the project's own
outward-facing hosting, not something users install, so the base stays clean.
Removed from base: containers/{getlibreportal,weblibreportal}, the
scripts/app/containers/<app>/<app>_publish.sh tool functions, and their entries in
webui_tools.sh; regenerated the sourced-file arrays; dropped the dead .gitignore
docroot lines. scripts/release/make_release.sh stays here (it builds the base
release). docs/DEVELOPMENT.md now points publishing at LibrePortal-Infra.
LibrePortal-Infra overlays onto an install and picks up releases/catalogue from the
base tree — see its README.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Surface the publish step through the existing Tools system (apps-tools.json -> Tools
tab + 'libreportal app tool <app> publish'), so the docroot can be (re)built from
the WebUI instead of a manual cd + script.
- webui_tools.sh: declare a 'publish' tool (no inputs) for getlibreportal + weblibreportal.
- scripts/app/containers/getlibreportal/getlibreportal_publish.sh (appGetlibreportalPublish):
runs the host's publish.sh into the served data dir, as the container user (owns it).
- scripts/app/containers/weblibreportal/weblibreportal_publish.sh (appWeblibreportalPublish):
builds Eleventy as the manager (owns the install tree), then syncs the result into
the container-user-owned docroot — handling the build-vs-write owner split.
- Both guard for the build prerequisites (repo source / npm / dist) and fail with a
clear message; regenerated the sourced-file arrays.
Honest status: scaffolding only — wiring verified (dispatch names match, files sourced,
JSON valid) but the end-to-end tool RUN is untested, and it's build-box-only (needs the
repo checkout + npm + a built dist/). These hosting apps are dev-only and headed for a
separate repo; this just sets the automation up so it's ready to iterate on.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move the loose root-level site/ into a proper containers/weblibreportal app
(mirrors getlibreportal): the Eleventy source + nginx serving ./data via publish.sh
(npm run build -> docroot). Fix gen-data.mjs repoRoot (now ../../.. from
containers/weblibreportal/scripts) so it still finds containers/ for the catalogue.
Decouple the two hosts:
- weblibreportal -> the website (libreportal.org)
- getlibreportal -> downloads only (install.sh + signed release channels); its
publish.sh no longer builds the site, and its config text updated to match.
Both are dev-only project hosting and will move to a separate repo later; for now
they live under containers/ as normal apps. ignores updated for their built
docroots; dropped the dead 'site export-ignore'.
Verified: gen-data builds the catalogue from the new location (33 apps), and
weblibreportal/publish.sh produces a docroot with index.html.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Redo the download/website host as a normal app under containers/ (dogfooded — the
project hosts its own downloads on LibrePortal), instead of the bespoke repo-root
thing. Modeled on speedtest: standard getlibreportal.{config,sh,svg} +
docker-compose.yml (tagged template) so it plugs into the app scan + install
dispatch (installGetlibreportal) like every other app. nginx serves ./data (the
app data dir) — no special /web.
- getlibreportal.config: features category, public (login=false — it's a download
host), no backup (regenerable), healthcheck on.
- docker-compose.yml: nginx:alpine, ./data:ro docroot + ./nginx.conf, traefik tags.
- nginx.conf: install.sh + latest.json no-cache; tarball/.sha256/.minisig immutable.
- publish.sh: assembles the docroot (built site + install.sh + dist/<channel>) into
a target data dir; run on a full repo checkout (site/ + dist/ are host-side).
- exec bits set on the run-directly scripts (make_release.sh, install.sh, publish.sh).
- .gitattributes: dropped the stray 'getlibreportal export-ignore' (the no-slash
pattern would also have excluded containers/getlibreportal — the app must ship);
data/ gitignored.
Verified: app discovered by the site catalog (32 apps), installGetlibreportal matches
the dispatch name, and the full release->publish flow yields a docroot with the
website + install.sh + the signed/checksummed stable channel. The actual app-install
run + DNS/TLS for get.libreportal.org are operational steps (need a real host).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The sha256 only proves a download is intact; a compromised host could swap the
tarball + its checksum. Add minisign signatures, which prove authenticity (the host
can't forge them without the offline secret key). Ships INACTIVE behind a REPLACE_ME
placeholder, so installs work until a real key is generated; then it's REQUIRED.
- make_release.sh: signs the tarball when LP_MINISIGN_SECKEY is set -> <tarball>.minisig.
- libreportal.pub: the public key (placeholder), ships in the tarball and is installed
to the ROOT-OWNED footprint (/usr/local/lib/libreportal/libreportal.pub) by init.sh
-> the manager can't swap it to accept forged updates. footprint_version -> 2.
- install.sh: LP_MINISIGN_PUBKEY constant; once non-placeholder, downloads + verifies
the .minisig (minisign -P) and REFUSES on invalid/missing (auto-installs minisign if
needed). --no-verify-signature is a dev-only escape hatch.
- fetch.sh (update path): verifies against the footprint .pub (minisign -p), refuses on
invalid/missing.
- docs/DEVELOPMENT.md: keygen (minisign -G), paste pubkey into libreportal.pub +
install.sh, keep the secret key offline, sign builds via LP_MINISIGN_SECKEY, bump
footprint_version on key rotation.
Verified end-to-end with a real throwaway key: good signature accepted; tampered,
wrong-key, and missing-signature all refused; placeholder skips (sha256 still enforced).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
A manager-run 'update apply' refreshes code/apps/WebUI but CANNOT rewrite the
root-owned footprint (helpers/wrapper/uninstall/unit/sudoers) — that immutability
is the de-sudo boundary. Previously a release that changed those would silently
leave them stale. Make it explicit:
- init.sh: footprint_version=1 constant, baked at install into
/usr/local/lib/libreportal/.footprint_version (root:root 0644) by initRootHelpers.
Bump it whenever a root component changes.
- make_release.sh: publishes footprint_version in latest.json.
- fetch.sh: lpInstalledFootprintVersion (marker) + lpReleaseLatestFootprint (manifest).
- check_update.sh: 'update apply' REFUSES when the release's footprint_version
exceeds the installed one, directing to a root re-install (which fetches +
re-bakes everything atomically). No half-applied updates.
- webui_system_update.sh: badge sets footprint_update_needed + clears can_update so
the WebUI won't offer a one-click apply for a footprint-bumping release.
- docs/DEVELOPMENT.md: the bump rule + the footprint exception explained.
Verified: manifest carries footprint_version; drift decision correct both ways
(no marker/older -> needs re-install; equal -> no drift).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Remove the redundant repo-root uninstall.sh (it duplicated libreportal-uninstall).
init.sh now GENERATES the libreportal-uninstall launcher into the fixed footprint
(/usr/local/lib/libreportal/uninstall.sh + the /usr/local/bin symlink) — same
pattern as the CLI wrapper, so the on-box command survives without a separate repo
file. The launcher just runs the engine's uninstall ($script_dir/init.sh baked in,
/root/init.sh fallback).
This resolves the install/uninstall asymmetry: a bootstrap (install.sh) exists only
because install faces a bare box with no code yet; uninstall always runs the engine
that's already installed, so it needs no bootstrap — just a generated door into
init.sh. Repo root install/uninstall surface is now init.sh + install.sh.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The docs were telling users to run /libreportal-system/install/uninstall.sh — a
hardcoded data path, wrong for any custom --system-dir, contradicting the whole
relocatable design.
Fix it the way the CLI already works: install uninstall.sh to the FIXED footprint
(/usr/local/lib/libreportal/uninstall.sh) and symlink it onto $PATH as
'libreportal-uninstall' (initLibrePortalCommand). It self-resolves the real data
roots from the systemd unit, so the command is the same everywhere regardless of
where data lives. Teardown removes the new symlink; FOOTPRINT.md lists it.
Docs now say 'sudo libreportal-uninstall' — no data path. (Dev-from-clone still
uses ./uninstall.sh / ./init.sh uninstall.)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Bug: runFullUninstall used the derived $docker_dir/$containers_dir/$backup_dir,
but a bare 'init.sh uninstall' on a CUSTOM-location install has no LP_*_DIR in
scope and no /docker marker — so it defaulted to /libreportal-* and would MISS the
real data (e.g. /mnt/ssd), leaving it behind.
Fix: libreportalReadBakedRoots reads the authoritative baked record from the
systemd unit (Environment=LP_SYSTEM_DIR/CONTAINERS_DIR/BACKUPS_DIR + User=<manager>)
and runFullUninstall re-derives from it before removing anything. Legacy units
(no LP_*_DIR) fall through to the derive defaults + /docker compat shim.
Add top-level uninstall.sh: a root-only convenience that finds the installed
init.sh (via the unit's system root, then common locations) and runs it —
'sudo ./uninstall.sh [--skip-docker-images]'. Verified the unit parsing extracts
custom roots/manager and the discovery picks the right init.sh (without running the
destructive teardown).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Tidy the repo root (README + LICENSE stay there per convention; everything else
moves):
- CONTRIBUTING.md, PROMISE.md, FOOTPRINT.md -> docs/ (alongside USER.md/DEVELOPMENT.md)
- update the references: README links, the website site.json raw URLs, init.sh's
'see FOOTPRINT.md' comments -> docs/FOOTPRINT.md; drop the now-redundant
CONTRIBUTING.md export-ignore (docs/ is already export-ignored).
Refresh FOOTPRINT.md: it claimed 'everything lives under /docker', which is no
longer true. Now describes the three relocatable roots (system/containers/backups)
and makes explicit that the roots + manager name are baked into the helpers/unit/
wrapper at install (the privilege boundary) while this out-of-root footprint stays
fixed by design. Uninstall sketch + sudoers/unit rows updated for the configurable
manager. CONTRIBUTING/PROMISE were already current — left as-is.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Two guides covering what wasn't written down:
- USER.md: install (the install.sh one-liner), placing the three roots on separate
disks/external drives, channels, updating, backups (REQUIRE_MOUNT), uninstall.
- DEVELOPMENT.md: the install-mode/roots/users model + key files; running a dev copy
(local/git); cutting stable/edge releases (bump VERSION -> make_release.sh ->
dist/<channel>/{tarball,.sha256,latest.json} -> publish); testing a release
locally via LP_RELEASE_BASE_URL + python3 -m http.server (incl. checksum-refusal);
how release updates work; conventions.
README Quick start updated to the release flow + a docs pointer. docs/ is
export-ignored so it doesn't bloat release tarballs.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The git-era recovery commands now do the right thing in release mode instead of
attempting a clone:
- gitReset (libreportal reset / update reset) and runReinstall (CLI/system reset,
missing-files recovery): a release branch re-fetches the verified tarball via
lpFetchRelease, then refreshes /root/init.sh + ownership.
- the CLI wrapper's clone_and_install (libreportal reset): sources fetch.sh and
re-fetches the release; falls back to directing the user to the install.sh
bootstrap if the helper isn't present.
git/local behaviour unchanged. Wrapper still bakes cleanly (no placeholders left).
Phases A–D (release build, bootstrap installer, fetch abstraction, release-aware
install + update + recovery) are complete and locally verified. Remaining: phase E
(host install.sh + channels + tarballs on get.libreportal.org) and a real fresh
install on a throwaway box.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The admin landing (overview) and the tools pages (ssh-access, system) call
populateSidebar() without first loading window.configData. On a cold admin
visit — e.g. navigating straight from the dashboard — configData is undefined,
so populateSidebar() bails early and the sidebar renders empty. Visiting
Backups happened to set window.configData, which is why returning to admin
afterward showed the sidebar.
Load (cached) config data up front in renderConfig before any branch renders so
the sidebar always has its categories. The config-category path's later
loadConfig is now a cache hit.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Make the WebUI updater work off release versions, not git commits, in release mode
(git/local paths untouched):
- webui_system_update.sh: a release branch resolves latest_version from the channel
manifest (lpReleaseLatestVersion), computes update_available via lpVersionGt vs
the local VERSION, reuses the same throttle + the same update_status.json schema
(source="release"); reuses last-known latest when throttled so the badge
doesn't flicker.
- check_update.sh webuiRunUpdate: a release branch version-compares and, if newer,
lpFetchRelease (download + checksum-verify) the new tarball + dockerInstallApp
redeploy + regen. No config-backup dance — lpFetchRelease replaces only the
install tree; configs/logs are in the separate system tree.
Verified against a local server: latest-version read + the no-update / update-
available decision (0.2.0==0.2.0 no; 0.3.0>0.2.0 yes). Remaining: route the
reset/reinstall recovery paths through the release fetch.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Adding scripts/release/ made generate_arrays emit files_release.sh, which would
source the build tool make_release.sh at runtime (it runs git archive at load —
wrong + harmful). Skip release/ in the folder scan (like system/ and unused/),
remove the orphaned files_release.sh, and regenerate the meta-list. fetch.sh stays
sourced; make_release is in no array.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
scripts/source/fetch.sh (sourced at runtime via files_source.sh):
- lpFetchRelease [ver]: resolve channel manifest -> download tarball -> VERIFY
sha256 (refuse on mismatch/absence) -> replace the install tree ( is
code-only now; configs/logs live in the separate system tree, so no backup
dance). Host/channel from LP_RELEASE_BASE_URL/CFG_RELEASE_BASE_URL + channel.
- lpFetchSource: dispatch release|git|local.
- lpVersionGt: numeric dotted semver compare (used by the updater + badge).
init.sh initGIT is now release-aware: the bootstrap (install.sh) stages+verifies
the code and sets LP_ALREADY_FETCHED=1 (skip re-fetch); a direct release run sources
fetch.sh; a bare /root reinstall is directed to install.sh. install.sh exports
LP_ALREADY_FETCHED + LP_RELEASE_BASE_URL on hand-off. validateUnattended already
accepts release (git-url is gated on git mode).
Config: CFG_INSTALL_MODE default -> release, + CFG_RELEASE_BASE_URL / CFG_RELEASE_CHANNEL
(add-only reconcile preserves existing installs' git/local mode).
Verified: lpVersionGt across cases; lpFetchRelease downloads+verifies+extracts a
clean tree against a local server. The updater + reset/reinstall release paths are
phase D.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
curl -fsSL https://get.libreportal.org/install.sh | sudo bash [-s -- FLAGS]
Self-contained bootstrap (downloaded and run before any LibrePortal code exists):
resolves the channel manifest (<base>/<channel>/latest.json), downloads the
release tarball, VERIFIES the sha256 (manifest or .sha256 sidecar) and refuses on
mismatch, extracts to <system>/install, then hands off to init.sh unattended with
the relocatable flags (--system-dir/--containers-dir/--backups-dir/--manager-user/
--allow-home). curl-or-wget, sha256sum-or-shasum; honours LP_RELEASE_BASE_URL for
testing against a local/file server.
Dev modes preserved: --local=PATH and --git-url= (with --git-user/--git-token).
Generates a random manager/WebUI password if none given. --dry-run stages + verifies
without installing.
Verified against a local http server: dry-run resolves→downloads→verifies→extracts
and reports the correct init.sh handoff; a corrupted tarball is refused. The
end-to-end exec needs init.sh release-mode awareness (phase C) + a throwaway box.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
scripts/release/make_release.sh builds a versioned, checksum-verified release
artifact from the committed tree via 'git archive' — the basis for installing
LibrePortal without git/auth. Output lands in dist/<channel>/ laid out as the
hosting will serve it (libreportal-<ver>.tar.gz + .sha256 + latest.json), so it
can be served locally for testing via LP_RELEASE_BASE_URL.
.gitattributes marks dev-only trees export-ignore (scripts/unused, scripts/release,
site, .claude, CONTRIBUTING.md, the git meta files) so they never ship; validated
the archive includes init.sh/start.sh/scripts/configs/containers/VERSION and
excludes all of the above. dist/ gitignored.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
sudo_user_name (the real manager — owns the install, runs the runtime, baked as
__MANAGER__ into the root helpers) was hardcoded to 'libreportal'. Make it
configurable, consistent with the relocatable roots:
- --manager-user=NAME flag + LP_MANAGER_USER env (default libreportal); resolved
early in init.sh and in scripts/source/paths.sh (so the standalone processors
get it too), validated as a real Linux username in libreportalValidatePaths.
- Baked everywhere it must be stable: the helpers + CLI wrapper (CHECK_USER now
__MANAGER__, exports LP_MANAGER_USER) via the install-time sed; the systemd unit
exports LP_MANAGER_USER=<manager>. User creation (initUsers), the sudoers
drop-in, and ~35 call sites already used $sudo_user_name, so they follow.
- Fix the stray manager-name literals: install_crowdsec.sh chown, the
check_install_type fallback. (Brand/identity strings like the backup
engine:libreportal tag are left — they're not the username.)
Verified: resolves default/env/flag; wrapper bakes a custom name (admin) with no
placeholders left; validation rejects invalid usernames. The footprint paths
(/etc, /usr/local) stay fixed by design.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
CFG_DOCKER_MANAGER_USER / installDockerManagerUser was a chrooted SFTP file-access
user — unrelated to the LibrePortal control-plane manager (sudo_user_name), and
the source of the 'two managers' confusion. It was permanently-off dead code: the
gate CFG_DOCKER_MANAGER_ENABLED and the CFG_DOCKER_MANAGER_USER/_PASS keys are
defined in no config template, so it never ran. Its SSH-key-management sibling
(unused/ssh_manager.sh) was already retired; admin host SSH access is handled by
the current /ssh page + scripts/ssh/host_access.sh.
Move install_user_manager.sh / uninstall_user_manager.sh / check_manager.sh to
scripts/unused/manager/ (recoverable, matches the graveyard convention — not
deleted, in case the SFTP-user idea is rebuilt cleanly later), drop the two call
sites (start_preinstall.sh, check_requirements.sh), regenerate the arrays.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Audit follow-up — after a full-repo sweep, the only remaining functional /docker
refs are intentional (the legacy compat shim + the env-overridden legacy-safe
backend default). Fix the last user-visible/stale ones:
- config-options.js: backup PATH_MODE 'auto' label no longer hardcodes
/docker/backups (the path is relocatable) — describes the behaviour instead.
- config.js / setup-detector.js / webui_install_image.sh: refresh comments that
named /docker to the relocatable system/containers roots.
No behaviour change. Active container app scripts already use $containers_dir;
the remaining /docker hits across the tree are docker-compose.yml filenames,
/var/lib/docker, the docker binary, relative array paths, docs/site, and the
unused/ graveyard.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
backupLocationLocalGuard (engine-agnostic, in location_paths.sh), wired into the
dispatcher before init, readiness, and every backup write (engineInitLocation /
engineEnsureLocationReady / engineBackupApp):
- Filesystem warning: the ownership model chowns the repo to the backup user, which
needs POSIX permissions — warn (non-fatal) on FAT/exFAT/NTFS via findmnt FSTYPE.
- Mount-presence refusal: a location with CFG_BACKUP_LOC_<idx>_REQUIRE_MOUNT=true
(an external/removable disk) is refused when its path isn't on a real mount
(findmnt TARGET is '/' or unknown) — so an unplugged drive never silently fills
the system disk. Opt-in; default false leaves on-disk locations unaffected.
New REQUIRE_MOUNT field documented in the location.config template (location_add.sh)
so it surfaces on the Locations page. Verified: REQUIRE_MOUNT+unmounted refuses;
default allows; non-local no-ops.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Make the three roots selectable at install and bake them into the CLI wrapper
(the last /docker-hardcoded consumer).
- init.sh: --system-dir= / --containers-dir= / --backups-dir= flags (=form keeps
the single-token shift logic), plus --allow-home; LP_*_DIR env also honored.
Re-derives paths after flag parsing.
- libreportalValidatePaths (run only in the install flow): each root must be a
non-root absolute path outside protected system trees; the three must not nest
(except the legacy /docker compat layout); a containers/backups root inside a
human home is refused unless --allow-home (rootless o+x traversal = privacy
trade-off). The root helpers re-check at runtime (defence in depth).
- CLI wrapper: a baked bootstrap (the same __ROOT__ placeholder mechanism as the
helpers) exports LP_*_DIR and derives docker_dir/configs_dir/script_dir; every
/docker literal in the heredoc now resolves from those at runtime. init.sh seds
the placeholders into the root-owned wrapper after writing it.
The scoped sudoers needs no change (it references only the fixed helper paths +
system binaries, never a data root). Custom locations verified end-to-end:
generate+bake the wrapper with /mnt/* roots → syntax OK, no placeholders left,
paths resolve. Live box untouched (wrapper/helpers only change on reinstall).
Phase 3b (external-drive guards) + phase 4 (verify) follow.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Split the single tree into three owner-isolated roots and fix the backup
permission failure (restic, running as the container user, could not write the
manager-owned /docker/backups).
Ownership helper (libreportal-ownership), rewritten for three baked roots:
SYSTEM_DIR (manager) CONTAINERS_DIR + BACKUPS_DIR (container user)
- reconcile now drives each tree to its single owner; backups + the WebUI dir go
to the container user (the actual fix). The container user reaches only the
WebUI bind-mount sources (configs/webui/*) via a scoped _webui_bind_access —
traverse the system root + configs, read configs/webui only, nothing else.
- defence-in-depth: refuse dangerous/relative roots even if mis-baked; new
backups-top action.
Baking: init.sh initRootHelpers now seds __SYSTEM_DIR__/__CONTAINERS_DIR__/
__BACKUPS_DIR__ (alongside __MANAGER__) into every helper at install — the trust
boundary stays root-controlled. svc/socket/appcfg helpers updated to derive from
the baked SYSTEM_DIR; the svc unit now exports LP_*_DIR so the processor resolves
roots authoritatively. A baking-safe '*"__"*' sentinel check survives the sed.
Install/uninstall: initFolders creates the three roots; initContainerLayer hands
containers + backups to the container user; uninstall removes all three
(idempotent on legacy single-tree installs). Remaining functional /docker
literals in init.sh (config reads, setupConfigsFromRepo, uninstall) parameterised.
Compose: the WebUI's two relative ../../configs mounts (the only cross-tree
relative mounts in the tree) are now absolute, filled at generation via a new
CONFIGS_DIR_TAG; CONTAINERS_DIR_TAG likewise for the LP_CONTAINERS_DIR env.
Live box unaffected: installed helpers + the live compose only change on reinstall/
rebuild (both of which fill the tags); the CLI-wrapper heredoc paths are baked in
phase 3.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Introduce scripts/source/paths.sh as the canonical path resolver for three
independently-relocatable roots:
LP_SYSTEM_DIR manager-owned control plane (configs/logs/install/db/ssl/ssh/migrate)
LP_CONTAINERS_DIR container-user-owned live app data
LP_BACKUPS_DIR container-user-owned backup repos (own mount-able)
Roots come from the environment when set (install bakes them; CLI/app inherit
from init.sh), else default to /libreportal-*. A transitional compat default
keeps EXISTING installs (legacy single /docker tree, by config marker) on /docker
until a deliberate reinstall, so deploying this never strands a running box.
- init.sh derives the same vars inline (self-contained for the bare /root/init.sh
reinstall case); paths.sh mirrors it for the standalone task/check processors,
which now self-locate their scripts dir and source it.
- Replace functional /docker literals with the derived vars across runtime,
install, backup, crontab, crowdsec/restic, headscale, and reinstall paths;
clean the inert '== /docker/containers/*' guard fallbacks to the variable form.
- backend: CONTAINERS_DIR now from LP_CONTAINERS_DIR (compose env, filled at
generation via a new CONTAINERS_DIR_TAG), legacy-safe default for un-recreated
containers.
- backup default path falls back to the backups root; exclude paths.sh from the
sourced-file arrays (bootstrap file, sourced explicitly).
The CLI-wrapper heredoc + root helpers still reference /docker; those get baked
in phase 3. No layout/ownership change yet (phase 2).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The root-owned helpers all live in the same fixed dir, so printing the
full /usr/local/lib/libreportal/... path on each success line was long and
repetitive. Use the bare helper name, matching the error branch below.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
When the task processor service is down the rootless daemon socket is
absent, so the pre-install `docker images` probe printed a raw daemon
connection error to the terminal. The surrounding notices already convey
the meaningful state (service not running → image not setup), so the raw
error was noise.
Capture the probe output and redirect its stderr to libreportal.log
instead of the terminal, keeping the detail for diagnostics.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Same class of bug as the topbar partial: icon and data-file references were
relative (icons/apps/x.svg, data/apps/...), so on deep path routes (/app/<name>,
/admin/config/x) the browser resolved them against the route dir and the SPA
catch-all served index.html with HTTP 200 instead of 404 — broken images and
silently-wrong JSON.
Make every reference absolute (anchored on the quote/backtick so already-absolute
/icons paths are untouched):
- JS: all icons/ and data/ literals + templates across components/utils/system
- html/topbar.html: logo <img>
- generators: webui_config.sh and webui_create_app_categories.sh now emit
/icons/... into apps.json / apps-categories.json (regenerated on install)
- updated the two icon-path comments to match
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
userdel does NOT remove /var/spool/cron/crontabs/<user>, so across an
uninstall->reinstall the manager's uid can be recycled (e.g. 1001 -> 1003)
while the old spool file stays owned by the dead uid. The spool dir is
sticky (1730), so the new manager can't rename its temp over the
old-uid-owned file → "crontab: crontabs/libreportal: rename: Operation
not permitted", and the crontab silently never updates (the "added"
success message doesn't check the rename). Same class as the stale
easydocker spool left by the pre-rename migration.
Two fixes:
- runFullUninstall removes each torn-down user's cron spool (+ the legacy
easydocker one) so teardown stops leaving orphans.
- initUsers defensively drops a manager cron spool owned by a different
uid (recycled) before the manager-run crontab setup runs — fixes an
already-dirty box and any uid drift, in both modes.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Path-based routes (e.g. /app/<name>) made the relative fetch('html/topbar.html')
resolve to /app/html/topbar.html. The SPA catch-all returns index.html with HTTP
200 instead of 404, so response.ok passed and index.html got injected as the
topbar, leaving #nav-app-center absent -> 'Nav element not found' in setActiveNav.
Make the topbar fetch and the loadConfig fetch absolute, and switch the remaining
relative topbar nav hrefs (index/dashboard/tasks .html) to absolute paths so the
SPA click interceptor routes them instead of doing a real browser navigation.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The previous commit handed /docker/containers to the container user but
left /docker itself at initFolders' 750 (manager-only) during the install
— so the container user couldn't traverse INTO /docker to reach its now-
owned containers/, and the boot scan still hit "find:
'/docker/containers/': Permission denied" (the dir's documented rootless
mode is 751, but the reconcile that sets it runs later). initContainerLayer
now adds the o+x traversal bit to /docker (→ 751) alongside the
containers/ handover, so the boot scan can both enter /docker and read
containers/.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Reverts the 2>/dev/null band-aids and fixes the root cause. The
manager-run install boot scans app configs under /docker/containers AS
the container user (runFileOp). But init.sh's initFolders creates that
dir manager-owned, and the handover to the container user happened later
(start_preinstall), AFTER the boot scans — so the scans ran as the
container user against a dir it didn't own yet: "find:
'/docker/containers/': Permission denied" (cosmetic; the dir is empty
that early, but it's the wrong ownership at the wrong time).
Add initContainerLayer() to init.sh's root phase (after initGIT +
initUpdateConfigs, before the manager-run handoff): rootless-only, it
creates the docker-install user if missing and chowns /docker/containers
to it (751). The later rootless setup is now idempotent — it finds the
user existing and just (re)asserts its password + daemon config (moved
updateDockerInstallPassword out of the create-only branch). Rooted is
unaffected (containers stay manager-owned, which the manager reads).
Result: by the time the boot scans run, /docker/containers is owned by
the user doing the scanning — no permission error, nothing suppressed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Early in an install the docker-type config isn't loaded yet, so runFileOp
falls back to the manager, which can't list the container-owned (751)
/docker/containers/ dir. Two best-effort scans then leaked
"find: '/docker/containers/': Permission denied" to the install output
(x3 per run): scan_files.sh's app_configs scan and the application config
reconcile. No app configs exist that early on a fresh install, so the
empty result is correct — just suppress the find stderr (the -print0
output still flows). Cosmetic only; doesn't change what's enumerated once
the config is loaded.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Both used the pre-migration query/.html URL form through navigation that
no longer exists, so they landed on a not-found / wrong page:
- setup-wizard handoffToTasks: navigated to `tasks.html?task=<id>` via the
never-defined window.router, falling back to a *relative*
window.location.href. From any non-root path that resolves under the
current path (e.g. /admin/config/tasks.html → matches the /admin*
route), so the first-install "x of x installing" hand-off hit a
not-found task page. Now navigates to the path-based
`/tasks/all?task=<id>&from=setup` via window.navigateToRoute (absolute
full-load fallback).
- apps-manager getNavigationButton / handleNavigation: the "Install
<Service>" buttons on config requirement fields used
`app.html?app=<name>` with a relative window.location.href; from the
/admin/config/* pages they render on, that resolved to
/admin/config/app.html (wrong route). Now `/app/<name>` via
navigateToRoute.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
A full uninstall tears down the rootless daemon and removes the
docker-install user's home, which destroys the WebUI image AND the build
cache — so every reinstall's `docker build` runs from scratch (slow,
re-pulls the base image + reinstalls deps). On a slow local box that
dominates the iteration loop.
--skip-docker-images on `init.sh ... uninstall` preserves the rootless
docker layer: it still removes stale containers, the control plane,
manager user, footprint and /docker, but keeps the daemon running, the
docker-install user + home (image/layer cache), and the rootless sysctl
drop-in. The following reinstall then finds rootless already set up and
rebuilds the WebUI image from cache — fast. No effect on install.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
On a fresh install the requirement checks run before the things they
probe exist, leaking raw command stderr:
- check_install_type.sh: `$( (id -u "$user") )` printed
"id: 'dockerinstall': no such user" to the terminal AND — since id's
error goes to stderr, not the captured stdout — the next line's
`[[ "$ISUSER" == *"no such user"* ]]` could never match, so the
rootless-user-absent branch was dead. Add `2>&1` (matching siblings on
lines 25/31): no leak, and the check now works.
- grep on $sysctl (the rootless marker conf, absent until rootless is set
up) printed "grep: /etc/sysctl.d/99-libreportal-rootless.conf: No such
file or directory". Add -s to the four $sysctl greps
(check_docker_rootless, rootless_start_setup, rootless_docker x2);
"marker absent" is still detected (non-zero exit), just without the
file-not-found message.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The two docker-type-switcher finds run mid-switch, BEFORE
reconcileDockerOwnership, so containers/ is still owned by the OLD mode's
container user while CFG_DOCKER_INSTALL_TYPE is already the target. A
plain runFileOp resolves to the target user, which can't list the
old-mode-owned (751) dir under rootless — so enumerate as the old-mode
owner instead:
- switchMigrateBackupApps: move the find inside the existing
old_mode/resolveDockerInstallUser window (runFileOp now resolves to the
old owner). It previously ran as the manager and silently enumerated
nothing under rootless, so no app got backed up before the switch.
- dockerSwitcherUpdateContainersToDockerType: take old_mode as an arg,
flip CFG to it only for the find (restore before the per-app socket
scan + restart, which need the new daemon). Callers in swap_docker_type
pass $docker_type. The two former rooted/rootless branches were
byte-identical and are collapsed.
NOTE: the full rooted<->rootless switch round-trip is still unvalidated
on the VM (needs a stateful app + an enabled backup location); this fixes
the container enumeration, not yet the end-to-end migration.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Bare `find "$containers_dir"` runs as the manager, but under rootless
containers/ is dockerinstall-owned 751 (traversable, not list-readable by
the manager) -> "find: /docker/containers/: Permission denied". For the
app-log generator that was cosmetic; for dockerComposeUpAllApps /
dockerComposeDownAllApps it silently enumerates nothing so no apps come
up/down. Route these through runFileOp find (dockerinstall in rootless,
manager in rooted — correct in both). The two docker-type switcher finds
are deliberately left: mid-switch the at-rest container owner can differ
from the target-mode user runFileOp resolves to, so they need mode-aware
handling rather than a blind swap.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
installLibrePortalImageWebUI copies the WebUI template into the
dockerinstall-owned containers/ dir, but on a fresh install the general
traversal/ownership reconcile (fixFolderPermissions -> runOwnership
traversal) runs LATER. So at copy time /docker is still 750
(untraversable by the container user) and containers/ may still be
manager-owned, and the copy fails ("tar: /docker/containers: Cannot
open: Permission denied"), cascading into the WebUI never starting on a
first install. Call fixFolderPermissions first so /docker is +x and
containers/ is owned by the container user before the copy.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Two pre-existing bugs a genuinely-clean rootless install exposes:
copyFolder picked the copy user by destination only: a manager-owned
source (e.g. the install dir) copied into the dockerinstall-owned
containers/ ran the cp AS dockerinstall, which can't read the source ->
"cp: Permission denied". The `local result=$(...)` then masked the
failure (local returns 0) so checkSuccess printed success. This broke
installLibrePortalImageWebUI: the WebUI dir wasn't populated, so
initializeAppVariables couldn't read libreportal.config ("No app name
provided"), compose tags were never substituted, and the WebUI container
couldn't start (user: "USER_DATA"). Fix: when source and destination
owners differ (manager -> container), bridge with a tar pipe — the
manager reads, dockerinstall writes — with pipefail so a read-side
failure is no longer masked.
start.sh created the per-run install log with `sudo touch` (root:root
644) but tee's to it as the manager -> "tee: Permission denied" -> every
install-*.log was empty. Fix: chown the log to the user running the
install so the tee can append.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
initRootHelpers ran inside initUsers, before initGIT copies the repo into
/docker/install — so it read helper sources from a not-yet-populated
$script_dir/scripts/system and skipped all 7 ("Root helper source
missing"). This was masked on every prior install because the old
deploy's `rm -rf /docker` left /usr/local/lib/libreportal/ intact, so the
helpers were simply never reinstalled. A genuine clean install (now that
the deploy uses the full uninstall) exposed it: the runtime ended up with
only the CLI wrapper, the scoped sudoers pointed at missing helper paths,
and the WebUI never came up.
The helpers are only needed at runtime (the install phase uses the broad
install-phase sudo), and nothing between initUsers and initGIT uses them,
so move the call to right after initGIT (before initLibrePortalCommand,
which already installs the wrapper to the same dir post-copy).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
runFullUninstall always prompted for `DELETE LIBREPORTAL`, so it couldn't
be driven non-interactively. Honor the existing global --unattended flag
(init_unattended_mode) to skip the prompt; an interactive `init.sh
uninstall` still requires it.
This lets the deploy helper do a clean teardown (`init.sh --unattended
uninstall`) for a full reinstall instead of `rm -rf /docker`. The brute
wipe left the task-processor systemd service running against a deleted
runtime dir; init.sh's idempotent service setup then saw an unchanged
unit and skipped the restart, so the reinstalled WebUI container was
never started. The uninstall stops the service and tears down the
rootless daemon + users in order, so the follow-up install behaves like
a true first install.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The install hands the heavy setup to the manager (completeInitMessage:
sudo -u libreportal 'libreportal run install') — creating the
docker-install user, rootless setup, apt, sysctl — which needs broad root.
initUsers was installing the SCOPED sudoers up front, so that handoff died
with 'sudo: a password is required' on useradd. Fix: initUsers installs a
temporary NOPASSWD: ALL for the install phase; completeInitMessage calls
the new initScopedSudoers to tighten to the runtime allowlist only after
the install succeeds (on failure, broad sudo is left so the manual
'libreportal run install' retry works). This restores the documented
'kill NOPASSWD:ALL AFTER the runtime is set up' ordering.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The validation teardown left /home/libreportal orphaned: userdel -r skips
the home when the user still has a live session/processes, and the manager
only got a pkill (not a loginctl terminate) before userdel. Now both users
get disable-linger + terminate-user + pkill before userdel -r, plus an
explicit rm -rf /home/<user> backstop.
Signed-off-by: librelad <librelad@digitalangels.vip>
A single 'sudo bash init.sh uninstall' that permanently removes the whole
LibrePortal footprint, behind a typed 'DELETE LIBREPORTAL' confirmation:
- stops + removes the task-processor service
- best-effort graceful container removal, then tears down the rootless
docker setup + the install user's session (linger/terminate/pkill)
- removes the out-of-/docker footprint (/usr/local/lib/libreportal +
/usr/local/bin/libreportal, /etc/sudoers.d, the systemd unit, the
sysctl drop-ins, restic/kopia/ufw-docker, /root/init.sh)
- rm -rf /docker
- removes the libreportal + dockerinstall users + subuid/subgid ranges
Runs as root (the entrypoint root-check enforces it — and the scoped
sudoers can no longer self-remove anyway); self-contained (only init.sh's
inline helpers, so it works as it deletes /docker); ordered so containers/
daemon stop before the users are removed. Leaves docker/compose/apt deps
and SSH config in place (no lockout). Mirrors FOOTPRINT.md.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Organise the system footprint outside /docker:
- All LibrePortal executables now live together in /usr/local/lib/libreportal/
(root:root): the 7 root helpers AND the CLI wrapper. /usr/local/bin/libreportal
becomes a symlink onto $PATH. run_privileged._runRootHelper, init.sh
(initRootHelpers + scoped-sudoers Cmnd_Alias + command setup) all point there.
The wrapper is now root-owned too (manager can't tamper with its entrypoint).
- Fix a real bug: rootless sysctl settings were written to /etc/sysctl/99-custom.conf,
a dir does NOT read, so net.ipv4.ip_unprivileged_port_start /
kernel.unprivileged_userns_clone never persisted across reboot. Moved to
/etc/sysctl.d/99-libreportal-rootless.conf (the existing
reload now actually applies them). Consistent libreportal* naming.
- Drop dead fqdn_file=/root/libreportal-fqdn.txt global (never used).
- Add FOOTPRINT.md: a manifest of every file LibrePortal places outside /docker
(doubles as an uninstall checklist).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Bring the remaining deferred subsystems under the scoped sudoers, and drop
the one that's redundant.
Backup engines + app configs -> root-owned helpers (same pattern as
ownership/dns/ssh/socket/svc):
- scripts/system/libreportal-bininstall: install <restic|kopia> — does the
whole pkg-manager/signed-download install itself for a fixed, validated
engine name (no blanket sudo apt-get/install). restic_install/kopia_install
call it.
- scripts/system/libreportal-appcfg: {adguard-auth <user> <bcrypt>|
crowdsec-priority|owncloud-config <public> <host> <ip> <public_ip>} —
faithful ports of the AdGuard yaml / CrowdSec bouncer / ownCloud config.php
rewrites, fixed paths + validated args. adguard_auth/crowdsec_fix_priority/
owncloud_setup_config call it.
- run_privileged: runBinInstall / runAppCfg; init.sh installs + allowlists both.
Retire standalone (host-level) WireGuard — it's a duplicate of the
containerized containers/wireguard app (+ headscale mesh), its slirp4netns
speed rationale is largely moot with a better rootless net backend / typical
WAN-bound throughput, and it was the heaviest host-root subsystem (apt +
sysctl + iptables + /etc/wireguard), the worst fit for the rootless/
least-privilege direction:
- moved scripts/wireguard/ + manage_wireguard.sh + check_wireguard.sh to
scripts/unused/; dropped the install-path call, the Tools menu 'w' entry,
and the requirement check; removed the half-built libreportal-wg helper.
- generate_arrays.sh now also skips system/ (root-owned helpers, never
sourced); arrays regenerated (files_wireguard.sh pruned).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
installLibrePortalImageWebUI copyFolder's the template docker-compose.yml
(raw #LIBREPORTAL|TAG|VALUE placeholders) over the runtime one on every
WebUI build — including rebuilds/updates. On a fresh install the following
dockerInstallApp substitutes them, but on a rebuild (libreportal already
installed) nothing did, so the at-rest compose kept raw placeholders and a
plain 'docker compose' against it failed ("invalid boolean:
HEALTHCHECK_DATA", etc.) — it only worked because up_app.sh self-heals at
CLI start time. Re-run the tag processors (initializeAppVariables +
dockerConfigSetupFileWithData, the same heal up_app.sh uses) right after
the copy when libreportal is already installed, so the runtime compose is
always fully substituted at rest.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the NOPASSWD: ALL drop-in with a validated, scoped grant:
- (dockerinstall) NOPASSWD:SETENV: ALL (data plane; rootless-confined)
- (root) NOPASSWD: the 5 root-owned /usr/local/sbin/libreportal-* helpers
+ a fixed system-binary allowlist (systemctl/ufw/ufw-docker/nft/sysctl/
loginctl/service)
No bash/su/tee/cp/chmod/chown/sed/mv/rm/install — none of the
root-equivalent primitives. Also: drop '-G sudo' from the manager useradd
(privileges come from the user-specific drop-in, not group membership),
and defensively remove legacy broad grants on re-run (a NOPASSWD: ALL line
appended to the main /etc/sudoers + sudo-group membership).
Validated live end-to-end as the manager: app lifecycle, webui generate,
ownership reconcile, ssh/dns/socket/svc helpers, task service, data-plane
drop (incl. -E for backups) all denial-free; sudo bash / sudo cat shadow /
arbitrary sudo chown all denied.
Residual (still raw runSystem file-primitives, denied under the scoped
grant until they get helpers / docker-exec rework): owncloud/adguard/
crowdsec app-config edits, wireguard-standalone, restic/kopia binary
self-install. These are opt-in/deferred features.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
start.sh sources init.sh for its function defs at runtime (Model A). The
top-level install-mode auto-detect + initUpdateConfigOption write ran on
every source, rewriting CFG_INSTALL_MODE via 'sudo sed' on the
manager-owned config — denied under the scoped sudoers (the last
per-command 'a password is required'), and spurious '"Auto-detected ..."'
noise. Gate both on BASH_SOURCE==$0 (executed directly only); also drop
the needless sudo from initUpdateConfigOption (config is manager-owned).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The scoped sudoers grants the manager (root) and (dockerinstall) but NOT
(itself), so the many 'sudo -u $sudo_user_name <cmd>' calls (crontab,
git/update, reinstall, swapfile, …) failed with 'a password is required'
once per CLI command. runAsManager runs the command plainly when already
the manager (the runtime case) and only sudo -u's when root (install
time), so it's correct in both contexts and needs no sudoers self-grant.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
scan_files used 'sudo find' to enumerate config files to source. Under the
scoped sudoers that's denied, so NO configs got sourced -> CFG_DOCKER_INSTALL_TYPE
ended up empty -> runFileOp/runFileWrite fell back to the manager branch and
every container-path write failed. Root cause of the 'sudo: a password is
required' + 'tee: Permission denied' storm when running under the scoped grant.
- configs/ scan (manager-owned): plain find
- app_configs scan (/docker/containers, docker-install-owned, not list-readable
by the manager): runFileOp find (enumerate as that user; manager still sources
each .config, which is o+r). 'containers' install templates stay plain find.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move the last runtime-critical root file-primitive subsystems behind
root-owned helpers so the type switcher + task service work under a scoped
sudoers:
- scripts/system/libreportal-socket: {rootless|rooted} {on|off} chmod of
the docker sockets (paths computed from config, not caller-supplied;
exit 3 = absent so the *_found flags come from its exit code)
- scripts/system/libreportal-svc: GENERATES + installs the systemd unit
from config (mode/uid/baked manager) — never accepts unit content from
the caller (arbitrary unit = root). Idempotent install/enable/restart.
- ownership helper: add db-own + app-file <app> <relpath> actions
- run_privileged: runSocket / runSvc
- set_socket_permissions -> runSocket; webui_install_systemd -> runSvc
(+ crontab cleanup runs as the manager directly, no sudo -u self)
- before_start: db chown -> runOwnership db-own; traefik cert/yml ->
runOwnership app-file (retires updateFileOwnership/changeRootOwnedFile)
- init.sh installs all five helpers
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Two more runtime root file-primitive subsystems moved behind self-
validating root-owned helpers so the scoped sudoers needn't grant blanket
sudo sed/tee/cp on /etc (which is root-equivalent — sudo arg wildcards
match across '/', so even path-scoped entries are bypassable):
- scripts/system/libreportal-dns: {clear|add <ip>} — edits /etc/resolv.conf
only, validates the IP argument
- scripts/system/libreportal-ssh-access: authorized_keys + sshd
PasswordAuthentication management, with the lockout guards moved INTO the
helper (the trust boundary) so a compromised manager can't bypass them
- run_privileged: _runRootHelper dispatcher + runResolv / runSshAccess
(runOwnership now uses it too)
- init.sh: initRootHelpers installs all three helpers root:root 0755 with
the manager name baked in
- setup_dns -> runResolv (+ ping de-sudo'd, works unprivileged); host_access
+ webui_ssh_access -> runSshAccess
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Under Model A the runtime runs as the manager, so establishing the
/docker ownership model needs root. Granting the manager a blanket
'sudo chown'/'sudo chmod' in the scoped sudoers would be root-equivalent
(chown /etc/sudoers, ...). Introduce a self-contained, root-owned helper
that performs only a FIXED set of reconciles on FIXED LibrePortal paths,
with owners derived from config + a baked manager name (never the caller)
and a strictly-validated app-name argument.
- scripts/system/libreportal-ownership: the helper (actions: reconcile,
traversal, containers-top, app-perms, webui, taskdir, app-data-nobody)
- run_privileged: runOwnership wrapper (sudo the installed helper; run the
bundled copy directly when already root mid-install)
- init.sh: installOwnershipHelper bakes the manager name and installs it
root:root 0755 to /usr/local/sbin (manager can't modify it)
- libreportal_folders/app_folder/app_update_specifics/task processor:
delegate the ownership chowns to runOwnership instead of runSystem chown
This removes chown/chmod-on-/docker from the runtime sudo surface, a
prerequisite for a non-root-equivalent scoped sudoers.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The CLI wrapper already runs as the manager (libreportal) but then did
'sudo ./start.sh', so the whole runtime executed as root — the reason
NOPASSWD:ALL was load-bearing. Drop that sudo so start.sh runs as the
manager; also drop the now-redundant sudo from the wrapper's own
manager-owned ops (config sed, /docker/configs + /docker/install
mkdir/cp/chown/rm, 'sudo -u libreportal' git clone, chmod). Only the
'cp -f init.sh /root/' copies stay root.
Running as the manager surfaced data-plane writes that only worked under
root; fixed to be owner-correct:
- webui_system_metrics: .metrics_{cpu,net}_prev state via runFileWrite
- atomicWriteWebUI: path-aware temp+chmod+mv (atomic same-dir rename as
the path owner) instead of bare >/mv
- webui_app_config last_update trigger via runFileWrite
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- docker_run: in rooted mode run docker AS the manager via the docker
group (no sudo); the type=='sudo' branch was unreachable dead code
- 8 db helpers: fix 'command -v sudo sqlite3' guard to 'command -v
sqlite3' (bodies already query via runInstallOp)
- restic/kopia single-file dump: write target_file via runBackupOp tee
(as the backup user, matching the snapshot-restore path) instead of
root tee
- adguard auth: root-owned scratch via runSystem mktemp
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The borg/restic/kopia engines all dropped to the dedicated backup user
via scattered 'sudo -E -u $docker_install_user'. Centralize that into a
single runBackupOp helper so the backup subsystem has one audit point and
the scoped sudoers needs only the (dockerinstall) drop rule.
Also:
- owncloud config heredoc tees -> runSystem (container-UID file)
- webui_display_logins: fix the broken 'command -v sudo sqlite3' guard
to 'command -v sqlite3' (body already runs sqlite3 via runInstallOp)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
tagsManagerUpdateUniversalTag did a bare 'sed -i "$file_path"' — works only
because start.sh runs as root today; under Model-A-as-manager the manager
can't create sed's temp file in the dockerinstall-owned containers dir
(permission denied). Make the in-place edit run AS the file's owner: classify
by path (containers/<app> -> runFileOp, manager configs/templates ->
runInstallOp), like createTouch. The awk read stays unescalated (config/compose
are world-readable). Unblocks running the whole app as the manager for tag ops.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Wireguard standalone touches /etc/wireguard + sysctl exclusively (genuine
root) -> runSystem for all its mkdir/chmod/sed/rm/grep/tee/qrencode. Traefik
dynamic configs live under containers/traefik (docker-install-owned) ->
runFileOp/runFileWrite (whitelist.yml, protectionauth.yml, the router-rewrite
awk|tee|mv in port_subdomains). sudo -u drops left.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- copy_build_context: rsync/cp/rm -> runFileOp (writes the deployed tree AS the
container owner with --no-owner); drop the now-redundant runSystem chown.
- setup_lock: .setup_complete is in the docker-install-owned frontend/data ->
runFileOp touch/chmod/rm (drop the chown).
- tags_processor_docker_installation 'user:' enable + update_compose_yml
jail.local -> runFileOp (deployed compose/config under containers).
- crontab_clear: clear the manager's own crontab via runInstallOp.
- reinstall: cp init.sh to /root -> runSystem (genuine root path).
- create_successful_run_file: drop the pointless sudo echo -> runInstallWrite to
/docker/run.txt.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
scanConfigsForRandomPassword iterates $configs_dir (manager-owned), so the
placeholder grep/sed/awk on the config file -> runInstallOp. The bcrypt export
log ($containers_dir/bcrypt.txt) is docker-install-owned, so its touch/chmod/
sed/grep/append -> runFileOp/runFileWrite (NOT runInstallOp). Covers all
password_replace*/password_user_replace/password_update_all and bcrypt/*.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The old copy/move helpers ran 'sudo cp/mv X Y; sudo chown $user_name Y' (root +
arbitrary chown). Rework them to write AS the destination's owner — no root, no
chown — classifying by dest path like createTouch: /docker/containers/<app> ->
runFileOp (docker install user), manager-owned control plane -> runInstallOp.
The $user_name arg is now advisory (the path decides). Covers copyFile/copyFiles/
copyFolder/copyFolders/moveFile; copyResource is always containers -> runFileOp;
createFolders' non-container branch -> runInstallOp; updateFileOwnership (an
arbitrary user1:user2 chown) -> runSystem. Confirmed by callers (containers vs
$docker_dir/backup_install_dir/configs dests). Removes a class of root data ops
+ arbitrary-chown from the runtime.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
app_generate operates on the manager-owned install template -> runInstallOp
(cp/mv/sed); drop sudo on the interactive editor. localDnsApplyPihole edits
containers/pihole/.../custom.list (docker-install-owned) -> read via runFileOp,
build in a manager /tmp scratch, write back via runFileWrite.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
htpasswd -bnBC just computes a bcrypt hash to stdout (no file/root access), so
the sudo was unnecessary — drop it in the adguard/focalboard/invidious auth
helpers and password_hash. (App-config file edits owned by container UIDs —
owncloud config.php/adguard yaml — are deferred as category-3 cross-owner work
for the root-owned ownership helper.)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The git-update backup helpers operate on the manager-owned $backup_install_dir:
use_git_backup unzip + config_git_check find -> runInstallOp; install_git_backup
standalone find -> runInstallOp (drop the nested -exec sudo rm), and its
cd && find | xargs rm pipeline drops its sudos (manager owns the dir). The
many 'sudo -u $sudo_user_name git/rm/zip' calls stay (already least-privilege).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The backup engine already drops to the backup user (sudo -E -u
$docker_install_user) and backupLocationOwner == $docker_install_user, which is
exactly what runFileOp/runFileWrite resolve to in both modes. So convert the
raw-sudo data ops (mkdir/chmod/rm/find/cat/grep/mv/chown/tee on backup repos,
location configs, keys, manifests) to runFileOp/runFileWrite — creating files
as the owner directly, no root chown. backup_verify creates its scratch as the
backup user (runFileOp mktemp) instead of chown-after. Binary installs
(kopia tar/install, borg dnf) -> runSystem. The 44 sudo -u engine drops stay
(already least-privilege; the scoped sudoers will grant them).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Model A prototype (run start.sh AS the manager, escalate only via helpers):
- check_root.sh: accept the manager user, not root-only (init.sh keeps its own
install-time root check).
- init.sh: guard the top-level root-check + installer entrypoint with
BASH_SOURCE!=$0 so it runs ONLY when init.sh is executed directly; when
start.sh sources it as the manager the entrypoint (and its root check) no
longer fires.
Also: convert bare daemon-touching 'docker' calls (no helper -> hit the
nonexistent /var/run socket in rootless) to runFileOp docker across
app_status, app_health_*, network_prune, ip_is_available, check_docker_network,
backup_db (db dumps) and crontab_check_processor. cd&&compose rooted-branches
and 'docker compose --version' checks left as-is (rooted-only / no daemon).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
New 'System' admin page (sidebar Tools group) rendering the metrics the
collector now produces:
- live ring gauges for CPU, memory, disk and load
- SVG trend charts (CPU/mem/disk/network) with 1h/6h/24h range toggle
- host info + swap + docker summary strips
- per-app table: CPU/mem bars, network, status, CPU sparkline
Charts are hand-rolled SVG in charts.js (LPCharts) — no third-party libs or
CDN calls — themed entirely from the active theme's CSS variables. The
Overview System card now links here.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add webui_system_metrics.sh, run each minute from webuiSystemUpdate:
- whole-server snapshot (metrics.json): CPU% + load, memory + swap,
per-mount disk + inodes, network rx/tx rate, docker summary
- capped ring buffer (metrics_history.json, 24h default) for trend charts
- per-app docker stats grouped by compose project (metrics_apps.json)
plus a short per-app history (metrics_apps_history.json) for sparklines
CPU% and network rate use stateful deltas stashed beside the JSON; all
host metrics read from /proc and docker via runFileOp, so it works rootless.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Container-plane docker now routes through the mode-aware helpers instead of
sudo: simple calls (exec/ps/run/build/images/inspect/port/logs across ~15
app/check scripts) -> runFileOp docker (rootless socket as the install user;
rooted via the docker group). The cd && docker compose paths drop the sudo on
the rooted branch (the rootless branch already used dockerCommandRunInstallUser
-- byte-identical now, manager-ready later); gluetun, which had no rootless
branch, now uses dockerCommandRun so force-recreate works in both modes.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The apps SQLite DB ($docker_dir/$db_file) is owned by the manager user, so
read/write it AS the manager via runInstallOp instead of sudo (root). 48 call
sites across 28 scripts. In rooted this drops root->manager (correct owner);
in rootless it's the manager too (using runFileOp/dockerinstall here was the
'unable to open database' bug). The broken 'command -v sudo sqlite3' check
lines are left untouched (separate pre-existing issue).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Foundation for a scoped sudoers: route every genuine system-admin command
(systemctl/ufw/ufw-docker/nft/apt/apt-get/pacman/sysctl/useradd/usermod/
service/wg/wg-quick/cscli/loginctl) through runSystem instead of raw sudo
across 28 active scripts. runSystem is 'sudo "$@"' so this is byte-identical
in every mode (safe on live installs) — it just collects all real-root use at
one chokepoint that will define the eventual /etc/sudoers.d allowlist.
Also: revert a crowdsec advice message the sweep wrongly rewrote (the admin
types sudo, not runSystem), and give crontab_check_processor.sh the same
startup bootstrap as the task processor — it runs standalone via cron and
already used runFileOp/runFileWrite (undefined there), so it was silently
broken; now it sources the helpers + docker-type config.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
install_user_manager appended a 'Match User' block to sshd_config with no
marker guard, unlike the rootless .bashrc block beside it. The enclosing
'if ! userExists' gate hides it today, but a user delete+recreate would append
a second block. Guard on the '### LibrePortal Manager User Start' marker so the
append is idempotent.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
installLibrePortalWebUITaskService only wrote the unit if it didn't already
exist, so env/User/mode changes never reached an existing install and a
docker-type switch couldn't update the service. Make it converge: compute the
desired unit for the current mode and only rewrite + daemon-reload + restart
when it actually differs (otherwise just ensure enabled+running, no restart, so
routine re-runs don't bounce the processor and kill in-flight tasks). The
docker-type switcher now calls this idempotent setup (replacing the one-shot
restart helper), so a swap updates the env AND restarts in one step.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The rootless task-processor service env used id -u $sudo_user_name (the
manager, e.g. 1001) for DOCKER_HOST/XDG_RUNTIME_DIR, but the rootless daemon
runs as the docker install user, so its socket lives at
/run/user/<install-user-uid>/docker.sock (e.g. 1002). The manager-uid path
doesn't exist. Use id -u $CFG_DOCKER_INSTALL_USER so the env matches the
actual rootless socket (same values dockerCommandRunInstallUser uses).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The task processor reads CFG_DOCKER_INSTALL_TYPE once at startup to decide how
runFileOp writes into the task dir (rootless -> as the docker install user,
rooted -> as the manager). After a rooted<->rootless swap a running instance
keeps the old mode and writes task files wrong. Add
restartLibrePortalWebUITaskService and call it at the end of both switch
branches so the processor re-sources the new mode. The switch is a CLI
one-shot, not a processor task, so the restart won't interrupt it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The dir-ownership chown used runFileOp (the unprivileged dir owner), which
can't reclaim files a prior run left root/manager-owned — leaving a root-owned
task_processor.log the daemon then couldn't append to. Use runSystem (root) so
ownership is actually established.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
systemd launches the processor standalone, so it never sourced the LibrePortal
function library — runFileOp/runFileWrite were 'command not found' at runtime,
so it couldn't write its log, create its lock (flock died on a bad fd), or
update task status. Every task stayed queued and looped forever, and the setup
'finalize' never ran.
Source the privilege helpers (run_privileged.sh, docker_run_install.sh,
check_install_type.sh) + read the docker-type config at startup so runFileOp
knows rooted vs rootless. Also create the lock and per-task log via runFileOp
(world-writable) so the manager-user processor can open/append them in the
docker-install-owned task dir.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
updateTaskFields wrote its temp with a plain 'jq … > "$tmp"' shell redirect,
which runs as the processor's own user (the manager). But TASK_DIR is owned by
the docker install user and the manager can't create files in it, so the
redirect failed and the status write silently no-op'd — every task stayed
'queued', got reprocessed in an endless loop, and follow-on tasks (e.g. the
setup 'finalize' after 'config') never ran. The fix mirrors writeAtomic:
capture jq's output, write the temp through runFileWrite (the privileged
helper), then chmod + atomic mv.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerCopyBuildContext rsync'd the install template into the container dir
with -a, which preserves owner/group — so the deployed WebUI tree (frontend/
included) inherited the repo clone's owner (the human user, uid ~1000) on
every install. The trailing chown used the $docker_install_user global, which
is stale/empty in this context, so it silently no-op'd and uid 1000 survived
(visible as frontend/ owned by 1000 with the template's mtime).
Add --no-owner --no-group so the copy doesn't carry source ownership, and
chown via the config-authoritative dockerContainerOwner (rooted -> manager,
rootless -> docker install user) through runSystem. The deployed tree now
lands owned by the mode's container owner.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The remaining WebUI generators built JSON into a temp file inside the
output dir then placed it with mv/sudo mv + a createTouch that can't re-own,
so in rootless they produced root/libreportal-owned data and 'touch:
Permission denied' spam. Two problems: the temp lived in the (now
dockerinstall-owned) output dir, which the cron updater — running as
libreportal — can't write; and the final file landed wrong-owned.
Move each temp to mktemp (/tmp, writable by whoever runs the updater) and
place the result via runFileWrite (writes as the container owner:
dockerinstall in rootless, manager in rooted), dropping the redundant
createTouch; convert the dir mkdirs to runFileOp. Covers apps
(services/config/tools/app_status/gluetun/config_patch), categories
(app/config-categories/field-mappings), config (configs.json) and system
(info/memory/disk/update). The logs file is handled by the now mode-aware
createFolders + createTouch.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The backup + ssh generators created their frontend/data dirs via plain/sudo
mkdir and wrote files via sudo tee/mv (root-owned), then called createTouch
(dockerinstall) which can't re-own a root file — so every write hit
'touch: Permission denied' in rootless and left root-owned data the
dockerinstall container/generators can't rewrite. Convert dir creation to
runFileOp mkdir and file writes to runFileWrite (both run as the container
owner: dockerinstall in rootless, manager in rooted), dropping the
temp/mv/createTouch dance. Also make the createFolders chokepoint mode-aware
(containers/ paths created via runFileOp) so it mirrors createTouch.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerCommandRunInstallUser sudo's to the unprivileged docker install user but
inherited the caller's cwd. At install time the caller is root in /root, which
that user can't enter, so cwd-sensitive tools failed — e.g. 'find: Failed to
change directory: /root' / 'Failed to restore initial working directory'
during the app scan (the scan still worked via the absolute start path, but
the errors are noise and could bite other commands). Add env --chdir to the
install user's HOME for both the argv and shell exec paths so every runFileOp
runs from a directory the user can access.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The firewall rebuild chose ufw-docker vs ufw from $EUID -eq 0 (am I root?)
rather than the docker mode. During a rootless install everything runs as
root, so it wrongly picked ufw-docker — which manages the rooted daemon's
DOCKER-USER chain that rootless never creates — and failed with 'Docker
instance libreportal doesn't exist'. (It was also inconsistent at runtime: the
non-root cron refresh always fell through to plain ufw.) Select by
CFG_DOCKER_INSTALL_TYPE so rootless always uses plain ufw (ports are published
on the host) and rooted always uses ufw-docker.
Also: ufw-docker needs the container name, not the app name — pass
service_name (e.g. libreportal-service) with an app_name fallback; route the
traefik-detect docker ps through runFileOp (was raw docker -> /var/run in
rootless); and move the ufw/ufw-docker sudo calls to runSystem.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The install/start paths and the switch reconcile managed /docker ownership
separately, so a fresh install produced different ownership than a post-switch
state — the root cause of the rootless 'touch: Permission denied' storm.
Consolidate onto the reconcile model:
- dockerContainerOwner(): single definition of the mode's container owner
(rooted -> manager, rootless -> config-authoritative docker install user).
- reconcileContainersTopOwnership(): owns + makes traversable the structural
containers/ top dir; now also run by the switch reconcile (previously only
the install pass set it, so a rootless->rooted switch left it stale).
- reconcileWebuiDirOwnership(): now uses dockerContainerOwner.
- reconcileDockerOwnership(): calls both helpers.
- fixFolderPermissions(): slimmed to the +x traversal bits; its ad-hoc
containers/ chown is now the shared helper.
- fixPermissionsBeforeStart(): drop changeRootOwnedFilesAndFolders (a
pre-de-sudo band-aid that only fixed root-owned files and ran contrary to
the don't-touch-third-party-data rule); reconcile the WebUI dir via the
shared helper instead. Delete the now-unused root_files_folders.sh and
regenerate the source arrays.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Three docker calls ran the binary directly (two plain, one sudo), so in
rootless they hit /var/run/docker.sock (the rooted socket, absent) and
printed 'Cannot connect to the Docker daemon' — the WebUI-image requirement
check, the system-disk WebUI generator (docker system df), and the
app-install fallback (docker ps). Route all three through runFileOp, which in
rootless runs as the docker install user with DOCKER_HOST set and is
argv-safe for --format, and in rooted runs as the manager via the docker
group.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
A fresh rootless install left /docker/containers/libreportal/frontend owned
by the manager (webui_install_image chowned -R to $sudo_user_name) while the
WebUI container and the host-side runFileOp generators run as dockerinstall.
So every generator touch under frontend/data and frontend/logs failed with
'Permission denied' (~27 in the install log). reconcileDockerOwnership chowns
the WebUI dir to the mode's container owner, but only runs on a mode switch,
not on a fresh install.
Extract that WebUI-dir chown into reconcileWebuiDirOwnership (rooted ->
manager, rootless -> the config-authoritative docker install user; runs as
root so it can chown either way) and call it from both reconcileDockerOwnership
and the fresh-install WebUI setup. A fresh install now lands the same
ownership a switch does, so the dockerinstall generators can write.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Switching rooted<->rootless re-maps every container's on-disk UIDs (rootless
offsets them by the subuid base), so a stateful app's data no longer lines up
in the new mode and a chown can't carry it. The portable carry is backup
(old mode) -> switch -> restore (new mode): restoreAppStart wipes and re-lays
each tree and re-owns it to the new mode's install user, which is exactly the
remap needed.
Wire that into dockerSwitcherSwap:
- switchMigrateBackupApps <old_mode>: before the switch, back up every
installed app except libreportal (reconcile already carries the control
plane). CFG_DOCKER_INSTALL_TYPE is already the target mode by the time the
switcher runs, so force it (and the resolved install user) back to the old
mode for the backups, else backupAppStart would talk to the not-yet-running
new daemon. Any backup failure aborts the switch before the daemon is
touched (nothing changed). No backup location enabled -> skip and keep the
manual warning.
- switchMigrateRestoreApps: after the new daemon is up, restore each captured
app best-effort, re-resolving the install user first so data is owned
correctly; failures are reported per app rather than blocking.
Subject to the existing backup-completeness limitation (restic-as-libreportal
can't read files owned by other UIDs unless the app declares container-side
file capture) — same caveat as the manual procedure this automates.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The switch-to-rootless branch passed the literal 'root' to mode-aware
helpers whose vocabulary is 'rooted'/'rootless' (matching
CFG_DOCKER_INSTALL_TYPE). Two were real no-ops: dockerComposeDownAllApps
root never matched dockerComposeDown's 'rooted' check (old rooted apps were
never composed-down before the switch), and dockerServiceStop root never
matched dockerServiceStop's 'rooted' check (the old rooted docker service
was never stopped/disabled). dockerServiceStart root was harmless only
because that function ignores its arg and reads CFG. Both no-ops were
silent until dockerComposeDown started reporting unknown modes. Align all
three to 'rooted'.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The single long isNotice was hard to read in both source and terminal
output. Break it into three lines (re-map warning / backup-restore guidance
/ app-data note).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The control-plane/app-install-user detail was noise on the success line;
keep it concise as 'Reconciled ownership for <mode>'.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerComposeDown printed the 'Docker Compose down <app>' header then could
fall through silently: when the effective install type (passed type arg or
CFG_DOCKER_INSTALL_TYPE) was empty/unrecognised no branch ran, and on a
non-Ubuntu/Debian OS the whole block was skipped. Collapse the duplicated
type=='' vs type!='' branches into one mode fallback and add notices for the
unknown-mode and unsupported-OS cases so the header always has a result line.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The rooted branch executed the command string as bare $command, which
word-splits without shell interpretation: pipes, redirects, && and quoted
Go templates were passed as literal argv to a single process. Nearly every
caller relies on shell syntax (docker ps | xargs -r ..., cd && docker
compose, --format='{{...}}', > /dev/null), so rooted mode silently
mishandled them — most visibly dockerStartAllApps after its pipe rewrite,
which failed with 'unknown shorthand flag: r'. Run via bash -c like the
rootless path so both modes share identical shell semantics. No caller uses
the sudo type arg.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerStartAllApps expanded $(docker ps -a -q) in the outer control-plane
shell, which has no DOCKER_HOST and so hit the nonexistent rooted socket at
/var/run/docker.sock. In rootless mode that connection fails, the
substitution returns empty, and 'docker restart' is then called with no
arguments. Push the whole pipeline into dockerCommandRun (matching
dockerRestartApp) and guard with xargs -r so it runs against the rootless
socket and no-ops cleanly when there are no containers.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The 'Update the .bashrc file' step printed its header but, when the rootless
block was already present, the if-guard skipped the whole body with no output
— looked like nothing happened. Add an else that notes it's already configured.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
ROOT CAUSE of the WebUI-dir misownership on rooted->rootless:
check_install_type.sh sets the lowercase $docker_install_user to the MANAGER
user in rooted mode (it's a mode-dependent 'container owner' var). reconcile
trusted it, so mid-switch it held the stale rooted value (=manager) and chowned
the rootless WebUI dir to libreportal -> WebUI Exited(137) -> dockerStartAllApps
retried forever (the 'switch hangs' symptom). Now read CFG_DOCKER_INSTALL_USER
straight from the live config file (authoritative, never polluted), falling back
to the CFG var then a hard default.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Previous fix still no-op'd the WebUI-dir chown: in the CLI/switch context the
path globals (containers_dir etc.) and the install-user vars can be unset,
making webui_dir a relative path the [[ -d ]] check skips, and the chown user
empty. Resolve everything with absolute-path fallbacks and read the install
user from the live config file when the vars are empty (never empty now), and
log what was reconciled (incl. a 'WebUI dir not found' notice) so a switch is
diagnosable.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Bug found via round-trip: after rooted->rootless the WebUI dir stayed
libreportal instead of dockerinstall, so the rootless WebUI Exited(137).
Cause: reconcile referenced $docker_install_user, which is unset in the
CLI/switch context (only $CFG_DOCKER_INSTALL_USER is, like the rootless
helper uses) -> chown to an empty user no-op'd. Use
${docker_install_user:-$CFG_DOCKER_INSTALL_USER} (and ${sudo_user_name:-libreportal})
so reconcile resolves the users reliably in any context.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Maintainer confirmed the intended model: the manager user (libreportal, in the
docker group) owns /docker in BOTH modes and runs things directly; root:root was
always an accident of un-de-sudo'd sudo. Rework the helpers accordingly:
- add runAsManager (run as the manager: plain when already it at runtime, else
sudo -u at install time) so files end up manager-owned, never root-owned.
- runFileOp/runFileWrite: rooted -> runAsManager (was sudo->root); rootless
unchanged (docker install user owns containers/).
- runInstallOp/runInstallWrite: always runAsManager (control plane is manager-
owned in both modes).
- runSystem unchanged (genuine root: apt/systemctl/ufw/sysctl).
All ~40 converted call sites inherit this via the helpers. reconcile's WebUI dir
now -> manager in rooted / docker install user in rootless.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Correction from the maintainer: /docker was always libreportal:libreportal;
root:root only ever appeared as an artifact of un-de-sudo'd sudo commands, not
by design. reconcileDockerOwnership now always assigns the control plane to the
manager user regardless of mode (was wrongly root:root for rooted). The deeper
implication — that the de-sudo helpers' rooted=sudo path also re-creates
root-owned files — is being confirmed before realigning.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Round-trip test exposed it: during a rooted stint the WebUI (root-in-
container) writes root-owned files into its data dir; back in rootless the
WebUI user (dockerinstall) can't manage them -> container Exited(137).
Since the WebUI is LibrePortal's OWN regenerable 0:0 component, reconcile now
also chowns containers/libreportal to the mode's container owner (root rooted
/ install user rootless). Validated: after this the WebUI returns to HTTP 200.
Third-party app data under containers/ is still untouched (backup/restore).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Mode switches change /docker ownership expectations, but the switcher only
ever fixed the socket — never file ownership — so a rooted<->rootless swap
left the control plane owned for the wrong mode (CLI + de-sudo helpers then
can't access it).
Add reconcileDockerOwnership (single source of truth): swaps ONLY the owner
of LibrePortal's control plane (configs/logs/scripts/DB + /docker top) to the
mode owner (root rooted / manager rootless). It never resets mode bits (only
adds o+x on /docker for traversal and o+r on the DB for the WebUI), and never
touches /docker/containers/** app data, backups/, or ssl/ssh keys. Wired into
both switch branches between container-retag and app-start.
App data is deliberately NOT chowned: container UIDs re-map across modes
(rootless subuid offset), so a chown can't carry e.g. Postgres data across —
that's a backup->switch->restore operation. Switcher now warns to back up
stateful apps before switching and restore after.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Per the confirmed ownership model: files under /docker/containers/<app>/ are
app data owned by the docker install user; everything else is the manager-
owned control plane. createTouch now picks runFileOp vs runInstallOp by the
file's location and creates it directly as the right owner — no more
chown-to-another-user (which needs root the unprivileged runtime lacks).
The $2 user hint is now advisory. (Generator content-writes into
frontend/data still need converting to runFileWrite — next.)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Disabling userland-proxy makes rootless dockerd require br_netfilter
(/proc/sys/net/bridge/bridge-nf-call-iptables), absent in the rootless
netns on Debian -> default bridge creation fails -> daemon won't start.
Drop the daemon.json userland-proxy=false write. Source-IP is preserved
at L7 by Traefik (X-Forwarded-For), so no real loss.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Reinstall test on Debian 12 surfaced three rootless-only breakages (rooted
was byte-identical/fine):
1. pasta blocked by Debian's passt AppArmor profile (DENIED ptrace read ->
can't open container netns -> rootless dockerd never starts). Default
CFG_ROOTLESS_NET back to slirp4netns (reliable); pasta stays selectable
for hosts that relax the profile.
2. de-sudo mis-assigned helpers by owner. /docker management layer (apps DB
chowned to libreportal by install_sqlite, /docker/logs) is MANAGER-owned,
not dockerinstall. Add runInstallWrite; move apps-DB sqlite3 -> runInstallOp
and /docker/logs appends -> runInstallWrite. Revert ownership-SETUP scripts
(libreportal_folders, app_folder) to runSystem — they must run as root to
establish ownership during install. Container files (/docker/containers/<app>)
stay runFileOp.
3. kernel hardening sysctls written to /etc/sysctl/99-custom.conf, which
'sysctl --system' does not read -> never applied. Write them to
/etc/sysctl.d/99-libreportal-hardening.conf instead.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
FIX: dockerCommandRun rooted path is 'sudo $command' (unquoted word-split),
so 'docker ps --format "{{.Names}}"' was passing the format with LITERAL
quotes -> docker emitted '<name>' and the downstream grep never matched
(broken in rooted too). Switch all docker invocations to runFileOp, which
preserves args via "$@" in both modes (and runs as dockerinstall against
the rootless socket). Fixed monitoring.sh, dashy, tags_processor_network_mode.
Convert: jitsimeet (rm/wget/unzip/mv/sed/tee/gen-passwords on /docker ->
runFileOp/runFileWrite), authelia (config sed/mkdir/chmod/chown/secrets tee
-> runFileOp/runFileWrite; docker exec -> runFileOp docker, preserving
--password), reset_git (cp->/root runSystem, install-dir chown runInstallOp;
kept sudo -u manager). check_update/update_git_check need no change (all
sudo -u manager git, already least-privilege).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- restic_install, crowdsec_update/verify_firewall/fix_priority: pure host
ops (apt/cscli/nft/systemctl, /etc/crowdsec) -> runSystem.
- kopia_backup/borg_restore: ignore-file/target tee+chown+mkdir -> runFileOp/
runFileWrite; kept the 'sudo -E -u dockerinstall' engine calls as-is —
those already run as the unprivileged backup user (least-privilege; the
scoped sudoers will permit (dockerinstall)).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
firewall_initial_setup + firewall_clear_rules (ufw/ufw-docker),
host_access.sh (sshd/-T/-t, /etc/ssh, authorized_keys, systemctl reload),
set_socket_permissions (docker socket test/chmod), and webui_install_systemd
(systemd unit tee + systemctl) -> runSystem. These stay real-root in both
modes and define part of the eventual scoped allowlist. Left the
'sudo -u <manager> crontab' run-as-manager lines for a dedicated pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- New runInstallOp helper for manager install-dir/template ops (rooted:
sudo; rootless: run as the current manager user, which owns the tree).
- adguard.sh, traefik.sh: container-config sed -> runFileOp.
- crowdsec.sh: host crowdsec systemctl/apt-get -> runSystem.
- dashy_update_conf.sh: conf-file mkdir/chown/md5sum/tee -> runFileOp/
runFileWrite; docker ps/restart -> dockerCommandRun.
Deferred (cross-owner copy / temp-file across /tmp<->/docker, need rootless
env to bridge correctly): owncloud_setup_config.sh, adguard_auth.sh.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
All operate on /docker data-plane (DB at $docker_dir, compose files,
task dir /docker/.../frontend/data/tasks): sqlite3/find/sed/mkdir/chmod/
chown/mv/rm/mkfifo/truncate/install/tee -> runFileOp/runFileWrite. The
two systemctl enable/start calls in the check processor -> runSystem.
Dropped spurious sudo on text-only echo/grep/date in db_app_scan.
Byte-identical in rooted.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add -a/--append to runFileWrite so the pervasive /docker/logs log-append
idiom (`… | sudo tee -a $logs_dir/$docker_log_file`) routes through the
mode-aware helper instead of raw sudo.
Convert scripts/config/docker/docker_config_to_container.sh fully: all
ops target /docker app config + logs (data-plane), so md5sum/grep/chmod/
cmp/editor -> runFileOp and the log-appends -> runFileWrite -a.
Byte-identical in rooted.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Give dockerCommandRunInstallUser an --argv mode that execs arguments
verbatim (sudo -u <user> env ... "$@") instead of bash -c "$*", and
point runFileOp at it. The old $*+bash -c re-parse silently mangled
backslashes/quotes in args — e.g. sed scripts (\1, \( become 1, ( ) and
the sqlite3 .backup arg — so rootless data-plane ops with regex were
broken. Verified: the WG_DEFAULT_DNS sed now applies correctly as the
install user. All existing runFileOp callers pass plain commands, so the
switch is safe (and fixes the latent sqlite3 case).
Convert scripts/network/dns/setup_dns.sh: /etc/resolv.conf edits and
ping -> runSystem; the WG_DEFAULT_DNS compose-file sed -> runFileOp.
Byte-identical in rooted; correct in rootless.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Enabling unprivileged user namespaces for rootless widens the kernel
attack surface reachable by unprivileged users (a known source of LPE
CVEs). Pair it with three distro-portable, low-impact sysctls that close
the surfaces those exploit chains rely on: kernel.kptr_restrict=2 (hide
kernel pointers), kernel.yama.ptrace_scope=1 (block cross-process
ptrace), net.core.bpf_jit_harden=2 (harden the JIT). Added as a separate
guarded LIBREPORTAL KERNEL HARDENING block so it's clearly deliberate and
independently idempotent.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Split the rootless network-driver knob out of network_docker into a
dedicated advanced subcategory (configs/network/network_rootless), marked
**ADVANCED** at both the file-header and field level so it's hidden behind
the advanced toggle in the WebUI. Registered in the network .category
order after network_docker. Pure config relocation — the install script
reads CFG_ROOTLESS_NET as a sourced var, so no script change needed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Switch the rootless network stack from slirp4netns+builtin to pasta+
implicit (faster and propagates the real client source IP). The earlier
pasta+builtin attempt bricked the daemon because rootlesskit rejects
mismatched net/port-driver pairs; expose a single CFG_ROOTLESS_NET knob
(pasta default, slirp4netns fallback) and derive the matching port
driver in-script so an invalid combo can't be configured. Disable
userland-proxy in the rootless daemon.json (merged, not clobbered) so
containers see the real source IP. Both driver binaries are always
installed, so switching is a config flip + rootless re-setup.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
In rootless the container was set to user: <host-uid> (e.g. 1002:1002) with
group_add: <host-sub-gid> (e.g. 166528). Inside the daemon's user namespace
those are out of range — group_add made runc call setgroups() with an unmapped
GID ('setgroups: invalid argument'), so the WebUI container never started.
In rootless the container now runs as 0:0 (userns-root == the install user, which
owns the bind-mounts and the rootless socket) with socket gid 0. Rooted is
unchanged. Verified: libreportal-service comes up and talks to the rootless
socket.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The rootless dockerd override forced NET=pasta + PORT_DRIVER=builtin, which
rootlesskit rejects ('pasta requires port driver none or implicit'), so the
daemon failed to start every time (the real cause behind 'rootless socket not
found'). Use slirp4netns + builtin (valid, still skips the userspace
port-handler). Verified: daemon now comes up, docker Server 29.5.2 responds.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerCommandRunInstallUser ssh'd to <user>@localhost, but nothing set up an
SSH server/keys/authorized_keys, so every rootless setup command (daemon
install, systemctl --user) silently no-op'd. Replace with 'sudo -u <user> env
…' that sets XDG_RUNTIME_DIR / DBUS_SESSION_BUS_ADDRESS / DOCKER_HOST / PATH
explicitly; linger keeps the user systemd + /run/user/<uid> alive so
systemctl --user works. No SSH server, no keys, less attack surface, and
sudo -u to an unprivileged user is not a root escalation.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
useradd was missing its login-name argument (and -m), so it failed — silently,
because local result=$(...) swallowed the exit code and checkSuccess reported
success. The rootless install user was therefore never created, which cascaded
into 'invalid user dockerinstall' and a daemon that never came up. Pass the
username + -m (subordinate uid/gid ranges come from login.defs), unmasked.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
setupConfigsFromRepo / sync_configs_from_install used 'cp -a' of the template
over /docker/configs, so any fast/full deploy (which runs init.sh) silently
reset user config to template defaults — e.g. it flipped a live rooted box to
the new rootless template default and broke it. Use 'cp -an' (no-clobber):
fresh installs still get the full template, existing installs keep their values,
and new keys are still added by the add-only reconcile pass. This is also what
makes a rootless template default safe for existing rooted boxes.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Convert the backup/restore data-plane sudo calls (mkdir/chown/rm/sqlite3/tar/
gzip|tee) to runFileOp/runFileWrite. Rooted behaviour is identical (helper runs
sudo); rootless will run them as the unprivileged install user. Pilot subsystem
for the wider de-sudo. verify.sh's /tmp scratch ops left as-is (different
ownership domain, handled separately).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Single place that decides how a privileged op runs by Docker mode:
- runFileOp / runFileWrite: /docker data-plane ops — rooted uses sudo (identical
to today), rootless runs as the unprivileged install user (no root).
- runSystem: genuine system-admin ops, sudo in both modes, funnelled here so it
can later be confined to a scoped sudoers allowlist.
Call sites converted to these are byte-for-byte unchanged under rooted, so
existing/live boxes can't regress; rootless gets the de-privileged path.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The SSH-access feature's files_ssh.sh array was never registered in
files_source.sh, leaving it unsourced and blocking the deploy auto-merge.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
init.sh appended 'libreportal ALL=(ALL) NOPASSWD: ALL' straight to /etc/sudoers
— a malformed line there locks out sudo entirely. Move it to a validated
/etc/sudoers.d/libreportal drop-in (visudo -cf before install, 0440 root:root).
The grant is still broad; this is the single managed file we tighten to a
scoped command allowlist once the runtime no longer needs broad root. Only runs
at install, so existing boxes are untouched.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Containers now run unprivileged by default — a container breakout maps to a
sub-UID, not host root. Rooted remains available as a legacy opt-in. Existing
installs keep their current mode (config reconciliation is add-only); fresh
installs get rootless. The rootless path already handles unprivileged ports
(ip_unprivileged_port_start=0) and userns.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Convert the remaining sections off the legacy ?= query form to clean paths,
matching the Admin area:
/apps/<category> (was /apps?=<category>)
/app/<name>?tab=&task= (was /app?=<name>&tab=&task=)
/tasks/<category>?task= (was /tasks?=<category>&task=)
/backup/<tab> (was /backup?=<tab>)
Builders updated everywhere (sidebar, dashboard, notifications, tasks, apps,
app tabs, task-actions, setup watcher); parsers now read the resource from the
path with the legacy ?= kept as a fallback so old links/bookmarks still work
(server already serves index.html at any depth). Route table gains /apps* and
orders it before /app* (since '/apps' startsWith '/app'); active-nav and
config/apps data-loading recognise the new paths.
Tab/task remain ordinary query params (modifiers, not the primary resource).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the Admin area's ?= query URLs with clean, hierarchical paths that
mirror the breadcrumb:
/admin -> Overview
/admin/config/<category>-> Config / <category>
/admin/tools/ssh-access -> Tools / SSH Access
New /admin (+ /admin*) SPA route -> handleAdmin, which parses the path via the
shared window.adminPath / window.adminCategoryFromPath helpers and renders
through the existing ConfigManager. Legacy /config, /config?=<x> and /ssh now
redirect into the matching /admin path, so old links/bookmarks keep working
(server already serves index.html for any depth). Sidebar, Admin Overview,
dashboard link and top-nav now build /admin paths; active-nav + config data
loading recognise /admin across spa.js, topbar.js, router.js, data-loader.js.
Scope: Admin area only — /app, /apps, /tasks, /backup keep their existing ?=
URLs.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
With live dumps + container-side file captures the live app dir intentionally
differs from the snapshot (raw DB dirs and private trees are excluded, replaced
by dumps/captures), so the old source-vs-restored file-count check false-failed.
The scratch restore succeeding already proves restorability (restic hash-checks
every blob); keep a non-empty sanity check instead.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The container capture preserved the app's ownership (e.g. www-data 0640), so
restic still hit permission denied on the staging copy. chown the staging tree
to the backup user after capture (modes unchanged, so the owner reads fine);
real ownership is reapplied from the descriptor on restore.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- Sidebar now groups items: Overview at top, a 'Config' heading over the config
categories, and the existing 'Tools' heading over SSH Access.
- Breadcrumb reflects the group: config pages read 'Config' (was 'Admin'), SSH
reads 'Tools', Overview stays 'Admin'.
- SSH Access page restyled to the config page's section layout
(.config-category/.domains-wrapper sections) instead of backup-style cards, so
it matches the other Admin config pages.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Reads files the backup user can't see from the host (container-owned, e.g.
Nextcloud's www-data data dir) by streaming them out THROUGH the container
(docker exec tar) — no host root, no host read perms, works rooted + rootless.
Extracts to staging as plain files so restic keeps full dedup + per-file
restore (not a piped tar blob); the live path is excluded from the snapshot.
Restore streams the staging copy back through a throwaway in-namespace
container that recreates the tree with the app's uid:gid.
Declared via a libreportal.backup.files compose label; Nextcloud (html, 33:33)
is the first to use it. Live capture failure falls back to stop-snapshot-start.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add an Admin Overview as the Admin landing (default when you open Admin): an
ops/health board distinct from the user Dashboard. Four cards built from data
we already generate — Updates (update_status.json, with one-click update),
Backups (backup dashboard.json), SSH & Security (access.json), System
(disk/memory/system_info) — each with a Manage link into the right section.
Styled like the backup dashboard (tiles/status dots).
Wire-up: 'Overview' is the top sidebar item and the default category
(handleConfig + sidebar), rendered by AdminOverview into #config-section via a
renderConfig('overview') special case. Every Admin page now shows the same
'Admin' breadcrumb header (Overview, SSH Access, and the config categories) for
a consistent Admin → Section feel. User Dashboard gets an 'Admin overview →'
link.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Rename the Config top-nav to 'Admin' and move SSH Access into its sidebar
under a 'Tools' group, instead of a separate top-level nav item. SSH Access is
rendered by SshPage into the config main pane via a renderConfig('ssh-access')
special case; the sidebar item (config-sidebar.js) routes there. SshPage now
mounts into any container (defaults to #config-section). /ssh redirects to
/config?=ssh-access for old links; the standalone ssh-content.html is removed.
Declutters the top bar and gives system/admin features one home that scales
(updates, users, Connect settings can become sidebar entries later).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Focalboard runs as nobody (65534) but fixPermissionsBeforeStart hands the app
dir to the install user, so the server couldn't open its sqlite db on the newly
mounted data dir. Chown data/ to 65534 in appUpdateSpecifics and restart.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Focalboard writes its sqlite db and uploads under /opt/focalboard/data (its
working dir), but the compose mounted ./data:/data — an unused path — so the
database was never persisted and was lost on every container recreation. Mount
./data:/opt/focalboard/data so db + files survive, and declare the db for live
backup (data/focalboard.db).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Verified on a live install: gitea's app.ini sets [database] PATH=/data/gitea/
gitea.db with mount ./data/gitea:/data, so the host path is data/gitea/gitea/
gitea.db. (Created only after gitea's web setup completes; until then live
backup safely falls back to stop.)
focalboard intentionally still omitted — its db lives at /opt/focalboard/data/
focalboard.db inside the container, which the compose does not mount, so it
isn't persisted to the app dir and can't be backed up until that's fixed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Live-restore of a server DB (MariaDB/Postgres) raced the engine's first-init:
it starts a throwaway temp server, runs setup, then restarts the real one. The
old ping-based readiness passed against the temp server and the load hit the
restart, failing once.
- _backupDbWaitReady now requires a real query to succeed on two consecutive
checks, so the restart breaks the streak and we only proceed once the real
server is stably up.
- The dump load is retried (idempotent — the dump drops+recreates each object)
to ride past a final init bounce.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
New /ssh page (topbar nav + SPA route + SshPage controller + ssh-content.html
+ ssh.css). Reads data/ssh/access.json and lets the admin: paste a public key
to authorize a machine, remove keys, and toggle key-only login — all via
'libreportal ssh ...' tasks through the backend's lockout guards. Reuses the
backup key-card styles for a consistent look. This is the inbound counterpart
to the backup location key card (outbound): same paste-a-key model, opposite
direction.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Fresh, on-demand inbound SSH-access management for the host (replaces the old
maze). scripts/ssh/host_access.sh manages the install user's authorized_keys —
add a pasted public key (validated), list, remove — and toggles sshd password
login behind a lockout guard (won't disable passwords with no key; won't drop
the last key while passwords are off; sshd -t before reload, with backup).
New 'ssh' CLI category (status/key-add/key-remove/password-auth/generate) and
a webuiGenerateSshAccess snapshot (data/ssh/access.json: user, password_auth,
authorized keys as type+fingerprint+comment — public only) wired into the
regen chain. Nothing runs automatically; only explicit admin actions change
anything. WebUI page next.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Found while testing live backups end-to-end:
- Engine backup adapters logged to stdout, so the caller's $() snapshot-id
capture was polluted with log text — verify-after-backup then failed with
'no matching ID' on every run. Route their log lines to stderr so stdout is
only the id (restic/borg/kopia).
- 'libreportal app restore <app> --latest' (as the help advertises) and the
bare 'restore <app>' both failed: --latest was passed to restic verbatim and
unset args arrive as the literal 'empty'. Normalise both to 'latest'.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The old inbound-admin-SSH layer was effectively dead: gated on config flags
that don't exist (CFG_SSHKEY_*_ENABLED, CFG_REQUIREMENT_SSHREMOTE), its
authorized_keys installer was unwired, and its download path (sshdownload
container) was already retired. What remained reachable was either a no-op or
a lockout footgun (disable-passwords with no working key install).
Remove it whole: scripts/ssh/*, the four SSH requirement checks, the SSH tools
menu, the dead webui SSH populater, and the unused ssh DB inserts; drop their
calls from the start/requirements/menu flows. A fresh, WebUI-driven admin SSH
access feature replaces it next.
Also make generate_arrays.sh self-healing: prune files_*.sh whose source
folder no longer exists (cleared the now-stale files_ssh.sh + an orphan
files_api.sh) so removed areas don't linger in the sourced set.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
When a location uses SSH key auth, show a key card: paste an existing private
key, or 'Generate keypair', then the card displays the public key to copy into
the remote server's authorized_keys (with Copy/Delete). Wires to the
ssh-key-set/generate/delete CLI; key mutations refresh locations.json so the
card reflects state immediately. applySshAuthVisibility toggles the card vs the
password field by auth mode. Private key only ever flows in (base64); only the
public key is ever shown.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Expose the existing location_ssh.sh key store through the backup CLI:
'backup location ssh-key-set|ssh-key-generate|ssh-key-public|ssh-key-delete <idx>'
(the WebUI runs these as tasks). The locations generator now emits
ssh_key_exists + ssh_public_key (public key only — the private key never
leaves the per-location ssh.key file), so the editor can show the key state.
Also fix the stale SSH_AUTH label (~/.ssh/id_rsa -> managed per-location key).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Every backup-scope app now carries CFG_<APP>_BACKUP_STRATEGY=auto, so the
Backup Strategy dropdown appears in each app's Advanced tab — not just the
DB apps.
To keep it honest, the 'live' option is hidden where it isn't safe:
- apps.json generator emits backup_live_capable per app (from compose backup
labels: a dumpable DB, or a live-safe marker).
- apps-manager filters the live option out of the strategy select when the
current app isn't live-capable, so apps like gitea/focalboard (a DB we don't
yet dump) never offer it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds CFG_<APP>_BACKUP_STRATEGY (default auto) so an app's backup strategy can
be overridden from its Advanced config tab, taking precedence over the global
default. Added to the 10 live-capable apps, so the dropdown's 'live' option only
appears where it actually works.
- backupResolveStrategy now checks the per-app override before the global value.
- backupAppLiveCapable / backupAppStrategyOptions expose capability + the valid
option set; predicate helpers hardened with explicit returns so they behave
identically with or without shell errexit.
- BACKUP_STRATEGY field mapping (select, advanced) renders the dropdown.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
New script files are sourced from the committed files_*.sh arrays (built by
generate_arrays.sh), not a live tree scan — and quick deploys don't rerun
generate_arrays. So the schema generator added last commit was never loaded
live: webuiGenerateBackupSchema was undefined, breaking the webui_updater
backup chain at that step (skipping the passwords regen after it) and leaving
schema.json un-generated.
Regenerate the arrays so the file is registered; deploy now sources it and
'webui generate all' rebuilds schema.json on its own.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The per-type field map lived hardcoded in backup-page.js. Add a
webuiGenerateBackupSchema generator that emits the type -> ordered field list
to data/backup/generated/schema.json (wired into the backup regen chain and
the CLI 'webui generate backup'). The editor fetches it into this.locSchema
and reads it via locFieldsForType; BACKUP_LOC_FIELDS_BY_TYPE stays only as a
fallback if the fetch fails.
Keeps the data-in-generators pattern consistent — the schema now has one
backend source of truth. The dynamic show/hide behaviors (SSH auth, path
mode, engine filtering) remain frontend logic by nature.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add libreportal.backup.db labels for the SQLite apps with confirmed db paths:
vaultwarden, linkding, trilium, headscale, authelia. These are dumped live via
sqlite3 .backup and rehydrated before start on restore.
gitea and focalboard are intentionally left out until their sqlite paths are
confirmed on a live install — a wrong path would just fall back to stop, but
there's no point shipping a descriptor that always falls back.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Type is the choice that determines which other fields appear, so it should be
the first thing you pick; Name is just a label. Reorder the Connection-tab
fields (and the Add-location dialog) to Type → Name.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- Add libreportal.backup.db labels to the MariaDB/Postgres apps (nextcloud,
owncloud, bookstack, mastodon, invidious) so they back up live + consistent.
- If a declared dump cannot be taken (DB down, wrong path), the backup falls
back to stop-snapshot-start for that run instead of snapshotting torn data —
a misconfiguration degrades to 'safe with downtime', never to 'unsafe'.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Adds a logical-dump path so apps with a database can be backed up with zero
downtime and full consistency, instead of stopping the container.
- backup_db.sh: dump each declared DB live (mysqldump --single-transaction /
pg_dump / sqlite3 .backup), exclude the raw data dir from the snapshot, and
replay the dump on restore (pre-start rehydrate for sqlite, post-start load
for server engines).
- Databases are declared via a 'libreportal.backup.db' compose label so the
metadata travels with the app in the snapshot.
- New 'auto' strategy (now the default): live where a DB is dumpable or the app
is marked live-safe, stop-snapshot-start otherwise. Explicit stop/pause/live
remain as overrides.
- restic/borg/kopia adapters honour an exclude list on the live path.
- Manifest records the resolved per-app strategy and dumped databases.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- Remove the per-tab 'How LibrePortal connects…' description lines; the tab
labels already say what each panel is, and the paragraphs read as misplaced
titles.
- Give the tab panels even, comfortable padding (tabs-content padding zeroed so
the panel owns it) instead of the cramped 2px sides.
- Round the tab strip's top corners (.tabs-list) so the strip + content read as
one card — .tabs-content already rounds the bottom, leaving the top square.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Automatic path mode hardcoded /docker/backups/<id>, baked into the Path Mode
dropdown label. Add a CFG_BACKUP_DEFAULT_PATH option in the Backup Engine
config ("Default Backup Location", default /docker/backups) and have
backupLocationResolvedPath build the auto path from it (<base>/<id>, trailing
slash tolerated). Defaults to the old path, so existing auto locations are
unchanged.
Path Mode's option is now just "Automatic" (no inline path); its tooltip
points at the Default Backup Location config option instead.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Reuse the shared .tabs-wrapper/.tab-button/.tab-panel components (same as an
app's Config/Tasks tabs) for the location editor instead of bespoke tab CSS:
emoji + label buttons, equal-width strip, accent active state. Panels toggle
via the .active class like the rest of the UI; only the panel padding is
trimmed so it nests inside the backup row.
Also drop the now-dead 'No advanced options' empty state — every type has at
least Engine + append-only in the Advanced tab.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The backup engine is an implementation detail — LibrePortal picks a sensible
default and handles it — so it doesn't belong next to Name/Type on the
Connection tab. Add ENGINE to LOC_ADVANCED_SUFFIXES and mark it **ADVANCED**
in the location.config template + seed so it's metadata-driven.
Since the engine select now lives in the Advanced tab while SSH-auth and
path-mode stay on Connection, refreshInlineTypeFields re-applies the dynamic
behaviors (engine filtering, SSH/path visibility) against the shared
.task-details scope rather than a single panel.
Also fixed the live per-location engine label (restic -> Restic) which now
surfaces in the dropdown via the generator-emitted options.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The expanded location row was one long form. Split it into tabs so it opens
showing only the Connection fields. Retention moves from a stacked section
into its own tab, and the advanced overrides (URI/SSH port/append-only) get
their own tab instead of the inline disclosure from the previous pass.
Field grouping is metadata-driven: locFieldGroups partitions a type's fields
into Connection vs Advanced via the configs.json "advanced" flag (with
LOC_ADVANCED_SUFFIXES as the legacy fallback). Type changes rebuild both the
Connection and Advanced panels since advanced fields are type-dependent too.
Save still reads every field across all panels (hidden tabs stay in the DOM).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The Locations editor now renders field metadata from configs.json
(window.configData) instead of relying on the hardcoded BACKUP_LOC_FIELD_DEFS,
which drops to a fallback. Fields flagged advanced (URI override, SSH port,
append-only) move out of the main grid into a full-width "Advanced"
disclosure that's collapsed by default, so the common case stays simple.
Also load the unified config once on the backup page into window.configData
(metadata) + a flat window.systemConfigs (values). Previously systemConfigs
was only populated after a save — and with the full nested JSON, while the
code reads it as a flat map — so default-engine lookups and save-time change
detection silently misbehaved on first load. Both are now correct.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The config generator only scanned flat per-category files, so the dynamic
CFG_BACKUP_LOC_N_* keys carried no titles/descriptions/options — the Locations
editor had to hardcode that metadata in backup-page.js. Add a pass that
descends into configs/backup/locations/<n>/location.config and emits each key
(value/title/description/options) into the config map, plus an "advanced"
flag parsed from a **ADVANCED** token in the field comment (stripped from the
user-facing description).
These keys use subcategory "backup_locations", which isn't in any category's
subcategory_order, so the generic /config page ignores them — only the custom
Locations editor consumes them. URI, SSH port, and append-only are marked
advanced. Verified: configs.json stays valid JSON and /config subcategories
are unchanged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The config-grid toggle box used the input's 12px vertical padding, but its
24px pill made it render 48px tall vs the inputs' 44px, so it sat too tall
to read as inline. Trim vertical padding to 10px so the box is 44px.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- Display the restic engine as "Restic" to match BorgBackup/Kopia. The
lowercase name lived in scripts/backup/engines/restic.json (drives the
location-row engine pill, per-location engine select, and engine modal),
the hardcoded per-location dropdown options, the engine-list fallback, and
the config-option metadata. All set to "Restic".
- In each location's Engine dropdown, float the system-default engine
(CFG_BACKUP_ENGINE) to the top and tag it "(default)", mirroring the
retention-preset pattern.
Repo config metadata is the install template (add-only reconciliation), so
the live /docker/configs/backup/backup_engine label was updated in place too
for the global Configuration-tab dropdown on this install.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
On installs migrated from EasyDocker the spool file
/var/spool/cron/crontabs/<user> can be left owned by a defunct UID. The
sticky bit on the spool directory then blocks the current install user from
replacing it, so every `crontab -` write failed with
"rename: Operation not permitted" while the scripts still printed success.
crontabClear now removes the crontab as root (`crontab -u <user> -r`), which
bypasses the sticky bit and clears the stale file; the setup steps recreate
it owned by the install user, so the next crontab refresh self-heals.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The WebUI data snapshots (locations.json, dashboard.json, snapshots_*.json,
etc.) are regenerated on every wizard/config change. Each file emitted two
extra success lines via createTouch — "Touching <file>" and "Updating
<file> with <user> ownership" — which spammed the output around the genuinely
useful "... JSON regenerated" line.
Add an optional "silent" flag to createTouch (third arg; default keeps the
existing loud behaviour for interactive install flows) and pass it from every
WebUI data generator/task. Touch + chown still run; only the logging is
suppressed for these background regenerations.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
WebUI-driven commands (`setup finalize`, `backup`, restore) ran with an
empty $docker_install_user because cliInitialize only called
checkInstallTypeRequirement for the `app` category. The backup engine then
ran `sudo -E -u "" restic init`, which sudo rejects with a usage dump —
surfacing as "Failed to initialize Local disk" in the setup wizard.
Factor the user resolution out of checkInstallTypeRequirement into a
side-effect-free resolveDockerInstallUser (rooted -> sudo_user_name,
rootless -> CFG_DOCKER_INSTALL_USER, with fallbacks so it is never empty)
and call it at the cliInitialize chokepoint so all command categories get a
valid install user, not just app.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Divider: .domains-divider was a bold 2px accent bar under every section header,
which read as a stray line. Drop it to a subtle 1px low-opacity neutral rule so
it separates without shouting.
Toggle: the boxed config toggle (.checkbox-label) used a different radius (10px),
fill (0.04) and border (0.10) than the .form-control inputs beside it (8px /
0.05 / 0.20), so it looked off and out of line. Match it to the input field box
exactly so toggles and inputs read as the same surface. The app-config
borderless toggle override is unaffected.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Give the per-location Save changes / Delete location row some breathing room
from the bottom of the expanded card.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Buttons: the per-location Save changes / Delete location buttons had no icons,
unlike the apps-config action buttons. Add a save (floppy) icon and a trash
icon so they match the reference; colour comes from the nebula button groups
they already belong to.
Theme refactor: move the theme-specific [data-theme="nebula"] button/topbar/CTA
rules out of the shared css/themes.css and into themes/nebula/theme.css, where
the README says theme overrides belong. css/themes.css keeps only the generic,
non-theme-scoped defaults (solid status/accent buttons, danger-zone,
warning-banner) shared by dark-blue/light. No behaviour change: the nebula file
loads after css/themes.css so the moved rules still win.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Two follow-ups to the button restyle:
- On the nebula theme, primary/danger CTAs are translucent (rgba accent/danger
fill + white text + border), not the solid generic .btn-primary. The earlier
change only added the backup classes to the generic groups, so on nebula the
Add location / Save changes / Delete buttons fell back to a solid fill with
dark text. Add .backup-primary-btn and .backup-danger-btn to the
[data-theme="nebula"] groups too, so they match the config-page buttons.
- The per-location action row used justify-content: space-between, throwing the
two buttons to opposite edges. Switch to flex-start with a gap (like
.config-actions) and put Save changes (primary) before Delete location.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The backup pages' primary buttons (Add location, Save changes) and Delete
location button used a local gradient + glow in backup.css, so they didn't
match the flat solid-accent buttons on the config page. Add .backup-primary-btn
and .backup-danger-btn to the shared nebula button groups in themes.css
(.btn-primary / .btn-uninstall) so they get the same solid accent/danger fill
and hover with !important across themes, and drop the local gradient/shadow/lift
from backup.css. Top-right and bottom-of-page backup buttons now match config.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add a reusable Dismissible helper that persists 'hide this permanently' state server-side in data/ui-state.json via the existing authenticated /read-file + /write-file endpoints. It's a direct file write — no task is created (nothing in the task manager) and no system scan runs — so it sidesteps the heavyweight config_update path entirely and works across browsers/devices. The backup config-backup warning now dismisses through Dismissible instead of localStorage; any future notice can opt in with Dismissible.isDismissed(id)/dismiss(id).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Drop the Export button from the config-backup warning banner — it's now just the alert + dismiss (x). On the Configuration tab the top-right primary action becomes an 'Export' dropdown (first item: Repository Passwords, reusing the existing export-passwords action) so more export types can be added later. Other tabs keep Backup all apps / Add location. Menu opens from the trigger and closes on outside click, item click, or tab switch.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add a large amber alert-triangle icon to the 'keep your config backed up offline' banner and a close (x) button in its top-right. Dismissal is stored in localStorage (libreportal:backup-config-warning-dismissed) — a per-browser UI nudge, not server config — and hides both the banner and its divider until cleared.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Drive the section title from the filename instead of the JS display-override
added earlier: rename configs/backup/backup_advanced -> backup_engine, update
the category SUBCATEGORY_ORDER and the file's header comment, and revert the
formatSubcategoryName override. The CFG_BACKUP_* keys are unchanged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Empty tabs: generateSimpleTabsAndContent gated tabs on a hasFields heuristic
that drifted from what generateConfigFields actually emits, so a category like
Network could show a tab whose body only read "No configuration options
available". Render each category's fields first and emit the tab only when the
output is non-empty, keeping tabs and content in lockstep.
Rename: the backup_advanced subcategory now displays as "Engine" via a
display-name override in formatSubcategoryName. File and CFG_BACKUP_* keys are
unchanged, so saved values are unaffected.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The location editor's Type, Path Mode, Engine and SSH Auth selects rendered
with no options. The config generator only scans flat per-category files and
never descends into configs/backup/locations/<n>/, so configData carries no
options for CFG_BACKUP_LOC_<n>_* keys — and the hardcoded fallbacks had been
removed in favour of generator-emitted ones.
Resolve these four dropdowns by suffix in ConfigOptions.getSelectOptions with
their static option lists (labels mirror location.config), so every location
works regardless of index — including locations added after install. The
global CFG_BACKUP_ENGINE/STRATEGY selects still come from the generator.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Location config dropdowns (Type, Path, etc.) live inside .task-item cards,
whose :hover applies transform: translateY(-2px). A transformed element
becomes the containing block for position:fixed descendants, so the popup —
previously a child of the card — was positioned with viewport coords against
the card instead of the viewport (wrong placement) and perturbed layout
(content shifted left).
Portal the popup to <body> on open and detach on close, so position:fixed is
always relative to the viewport regardless of any transformed/overflow
ancestor. flip-up styling moves onto the popup element and the topbar's wider
popup is carried via a class, since the popup no longer nests in the wrapper.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The retention "Backup style" dropdown hardcoded "(default)" into two preset
labels, so both the global and per-location selectors showed two "(default)"
tags, and the global selector listed "Inherit global retention" — which has
nothing to inherit at the global level.
Apply "(default)" dynamically to the scope's actual default (self-hosting
globally, inherit-global per-location) via a shared retentionPresetOptions
helper that also floats that option to the top. All presets stay in the list;
inherit-global remains omitted from the global scope.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
CFG_BACKUP_VERIFY_DATA_PERCENT had no field-type case in config-shared.js so
it fell through to the default text input — no up/down stepper and no bounds.
Render it as a number input (min=1, max=100, % unit), which the custom-number
enhancer picks up automatically for the up/down controls.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Application backups were driven by one crontab entry per app, each offset by
id * CFG_BACKUP_CRONTAB_APP_INTERVAL minutes. That minute offset is written
straight into cron's 0-59 minute field, so past ~20 apps it overflowed into
an invalid entry that silently never fired, and the fixed spacing could not
serialize backups that ran longer than the gap.
Replace it with a single daily entry (`libreportal backup scheduled`) that
enqueues a backup task per enabled app. The existing systemd task processor
drains them serially — no minute overflow, real serialization, and backups
are now visible/cancellable in the Tasks UI. Per-app enable is read from
CFG_<APP>_BACKUP at schedule time instead of being mirrored into crontab.
Removes the stagger machinery (timing/setup/check/remove scripts), the
now-unused cron_jobs table + insert, and the CFG_BACKUP_CRONTAB_APP_INTERVAL
config knob and its WebUI field.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Both Backup style controls (global Configuration retention and per-location) now surface the selected preset's description via a tooltip on the label, updated when the preset changes, instead of an always-visible hint line. Also makes the per-location tooltip accurate for non-default presets (it was previously fixed to the inherit text).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Make the expanded location editor read like /config: Connection and Retention now use the section header + .domains-divider layout, and Connection gets a description. Move the retention 'Backup style' guidance into a tooltip and drop the always-visible hint line below it. Move the Enabled toggle out of the Connection fields into the collapsed location row header so a location can be enabled/disabled without expanding it; setLocationEnabled persists the change via the same config_update routing as saveSection.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Render a backup location's connection fields in a single .config-fields grid like /config's renderer, instead of chunking every 3 fields. The chunking left ragged blocks whose columns stopped lining up once any field was hidden (PATH_MODE/SSH/etc.) — the grid handles row layout and drops hidden fields cleanly. Also add a config-divider below the "Keep your config backed up offline" warning banner.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add a config-divider after the header-only Danger Zone banner on the features page so a line separates it from the feature fields. Drop the now-redundant 24px top padding on .config-actions since the divider above the Save/Reset buttons already provides that spacing.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Emit a config-divider before the Danger Zone / "Show Advanced Options" container in the shared renderer, so a line separates the regular fields from the advanced toggle — mirroring the dividers above Save/Reset and inside #advanced-sections. Applies to every config category.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Propagate the ✓ Success / ✗ Error / ! Notice / ❯ Question glyphs (from markers.sh) through the rest of the pipeline: swap the inlined helpers in init.sh and generate_arrays.sh, and replace raw echo -e "${RED}ERROR:${NC}" calls with the isX helpers in config_check_missing.sh, check_success.sh, initilize_files.sh, and reset_git.sh.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the ALLCAPS "SUCCESS:/NOTICE:/ERROR:/QUESTION:/OPTION:" prefixes
with distinct per-status glyphs and calmer title-case words:
✓ Success ! Notice ✗ Error ❯ Question ❯ Option
The portal chevron ❯ marks the interactive prompts. Distinct glyph + word
stays readable with no colour and greppable in logs. Display-only; nothing
parses these prefixes.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Swap the ### hash headers (isHeader) for a ╔═╗ ║ ╚═╝ double-line box and
wrap the LibrePortal logo in a matching 52-wide box. Build the rule with
printf-repeat and fixed pad widths instead of tr/${#} so multibyte box
chars stay aligned regardless of locale. Mirrors the credentials panel.
Applied to all three copies (markers.sh, init.sh, generate_arrays.sh).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add a thin divider above the Save/Reset buttons, and one at the top of
#advanced-sections so a line appears between the "Show Advanced Options"
toggle and the advanced fields only when they're revealed. Shared config
renderer, so it applies to every config category (backup included).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replaces the slow, interactive per-variable scan with a deterministic
reconcile: each live config is rebuilt from its (freshly-cloned) template —
keeping the user's existing values, adding new template keys
(CFG_REQUIREMENT_CONFIGS_AUTO_UPDATE), and dropping keys the template no
longer defines (new CFG_REQUIREMENT_CONFIGS_AUTO_DELETE, default true).
Structure/order/comments follow the template; non-interactive; atomic with a
.bak; refuses to act on a missing/empty template so a broken clone can't wipe
a config. Applies to both general and per-app configs.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The static per-app CFG_<APP>_HOST_NAME is gone. host_setup (the app's
canonical FQDN, feeding the legacy single DOMAINSUBNAME_DATA used by app env
vars, the app URL and trusted-domains) is now derived from the app's primary
Traefik port's subdomain: first recommended port, else first Traefik port;
@/root -> apex, set -> sub.domain, empty -> app-name. Removes HOST_NAME from
all app configs, the config-form field mapping (Hostname), the dead
headscale stub, and wireguard.sh (now uses host_setup). Completes the move to
dynamic per-port subdomain routing.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The Subdomain field's help text still said it inherits CFG_HOST_NAME and that
the label-generation refactor was pending — both untrue now that per-port
subdomain routing has shipped. Reword to: empty -> app-name default, @ ->
domain apex, multi-level supported.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
setupLocalDnsRewrites points every configured domain at the server's LAN IP
inside the self-hosted resolver, so app subdomains resolve locally and hit
Traefik directly (valid certs, no router hairpin). AdGuard gets a wildcard
rewrite per domain via its REST API; Pi-hole gets per-host A records in the
supported, mounted custom.list (no wildcard support there). Safe by
construction: idempotent, guarded by installed-checks, cannot corrupt the
resolver. Hooked into the Apply-DNS actions and resolver install. Also drops
the dead HOST_NAME read from the setupDNSIP stub.
NOTE: needs a live smoke-test — the AdGuard API call and Pi-hole reload
can't be exercised without the running containers.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
An empty subdomain previously resolved to the domain apex, which would
collide on the root for any unconfigured Traefik port. Treat empty as the
app-name default (matching legacy behaviour); apex is reachable only via the
explicit @ / root sentinel.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Replace the static one-host-per-app model with per-port routers: each
Traefik-managed port carries a subdomain (12-col PORT format) and gets a
DOMAINSUBNAME_TAG_<n> host, so one container can serve unlimited hosts.
tagsProcessorPortSubdomains stamps per-port hosts (subdomain @/empty = apex,
multi-level allowed); tagsProcessorPortRouterBlocks comments out
# TRAEFIK_PORT_<n>_BEGIN/END blocks for non-Traefik ports so unfilled
placeholders never ship (mirrors GLUETUN_OFF). Convert all 27 router apps
(subdomains seeded from HOST_NAME; headscale admin. prefix -> subdomain).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
HOST_NAME was read but ignored — the FQDN was built from app_name, so 8
apps (vault, cloud, search, notes, social, meet, board, bookmark) routed at
the wrong host and Traefik disagreed with DNS. Build host_setup from
HOST_NAME (falling back to app_name); treat HOST_NAME="@"/"root" as the
domain apex (root-of-domain hosting, previously impossible). Document @ in
the Hostname field tooltip.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
A nebula-themed marketing/get site under site/, matching the dashboard
(aurora background, glass cards, system fonts — no Google/third-party
calls). The app grid + category filters are generated from the repo:
scripts/gen-data.mjs reads each containers/<app>/<app>.config and emits the
data Eleventy renders. `npm run build` -> static site in dist/.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerInstallApp built the installer name by upper-casing only the first
letter of the slug (libreportal -> installLibreportal), which can't match
camelCase installers like installLibrePortal. After the EasyDocker ->
LibrePortal rename this broke `libreportal` installs with
"installLibreportal: command not found".
If the naive name isn't a defined function, resolve it case-insensitively
against the function table (compgen -A function), and fail with a clear
message if nothing matches. Works for any compound brand/app name.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Keep just the wordmark + portal; the underline read poorly.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The raised (‾▔) divider read strangely; go back to the low _▁ step-ticks
the prior look used and restore the leading blank line. Keep the divider
extended to the end of the final letter.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Raise the underline to high marks (‾▔) so it tucks under the wordmark,
extend it to reach the end of the final letter, and remove the leading
blank line so the banner starts flush.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Add a small left gap before the wordmark and a step-tick underline
(_▁ repeated) matched to the logo width.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The portal between Libre/Portal was a closed ring ("just a circle"); give it
two feet (╨─╨) and a touch more breathing room on each side.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The startup banner (displayLibrePortalLogo in init.sh/start.sh and the
generate_arrays.sh splash) still rendered the old "EASY DOCKER" figlet art.
Swap it for a LibrePortal wordmark — Calvin S mixed-case "Libre"/"Portal"
with a small framed portal glyph between the two words.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Surface when LibrePortal is behind upstream and let users update from the
WebUI, reusing the proven git-update path instead of reinventing it.
Detection (host): webuiSystemUpdateCheck writes
frontend/data/system/update_status.json from a throttled git fetch +
behind-count + VERSION compare, off the existing per-minute
`webui generate system` cron. A new /VERSION file is the canonical version.
Display (frontend): update-notifier.js/.css render a global topbar badge
(every page) and a dashboard banner (prominent when behind, subtle "up to
date" with a manual check otherwise), plus a details panel.
Actions go through the task pipeline:
- `libreportal update apply` -> webuiRunUpdate (non-interactive: guards,
forced check, gitPerformUpdate, then dockerInstallApp libreportal)
- `libreportal update check` -> forced recheck
gitFolderResetAndBackup's body is extracted into gitPerformUpdate (no exit)
so the WebUI path can reuse it; the interactive CLI flow is unchanged.
Detection JSON verified against the repo (up-to-date and behind cases).
webuiRunUpdate's re-clone + redeploy still needs validation on a live host.
The latest-version source is git for now and is the single swap point for
get.libreportal.org later — the JSON contract and frontend stay unchanged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
"Cloud" clashed with the project's anti-surveillance ethos and misdescribed
the service (it deliberately holds no plaintext). "Connect" reflects what it
actually does — reach and protect your self-hosted apps without us holding
your data.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- Add "Why LibrePortal" (privacy motivation) and "Acknowledgments"
(built from scratch since 2023, inspired by bmcgonag/docker_installs)
- Reframe LibrePortal Cloud with the plain-English "courier with a sealed
box" model: we only handle encrypted data, keys stay with the user
- Drop "Managed hosting — we run LibrePortal for you"
- Update Support to free community + optional paid priority
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
2026-05-21 21:56:43 +01:00
955 changed files with 45827 additions and 23600 deletions
CFG_BACKUP_ENGINE=restic # Default Backup Engine - Fallback engine for new locations (each location can override) [restic:restic|borg:BorgBackup|kopia:Kopia]
CFG_BACKUP_STRATEGY=stop-snapshot-start # Backup Strategy - How containers are quiesced before snapshotting [stop-snapshot-start:Stop → snapshot → start (safe default)|pause-snapshot-unpause:Pause → snapshot → unpause (less downtime)|live:Live — snapshot while running (only with DB dump hooks)]
CFG_BACKUP_ENGINE=restic # Default Backup Engine - Fallback engine for new locations (each location can override) [restic:Restic|borg:BorgBackup|kopia:Kopia]
CFG_BACKUP_DEFAULT_PATH= # Default Backup Location - Base directory for locations set to Automatic path mode; each location lives in its own numbered subfolder (<path>/<id>). Empty = the LibrePortal backups root (own mount-able).
CFG_BACKUP_STRATEGY=auto # Backup Strategy - How containers are quiesced before snapshotting [auto:Automatic — live where safe, stop otherwise (recommended)|stop-snapshot-start:Stop → snapshot → start (always safe)|pause-snapshot-unpause:Pause → snapshot → unpause (less downtime)|live:Live — snapshot while running (force)]
CFG_BACKUP_VERIFY_AFTER=true # Verify After Backup - Run integrity check after each backup
CFG_BACKUP_VERIFY_DATA_PERCENT=5 # Verify Data Sample % - Percentage of repo data to checksum-verify weekly
CFG_BACKUP_PARALLEL_REPOS=true # Parallel Repos - Push to all enabled locations in parallel
CFG_INSTALL_NAME=Change-Me # Installation Name - The name for your LibrePortal instance
CFG_TIMEZONE=Etc/UTC # System Timezone - Timezone for scheduled tasks and logging timestamps
CFG_INSTALL_LEVEL=beginner # Experience Level - Beginner hides technical detail and skips advanced setup steps. Advanced reveals everything by default. Set during the first-run wizard; can be flipped any time via the Advanced toggle in the WebUI. [beginner:Beginner — simple|advanced:Advanced — show everything]
CFG_DOCKER_INSTALL_TYPE=rooted # Docker Installation Type - Security based setup rooted or rootless Docker installation [rooted|rootless]
CFG_DOCKER_INSTALL_TYPE=rootless # Docker Installation Type - rootless (default, recommended): containers run unprivileged so a breakout isn't host root; rooted: legacy, containers run as root [rootless|rooted]
CFG_DOCKER_INSTALL_USER=dockerinstall # Docker Install User - Username for Docker installation operations
CFG_DOCKER_INSTALL_PASS=RANDOMIZEDPASSWORD2 # Docker Install Password - Password for Docker install user
CFG_INSTALL_MODE=local # Installation Mode - Method used for installation of LibrePortal
CFG_INSTALL_MODE=release # Installation Mode - How LibrePortal is fetched + updated. Hidden by default — only shown when Developer Mode is on (click the LibrePortal logo 10 times) or when the install is already on git/local (auto-enables Developer Mode). **DEV** [release:Release - Stable|git:Git clone|local:Local folder]
CFG_RELEASE_BASE_URL=https://get.libreportal.org # Release Host - Base URL serving the release channels (override for self-hosting) **ADVANCED**
CFG_RELEASE_CHANNEL=stable # Release Channel - Pick the release channel for the tarball installer. Stable is the recommended default; Edge ships from main and may contain in-flight changes. **DEV** [stable:Release - Stable|edge:Release - Bleeding Edge]
CFG_DEV_MODE=false # Developer Mode - Reveal advanced developer / dev-install options across the WebUI. Auto-enables when the install is already on git/local. Easter egg: click the LibrePortal logo 10 times to toggle. **ADVANCED** [true:On|false:Off]
CFG_GIT_USER=changeme # Git Username - Git username for repository authentication
CFG_GIT_KEY=changeme # Git Access Key - SSH key or API key for Git repository access
CFG_GIT_UPDATES=true # Auto Check Updates - Check for Git repository updates automatically
CFG_GIT_AUTO_UPDATES=true # Auto Apply Updates - Automatically apply Git updates when available
CFG_GIT_UPDATES=true # Auto Check Updates - Check for updates automatically
CFG_GIT_AUTO_UPDATES=true # Auto Apply Updates - Automatically apply updates when available
CFG_REQUIREMENT_CONFIG=true # Configuration Management - Install the configuration management system. Disabling this on an existing install will brick the system. **ADVANCED** **DEV**
CFG_REQUIREMENT_COMMAND=true # Command Line Tool - Install the libreportal command line tool. Disabling this on an existing install will brick the system. **ADVANCED** **DEV**
CFG_REQUIREMENT_WEBUI=true # Web Interface - Install the LibrePortal WebUI. Disabling this on an existing install will brick the system. **ADVANCED** **DEV**
CFG_REQUIREMENT_WEBUI_SERVICE=true # Web Task Service - Install the task-processor systemd service that backs the WebUI. Disabling this on an existing install will brick the system. **ADVANCED** **DEV**
CFG_REQUIREMENT_DATABASE=true # Database Support - Install and configure database support for application data storage. Install-time choice only — flipping post-install will not retrofit. **ADVANCED** **DEV**
CFG_REQUIREMENT_DOCKER_CE=true # Docker CE - Install Docker Community Edition instead of the distro default. Install-time choice only — flipping post-install does not swap Docker. **ADVANCED** **DEV**
CFG_TEXT_EDITOR=nano # Text Editor - Default text editor for system operations [nano|vim]
CFG_REQUIREMENT_CRONTAB=true # Scheduled Tasks - Install scheduled tasks and automated maintenance jobs
CFG_REQUIREMENT_CONFIGS_CHECK=true # Config Validation - Validate configuration files on startup for errors and consistency
CFG_REQUIREMENT_CONFIGS_AUTO_UPDATE=true # Auto Config Updates - Add new config options from the template (non-interactive)
CFG_REQUIREMENT_CONFIGS_AUTO_DELETE=true # Auto Config Deletes - Remove config options no longer present in the template
CFG_REQUIREMENT_MISSING_IPS=false # IP Configuration Check - Check for and alert about missing IP configurations
CFG_REQUIREMENT_CONTINUE_PROMPT=false # Continue Prompts - Show continue prompts during installation for user confirmation
CFG_REQUIREMENT_CONTINUE_ON_ERROR=true # Continue On Error - Log failures to error_report.log and continue instead of aborting (on by default to surface issues; turn off for strict abort once clean)
CFG_REQUIREMENT_SUGGEST_INSTALLS=false # Install Suggestions - Enable application suggestions and recommendations during installation
CFG_REQUIREMENT_SUGGEST_METRICS=true # Metrics Suggestions - Offer Prometheus and Grafana during first install (requires Install Suggestions enabled)
CFG_REQUIREMENT_WHITELIST_PORT_UPDATER=true # Auto Port Whitelisting - Update the port whitelist automatically when applications are installed or removed
CFG_ROOTLESS_NET=pasta # Rootless Network Driver - Network stack for rootless Docker. pasta (default): actively maintained, preserves the real client source IP on inbound connections, lower idle CPU; slirp4netns: legacy fallback, maintenance-only upstream. The matching rootlesskit port driver is selected automatically. On Debian, the installer also applies the local AppArmor override pasta needs (see scripts/docker/install/rootless/rootless_apparmor.sh) so this is a single-toggle switch. **ADVANCED** [pasta:Pasta (default, actively maintained)|slirp4netns:slirp4netns (legacy fallback)]
CFG_HOTFIX_AUTO=security-breakage # Hotfix Auto-Apply - Which signed hotfix severities apply automatically on the update check [security-breakage|all|off]
isError "AdGuardHome admin endpoint did not respond on $adguard_setup_url within $((adguard_max_attempts *2))s — open the URL and complete setup manually, then re-run the installer to apply the post-setup tweaks."
# 422/403 here typically means setup was already done on a
# previous install; the post-setup tweaks below are still
# safe to run against the existing yaml.
isNotice "AdGuardHome /control/install/configure rejected the request — assuming it's already configured. If this is a fresh install, complete setup manually at $adguard_setup_url."
fi
# Small breather so AdGuardHome finishes flushing AdGuardHome.yaml
# to disk before the sed edits below touch it.
#sleep 3
fi
#result=$(sudo sed -i "s/address: 0.0.0.0:80/address: 0.0.0.0:${usedport2}/g" "$containers_dir$app_name/conf/AdGuardHome.yaml")
#checkSuccess "Changing port 80 to $usedport2 for Admin Panel"
#result=$(sudo sed -i "s/port: 53/port: ${usedport3}/g" "$containers_dir$app_name/conf/AdGuardHome.yaml")
#checkSuccess "Changing port 53 to $usedport3 for DNS Port"
#result=$(sudo sed -i "s/port_https: 443/port_https: ${usedport4}/g" "$containers_dir$app_name/conf/AdGuardHome.yaml")
#checkSuccess "Changing port 443 to $usedport4 for DNS Port"
#result=$(sudo sed -i "s/port_dns_over_tls: 853/port_dns_over_tls: ${usedport5}/g" "$containers_dir$app_name/conf/AdGuardHome.yaml")
#checkSuccess "Changing port 853 to $usedport5 for port_dns_over_tls"
#result=$(sudo sed -i "s/port_dns_over_quic: 853/port_dns_over_quic: ${usedport5}/g" "$containers_dir$app_name/conf/AdGuardHome.yaml")
#checkSuccess "Changing port 853 to $usedport5 for port_dns_over_quic"
# NOTE: We deliberately do *not* force `tls.enabled: true` here.
# That section configures encrypted DNS (DoT/DoH/DoQ) and AdGuardHome
# crash-loops on startup with `[fatal] creating dns server: parsing
# tls key pair: tls: failed to find any PEM data in certificate input`
# if `enabled: true` is set without a real certificate pair pointed
# at by `certificate_path` / `private_key_path`. The admin user can
# opt into encrypted DNS from Settings → Encryption once they've
# provided a cert.
if[[$public=="true"]];then
result=$(sudo sed -i "s|allow_unencrypted_doh: false|allow_unencrypted_doh: true|g""$containers_dir$app_name/conf/AdGuardHome.yaml")
checkSuccess "Setting allow_unencrypted_doh to false for Traefik"
fi
result=$(sudo sed -i "s|anonymize_client_ip: false: false|anonymize_client_ip: true|g""$containers_dir$app_name/conf/AdGuardHome.yaml")
checkSuccess "Setting anonymize_client_ip to true for privacy reasons"
# Force the admin web bind back to 0.0.0.0:3000 inside the container.
# The docker-compose mapping is `<host_port>:3000`, so the container
# MUST listen on 3000 internally for the host port to reach it. After
# the install API call AdGuardHome sometimes ends up bound to 0.0.0.0:80
# (its build-time default) instead of the port we sent — which is
# exactly what causes "unable to connect" on the host port.
isSuccessful "AdGuardHome admin UI is reachable on $adguard_setup_url (HTTP $adguard_health_code)"
break
fi
sleep 1
((adguard_health_attempts++))
done
if((adguard_health_attempts >= 20));then
isError "AdGuardHome admin UI did not respond after restart on $adguard_setup_url. Check the container logs (\`docker logs adguard-service\`) and the conf/AdGuardHome.yaml bind address."
fi
((menu_number++))
echo""
echo"---- $menu_number. Running Application specific updates (if required)"
isError "AdGuardHome admin endpoint did not respond on $adguard_setup_url within $((adguard_max_attempts *2))s — open the URL and complete setup manually, then re-run the installer to apply the post-setup tweaks."
# 422/403 here typically means setup was already done on a
# previous install; the post-setup tweaks below are still
# safe to run against the existing yaml.
isNotice "AdGuardHome /control/install/configure rejected the request — assuming it's already configured. If this is a fresh install, complete setup manually at $adguard_setup_url."
fi
fi
local result
if[["$public"=="true"]];then
result=$(runFileOp sed -i "s|allow_unencrypted_doh: false|allow_unencrypted_doh: true|g""$containers_dir$app_name/conf/AdGuardHome.yaml")
checkSuccess "Setting allow_unencrypted_doh to false for Traefik"
fi
result=$(runFileOp sed -i "s|anonymize_client_ip: false: false|anonymize_client_ip: true|g""$containers_dir$app_name/conf/AdGuardHome.yaml")
checkSuccess "Setting anonymize_client_ip to true for privacy reasons"
# Force the admin web bind back to 0.0.0.0:3000 inside the container.
# The docker-compose mapping is `<host_port>:3000`, so the container
# MUST listen on 3000 internally for the host port to reach it. After
# the install API call AdGuardHome sometimes ends up bound to
# 0.0.0.0:80 (its build-time default) — exactly what causes "unable
isSuccessful "AdGuardHome admin UI is reachable on $adguard_setup_url (HTTP $adguard_health_code)"
break
fi
sleep 1
((adguard_health_attempts++))
done
if((adguard_health_attempts >= 20));then
isError "AdGuardHome admin UI did not respond after restart on $adguard_setup_url. Check the container logs (\`docker logs adguard-service\`) and the conf/AdGuardHome.yaml bind address."
fi
}
adguard_install_message_data()
{
# Echo the admin user + password as space-separated tokens so they
# become $username $password positional args to menuShowFinalMessages.
if ! sudo nft list tables 2>/dev/null | grep -qiE 'crowdsec';then
if ! runSystem nft list tables 2>/dev/null | grep -qiE 'crowdsec';then
isNotice "CrowdSec nftables table not yet present after ${_wait}s. Bouncer may still be starting; re-run the verification Tools action in a minute if rules don't appear."
else
local cs_prio ufw_prio
cs_prio=$(sudo nft list ruleset 2>/dev/null | awk '/table .* crowdsec/{flag=1} flag && /priority/{match($0,/priority [-0-9]+/); print substr($0,RSTART+9,RLENGTH-9); exit}')
ufw_prio=$(sudo nft list ruleset 2>/dev/null | awk '/chain ufw[a-z0-9-]*input/{flag=1} flag && /priority/{match($0,/priority [-0-9]+/); print substr($0,RSTART+9,RLENGTH-9); exit}')
cs_prio=$(runSystem nft list ruleset 2>/dev/null | awk '/table .* crowdsec/{flag=1} flag && /priority/{match($0,/priority [-0-9]+/); print substr($0,RSTART+9,RLENGTH-9); exit}')
ufw_prio=$(runSystem nft list ruleset 2>/dev/null | awk '/chain ufw[a-z0-9-]*input/{flag=1} flag && /priority/{match($0,/priority [-0-9]+/); print substr($0,RSTART+9,RLENGTH-9); exit}')
if[[ -z "$ufw_prio"]];then
isSuccessful "UFW not in nftables — no ordering needed (CrowdSec prio: ${cs_prio:-?})."
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
