When resolving a conflict, the last_sync_source is now set to the
winner's actual source name (koreader, web, etc.) rather than always
'manual'. This prevents subsequent saves from re-triggering conflicts.
Also removes the strict oneof validation on the winner field since
the source name is dynamic.
Web reader can now send surrounding text at current reading position.
Stored in reading_progress.context_text for use as CFI resolution
fallback when converting epubcfi to CREngine XPointer.
Forward (push): When KOReader pushes a CREngine XPointer, convert it
to standard epubcfi before storing. Uses CFIConverter.ConvertCREToStandard
with context_text for text search fallback.
Reverse (pull): When KOReader pulls progress, convert stored standard
epubcfi back to CREngine XPointer via CFIConverter.ConvertStandardToCRE.
Returns as koreader_xpointer field in GetMetadata response.
Other changes:
- updateProgressForBook: pass context_text to SaveProgress
- enqueueProgressForBook: pass context_text to queue
- KOReaderProgressData: add KoreaderXPointer field
- New convertCFIToXPointer helper method
- Wire libraryService in main.go for EPUB path resolution
- SaveProgressRequest: add ContextText field
- SaveProgress: carry existing ContextText from DB, overwrite when provided
- ProgressUpdate: add ContextText field for checkpoint sync
- SyncQueueProcessor: serialize/deserialize context_text in sync data
- buildProgressSnapshot: include context_text in snapshot data
Stores surrounding text (~100 chars) at the reader's current position.
Used as fallback for CFI resolution when converting between epubcfi
and CREngine XPointer formats.
Updates:
- schema.sql: add context_text TEXT column, update stored procedure
- queries.sql: add context_text to GetUniversalProgress and
UpdateUniversalProgress queries
- Regenerate sqlc Go code (models.go, querier.go, queries.sql.go)
Implements ConvertCREToStandard which converts CREngine XPointers
(e.g. /body/DocFragment[6]/body/div/p[47]/text().2399) to standard
epubcfi format (e.g. epubcfi(/6/12!/4/2[id]/4/1:7)).
Key components:
- indexChildNodes: faithful port of foliate-js's epubcfi.js algorithm
for computing CFI-compatible child node indices including virtual
positions, null positions between adjacent elements, and text chunks
- preprocessXHTML: converts XHTML self-closing tags (e.g. <a id="x"/>)
to open/close pairs so Go's HTML parser produces the same DOM as the
browser's XHTML parser
- buildCFI: walks up from a text node to body, computing CFI indices
at each level using indexChildNodes
- findTextInNode: regex-based whitespace-flexible text search for
context_text fallback positioning
- convertByPercentageOffset: estimates position via book-wide character
counts when no context_text is available
- ConvertCREToStandard: orchestrates text search → percentage fallback
Supports CREngine XPointer format, CREngine fragment ID format
(#_doc_fragment_N_anchor), and includes round-trip test coverage for
1984 and Crime and Punishment EPUBs.
Previously, when HTMX form submissions (profile update, password change)
returned a successful JSON response like {"message": "profile updated
successfully"}, the raw JSON was swapped into the target div as plain text.
The htmx:afterSwap listener in toast.ts only handled error responses.
Extend it to also intercept successful 2xx JSON responses that contain a
"message" field, showing a green success toast and clearing the raw JSON
from the target element. Only JSON responses are intercepted (checked via
Content-Type header), so legitimate HTML swaps are unaffected.
The timezone feature has been fully implemented across the codebase
(database queries, API handlers, profile form, reader settings).
This planning document is no longer needed.
The cases.Title caser panicked with 'slice bounds out of range' when
processing certain Unicode characters that expand during case transformation
(e.g. ß → SS). This panic crashed the entire server during scanning, causing
WebSocket disconnections and failed scan requests.
On viewports below 970px the header now collapses to a compact bar with
only the Bookhoard logo and a hamburger button. Clicking the hamburger
reveals a slide-down panel containing:
- Full-text search input (wired to the existing debounced search API)
- Navigation links (Library, All Books, Series, Collections, Progress, Devices)
- Collapsible theme switcher with all 7 themes and 4 bookshelf backgrounds
- User section: profile/admin/logout when logged in, inline login form when logged out
Changes:
- templates/header.templ: add hamburger button, mobile panel with all controls,
hide desktop search/theme/user controls below nav breakpoint
- web/src/header.ts: add mobileMenuOpen state to Alpine header component
- web/src/search.ts: refactor initializeSearch to wire both desktop and mobile
search inputs, track active input for results container placement
- tailwind.config.ts: add custom 'nav' screen breakpoint at 970px so the
mobile menu activates before the search bar becomes unusable
- web/static/style.css: rebuilt with new nav breakpoint utility classes
Upgrade from templ v0.3.1001 to v0.3.1020. Generated code changes include
JoinStringErrs -> ResolveAttributeValue and removal of manual EscapeString
calls (now handled internally by ResolveAttributeValue).
Add --mount=type=cache for Go build cache to speed up repeated builds.
Remove the -a flag which forces full rebuilding of all packages and is
unnecessary when using cache mounts.
npm ci requires a package-lock.json and installs exactly what it specifies.
npm install resolves from package.json directly, making the Docker build
fully self-contained — you can clone the repo and run docker build with no
local Node tooling or pre-existing lockfile.
PDFs failed to open with error:
Invalid factory url: "http://localhost:8765/static/undefined"
Root cause: foliate-js's pdfjsPath() uses new URL(dynamicPath, import.meta.url)
to resolve runtime asset paths (standard_fonts/, cmaps/). Vite transforms this
pattern into a static asset map lookup at build time, but can only resolve
known static file paths — not dynamically-constructed directory paths. The
lookup returns undefined, producing a broken URL.
Fix (two parts):
1. foliate-js fork (commit d164d6f): Export an overridable config.pdfjsPath
function. Module-level code (worker, CSS) continues using import.meta.url
directly (works fine with Vite for static filenames). The makePDF function
uses config.pdfjsPath for runtime paths, allowing consumers to override it.
2. Bookhoard changes:
- Update foliate-js dependency to d164d6f
- Override config.pdfjsPath in reader.ts to resolve to /static/vendor/pdfjs/
- Add Vite plugin (pdfjsAssets) that copies standard_fonts/ and cmaps/ from
node_modules to the build output during vite build (the standard approach
used by react-pdf and other pdfjs-dist consumers)
- Remove manual cp commands from build:ts scripts
Replace hardcoded 'podman' with auto-detected CONTAINER_RUNTIME variable
that prefers docker and falls back to podman. Override with:
CONTAINER_RUNTIME=podman make rebuild-app
Remove the ensure_healthy and compose_up macros that worked around
podman-compose hanging on non-systemd systems (e.g., Void Linux with
runt). These are no longer needed — healthchecks are now handled by
plain wait loops in the targets themselves, and compose up -d no longer
blocks on health conditions.
Regenerated database code after sqlc version upgrade from v1.30.0 to
v1.31.1. No functional changes — only the version header in generated
files was updated.
Files: db.go, models.go, querier.go, queries.sql.go
Podman relies on systemd timers to schedule automatic healthchecks. On
non-systemd systems (e.g., Void Linux with runit), healthchecks never
fire, which causes podman-compose to hang forever waiting for
service_healthy conditions that never resolve.
Add two Make macros to handle this transparently:
- compose_up: runs podman compose up -d normally on systemd, but with
a 15-second timeout on non-systemd to create containers without
hanging. Supports passing compose flags via $(call compose_up,args).
- ensure_healthy: on non-systemd systems, waits for the database to
accept connections, manually triggers its healthcheck, starts the app
container, waits for the app health endpoint, and triggers its
healthcheck. On systemd systems, the runtime check is skipped entirely
(zero overhead).
Both macros use a runtime shell check for /run/systemd/system, so the
same Makefile works identically on all systems without parse-time
conditionals.
Applied to all compose-up targets: up, rebuild, rebuild-force,
rebuild-force-db, rebuild-app, rebuild-app-force, test-integration,
and test-env-up.
Refs: https://github.com/containers/podman/pull/27033
- helpers.go: Promote getText() from a local closure in frontend.go
to a package-level function so it can be used by resolveLibrary.
Add resolveLibrary(c, cfg, user.ID) helper that:
1. Reads library_id query param (explicit navigation wins)
2. Falls back to selectedLibrary cookie — validates __all__
sentinel or real UUID, rejects garbage values silently
3. Falls back to user's first visible library
Returns LibraryResolution struct with LibraryID, IsAll, LibUUID,
Libraries, and FirstID — eliminating repeated boilerplate across
all SSR routes.
- frontend.go: Replace manual library resolution boilerplate in 5
SSR route handlers (series, tags/detail, bookshelf, dashboard,
collections/:id) with resolveLibrary(). Each route now gets cookie-
aware library selection for free. Collection detail correctly
handles All Libraries mode for both system and user collections.
Dashboard no longer makes a redundant second GetUserVisibleLibraries
call.
- storage.ts: Centralize ALL_LIBRARIES = "__all__" sentinel constant.
setSelectedLibrary() now writes both localStorage and a cookie
(selectedLibrary, Path=/, SameSite=Lax, max-age=365d). The
sentinel "__all__" is used in both storage mediums — empty strings
are never stored. getSelectedLibrary() maps __all__ back to "".
Cookie enables server-side rendering to read the stored library
selection without access to localStorage.
- library-switcher.ts: Import ALL_LIBRARIES and setSelectedLibrary/
getSelectedLibrary from storage.ts instead of managing localStorage
directly. Remove local constants.
- dashboard.ts: Remove duplicate localStorage.setItem call that was
overwriting the __all__ sentinel with raw empty string. Fix
reloadPage() and scan-complete handler to work with empty libraryId.
openDashboardSettings/saveDashboardSettings show clear messages for
All Libraries mode.
- collections.ts: Remove library switcher initialization from the
collections list page — the list page no longer has a switcher.
- series.ts: Rewrite to use initLibrarySwitcher from library-switcher
module and switchWithTransition for navigation. Series card links
no longer include library_id in their URLs.
- bookshelf.ts: Autocomplete fetch calls handle empty libraryId
correctly for All Libraries mode.
- search.ts, collection-rules.ts: Use setSelectedLibrary() and
getSelectedLibrary() from storage.ts instead of direct localStorage
access.
Regenerate all templ-generated Go files. These changes are caused
by running templ generate with a slightly different CLI version
(v0.3.1001) than the go.mod dependency (v0.3.1020), resulting in
minor formatting/import diffs across all templates. No functional
changes.
- bookshelf.templ: Fix form field name from "library" to "library_id"
to match the handler's QueryParam("library_id"). Add "All Books"
as the default option in the library filter dropdown. The bookshelf
uses its own inline filter, NOT the universal library switcher.
- collections.templ: Remove @LibrarySwitcher from the collections list
page — collections are not library-specific, so the switcher was
misleading. Fix data-id interpolation bug where {collection.ID} was
rendered as literal text instead of being interpolated.
- series.templ: Replace inline library selector with the universal
@LibrarySwitcher component. SeriesCard links no longer include
library_id since series detail always shows all books.
- dashboard.go: library_id query param is now optional. Empty/missing
library_id is passed as pgtype.UUID{Valid: false} to the service
layer, enabling All Libraries mode.
- series.go: library_id is optional for series listing. GetSeriesBooks
no longer receives a libraryID — it always returns all books in a
series regardless of library.
- collections.go: Restructure GetCollection to handle system
collections (query_type != "") with an optional libraryID. When
libraryID is empty (All Libraries), GetDashboardSections receives
pgtype.UUID{Valid: false} so no library filter is applied.
- dashboard_service.go: Change libraryID parameter from uuid.UUID to
pgtype.UUID across GetDashboardSections, GetDashboardPreferences,
and all helper methods. pgtype.UUID{Valid: false} now signals
"no library filter" (All Libraries), which gets passed through
to sqlc.narg() in the SQL layer.
- series_service.go: Drop libraryID parameter from GetSeriesBooks
entirely. Series are not library-specific — all books in a series
are shown regardless of which library they belong to.
Convert 12 SQL queries to use sqlc.narg('library_id') instead of
direct @library_id parameters. This allows passing a NULL/invalid
pgtype.UUID to mean "no library filter" (i.e., All Libraries),
making the SQL layer correctly handle the optional filter via:
(sqlc.narg('library_id')::uuid IS NULL
OR mi.library_id = sqlc.narg('library_id')::uuid)
Also remove the library_id filter from GetSeriesBooks entirely —
a series is a series regardless of library.
Queries affected:
- GetDashboardSections, GetRecentlyAdded, GetInProgress
- GetHighestRated, GetMostRead, GetAbandonedBooks
- GetLeastRead, GetBooksByTag, GetCollectionItemsForDashboard
- SearchMediaItemsUnified, GetSeriesCardsData
Generated code (queries.sql.go, querier.go) regenerated via sqlc.
Wire up the shared library switcher module on dashboard, collections
list, collection detail, and collection rules pages. All pages use SSR
for initial load and AJAX with fade transitions on library switch.
web/src/collections.ts:
- Add initCollectionsPage() that auto-detects list vs detail page
by checking for #collection-data element
- Collections list: onSwitch fetches /api/collections?library_id=X and
re-renders the grid with per-library book counts
- Collection detail: onSwitch fetches /api/collections/:id?library_id=X
and re-renders the books grid
- Add renderCollectionsGrid() and renderCollectionBooks() with
Alpine.initTree() calls for dynamic content
- Collection cards now link with ?library_id= from selected library
- Update hidden #collection-data data-library-id on switch
web/src/dashboard.ts:
- Replace standalone switchLibrary() with initLibrarySwitcher() +
switchWithTransition() from shared module
- Extract fetchAndRenderSections() helper shared by onSwitch callback,
reloadPage(), and saveDashboardSettings()
- Remove inline #library-select change listener and switch-library
data-action handler (now handled by shared module)
- Scan-complete event handler unchanged (independent incremental logic)
web/src/collection-rules.ts:
- Update backToCollection() to preserve library context by appending
?library_id= from localStorage selectedLibrary key
Replace inline library switcher HTML with shared LibrarySwitcher component
across all collection pages and the dashboard.
templates/collections.templ (Collection):
- Update signature to accept libData []LibraryData, currentLibraryID
- Add @LibrarySwitcher(libData, currentLibraryID) after header
- Change x-init to initCollectionsPage() for unified initialization
templates/collections.templ (CollectionDetail):
- Update signature to accept libData []LibraryData
- Add @LibrarySwitcher(libData, libraryID) after header
- Fix broken "Back to Collections" button: replace non-existent
backToCollections Alpine method with a plain <a href="/collections"> link
- Change x-init to initCollectionsPage() for unified initialization
templates/dashboard.templ:
- Replace 46-line inline sticky library selector (lines 20-66) with
@LibrarySwitcher(libData, currentLibraryID, DashboardActions())
- Dashboard-specific settings and refresh buttons extracted into the
DashboardActions sub-component via the variadic actions parameter
Update frontend route handlers for /collections and /collections/:id
to fetch user-visible libraries and pass libData + currentLibraryID
to templates, enabling the library switcher dropdown.
/collections handler:
- Fetch GetUserVisibleLibraries for the current user
- Derive currentLibraryID from query param, falling back to first library
- Convert to []templates.LibraryData and pass to Collection template
/collections/:id handler:
- Fetch GetUserVisibleLibraries alongside existing book fetching
- Pass libData to CollectionDetail template alongside existing libraryID
- Refactored to use shared libraryID variable across system/user paths
Add optional library_id query parameter support to GetCollections and
GetCollection API handlers for library-scoped book filtering.
GetCollections (GET /api/collections?library_id=X):
- When library_id is provided, include per-library book_count in the
response by querying GetCollectionItemsForDashboard for each collection
- When omitted, returns all collections as before (backward compatible)
- Added BookCount field to CollectionResponse struct
GetCollection (GET /api/collections/:id?library_id=X):
- System collections (non-empty QueryType): uses DashboardService to
fetch library-scoped sections, matching the existing SSR handler logic
- User collections: uses GetCollectionItemsForDashboard for
library-filtered results, excluding soft-deleted items
- When library_id is omitted, returns all books as before
Add reusable library switcher infrastructure that can be used across
dashboard, collections list, and collection detail pages.
New files:
- templates/library_switcher.templ: Shared LibrarySwitcher component
with variadic actions slot for page-specific buttons (e.g. dashboard
settings/refresh). Includes DashboardActions sub-component.
- web/src/library-switcher.ts: Shared module providing:
- initLibrarySwitcher(): syncs dropdown with localStorage, attaches
change listener with configurable onSwitch callback
- switchWithTransition(): generic fade-out -> spinner -> fetch ->
fade-in transition used by all pages
- getCurrentLibraryId(): reads "selectedLibrary" from localStorage
Modified:
- web/src/main.ts: import new library-switcher module
- web/src/types/api.d.ts: add book_count field to CollectionData
Tests were deleting the development admin user, causing ON DELETE SET NULL
to cascade and set created_by_admin_id to NULL on all libraries.
- test_helpers: skip deletion of testuser@tests.bookhoard.internal
- sync_integration_test: use test-sync% prefix for isolated test data
The uploads/ directory is bind-mounted at runtime via docker-compose and
should not be copied into the Docker build context. This was slowing down
builds and including potentially large media files in the context.
When a scan completes, dynamically update the dashboard carousels instead of
requiring a full page reload:
- Listen for bookhoard:scan-complete custom event dispatched by header
- Fetch updated sections from /api/dashboard/sections
- Diff existing book cards by data-media-item-id attribute
- Prepend new items to carousel tracks (afterbegin) to match API sort order
- Create entirely new section DOM for sections that don't yet exist on page
- Remove 'No items' placeholder when items are added
- Scroll carousel to left (scrollLeft=0) so newly prepended items are visible
Also:
- Extract renderSectionHTML() helper from renderDashboardCollections() for reuse
- Add data-media-item-id attribute to book card template for DOM diffing
- Add diagnostic console.log statements for debugging scan-complete flow
Add a scan progress indicator to the header that shows during library scans:
- Spinning SVG icon next to the BookHoard title
- Percentage display during active scans
- Dispatches bookhoard:scan-complete custom DOM event on window when scan
finishes, enabling other components (dashboard) to react without polling
- Auto-resets progress display after 3 seconds
- Uses WebSocket pub/sub via addListener/removeListener with cleanup on
header element removal
Replace the single-listener createWebSocket pattern with a pub/sub model
using addListener/removeListener. This allows multiple components (header
spinner, dashboard refresh) to subscribe to WebSocket messages independently
without clobbering each other's handlers.
- Maintain a Set of message listeners
- Auto-connect on first addListener, auto-disconnect when last listener removed
- Retain reconnect logic with configurable delay
The MessageTypeScanComplete constant existed but was never actually sent by
the worker. This meant the frontend had no way to know when a scan finished.
- After a JobTypeScan completes, broadcast scan_complete to the job's user
via WebSocket ConnectionManager
- Includes job_id, files_scanned, new_items, and errors in the payload
- Only broadcasts for JobTypeScan (not other job types) when connManager
is available and job.UserID is set
When an admin was deleted, the ON DELETE SET NULL foreign key would set
created_by_admin_id to NULL on all their libraries. This caused the scanner
to fail to find an admin ID for broadcasting scan-complete WebSocket messages.
- On admin deletion, reassign all libraries and media items to the next admin
- Prevents created_by_admin_id from ever being NULL on active libraries
- Uses new ReassignLibraries and ReassignMediaItems DB queries
AllowedExtensions in Go was the intended single source of truth for library
type file extensions, but it was never synced to the database. This caused
missing extensions like .pdf for manga to be absent from library_types.
- Add SyncAllowedExtensions() to sync Go AllowedExtensions map to DB
- Call SyncAllowedExtensions() from cmd/server/main.go on startup
- Ensure .pdf is included in manga extensions
The root cause of scanner failures in Podman containers was NOT that
inotify doesn't work through bind mounts (it does — same kernel, same
inodes). The real bug was SetFolders() only watching root directories.
Linux has no recursive inotify — every subdirectory must be added
individually to the watcher.
Changes:
- SetFolders() now walks all subdirectories and adds each to the watcher
(same approach as Audiobookshelf/Kavita)
- Remove broken mtime-based detection: seedDirectoryMtimes,
pollDirectoryChanges, detectChangedRoots, checkDirectoryMtimes,
SyncFilesystemWithDatabase — all unreliable in container overlay mounts
- Replace StartPolling with startBackupScan: enqueues full JobTypeScan
every 5 minutes (down from 30) as a safety-net fallback
- enqueueLibraryScan() sets job.UserID from admin ID so the worker can
broadcast WebSocket messages
- performInitialScan() sets job.UserID for the same reason
- Add [WATCHER] prefix logging to all fsnotify event loop messages
- Add defense-in-depth: fallback to GetFirstAdmin() when library has
no created_by_admin_id (NULL from test cleanup)
- Fix processDirectoryScanJob to use prefix-match (GetLibraryByFolderPathPrefix)
- Fix nil context panic: all jobs now set Context: context.Background()
- Remove mtime-related tests; update default interval test from 30m to 5m
The created_at column stores file modification time (intentional for preserving
original metadata), but this makes 'Recently Added' sorting unreliable for
imported files. Add imported_at column that records the actual database insert
timestamp.
Changes:
- Add imported_at TIMESTAMPTZ column to media_items (nullable)
- Update GetRecentlyAddedItems to sort by imported_at DESC NULLS LAST first
- Add ReassignLibraries and ReassignMediaItems queries for admin deletion handover
- Add SyncLibraryTypeExtensions query for startup extension sync
- Update all media_items SELECT queries to include imported_at column
Podman rootless containers with overlay storage do not propagate inotify
events through bind mounts, making the fsnotify file watcher ineffective.
This caused new files added on the host to go undetected until the
5-minute full-filesystem-walk polling fallback caught them.
Add a lightweight directory mtime polling mechanism that runs every 10
seconds, checking stat() on all subdirectories under watched library
folders against a cached mtime value. When a directory's mtime changes
(indicating files were added/removed/renamed), it feeds into the existing
markDirectoryDirty() → processDirtyDirectories() → job queue pipeline.
Changes:
- Add dirMtimes cache + mutex to MediaScanner struct
- Add seedDirectoryMtimes() to populate cache on startup (prevents
false-positive flood on first poll)
- Add pollDirectoryChanges() goroutine (10s ticker) and
checkDirectoryMtimes() (walks directories, compares mtimes)
- Launch mtime poller from WatchChanges() alongside existing goroutines
- Rename StartPolling logs to [ORPHAN-CLEANUP] to clarify its role
- Change default poll interval from 60s → 30m (new file detection now
handled by the fast mtime poll; full sync focuses on orphan cleanup)
- Update GetScanSettings default from 60 → 1800 seconds
- Add 5 tests: seed cache, skip nonexistent, detect new dir, skip
unchanged, detect modified dir
Expected result: new files detected in ~20 seconds (10s poll + 10s
debounce) regardless of inotify/container support.
Reverts generated Go template files from templ v0.3.1020 back to
v0.3.1001 output. Changes include filename path prefix adjustments
(admin_library.templ → templates/admin_library.templ) and attribute
handling differences (ResolveAttributeValue → JoinStringErrs + EscapeString).
Remove 185 binary files (169 CMaps + 16 standard fonts) from git
tracking. These are build artifacts copied from
node_modules/@bookhoard/foliate-js at build time and should not be
version-controlled.
Changes:
- Remove web/static/vendor/pdfjs/ from git (169 cmap files + 16
standard font files)
- Add web/static/vendor/ to .gitignore
- Drop CJK cmap copying from build scripts — the app is English-only
and CJK support can be re-added later if needed (saves ~1.7MB in
the container image)
- Update all three build scripts (build:ts, build:ts:dev,
build:ts:watch) to copy only standard_fonts/ from node_modules
- Remove cMapUrl from reader.ts PDF config since we no longer ship
cmaps
- Keep standardFontDataUrl pointing to the build-copied fonts which
are needed for PDFs with non-embedded standard fonts (Helvetica,
Times, Courier, etc.)
The postgres container was being healthchecked every 5s which is
aggressive for a database, especially on slower machines or under load.
The bookhoard service was checked every 10s.
- Increase postgres healthcheck interval from 5s to 30s
- Add start_period: 10s to postgres to give it time to initialize
before healthcheck failures count against retries
- Increase bookhoard healthcheck interval from 10s to 30s
templ v0.3.1020 generates different code than v0.3.1001 (uses
ResolveAttributeValue for attribute handling). Regenerated all 33
_templ.go files to match the new runtime.
Newer templ generates ResolveAttributeValue calls that don't exist in
v0.3.1001 runtime, causing Docker build failures ("undefined:
templ.ResolveAttributeValue"). Pin templ CLI version in Dockerfile to
match go.mod instead of using @latest.
Also updated local templ CLI to v0.3.1020 to match.
john-okeefe