Chapters are parsed in XML mode, which resolves only the five predefined XML entities and numeric
character references. EPUB2 XHTML may use any named HTML entity through its DTD, so ` ` and its
kin were indexed as literal text while kepubify wrote the decoded character. The two text indexes
then never matched, which disabled structural resolution for every annotation in the chapter and
failed each one with kepub_text_mismatch, even though the Kobo position was exact. Resolve the
references on the raw markup before parsing, where an escaped ` ` is still distinguishable
from a real reference, and read each one back through an attribute value so an unknown reference is
never partially matched against a legacy entity.
Rebuild canonical positions from the stored kobo_span position on position retry as well. Retry only
ran the cfi backfill, which derives cfi from an xpointer and therefore had no candidate when the
xpointer failed too, leaving an affected annotation pending forever with no way to recover from the
UI. Route the rebuild through a registry that the Kobo module registers into, since it already
imports the annotation module and the dependency cannot run the other way.
Fixes#944
The lockout was checked before the password was verified, so a correct password on a locked
account returned the same generic invalid-credentials error as a wrong one. Users had no way to
tell the two apart, and the natural response was to keep retrying a password that was already
right.
Completing a password reset did not clear the lock either, so the one recovery path the UI
advertises appeared to fail for up to fifteen minutes after a successful reset.
Verify the password first and report the lockout, with the remaining wait, only when the supplied
password was correct. Someone guessing wrong passwords still gets the generic failure, so account
existence is not leaked. Clear the lock whenever a password is reset or changed, and give admins
an explicit unlock action for the case where waiting is not acceptable.
* feat(metadata): add hardcover edition selection and refresh pinning
Mirrors the StoryGraph edition picker for Hardcover, but backed by real
GraphQL edition data instead of scraping. An edition can now be chosen
either from the Hardcover settings page's new "Linked books" section or
from the metadata downloader/refresh flow, and once chosen it pins
future automatic metadata refreshes to that exact edition instead of
silently re-deriving one by ISBN every time.
book_metadata.hardcover_edition_id is shared across all users, while
hardcover_book_state.hardcover_edition_id is scoped per (user, book).
Cross-writing between them is one-directional and non-destructive per
flow: the settings page always updates the acting user's own sync
state and only back-fills the shared field if it is empty; the
metadata downloader always updates the shared field and only updates
the acting user's own sync state if they already have one. The
propagation uses the existing achievement event bus rather than a
direct module import, since HardcoverModule already depends on
BookModule.
* fix(metadata): address hardcover edition selection review findings
Responds to CodeRabbit review on PR #936:
- Surface Hardcover edition-lookup failures as errors instead of an
empty list, so a provider outage cannot be mistaken for "no
editions". Updates the client to throw and the UI to show a
retryable error state.
- Validate that a manually-picked edition id actually belongs to the
book's matched Hardcover book before persisting it, rejecting
mismatches with a 400 instead of trusting an unvalidated client
input.
- Require an exact safe-integer string before parsing a propagated
hardcoverEditionId, so a value like "200abc" cannot be silently
truncated into a valid-looking edition id.
- Add durationMs to the edition-propagation failure log per the
project's logging conventions.
- Honor a pinned edition in the ISBN search() path too, not only in
lookupById(), so a refresh that falls back to search still returns
the previously chosen edition when it's present.
* test(metadata): strengthen hardcover edition-pin fallback assertions
The fallback tests for search() and lookupById() only used a
single-edition fixture, so they would have passed even if the
edition-pin filtering never ran. Use two-edition fixtures and assert
on which edition is actually selected instead.
* i18n(hardcover): use i18n keys for Linked books strings
HardcoverLinkedBooks.vue hardcoded its UI text instead of using
useI18n like its sibling components, breaking the ability to
translate the Hardcover settings page and non-English catalogs.
* fix(hardcover): harden edition selection and bound the linked-book list
Validate a picked edition against the book it belongs to rather than against a
capped, unordered list. Editions past the cap stay selectable, and two calls
can no longer disagree about which window they saw.
- Order the edition query by id and fetch one row past the cap, so truncation
is detectable instead of silent. Verified against the live Hardcover API:
both "The Name of the Wind" and "Pride and Prejudice" exceed the 50 cap.
- Report truncation for the linked-book list and the edition list, and say so
in the UI rather than presenting a partial list as complete.
- Bound and order the linked-book query. It backed a settings list with the
unbounded sync query; sync callers keep their previous unordered behaviour.
- Throw for "not connected" and "not matched" instead of returning an empty
list, which the UI could not tell apart from a genuinely empty catalogue.
- Drop release_year from the edition query. Hardcover never populates it, so
it was dead weight in the payload.
- Add a tuned --success token. The status pill hardcoded text-green-600, which
is untuned in dark mode; the light value is set for 4.5:1 on card and page.
- Expose row expansion with aria-expanded and aria-controls, announce load
failures through a live region, and keep the edition list open after a
switch so the refreshed marker lands where the user is looking.
---------
Co-authored-by: neon <18559658+neonsolstice@users.noreply.github.com>
A tab that slept past the 15 minute access-token lifetime woke up holding a token every consumer
then handed to the server independently. socket.io reconnected all seven namespaces with it, each
gateway rejected the handshake and closed the socket, and because a server-initiated close reports
`io server disconnect` the client never reconnects from it, so live updates stayed dead until the
page was reloaded. The dashboard widgets fetch once on mount with no retry, so whatever failed in
that window sat on "Failed to load" for the life of the page.
Make a valid token the thing consumers ask for rather than something they hope is still good.
Decode `exp` on the client and renew inside a 30 second skew before the token is used, so websocket
handshakes, uploads and `api()` requests no longer discover the expiry one 401 at a time. Route
every authenticated namespace through a shared factory that resolves the token per handshake and,
on an auth rejection, refreshes and reconnects with bounded exponential backoff.
Let the server say which rejections are worth retrying. Gateways now answer an authentication
failure with an `unauthorized` event before closing, and stay silent on a permission denial, which
a fresh token would not fix. Classifying that needed the gateways to stop throwing bare `Error` for
both cases, so the handshake paths now throw `UnauthorizedException` or `ForbiddenException` as the
project's conventions already require, and jsonwebtoken's own `TokenExpiredError` counts as an
authentication failure since it is the single most common way a slept-through handshake fails.
Collapse the twelve identical widget composables onto one loader that reloads a failed widget once
a rejected request is followed by a successful refresh.
Closes#942
Embedding metadata rewrites the book file, so its hash moves even though the book itself is
unchanged. Kobo reconcile reads a moved hash as new content and re-delivers the book as a
NewEntitlement, which makes the device rebuild its local entry and orphan every annotation
attached to the old one.
Carry the post-write hash into the device snapshots that recorded the pre-write hash, so
reconcile stays on the metadata-only path. The device keeps the bytes it already has and never
receives the rewritten file; a content change from any other source still triggers a download.
Ref #925
Normalize heterogeneous language metadata at the Kobo boundary and include validated ISBN values so device details remain accurate. Version metadata hashes to refresh previously synchronized books.
Fixes#920
Extract dominant color from book covers to apply ambient background tinting in BookDetailView, add cover tinting toggle to display settings, and test cover tint extraction.
* fix(kobo): stop offering plugin updates to devices that crash applying them
Plugin 1.3.x calls Device:unpackArchive, which KOReader removed in
koreader/koreader@751b4978 (July 2026). The call is unguarded and the whole apply path runs
without a pcall inside a ConfirmBox callback, so the missing method raises on UIManager's event
loop and takes KOReader down. Plugin 1.4.0 fixed the extraction, but the broken updater is the
code that would have to install that fix, so it can never self-deliver: every device still on
1.3.x crashes on every update attempt, forever.
Gate the offer server side instead. GET /koreader/plugin/version now resolves the calling user
and withholds pluginVersion when any of their devices last reported a plugin older than 1.4.0.
Those clients already treat "unknown" as "the server could not tell me" and show a message
rather than the update prompt, so the crashing path becomes unreachable without touching the
plugin on the device.
Two limits are inherent to a server-only mitigation and are not bugs. The endpoint carries no
device id, only the user, so one stale device parks self-update for all of that user's devices
until it is updated by hand or removed. A device that has never completed a sweep has no
recorded version, so it is not gated. Neither is closable until the plugin reports its own
version on the version check.
Surface requiresManualUpdate per device in the KOReader settings so the affected devices say
what to do, since the 1.3.x dialog string is baked into the installed plugin and cannot be
changed from here.
Closes#905
* fix(kobo): keep blank plugin versions inside the self-update gate
A device reporting an empty pluginVersion produced a falsy blocker, so the truthiness check
opened the gate and handed out the real version. The DTO allows an empty string, so this was
reachable. Also move the settings spec fixtures above the manual-update cutoff, since a
sub-1.4.0 version with requiresManualUpdate false describes a state that cannot occur.
Include completion timestamp in enrichment status payloads and track card dismissal in useMetadataCardDismissal to prevent finished notifications from reappearing across page navigations.
Closes#895
Stream ZIP archive entry rewrites sequentially in writeZipArchive, opening sources lazily to keep file descriptors and memory flat during file metadata operations.
Closes#893
Rewrite reading_state in Kobo initialization resources so device progress updates reach BookOrbit endpoints instead of defaulting to storeapi.kobo.com.
Closes#883
Extract fuzzy title and author matching into HardcoverImportFuzzyIndex with token indexing and yield pauses to prevent event loop lockups during large library imports.
Closes#881
* fix(metadata): split embedded subtitle out of long hardcover titles
Hardcover's subtitle field is legacy: it is fed by their data providers but
ignored and unmaintained, so librarians follow the convention of embedding
the subtitle in the title after a colon. Imports from search and from
editions without their own subtitle therefore stored the full
"Title: Subtitle" string as the title while the subtitle stayed empty.
Mirror Hardcover's own display rule, confirmed by their developers: split
the title on the first colon once it exceeds 60 characters. An API-provided
subtitle keeps precedence and leaves the title untouched, and the length
threshold protects short colon titles like "2001: A Space Odyssey" from
false splits.
Closes#867
* fix(metadata): trim hardcover titles that duplicate the api subtitle
Books like hardcover.app/books/quantitative-momentum carry the embedded
subtitle in the librarian-curated title and additionally expose the same
text through the legacy edition subtitle. The subtitle precedence rule
then skipped the split and the imported title kept the duplicated
subtitle. When the part after the first colon matches the API subtitle
(case-insensitive, trimmed) and the title is over the 60-character
threshold, trim the title and keep the API subtitle verbatim. Titles
whose subtitle genuinely differs stay untouched.
Ref #867
* refactor(metadata): tighten hardcover embedded-subtitle split helper
Narrow splitEmbeddedSubtitle to the non-optional title both call sites
already pass, which removes the unreachable `?? doc.title` fallbacks, and
rename the threshold constant so it no longer reads as an off-by-one:
titles up to 60 characters are left unsplit.
Treat a whitespace-only API subtitle as absent so it no longer blocks the
split, and cover two interactions that had none: the blank-subtitle case,
and a relevance query carrying the full combined title still ranking the
split candidate first.
---------
Co-authored-by: ohoeltke <hoeltke@oho-dt.de>
Co-authored-by: neon <18559658+neonsolstice@users.noreply.github.com>
Add index on books table for random dashboard scroller queries, batch dashboard scroller requests, instrument pg pool logging, and cover dashboard scroller batching in e2e tests.
Record how many books a series holds from Hardcover, ComicVine, RanobeDB,
Amazon and embedded ComicInfo metadata, storing it on the series row. Show
an ownership meter beside read progress on series detail, and extend gap
detection to report entries past the highest owned index instead of stopping
there. Allow the total to be set by hand in the book metadata editor.
Automated sources only ever raise the stored total so the result does not
depend on scan order; the manual field is the way to lower or clear it.
Closes#801
Kobo devices route every annotation request through us, including content BookOrbit does not
own, so a content id that resolves to no book is expected traffic rather than a sync failure.
Recording it surfaced a FAILED entry in the device sync history for each such request.
Keep returning 404: the device reconciles annotations by omission, so answering with an empty
set would make it delete its local annotations for that content.
Closes#831
Add a "Not on device" Browse tile listing recently added books not yet
downloaded here. On-device state is local to KOReader and has no
server-side filter, so the list is assembled by a bounded walk of the
newest books, capped by both a page budget and the server's ids limit.
It takes the slot of the Search tile, which duplicated the title bar
search button present on every non-detail page.
Add a folder picker to the bulk download settings, sharing one chooser
with the single-book dialog so both write the same download_dir.
Badge every Browse tile the server can count. Authors and series reuse
the cached totals the web sidebar already requests; libraries,
collections, smart scopes and in-progress get dedicated count queries
rather than the list methods behind the section endpoints, which
aggregate per-entity book counts a badge never reads. Badges are
right-aligned and abbreviated past 999 so the column stays fixed.
Replace the shared count path that paged and hydrated a book to read a
single integer, removing five queries per count on the dashboard, the
series read-through summary and the section endpoints.
Closes#869
Adds a server_fonts collection an administrator curates under Settings >
Admin, offered to every reader alongside their own uploads. Scoped CSS
family prefixes keep a user font and a server font of the same name
independent. Readers can hide individual server families from their own
picker via a new server-fonts user preference.
Validation, metadata parsing and storage are shared with per-user fonts
through an extracted upload pipeline rather than duplicated.
Address custom metadata fields as custom:<id> rules, mirroring the
existing custom sort fields, and resolve the value column from the
operator so buildWhere stays synchronous for every caller.
Closes#855
The catalog read-status enum omitted unread, so a device write could only
be undone from the web dashboard. Add it to the settable enum and the
plugin status dialog, and mark Unread as current when a book has no
status row.
Closes#853
Replaces the implementation from #818 with a job-backed flow. Moves are
planned up front and surfaced as a preview the user confirms, then run
through a planner/executor split against a book_move_jobs table, so
progress is trackable and partial failures roll back.
The dialog-based flow from #818 is superseded: its module, service, and
repository are removed here in favour of the job-based equivalents. Its
e2e coverage is replaced by a book-move suite that exercises preview
classification, the collision policies, file relocation, preserved user
state, and the books-before-book_files update ordering that the
composite folder foreign key constrains.
Also adds dashboard shelf layout configuration.
Ref #257
* feat(books): add moving books between libraries
Allow moving books to another library without re-uploading, so a staging
library (e.g. book-dock imports) can feed curated genre libraries. Re-parent
the existing book row instead of delete+recreate so reading progress,
annotations, collections, and Kobo sync state survive the move. Relocate
files by rebasing library-relative paths onto the target folder with
rollback on failure and a copy+delete fallback for cross-filesystem moves,
then schedule the target library's rename pattern. Exposed as a bulk
endpoint plus UI actions in the selection bar and book detail page.
Closes#257
* ci: allow container image dispatch from any branch
Enable building a GHCR image from a feature branch via workflow_dispatch
to test changes in a container before they land on main. The main tag is
still only updated by the CI-gated workflow_run path.
Ref #257
* fix(books): update book row before files when moving between libraries
book_files_book_folder_consistency_fk references books(id, library_folder_id)
with ON UPDATE CASCADE, so updating a file's library_folder_id while the book
still points at the source folder violates the constraint and the move fails
with a 500. Update the books row first so the new folder id cascades to the
files, then rewrite the file paths. Verified against a real Postgres instance
in addition to the reordered unit test expectations.
Ref #257
* test(books): add book move e2e suite
Unit tests mock the database and cannot catch constraint-order bugs like
the book_files_book_folder_consistency_fk violation fixed in b1c430df.
Cover the move endpoint against a real Postgres and filesystem: successful
cross-library move with row re-parenting, file relocation, and preserved
user state, plus disk-collision skip, foreign-folder rejection, and
format-restriction skip. Reuses the metadata-write e2e harness.
Ref #257
* fix(books): emit transfer event so library views refresh after a move
Moving books via the API updated the database and files but never told
connected clients, so the source and target library views kept stale book
lists and counts until a manual reload. Emit the existing book:transferred
websocket event (grouped per source library) after successful moves; the
library and book detail views already subscribe to it and refresh their
lists, counts, and buckets.
Ref #257
* fix(books): refresh views deterministically after moving books
The initiating library view only removed rows locally and relied on the
book:transferred websocket round-trip for counts, so deployments where
websockets are unavailable kept stale headers and sidebar badges until a
reload. Refresh the view's list and buckets directly after a successful
move, refresh the sidebar library counts on transfer events like scans
and uploads already do, and surface an error toast when a move is
confirmed with an empty selection instead of silently doing nothing.
Promote the move action from the overflow menu to a dedicated,
count-gated button in the selection action bar.
Ref #257
* fix(books): show move action for demo-restricted accounts like delete
The move button was additionally gated on the demo-restriction flag, so
accounts carrying the explicit demo_restricted permission saw the delete
button but no way to move books. Gate it on library_delete_books alone,
mirroring the delete action next to it; demo instances withhold that
permission anyway.
Ref #257
* style(server): align union type formatting with upstream main
The criss-cross merges of two upstream states left seven files with the
older prettier union-type layout, which fails lint against the current
formatter output. Restore the files byte-identical to upstream main.
* i18n(client): translate the move-to-library UI
The move action, dialog, and detail menu entries carried hardcoded
English strings from before the vue-i18n migration. Move them to
translation keys following the existing catalog structure and add
entries to all locale catalogs so the parity validator passes; Crowdin
can refine the non-English translations later.
Ref #257
* security(deps): update brace-expansion and engine.io past new advisories
Trivy fails the container build on CVE-2026-13149 (brace-expansion DoS)
and CVE-2026-59725 (engine.io DoS). Both fixes are semver-compatible
patch releases, so refreshing the lockfile resolves them without
manifest changes.
* security(deps): patch fastify static and find-my-way CVEs
Trivy blocks image builds on CVE-2026-15074 (route guard bypass in
fastify/static 9.x) and CVE-2026-47219 (HTTP2 DDoS in find-my-way
9.6.0). Bump the direct fastify/static dependency to 10.1.2 and add a
find-my-way override because nestjs/platform-fastify pins it to an
exact vulnerable version. The static plugin usage (register with root
and prefix plus sendFile fallback) is unchanged in v10; the peer range
warning from NestJS is expected until it allows v10.
* i18n(client): add BookBulkMove audit action label to all catalogs
Upstream's audit filtering feature derives its action options from the
AuditAction enum and requires a label for every member. Add the label
for the move feature's BookBulkMove action to all nine locale
catalogs.
* ci: restore main-only guard for container image dispatch
Revert the fork-local change that allowed workflow_dispatch container
builds from any branch. The guard matches upstream again so the branch
only carries the move feature. Fork builds continue from the
fork-build branch which keeps the relaxed dispatch.
This reverts commit b13d276a4becbba7be80723ff6708df23132b3fb.
* style(client): restore upstream formatting in untouched files
Merge conflict resolution left four files with stale prettier
formatting that differs from upstream while carrying no functional
change. Restore them byte-identical to upstream/main so the move
feature diff stays clean.
* fix(book-move): harden error paths raised in PR review
Address review feedback on bookorbit#818:
- Guard the applyMove call so a database error fails only that book
instead of aborting the batch, which discarded collected results and
skipped transfer events for books that had already moved
- Replace the bare catch around the source access check with one that
maps ForbiddenException to the access reason and logs unexpected
errors with a distinct failure reason
- Report an incomplete rollback in the outcome reason so callers know
a book may need manual verification
- Ignore backdrop and cancel clicks in the move dialog while a move is
running
- Catch thrown move requests in handleBulkMove, surface an error
toast, and keep the dialog open on transport errors so the user can
retry
- Wire onBooksMoved in SmartScopeView so scope totals and jump buckets
refresh deterministically after a move
- Extend the book-move e2e suite changedPaths with the library and
scanner modules and the metadata-write harness it depends on
* i18n(client): fill move keys with English source for Crowdin
Crowdin owns the target catalogs, so hand-written translations get
overwritten or conflict on the next export. Replace the translated
move dialog, menu, tooltip, and audit label values in all fifteen
target catalogs with the English source strings, keeping the
locale-specific ICU plural categories the parity validator requires.
Real translations happen in Crowdin.
* fix(book-move): harden access, filesystem, and audit behavior from review
Address the maintainer review on bookorbit#818, server side:
- Require editor-level access on the target library via a new
LibraryService.verifyUserAccessLevel backed by a repository lookup,
mirroring the LibraryAccessGuard ranking; viewer access could
previously write files into a library through this endpoint
- Skip books whose files are missing before touching the database
- Assert resolved target paths stay inside the target folder so a bad
relPath cannot escape the library via mkdir and rename
- Bound empty-directory cleanup at the library folder root and walk
nested directories upward so flat libraries keep their configured
folder and Author/Series trees do not orphan
- Bump books.updatedAt in applyMove so Kobo sync sees moved books, and
cancel the source library's pending unavailable-books notification
like the scanner's transfer path does
- Suppress file-watcher events for all touched paths through
SelfWriteRegistry while files move, closing the race against
missing-marking and duplicate detection
- Replace rename with link+unlink (COPYFILE_EXCL on cross-device
fallback) so concurrent moves resolving the same target fail with
EEXIST instead of silently clobbering each other
- Record the actual moved/skipped/failed outcome in the audit entry
instead of the requested count
* fix(book-move): use stable outcome codes and an accessible move dialog
Address the client-facing maintainer review items on bookorbit#818:
- Replace the free-form English outcome reasons with a stable code
union in MoveBookOutcome; the prose stays in server log lines and
the client maps codes to translated copy, so the UI no longer
renders raw server strings
- Localize the single-book move toasts on the detail page through the
new outcome keys (English source in all catalogs, translation via
Crowdin)
- Rebuild MoveBooksDialog on reka-ui DialogRoot/Content/Title like
ConfirmDialog, gaining role=dialog, focus trap and restore, Escape
handling, and aria wiring; closing stays blocked while a move runs
- Filter libraries without folders out of the target picker since
confirming one is a guaranteed 400
* test(book-move): assert database and disk state in no-op scenarios
The rejected and skipped e2e cases only checked the response before;
a partial move could have passed. Assert the 400 error message, the
unchanged libraryId and libraryFolderId, the source file remaining,
the destination staying absent, and that a colliding target file
keeps its original content.
* feat(book-move): stream bulk moves with per-book progress
The move endpoint processed an unbounded selection serially inside one
request, which cannot finish within proxy timeouts on target-scale
libraries and gave no progress. Convert it to the same SSE pattern as
bulk metadata refresh:
- The endpoint streams one event per book plus a final done summary
and stops processing when the client disconnects
- Target validation runs before headers are written so user errors
stay regular JSON error responses
- Audit is emitted manually with the actual outcome because the
interceptor never sees a response body on streaming handlers
- The bulk composable consumes the stream, reports progress through
the shared in-flight indicator, and removes moved books from the
view as their events arrive; the detail page reads the single
outcome from the stream
- The e2e suite parses the event stream and additionally asserts the
done summary and content type
* fix(book-move): explain skipped books in multi-library selections
Collection and smart-scope selections can span libraries, so the
books' own library legitimately appears as a move target and picking
it skips them with already_in_target. Surface that in the dialog with
a hint that is only shown when the view has no single current library
to filter out. Exact target filtering would need a library id on every
book card, which collapsed series cards cannot provide unambiguously.
---------
Co-authored-by: ohoeltke <hoeltke@oho-dt.de>
* fix(metadata): stop concatenating series and issue number into comicvine title
ComicVine search and auto-fill set title to a synthesized "Series #N -
Name" string instead of the issue's own name, inconsistent with how
ComicInfo.xml's <Title> is used elsewhere in the app (issue title only,
series and issue number in their own fields). title now carries just
the issue name, falling back to "<series> #<issue>" only when ComicVine
has no name for that issue.
The concatenated format is genuinely useful for telling issues apart
in the search-results list, so a new optional displayTitle field
preserves that exact UX there, populated only by the ComicVine mapper.
It's display-only and never read by any metadata-apply path, so the
list keeps its current look while the metadata that actually gets
persisted is cleanly separated.
filterAndRank now scores candidates against displayTitle when present,
falling back to title otherwise, so shortening title doesn't drop named
issues when users search by series name. Since displayTitle is only
ever set by the ComicVine mapper, this reproduces today's exact
ranking behavior for ComicVine and leaves every other provider
untouched.
Closes#841
* fix(metadata): leave comic titles unset when comicvine has no issue name
Address review feedback on the comicvine title mapping.
Issue #841 asked for the issue name alone, or nothing when comicvine has none for that issue.
Falling back to "<series> #<issue>" kept writing a synthesized title for unnamed issues, which
is the common case in a comic run, so the ComicInfo round trip only held for named issues. Make
MetadataCandidate.title optional and leave it unset instead: the pipeline already skips absent
fields, and the client already omits diff rows whose candidate value is empty, so an unnamed
issue now leaves any existing title alone. displayTitle still labels those candidates in the UI.
Comic titles no longer parse as "<series> #<issue>", which is how ComicVineProvider decided
between its precise volume-plus-issue search and the far looser general issue search. Carry
seriesName and seriesIndex on MetadataSearchParams and prefer them when the title does not
parse, so a refresh or a re-search of an already-matched comic keeps the structured path. A
title that does parse still wins, since that is a user typing an explicit target.
Also apply the display title in the diff panel's within-provider result picker, which renders
only when a provider returned several candidates and so needs the issue number most; match the
skip patterns against the same string that gets scored rather than title alone; and drop the
redundant cover alt text on the result card, where the headline already names the result.
---------
Co-authored-by: neon <18559658+neonsolstice@users.noreply.github.com>
Increase MAX_FONT_FILE_SIZE to 50MB to support large CJK and non-alphabetic fonts. Implement a lightweight binary font metadata parser to replace opentype.js buffer parsing, and update reader custom font composables to lazily manage blob URLs per font family.
Closes#793
Register Czech, Danish, Finnish, Hungarian, Korean, Russian, Simplified Chinese, Swedish, and
Ukrainian so the language picker, browser detection, and server preference validation accept them.
Catalogs hold English fallback values; Crowdin supplies translations through the normal export flow.
Give each catalog exactly the CLDR plural categories its locale defines rather than copying the
English branches. Czech, Russian, and Ukrainian gain `few` and `many` seeded from `other` because
the source has no matching text. Korean and Chinese drop `one`, which they have no rule for and
Crowdin cannot store. Those synthesized branches all start from the same English string, so machine
translation returns identical forms where the grammar differs and a human has to resolve them.
Ship Chinese as Simplified only under the bare `zh` locale so every zh-* browser resolves to it.
Traditional Chinese, also requested in #86, would arrive later as a separate zh-Hant catalog.
Use `sv-SE` for Swedish in the Crowdin config, because Crowdin has no bare `sv` language and would
have skipped the export. Tolerate a wrapped SUPPORTED_LOCALES array in validate-locales, which
prettier now produces since the list outgrew the print width, and extend the exhaustive ICU test to
reach Polish `other`, a category its counts never selected.
Add the repository activity embed to the README alongside the existing translation progress chart.
Repair the KOReader plugin source assertions that 40e366a4 left stale after reshaping the
dashboard streak block and the downloaded badge icon lookup, and watch `koreader-plugin/**` in
the server CI path filter so plugin-only changes run the server suite that reads those files.
Closes#86Closes#363Closes#576
* fix(metadata): extract and write back comicvine id for comics
ComicVine references were never read from a comic's embedded
ComicInfo.xml on import, even though every other provider id
(Goodreads, Amazon, Hardcover, Google Books, OpenLibrary, Kobo)
already round-trips through the same <Web>/Notes fields. This
affected Book Dock import, direct upload, and library scanning
alike, since they share the same extractor. Write-back had the
same gap: a comicvineId resolved via search/auto-fill was never
persisted back into the file.
Closes#840
* fix(metadata): keep comicvine id off the contested Web slot and source-check notes
ComicInfo has a single Web element and buildComicInfoXml writes the first
provider id in COMIC_INFO_PROVIDER_ID_KEYS order into it. The other six
providers are also in COMIC_INFO_MANAGED_NOTES_KEYS, so losing that slot costs
them nothing; comicvineId had no second channel, so a write-back for a book
that also carried a goodreads/amazon/hardcover/google/openlibrary/kobo id
dropped the id from the file entirely and clobbered any ComicVine URL that was
already there. Add comicvineId to the managed notes keys, which also makes the
existing managed-note read precedence reachable from our own writes.
ComicTagger emits "[Issue ID N]" for whichever talker produced the tag, naming
the source separately as "using info from <source>". Matching the bare marker
attributed Metron and other non-ComicVine issue ids to ComicVine and wrote them
back out as fabricated comicvine.gamespot.com URLs. Require the Comic Vine
source in the same sentence.
Also anchor the Web URL scan to a path segment so a segment ending in 4000-,
such as "14000-777", is no longer read as an issue id.
---------
Co-authored-by: neon <18559658+neonsolstice@users.noreply.github.com>
Add lastReadAt timestamp to KOReader catalog book items and introduce reading streak and goal statistics widgets to the KOReader plugin dashboard layout.
Extend book sorting to support custom metadata field references (custom:<id>). Join custom metadata values in BookSortBuilder based on field data type, resolve active field types in BookService, and expose custom fields in BookSortBuilder UI.
Closes#829
Books without an explicit user_book_status record are implicitly treated as unread across the app. Update BookQueryBuilder's readStatus filter handling so that querying for unread books also matches books lacking a status row.
Closes#815
Expose the "Visible to all users" toggle in the smart scope editor panel and allow updating the isPublic flag on existing smart scopes via the backend API.
Closes#805
* feat(koreader): add dashboard row configuration model
* feat(koreader): reuse catalog for dashboard source selection
* feat(koreader): configure dashboard rows on 1.4.0
* chore(koreader): preserve dashboard source line endings
* fix(koreader): keep BookOrbit tools menu loadable
* fix(koreader): keep dashboard requests out of rendering
* revert(koreader): restore stock 1.4.0 dashboard renderer
* revert(koreader): use stock single dashboard row
* fix(koreader): apply selected dashboard row source
* feat(koreader): add second dashboard row setting
* feat(koreader): render second dashboard row
* fix(koreader): fetch configured dashboard rows directly
* fix(koreader): honor configured dashboard row sources
* feat(koreader): support four configurable dashboard sections
* fix(koreader): use real catalog dashboard sources
* fix(koreader): select dashboard shelves from catalog
* fix(koreader): reuse catalog for dashboard source selection
* fix(koreader): close dashboard menu before catalog selection
* test(koreader): split dashboard configuration coverage
* chore(koreader): bump version to 1.5.0
* fix(koreader): address dashboard review feedback
* fix(koreader): address dashboard section review feedback
* test(koreader): cover dashboard behavior at runtime
* test(koreader): use Nest setup for dashboard DTO
* chore(ci): trigger pipeline
* fix(koreader): migrate untagged dashboard section settings
Plugin 1.4.x stored the configurable dashboard rows as a bare ordered
list with no schema marker, so gating the migration on schemaVersion == 1
never matched anything a device had actually written. Every upgrading
user who had ever picked a row was read positionally instead: the Stats
strip disappeared and Discover ended up in two slots, with a stored
SmartScope dropped entirely.
Treat an untagged stored list as the legacy shape and migrate it into the
slots those rows used to occupy. Cover it with the settings shapes 1.4.x
really wrote, rather than a marker nothing ever produced.
Also drop store() and primary(), which no caller uses now that slots are
written through storeAt(), along with the dead grow loop inside storeAt()
that normalize() already makes unreachable.
* fix(koreader): keep dashboard shelves in aspect, inside the page, and deduped
The four-slot renderer handed each shelf a height derived from the
leftover vertical budget while its cover cards kept the width computed
from the natural row, so any page with more than one shelf drew covers
out of aspect. It also had no answer for a page too small for the
configured slots: shelves clamped to a 48px floor and the layout simply
overflowed the available height.
Size shelves from their cards instead. A shelf that does not fit gets
more, narrower cards per row, so its height is always the height those
cards need. When even the narrowest row does not fit, drop slots the way
the pre-four-slot renderer did: the stats strip first, then shelves from
the bottom up.
Fetch each distinct shelf source once per dashboard load. Two slots
showing Discover previously rendered the same list twice, and two slots
on the same catalog filter each paid for their own request; a source that
fails is now also not retried by the next slot that wants it.
Drop what the four-slot rewrite left behind: the unused capabilities
require, the tall-layout helper and its constants, and the section
accessors and request helper that no longer have callers. Name the
per-shelf page size and the minimum cover width instead of repeating the
literals. Add setDashboardSections() so a whole configuration can be
applied in one refresh.
Add a layout spec that drives updateDashboardItems and asserts covers
stay in aspect, the drop order holds, and nothing is laid out past the
available height.
* fix(koreader): close the dashboard slot menu through its own API
Reset to Default wrote section_stale as a boolean where the renderer
reads a per-slot table, so the reset never actually marked anything
pending and stale books stayed on screen under the new headers until the
refresh landed. It also dereferenced current_context without a guard and
wrote the setting behind the catalog's back. Route it through the new
setDashboardSections() so the slots and the cache signature move
together in one refresh.
Close the invoking touch menu with closeMenu() rather than handing the
widget to UIManager:close(), which skips the close_callback that owns its
container.
Drop the section parameter from catalogDashboard() and the unused
catalogDashboardSection() helper: the plugin builds its shelves from the
ordinary catalog endpoints now and sends neither. The server keeps both
routes for devices still on 1.4.x, so its DTO test stays; rename it to
match the file it covers and drop the empty Nest module compile.
* revert(koreader): keep the plugin at 1.4.0
1.4.0 has not shipped yet: 1.3.0 is the last released plugin version, so
the dashboard slot work can land inside 1.4.0 rather than claiming a
version of its own. Devices in the field are on 1.3.0 and see a single
1.3.0 -> 1.4.0 step either way.
* fix(koreader): stop rotating upright read-status badges
buildReadStatusBadge turned every status icon a quarter turn, but only
two of the seven are corner folds. The dogear assets are drawn with their
right angle at the top left and the badge sits in the cover's top-right
corner, so those do need the turn. Bookmark, check and reload are upright
glyphs and were being drawn on their side: a want-to-read cover showed
its bookmark ribbon lying sideways like a chevron.
Rotate only the dogear icons.
---------
Co-authored-by: neon <18559658+neonsolstice@users.noreply.github.com>
The bundled binaries embedded golang.org/x/text v0.37.0, where a norm.Iter can enter an
infinite loop on input containing invalid UTF-8 bytes. Trivy flags this as HIGH and fails
the PR image scan on every branch opened since the advisory was published.
Rebuild all six binaries from the pinned upstream commit with x/text v0.40.0. The Go
toolchain was already current at 1.26.5 and is unchanged, so a toolchain-only rebuild would
not have cleared the finding; the vulnerable code lives in a module dependency. x/sync moves
to v0.22.0 as a side effect of the same resolution.
Keep the reported version string at v4-dev so rendition hashes stay stable and existing
cached kepub conversions remain valid.
Project manual book status changes onto Kobo reading states and mark device snapshot rows as unsynced. This ensures manual status updates made in the web UI reach registered Kobo devices on the next sync without overwriting device bookmarks.
Closes#812
Store per-user Kobo sync subscriptions for shared smart scopes in a new database table. This allows non-owner users to independently toggle Kobo sync on public smart scopes without modifying the owner's settings.
Closes#795
* i18n(client): add Spanish, French, Italian, and Polish locales
Add locale catalogs for Spanish, French, Italian, and Polish. Update locale types, Crowdin configuration, documentation, and test suites to support the expanded language options.
* i18n(client): sync Spanish, French, Italian, and Polish locale keys
Update Spanish, French, Italian, and Polish translation catalogs to include key additions for notification preferences, user management, and file naming settings.
* fix(client): correct expanded locale translations
Rework the app shell around a customizable sidebar: reorderable sections, sidebar zones, entity
popovers, and per-user preference persistence, backed by a dedicated browse counts endpoint so
entity totals stay cheap on large libraries.
Redesign the file naming settings with pattern fields, a token helper sheet, and a debounced live
preview. Rebuild the admin users table with server-side pagination, search, and filter DTOs, status
pills, confirmation dialogs, and magic link actions. Group notification preferences by category and
refine the account and privacy settings layouts.
Update theme tokens, icon references, and template event handler bindings across client features for
consistency, extend all locale catalogs, and add unit coverage for the new composables, components,
and server modules.
Add server support for writing metadata, embedded covers, and FictionBook XML structures directly into FB2 format files. Introduce migration 0058 to persist FB2 file write preferences per library, and update library creator UI and locale catalogs.
Add server support for writing metadata, covers, and EXTH records directly into MOBI/AZW format files. Introduce migration 0057 to persist Kindle file write preferences per library, and update the library creation UI and locale catalogs.
Add two-way bookmark sync support between KOReader plugin and server. Include migration 0056 for bookmark sync columns, server endpoints and services for bookmark exchange, and KOReader plugin bookmark management with comprehensive unit and E2E tests.
Add server endpoints and KOReader plugin settings to render configurable dashboard section rows, such as Smart Scopes or custom scrollers, replacing the hardcoded Discover section. Update the plugin capability checks to fall back gracefully on older server versions.
Lifecycle handlers previously blocked on network work, so closing or suspending a book stalled the
device on slow or offline links. Persist snapshots first and drain them through a lifecycle outbox
with connected-state gates and phase-aware retries, so the UI stays responsive and pending work
survives restarts.
Move persistence behind a state manager with atomic flushes, and chunk sweeps, bulk downloads, and
annotation exchange through a background queue with checkpoints and transfer policies, so large
libraries make progress incrementally instead of restarting from scratch after an interruption.
Server side, add bulk manifest and page stats endpoints with batched queries and catalog query
filtering so the plugin can reconcile many books per round trip rather than one request per book.
Ensure metadata score calculations are awaited when book metadata or covers are saved or updated, and add concurrency queuing for on-demand score recalculations.
Expand audit log entries to record detailed resource metadata (such as book titles on deletion), and add rich filtering by action, actor, and resource in the audit log viewer UI.
Closes#747
Allow users to adjust the gap between pages when reading comics in two-page spread mode. Include backend preference validation, shared settings defaults, translation keys, and reader settings UI controls.
Closes#740
Prefer the server account reading streak when provided by the dashboard endpoint, falling back to local KOReader statistics database data when offline or absent. This aligns the KOReader plugin day streak display with the web app.
Closes#711
Update dependency versions across root, server, and client package files and refresh the lockfile.
Upgrade @fastify/static to the patched major release and verify production static-serving compatibility.
neon
Dukko
Oliver
Gideon VanRiette
dependabot[bot]
Jadehawk
Chris Mansell