akai-tools/readest
2
0
Fork 0
forked from akai/readest
Code Pull Requests Activity
2,538 Commits 3 Branches 102 Tags
main
Commit Graph

2538 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Huang Xin c2ac207945 refactor(wordlens): rename "Word Wise" to "Word Lens" (#4633) 
"Word Wise" is a Kindle trademark, so rename the inline-gloss feature to
"Word Lens" throughout the product.

- User-facing strings → "Word Lens" across all 34 locales; brand translated
  for Chinese (zh-CN 单词透镜, zh-TW 單詞透鏡) and German (Word-Lens-Daten).
- Code identifiers: WordWise→WordLens, wordWise→wordLens, WORD_WISE→WORD_LENS.
- Files/dirs: src/services/wordwise→wordlens, WordWisePanel→WordLensPanel,
  wordwise{Ruby,Section}.ts, build/sync scripts, test dirs/fixtures,
  data/wordwise→data/wordlens.
- Storage paths: CDN base, R2 key, on-device cache dir, WORDLENS_R2_BUCKET env,
  pnpm wordlens:{manifest,sync}. manifest.json is path-agnostic so its
  sha256/bytes stay valid (verified).
- biome.json: point the formatter-ignore at data/wordlens so the generated
  one-line gloss packs aren't pretty-printed on commit.

Migration notes:
- Re-run `pnpm wordlens:sync` to upload packs to cdn.readest.com/wordlens/.
- Persisted view-settings keys renamed (wordWiseEnabled/Level/HintLang and
  wordWiseAutoDownload) — saved values reset to defaults once on upgrade.
- Cached packs under the old Data/wordwise/ orphan (harmless re-download).

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-17 19:09:29 +02:00
Huang Xin 9d0c5dc524 fix(koplugin): sync note deletions to Readest via tombstones (#4632) 
Deleting a highlight or bookmark in the koplugin never reached the
server. SyncAnnotations:push builds its payload from getAnnotations,
which walks only the live ui.annotation.annotations list, so a deleted
note can never appear in a push — the server kept the row and the next
pull resurrected it. The pull direction (server deletions → koplugin)
was already handled by removeDeletedAnnotations (#4119); this is the
missing push direction.

Capture a tombstone at deletion time instead. onAnnotationsModified
detects a removal (negative index_modified, with the deleted item at
items[1]) and records a deletedAt-stamped descriptor in the per-book
sidecar (readest_sync.deleted_notes). push folds those tombstones into
the payload and clears them only once the server accepts them, so a
failed push retries. Extracted buildNoteDescriptor so the deletion path
derives a note's id identically to the push walk.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-17 18:05:04 +02:00
Huang Xin 5e5564ef3f fix(search): show context for matches in italicized text (#4594) (#4631) 
Fulltext search showed only the matched word with no surrounding context
when the match fell inside inline-styled text (e.g. <i>/<em>). The root
cause and fix live in the foliate-js submodule's makeExcerpt
(readest/foliate-js#25); bump the submodule pointer to pick it up and add
a regression test covering both the simpleSearch and segmenterSearch paths.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-17 18:02:02 +02:00
Huang Xin b31699d052 fix(library): restore back button from series/author folders (#4437) (#4629) 
Inside a Series/Author library folder, the back arrow was a no-op after a
cold start. `GroupHeader.handleBack` deleted the `group` query param, leaving
an empty search string; `router.replace('/library')` with an empty search
silently no-ops under the Next.js 16.2 static export (every non-web build).
This is the same root cause as #3782, which was fixed for the breadcrumb
"All" button in #3832 — but the series/author back button never got the
workaround.

It only reproduces after a cold start, when `groupBy` comes from settings
(not the URL) and sort/order/view are at defaults, so `group` is the only
query param; that is why it could not be reproduced within a session.

Fix: set `group=''` instead of deleting it (mirroring
`handleLibraryNavigation`). The resulting `/library?group=` commits, and the
existing cleanup effect in page.tsx strips the trailing empty `group=`.

Verified on-device (Android, WebView 148, static export): tapping back inside
an author folder now returns to the main list.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-17 07:50:14 +02:00
Huang Xin 5861aead4b fix(android): move plugin @Command I/O off the main thread; fix WebView leak & #3297 blank screen (#4628) 
Backports the Android ANR/stability fixes from the julianshen fork,
adapted to upstream — notably preserving the cold-start shared-intent
replay queue and the #4559 dictionary-dispatch logic, neither of which
the fork kept.

- NativeBridgePlugin: run blocking @Command I/O (copy_uri_to_path,
  install_package, get_sys_fonts_list, and show_lookup_popover's
  queryIntentActivities) on Dispatchers.IO via a Main-dispatched
  pluginScope; startActivity hops back to Main; resolves are
  isActive-guarded; onDestroy cancels the scope and clears the static
  instance; the system-font scan is cached (@Volatile). The cold-start
  shared-intent queue (emitOrQueue / registerListener) is left intact.
- MediaPlaybackService: unmarshal the artwork Bitmap off the main thread
  (serviceScope + Dispatchers.Default), isActive-guarded; cancel the
  scope in onDestroy.
- ClipUrlController: hold the Activity via WeakReference and check
  isFinishing/isDestroyed before presenting, to avoid leaking the
  Activity/WebView during the up-to-30s clip window.
- MainActivity (#3297): on Android 14+ the window can gain focus before
  the WebView paints its first frame, leaving a blank screen. Force one
  repaint when both the window has focus and the WebView exists
  (whichever happens last).

Kotlin-only; not exercised by the JS/Rust test suites. Verified via
ktlint parse + a release `tauri android build` and on-device smoke test
(Xiaomi). The touch-event throttle and intent-handling rewrite from the
fork are intentionally NOT backported (they dropped touchmove forwarding
and the cold-start queue).

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-17 07:13:45 +02:00
Huang Xin fa120081a1 fix(library): never let a routine save shrink library.json (cold-start "Open with" wipe) (#4627) 
Opening a book via Android "Open with" on a cold start could clear the
entire library. `openTransient` built an ephemeral entry on top of the
not-yet-loaded (empty) store; the library page's `length > 0` cached-skip
then skipped loading the real library from disk; and a later
`saveLibraryBooks` persisted the empty/partial set — overwriting both
library.json and its .bak. Introduced by #4407 (transient "Open with"),
made reliably reproducible by #4527 (reliable cold-start delivery) — so
released v0.11.4, which lacks #4527, does not reproduce it.

Two layers of defense:

- saveLibraryBooks now merges with the on-disk library (union by hash,
  incoming wins), so a routine save is monotonic: it can add or modify
  rows (including `deletedAt` tombstones) but can never drop a book.
  Deliberate, authoritative rewrites (restore, tombstone GC, account
  reset) opt in via the new `{ replace: true }`. This layer alone makes
  the wipe impossible.
- openTransient loads the real library from disk before importing a
  transient book — also fixing the cold-start hash-match miss that
  re-imported already-imported books — and the library page's load-skip
  now gates on the store's `libraryLoaded` flag instead of `length > 0`.

Tests cover the merge floor (no-drop / no-wipe-on-empty / tombstone
preserved / incoming-wins) and both `{ replace: true }` paths.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-17 07:04:00 +02:00
Huang Xin d5c02e6253 feat(library): add Purge Data and fold detail actions into a More menu (#4615) (#4626) 
Resolves #4615. Re-importing updated serials left the app-generated
Books/<hash>/ folder (config.json reading progress/notes, nav.json, cover)
on disk after a normal delete, forcing a manual cleanup. "Purge Data" now
does a Cloud & Device delete AND wipes the whole directory in one action.

The book detail action row is redesigned to Edit · Download · Upload ·
Delete · More (hamburger):
- Goodreads, Share, and Export move into the hamburger "More" menu.
- Share is enabled only when signed in and the local file exists; Export
  is enabled when the local file exists (kept on every platform since the
  bottom-bar Send is mobile/macOS-only).
- Purge Data is the red entry in the Delete menu, behind a strong confirm.

Implementation:
- DeleteAction gains 'purge'; cloudService.deleteBook('purge') removes the
  in-place source file and removeDir's the whole Books/<hash>/ folder,
  clearing downloadedAt and leaving the tombstone + queued cloud delete to
  the page (mirrors 'both'/'local').
- The library page wires handleBookDelete('purge'); BookDetailModal adds
  the purge confirm config + share/export handlers and gates Share on auth.

Tests: cloud-service purge cases, BookDetailView More-menu + Purge tests.
i18n: 9 new keys translated across all 33 locales.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-17 06:28:21 +02:00
Huang Xin 5e217544f2 chore(config): add disableIncrementalCache to skip populating remote R2 incremental cache (#4623) 2026-06-17 03:51:17 +02:00
Huang Xin 6514d4aa58 build(web): standalone Docker image + drop Turbopack build cache (#4619) 
The Docker production-stage opts into Next.js `output: 'standalone'` via a
BUILD_STANDALONE env flag, so it ships only the traced runtime (server.js +
hoisted node_modules + static/public) and runs `node server.js` instead of
pnpm over the full source tree. The flag — and `outputFileTracingRoot`,
which traces from the monorepo root so workspace packages are included — is
set only in the Dockerfile build stage. Every other path keeps its original
output: Tauri `export`, local `build-web`, dev, and the Cloudflare/OpenNext
deploy (which forces standalone itself via NEXT_PRIVATE_STANDALONE).

Disable the experimental `turbopackFileSystemCacheForBuild`: a build
interrupted mid-compile leaves a partial cache that the next build
mishandles, fanning out workers until it exhausts host RAM. Remove the
pull-request CI step that cached `.next/cache` for it, now unused.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 21:21:24 +02:00
Huang Xin f6fbbf59f2 chore(deps): bump transitive deps for security advisories (batch) (#4620) 
Resolve a batch of transitive Dependabot alerts on the web lockfile via
pnpm-workspace overrides.

Bumped existing overrides:
- vite >=7.3.5 (#244 high, #245 med; GHSA path within 7.3.x) <- was 7.3.2
- dompurify >=3.4.9 (#249-#255; clears #252 <=3.4.6 by leaving the range)
- protobufjs >=7.6.3 <8 (#233, #247, #248); bounded <8 to stay on the
  patched 7.x line (a bare floor let pnpm jump to the 8.x major)
- ws >=8.21.0 (#241 high) <- was pinned 8.20.1

Added overrides:
- form-data >=4.0.6 (#246 high)
- js-yaml >=4.2.0 (#243 med)
- '@babel/core' >=7.29.6 (#242 low)
- '@opentelemetry/core' >=2.8.0 (#256 med)

Not auto-fixable here: #236 @ai-sdk/provider-utils (<=3.0.97, no recorded
fix; the installed 3.0.25 is pinned via a local patchedDependencies patch).

Verified: pnpm test (5685 passed), pnpm lint, pnpm build-web (exit 0).

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 21:19:36 +02:00
Huang Xin d6e59cedd7 chore(deps): bump esbuild to 0.28.1 and vitest to 4.1.x for security advisories (#4618) 
Resolve transitive Dependabot alerts on the web lockfile.

esbuild >= 0.28.1 (GHSA-gv7w-rqvm-qjhr #239, GHSA-g7r4-m6w7-qqqr #238):
Deno-module RCE via NPM_CONFIG_REGISTRY and Windows dev-server arbitrary
file read. Forced via a pnpm-workspace override (esbuild is a regular dep
of vite, so the override applies cleanly); bounded to <0.29 to stay on the
verified line. vite 7.3.3 declares ^0.27.0, but the 0.28 JS API is
unchanged for vite's usage -- verified by the full test run and web build.

@vitest/browser >= 4.1.8 (GHSA-g8mr-85jm-7xhm #240): Browser Mode CDP
bridge bypasses allowWrite/allowExec, enabling config overwrite -> RCE.
Bumped the vitest devDep family (vitest, @vitest/browser-*, coverage-v8)
from ^4.0.18 to ^4.1.8; resolves to 4.1.9.

Verified: pnpm test (5685 passed), pnpm lint, pnpm build-web (exit 0).

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 21:01:17 +02:00
Huang Xin d202d7a61e feat(library): add Clear Pending action to transfer queue (#4617) 
* feat(library): add Clear Pending action to transfer queue

Adds a "Clear Pending" button to the transfer queue panel that removes
only pending (including retry-pending) transfers, leaving in-progress,
completed, failed, and cancelled items intact. Wired through the store,
manager (with queue persistence), and useTransferQueue hook.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* i18n: translate Clear Pending and reading-statistics strings across locales

Adds translations for "Clear Pending" (transfer queue) and the two
reading-statistics sync-category strings ("Reading statistics" and its
description) across all 33 locales.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 20:02:11 +02:00
Huang Xin 480ab5b71e feat(hardcover): automatically sync progress and notes (#4614) 
Hardcover sync previously only ran when the user opened the reader menu
and tapped "Push Progress" / "Push Notes". Add an opt-in Auto Sync toggle
(default OFF) to the Hardcover settings so progress and notes are pushed
automatically while reading.

- useHardcoverSync: silent debounced (10s) auto-push of progress on page
  turns and of notes on annotation/excerpt changes, gated on
  enabled && autoSync === true; pending pushes flush on the existing
  sync-book-progress close event and cancel on unmount. Manual menu
  actions are unchanged (still loud).
- HardcoverSettings.autoSync flag (default false); existing connected
  users stay manual until they opt in.
- HardcoverForm: new "Auto Sync" toggle row.

Also backfills two untranslated strings surfaced by i18n:extract from the
reading-stats feature (#4606) across all locales, plus the new
"Auto Sync" key.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 19:25:01 +02:00
Huang Xin 757ed8066b feat(library): show series and number in list view (#4593) (#4612) 
In the library list view, surface each book's series and series number on
their own line, in addition to the description. Previously series info was
only visible by grouping by series or opening a book's details.

- Add `formatSeries(series, seriesIndex)` helper ("Series #N", trims the
  name, omits a zero/NaN/negative index) with unit tests.
- In list mode, render a dedicated single-line "Series #N" line above the
  description when the book has series metadata.
- Clamp every list line (incl. title) to one line and tighten the row gap
  to `gap-1` so the extra line fits the fixed-height row without clipping.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 17:07:27 +02:00
Huang Xin a30a310a17 fix(opds): handle entries with no downloadable format (#4599) (#4611) 
An OPDS entry with full metadata and a cover image but no acquisition
link — e.g. a Calibre book whose file was removed but kept for tracking
borrowed/loaned titles — was classified by foliate-js as a navigation
item whose href fell back to the cover image link. Tapping it loaded the
image, which is neither XML nor JSON, so the OPDS browser crashed with a
JSON parse error.

- Bump foliate-js to include the getFeed fix that classifies such
  metadata-only entries as publications instead of navigation.
- PublicationView: show "No downloadable format available" when an entry
  has no acquisition or stream links.
- loadOPDS: defense-in-depth — surface a clear message instead of a raw
  JSON.parse SyntaxError when a response is neither XML nor JSON.
- Add tests covering the Calibre no-format entry and a regression guard
  that a true navigation entry still classifies as navigation; add the
  two new UI strings across all locales.

Closes #4599

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 16:45:43 +02:00
Huang Xin f950685f22 chore(agent): update agent memories (#4610) 2026-06-16 15:52:29 +02:00
Huang Xin 35b02c4efc feat(statistics): KOReader-compatible reading stats with cross-device sync (#4606) 
Supersedes #3156. Adds a reading-statistics system whose canonical data model
is KOReader's own (book + page_stat_data), so stats round-trip losslessly
between Readest and KOReader.

- Storage: a cross-platform Turso statistics.db in KOReader's exact schema
  (web/Workers, desktop, iOS, Android) — replacing #3156's Node-only
  better-sqlite3 + statistics.json.
- Tracking: per-page reading events (time-on-page, idle-capped) flushed on
  page-change/idle/hide/close — the KOReader model — not session aggregates.
- Sync: legacy /api/sync extended with a stats type backed by self-contained
  Supabase tables (stat_books, stat_pages); union/longer-duration-wins merge
  keyed on book_hash. apps/readest.koplugin syncs through the same endpoint.
- Scale & robustness: per-tab singleton connection (avoids OPFS lock
  conflicts) + explicit WAL checkpoint; transactional bulk apply; chunked
  resumable push; client-driven paged pull with trailing-ms completion;
  paginated/scoped server merge.

Verified: 5668 unit tests, 155 koplugin busted tests, biome+tsgo + luacheck
all green; web OPFS DB verified live.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 10:42:39 +02:00
Huang Xin 359e406e51 docs(readme): make license badge static and modernize release badge (#4603) 
The license, release, and last-commit badges all query GitHub's API
through shields.io's shared instance, which intermittently fails with
"Unable to select next GitHub token from pool" when shields.io's own
token pool is rate-limited. A README author can't supply a token to the
hosted badges.

- License is fixed at AGPL-3.0, so use a static badge that makes no
  GitHub API call and can never hit the token-pool error.
- Switch the release badge from the deprecated `github/release`
  endpoint (which 301-redirects) to `github/v/release`.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 07:20:03 +02:00
Huang Xin f3c92f80d9 docs(readme): remove the Sponsors / TestMu AI section (#4604) 
Drop the Sponsors subsection and its lone TestMu AI logo from the
Support section.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 07:19:40 +02:00
Huang Xin e145eb835a feat(reader): open image gallery & table zoom on single tap (#4600) 
* feat(reader): open image gallery & table zoom on single tap

In reflowable EPUBs, a single tap on an image or table now opens the same
viewer a long-press opens, so the image gallery / table zoom is reachable by
both gestures. Fixed-layout books (PDF/comics/manga) keep tap-to-turn, and
long-press is unchanged everywhere.

Reuses the existing iframe-long-press -> handleImagePress/handleTablePress
flow via a new shared detectMediaTarget() helper (also adopted by the
long-press path so the two entry points can't drift). handleClick now takes
an isFixedLayout flag; the tap branch sits after the link/footnote/drag/
long-hold/Word-Wise guards so linked images still follow links and a
long-hold or double-tap won't double-trigger.

Context: #4584 (single taps stop registering after picture zoom on some
WebView builds) - this adds a second, independent way into the viewer rather
than fixing that root cause.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* refactor(reader): rename iframe-long-press message to iframe-open-media

The message is now posted for both a long-press (any book) and a single tap on
an image/table (reflowable books), so the long-press-specific name was
misleading. Rename the message type to `iframe-open-media` and the consumer
hook `useLongPressEvent` -> `useOpenMediaEvent`. The long-press detector
(`addLongPressListeners`/`handleLongPress`) keeps its name since it still
detects a long-press specifically.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 06:58:37 +02:00
loveheaven 675ee78bc9 perf(library): in-place re-import is a no-op on the same file path (#4597) 
Re-importing a folder via the in-place option used to reopen, parse,
and partial-MD5 every file before its byHash entry could short-circuit
the import. Worse, the byHash short-circuit treated every hit as "user
dropped a fresh file matching a known book", so it refreshed
createdAt/updatedAt/downloadedAt and cleared filePath/deletedAt. For a
user re-scanning the same external folder, that quietly rewrote sort
order and wiped soft-delete state. And once the import returned, the
ingest pipeline still ran group / tag / upload work — including a
path-derived empty groupId that silently clobbered manual
GroupingModal assignments on every re-scan.

This change adds an explicit byFilePath fast path at the top of
`ingestFile` so a re-scan returns the existing library entry verbatim,
before any I/O AND before any downstream side effect:

  byFilePath hit  -> the same on-disk source is being re-scanned in
                     place; the right answer is no-op. Don't open the
                     file, don't touch any timestamps, don't re-cover,
                     don't run the group / tag / upload steps.
  byHash hit      -> a different source path resolves to a known book
                     (e.g. the user dropped a copy from elsewhere, or
                     a soft-deleted book is being revived); the
                     existing "refresh timestamps + clear deletedAt"
                     behavior in importBook is correct here and is
                     left intact.

Implementation:

  - BookLookupIndex carries byHash / byMetaKey / byFilePath, with
    byFilePath built only from non-deleted books that have an absolute
    filePath. normalizeFilePathForIndex is the shared key function so
    shouldImportInPlace and the index agree on case-insensitive
    filesystems (macOS / iOS / Windows). osPlatform threads through
    buildBookLookupIndex and BaseAppService.importBook so the renderer
    and the importer compute the same key for the same file.

  - ingestService.ingestFile gains a byFilePath fast path before its
    importBook call: when `inPlace` was decided positive, the source
    is a real on-disk path (not a PSE stream / URL / content URI),
    and lookupIndex.byFilePath has a hit, return the existing Book
    directly. Returning here — rather than inside importBook — is
    deliberate: it skips the downstream group / tag / upload steps so
    a re-scan can never silently overwrite a manual GroupingModal
    assignment via a path-derived empty groupId.

  - ingest-service.test.ts covers both halves: an in-place re-import
    short-circuits importBook entirely (no call, existing object
    returned, createdAt / updatedAt / groupId / groupName all
    untouched); a copy-mode import (no external library folders) with
    the same byFilePath entry still goes through importBook so dedup
    falls back to byHash. import-metahash.test.ts retains the
    BookLookupIndex builder test that deleted and url-backed books
    are excluded from byFilePath.

Net effect: re-importing a folder of N already-imported books does
zero file opens, zero parses, zero MD5 passes, and leaves every book's
groupId / createdAt / deletedAt / cover untouched.
 2026-06-16 06:42:59 +02:00
Huang Xin bdfb595950 docs(readme): point donations to the unified donate.readest.com hub (#4601) 
donate.readest.com now lists every donation method, so the Support
section links there instead of enumerating GitHub Sponsors, Stripe,
and crypto separately.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-16 06:21:34 +02:00
Huang Xin 79496f88d7 feat(settings): move update & telemetry controls into Settings → Behavior (#4592) 
Relocate the update and telemetry toggles out of the library settings
menu into the Behavior (Control) panel, where global app settings live:

- New "Update" boxed-list (gated on hasUpdater): Check Updates on Start
  + Nightly Builds.
- New "Privacy" boxed-list: Help improve Readest (telemetry).
- Behavior section order: Update → Security → Privacy.
- Rename "Nightly Builds (Unstable)" → "Nightly Builds" and drop the
  "; may be unstable" note (the channel stays off by default).
- Updater dialog now shows the full version name (e.g.
  0.11.4-2026061506) instead of a parsed date.
- Extract + translate the new strings (Update, Privacy, Nightly Builds,
  Early daily builds) across all 33 locales.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-15 09:52:28 +02:00
Huang Xin 4908245042 feat(reader): Word Wise inline vocabulary hints (#4589) 
* feat(reader): Word Wise — inline native-language vocabulary hints

Kindle-style Word Wise: a short native-language gloss renders above difficult words
as you read (always-on ruby), gated by a CEFR vocabulary-level slider (A1–C2);
tapping a glossed word opens the existing dictionary.

- Pipeline: CEFR→frequency-rank difficulty, inflection-aware gloss index, pure
  offset-aware planner (EN regex + jieba for CJK).
- Rendering: <ruby cfi-skip>…<rt cfi-inert> injected per occurrence — CFI-transparent
  (verified), so highlights/bookmarks/progress are unaffected; kept out of TTS word
  offsets and find-in-book.
- Delivery: gloss packs are version-controlled in data/wordwise/, mirrored to R2, and
  downloaded on demand into local storage (sha-verified, single-flight) when enabled.
- Settings: a Word Wise sub-page under Settings → Language (enable, level, hint
  language, per-pack download/manage, auto-download toggle).
- Build tooling: scripts/build-wordwise-data.mjs (ECDICT / CC-CEDICT+HSK / WikDict +
  FrequencyWords, with lemmatization) and scripts/sync-wordwise-r2.mjs.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* data(wordwise): bundled gloss packs + manifest + attribution

13 frequency-trimmed gloss packs (en↔中文 + es/fr/de/pt/it/ru↔en, ~19 MB) generated
by build-wordwise-data.mjs from ECDICT (MIT), CC-CEDICT + HSK, and WikDict +
FrequencyWords (CC-BY-SA). Source of truth, mirrored to the CDN via `pnpm wordwise:sync`.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-15 07:56:00 +02:00
Huang Xin 51fede1a0d fix(rsvp): keep the audio toggle from overlapping transport on mobile (#4585) 
The read-along audio toggle + settings gear sat in an `absolute end-0`
cluster overlaid on the centered transport row. After the #3235 read-along
feature grew that cluster from a single gear (~36px) to ~81px (audio +
divider + gear), it covered the right end of the transport on narrow
phones, hiding the audio button behind the "skip forward 15" control.

Lay the playback controls out as a single full-width flex row: the audio
toggle moves to the far left and the settings gear stays far right,
symmetrically flanking the centered play button (justify-between on
mobile, justify-center on md+). Tighten the secondary buttons on mobile
(h-8, px-1.5) and add shrink-0 so the row fits without overlap; the
symmetry keeps the play button centered.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-14 21:18:07 +02:00
Huang Xin b76e3a3718 fix(nightly): publish latest.json via directory rclone copy (#4588) 
The assemble-manifest job promoted the manifest with single-file
`rclone copyto` + `moveto`. Before a single-file upload rclone issues a
CreateBucket probe (PUT /<bucket>), which the object-scoped RELEASE_R2_*
token can't satisfy -> 403 AccessDenied, so nightly/latest.json was
never published (the build legs and the stable release flow were fine
because they use a directory `rclone copy`, which PUTs the object
directly without that probe).

Mirror the release flow (upload-to-r2.yml): copy a one-file directory
into nightly/. R2 PutObject is atomic, so the .tmp + server-side move
added nothing. Verified against the live bucket with the current token:
directory copy -> 200 OK; single-file copyto -> CreateBucket 403.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-14 20:50:41 +02:00
Huang Xin aab721b219 feat(dictionary): lemmatize inflected words before lookup (#4574) (#4582) 
Dictionaries that store only base headwords (e.g. Oxford Dictionary of
English) miss inflected selections like `ran`, `mice`, `children`, or
`analyses` even though the lemma (`run`, `mouse`, `child`, `analysis`) is
present. Add a language-aware lemmatizer whose base-form candidates are
appended to the existing lookup candidate chain, after the exact/case
variants, so an exact/case match always wins and the lemma is only tried
once those miss.

- New pluggable `lemmatize/` registry keyed by primary language subtag;
  add a language by registering one lemmatizer, no caller changes.
- English lemmatizer: irregular-form table (suppletive verbs, irregular
  plurals/comparatives) + regular suffix rules (plural/past/gerund/
  comparative/possessive). Over-generates on purpose — the dictionary
  lookup is the validator, so bogus stems simply miss.
- Unknown/missing book language defaults to English (no-op on non-ASCII);
  an explicit non-English language with no registered lemmatizer is a
  no-op.
- Applies centrally to all definition providers (mdict/stardict/dict/slob
  and the online builtins) via `buildLookupCandidates`.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-14 18:41:39 +02:00
Huang Xin 131f83e15b fix(ci): correct nightly Linux AppImage collect path (#4581) 
The nightly Linux legs build with `cargo tauri build` WITHOUT `--target`
(no matrix `args`), so cargo emits bundles under `target/release/bundle/`
(host-target default) rather than `target/<triple>/release/bundle/`. The
macOS/Windows legs DO pass `--target`, so they legitimately get the triple
subdir — but the "collect artifacts" step reused `${rust_target}` for the
Linux AppImage path too, looking under
`target/x86_64-unknown-linux-gnu/release/bundle/appimage/` where nothing
exists. The build succeeded; only the collect step failed with
"missing artifact or signature for linux-x86_64-appimage".

Drop the `${rust_target}` subdir from the Linux AppImage path so it points
at the host-target default location where the bundle actually lands.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-14 18:10:54 +02:00
Huang Xin 0f0b4279a7 perf(reader): memoize global-annotation fan-out per section (#4575) (#4579) 
Highlighting recurring text (e.g. main-character names) as global
annotations made page turning very laggy. The `progress` effect
re-fans-out every global annotation across every rendered section on
EVERY page turn, and each pass re-walks the section DOM, recomputes
`view.getCFI()` for every occurrence, and tears down + recreates an SVG
overlay per match. The overlays already exist after the first pass, so
this is pure wasted work — profiled at ~25–45ms of synchronous
main-thread time per page turn for 6 names / 226 occurrences across 2
rendered chapters, multiplied on slower mobile hardware.

Memoize, per live section `Document`, which global notes have been
expanded (signature embeds `updatedAt`/style/color/text). Subsequent
page turns short-circuit to ~0ms. Keying on the `Document` makes
invalidation automatic: a re-rendered section gets a fresh document (and
fresh overlayer) so its overlays are rebuilt, while edits/recolors bump
`updatedAt` and toggling global off clears the memo.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-14 17:27:52 +02:00
Huang Xin 57501cc520 feat(updater): nightly update channel (Android/Windows/macOS/Linux) (#4577) 2026-06-14 16:33:53 +08:00
Huang Xin bfb85c2f68 feat(reader): sync paragraph mode & speed reader with TTS read-along (#3235) (#4576) 
* docs(reader): TTS-sync design spec for paragraph mode + RSVP (#3235)

Hardened via brainstorming + /autoplan (CEO/Design/Eng dual-voice review).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(tts): emit canonical tts-position event from TTSController (#3235)

Controller emits { cfi, kind, sectionIndex, sequence } alongside the existing
tts-highlight-mark/-word events. Monotonic sequence lets downstream consumers
(paragraph mode, RSVP — later slices) drop out-of-order positions. Additive;
existing events untouched.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(reader): containment+cursor CFI->index mappers for TTS sync (#3235)

RSVPController.syncToCfi + setExternallyDriven: containment match (fixes
mid-token skip), monotonic cursor + binary search (avoids O(N)-per-word jank,
no per-word getCFI), -1/no-op on no match (no silent jump to word 0), timer
suspension while externally driven.

ParagraphIterator.findIndexByRange: hinted + binary-search containment mapper
returning -1 on no match (never first()).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(tts): forward tts-position + tts-playback-state onto the app bus (#3235)

useTTSControl republishes the controller's canonical tts-position (tagged with
bookKey) via a dedicated listener — NOT inside the suppression-gated highlight
handlers, so page-follow suppression can't silently desync the modes. Adds
tts-playback-state (playing/paused/stopped) so RSVP can track playback without
the hook-local isPlaying. Verified by extending the real-foliate-view browser
harness.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(reader): paragraph mode follows TTS playback (#3235)

When paragraph mode + TTS are both active, the focused paragraph follows the
spoken position (sentence granularity, all engines). Section-generation contract
(stash cross-section position, apply after the iterator re-inits); sync-focus
path that does NOT arm isFocusingRef (avoids the relocate-eaten wrong-section
paragraph-0 bug); stale-sequence drop; decouple on manual nav, re-engage on next
playing. Start-alignment + visible indicator deferred to later slices.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(rsvp): speed reader follows TTS playback (#3235)

Edge word-boundary voices: RSVP shows the spoken word via syncToCfi. Non-Edge
(sentence-only) voices: sentence-paced estimator (clamp 60..600 wpm from voice
rate, hold at +60 words cap, snap to first word on each new sentence mark).
RSVP auto-advance suspended while TTS-driven. Decouple on manual nav via a
rsvp-manual-nav signal; re-engage on next playing. Cross-section positions
re-extract then apply. Pure decideRsvpTtsPosition helper unit-tested.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(reader): fixed-layout gate + ttsSyncStatus for TTS sync (#3235)

Gate sync to reflowable books (D7): fixed-layout reports 'unsupported' and
never engages. Both modes expose ttsSyncStatus (idle/following/syncing/
decoupled/unsupported) as the data source for the upcoming indicator. RSVPControl
now forwardRef-exposes the status via an imperative handle. Cross-bookKey events
ignored (regression-tested).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(reader): 'following audio' indicator for TTS sync (#3235)

5-state pill (following/syncing/decoupled, idle+unsupported render null) shown
top-center in the paragraph overlay and as a status row in the RSVP overlay.
Decoupled state is the tap-to-resume control; first decouple fires a one-time
toast. eink-bordered, glyph+text (no color-only), RTL logical props, touch
targets, safe-area top inset. RSVP 'plain' variant matches its themed surface;
non-Edge shows '· estimated'. New i18n keys need extraction before merge.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(rsvp): in-overlay TTS toggle + audio-paced speed control (#3235)

Voice-glyph audio toggle in the RSVP control row (trailing, by the gear) starts/
stops read-along from inside the full-screen overlay, start-aligned to the current
word (range validated against the live doc). While TTS-driven, the WPM control
shows a locked 'Audio pace' affordance that opens a compact rate picker; rate
changes go through a new tts-set-rate bus event reusing the existing throttled
setRate path. Pure buildRsvpTtsSpeakDetail helper unit-tested.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* test(reader): e2e paragraph mode follows TTS across a section boundary (#3235)

Real <foliate-view> browser e2e: with paragraph mode active, the focused
paragraph follows the TTS walk and re-targets to the new section after a Ch4->Ch5
boundary (proves no stuck wrong-section paragraph-0 / isFocusingRef trap).
Asserts on the owning section of the current range. Test-only.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* i18n(reader): translate TTS-sync strings across 33 locales (#3235)

Following audio / · estimated / Resume audio / Stopped following audio /
Play audio / Pause audio / Audio pace / Speed follows audio.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(reader): resolve TTS CFI anchors across iframe realms (#3235)

RSVP and paragraph follow silently failed to track the spoken word: the CFI
anchor from view.resolveCFI(...).anchor(doc) is a Range created in the book
iframe's realm, so 'anchor instanceof Range' (top realm) was always false
(cross-realm instanceof) -> resolveCfiToRange/applySyncCfi returned null ->
syncToCfi never advanced. Add isRangeLike() duck-type (cloneRange is unique to
Range) and use it at all 4 CFI-resolution sites. Confirmed live via CDP: before
= syncToCfi false (frozen); after = exact word map + RSVP follows Edge TTS at
~171 wpm (audio pace). Unit tests reproduce the cross-realm anchor (jsdom is
single-realm so the old code passed there but died in the app).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(rsvp): stop estimator/word fight + map transport to TTS play/pause (#3235)

Two read-along refinements (verified live via CDP with Edge TTS):

1. No more jump-ahead-then-snap-back flashing. Word-boundary engines (Edge) emit
   BOTH sentence marks and word boundaries; RSVP was routing sentence -> the
   estimator (self-paces ~190xrate, up to +60 words ahead) while word positions
   snapped it back. Now once a word position is seen, sentence positions are
   ignored and any running estimator is stopped, so words alone drive RSVP.

2. The RSVP transport (center play/pause, Space, center-tap) maps to TTS
   play/pause while read-along is engaged (tts-toggle-play), instead of RSVP's
   own suspended timer. Pausing TTS keeps RSVP suspended (no runaway); a full
   stop releases it.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(rsvp): keep indicator on pause + reach dict management from RSVP (#3235)

- Pausing read-along no longer dismisses the 'following audio' indicator / 'Audio
  pace' lock (layout shift). New 'paused' sync status keeps the indicator row and
  WPM lock present while TTS is engaged-but-paused; only a full stop clears them.
  Verified live via CDP: pause keeps the layout, no shift.
- Dict management is reachable from RSVP: the settings dialog is z-50, far below
  the full-screen RSVP overlay (z-[10000]), so it opened invisibly behind it.
  handleManageDictionary now exits RSVP first (position saved/resumable) so
  management shows over the reader.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(rsvp): show dict management over RSVP instead of exiting it (#3235)

Per feedback: opening dictionary management from the RSVP lookup popup no longer
closes the speed reader. The settings dialog is raised above the full-screen RSVP
overlay (z-[10000] -> SettingsDialog !z-[10050]) so it shows on top, and RSVP's
capture-phase keyboard handler bails while the settings dialog is open so its
inputs accept Space and Escape closes settings (not RSVP). Verified live via CDP:
management opens over RSVP, RSVP stays active behind it, Escape returns to it.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(dict): only apply drag-handle margin compensation when the handle shows (#3235)

The dictionary sheet header used -mt-4 to compensate for Dialog's drag handle,
but that handle is sm:hidden (shown only below sm). On sm+ the handle is
display:none, so -mt-4 pulled the header up into the top edge (broken layout
when the lookup renders as a sheet on a short/wide window). Mirror the handle's
breakpoint: -mt-4 sm:mt-0. Verified live via CDP.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(reader): in-mode TTS audio toggle for paragraph mode (#3235)

Paragraph mode already follows TTS, but there was no way to start read-along
from inside it. Add an audio toggle to the ParagraphBar (mirroring RSVP's): it
starts TTS start-aligned to the focused paragraph (range validated live, +
section index) and stops it. Track session-active vs playing so a pause keeps
the indicator ('paused' status) instead of collapsing to idle. Pure
buildParagraphTtsSpeakDetail helper unit-tested. Verified live via CDP: tapping
the icon starts audio from the focused paragraph and the focus follows speech.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(reader): highlight current TTS word/sentence in paragraph mode (#3235)

Paragraph mode follows TTS by advancing the focused paragraph, but the spoken
word wasn't highlighted within it like normal mode. The overlay renders a CLONE
of the paragraph, so the iframe's TTS highlight isn't visible there — reproduce
it on the clone with the CSS Custom Highlight API (no DOM mutation, spans inline
boundaries natively, leaves the fade-in animation untouched).

- TTSController already tags tts-position with kind word|sentence. The hook
  decides granularity: word boundaries (Edge) drive a per-word highlight; once
  seen, the coarse sentence event is skipped so the whole sentence doesn't
  flicker over the current word. Engines without word boundaries
  (WebSpeech/Native) fall back to the sentence highlight.
- Offsets are computed relative to the paragraph start (so they map 1:1 onto the
  clone's text) and tagged with the paragraph index so a stale highlight never
  paints the wrong paragraph. Cleared on stop / section change / disabled.
- The ::highlight() style mirrors the user's ttsHighlightOptions color+style.

Pure helpers (offset math, word/sentence decision, css builder) unit-tested.
Verified live via CDP: word highlight tracks Edge word-by-word and follows
across paragraph boundaries (news -> ... -> ladies), matching the TTS color.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-14 09:46:18 +02:00
Huang Xin cc618b8739 test(tts): add browser e2e for auto-advance across a chapter boundary (#4573) 
Mounts the real foliate <foliate-view> with sample-alice.epub, renders
the real useTTSControl hook with the real stores, and mocks only the
speech client. Starts TTS at the last paragraph of chapter 4 and
verifies the reading auto-advances into chapter 5, the page turns, and
the "Back to TTS Location" badge never appears (the TTS location stays
in view).

The mock client's speak() only needs to yield `end` — the real
TTSController drives forward() and the real view.tts walks the document
across the section boundary, so the cross-chapter navigation and badge
suppression are genuinely exercised rather than re-implemented in the
test.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-13 20:46:26 +02:00
Huang Xin 5a8f0873fa fix(library): refresh book cover after editing metadata (#4572) 
* fix(library): refresh book cover after editing metadata

Editing a book's cover in Book Details and saving showed the old cover
until a full reload, in two render paths:

- Library grid: handleUpdateMetadata mutated the book object in place,
  so the memoized <BookCover> compared fields off the same (mutated)
  reference and skipped re-rendering. Build a new book object via the
  new getBookWithUpdatedMetadata helper instead of mutating.
- Book Details view: BookDetailView renders cover/title/author from the
  modal's `book` prop, which the parent never re-passed after save.
  BookDetailModal now tracks the saved book locally (displayBook) and
  renders the view from it.

Adds a unit test for the immutable helper, a BookDetailModal regression
test (edit cover -> save -> view reflects it), and a sample-alice.txt
fixture for TXT import testing.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* chore(agent): add cover-refresh stale-render memory

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-06-13 19:56:25 +02:00
Huang Xin 4b0bbc77b0 fix(reader): open TXT files shared via "Open with" (#4571) 
* fix(reader): open TXT files shared via "Open with" by converting to EPUB

The Android "Open with Readest" (VIEW intent) transient path hands the
reader the original .txt file (its filePath points at the content:// URI),
unlike the managed library which stores the already-converted EPUB. The
DocumentLoader had no branch for a raw .txt, so open() returned
{ book: null } and initViewState crashed with
"TypeError: Cannot read properties of null (reading 'metadata')",
leaving the user stuck on the library splash.

Add an isTxt() check that converts the raw .txt to EPUB in-memory (the
same TxtToEpubConverter the import path runs) and parses that. The
converter emits a .epub-named file, so the importer's own
DocumentLoader.open() on the converted file is unaffected.

Verified on-device (emulator, warm + cold start): the TXT now opens and
renders in the reader instead of crashing.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(settings): allow adjusting highlight opacity in e-ink mode

Drop the isEink prop that disabled the highlight Opacity slider under
e-ink. Opacity is still meaningful on e-ink, so let users change it.
Removes the prop from HighlightColorsEditor, its ColorPanel call site,
and the test render helper.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test(send): mock fetch to fix flaky article conversion test

The article/page conversion paths fetch a favicon + author image for the
synthetic cover via globalThis.fetch. In jsdom that hit the real network:
a live fetch to the sample URL can hang up to faviconFetcher's 6s timeout,
exceeding the 5s test timeout and intermittently failing the suite. Stub
fetch so the cover falls back to its initial-letter tile (the pattern other
tests in this suite already use). Article test: ~5003ms hang -> ~80ms.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* chore(agent): update annotation-share-toolbar memory

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-06-13 18:07:28 +02:00
Huang Xin 67c22c770b feat(reader): Share intent + customizable annotation toolbar (#4014) (#4570) 
* docs(spec): annotation Share tool + customizable toolbar (#4014)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* docs(plan): implementation plan for Share tool + customizable toolbar (#4014)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(annotator): add 'share' annotation tool type and button (#4014)

* feat(annotator): add pure toolbar order/visibility helpers (#4014)

* feat(annotator): add annotationToolbarItems view setting (#4014)

* feat(annotator): add shareSelectedText ladder helper (#4014)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* feat(annotator): render Share tool and honor toolbar order in selection popup (#4014)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* feat(settings): add drag-and-drop annotation toolbar customizer (#4014)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* feat(settings): open the toolbar customizer from the Behavior panel (#4014)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* chore(i18n): extract and translate annotation share/toolbar strings (#4014)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* refactor(annotator): extract canShareText helper, preserve hidden Share on cross-platform edit (#4014)

Addresses final-review findings: de-duplicate the triplicated canShare
definition into share.ts::canShareText, trim ShareCapableService to the
fields actually read, and stop the toolbar customizer from dropping a
synced 'share' tool when edited on a non-share-capable device.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(settings): WYSIWYG drag-and-drop toolbar customizer (#4014)

Rework the customizer per live testing:
- Render 'In toolbar' as a faithful, content-width, start-aligned preview of
  the real selection popup (gray bar, icon-only buttons); 'Available' tools
  show as labeled chips.
- Multi-container dnd-kit pattern: in-place dragging (no DragOverlay, which a
  transformed modal offsets), pointerWithin collision so empty zones accept
  drops, live onDragOver reparent, itemsRef to dodge dnd-kit's drag-start
  handler-capture stale closure.
- Add 'Add all' (canonical predefined order) and 'Clear all' shortcuts.
- Align zone labels with the SubPageHeader breadcrumb.
- Empty toolbar now suppresses the selection popup entirely (no empty bar),
  while still allowing highlight-edit/notes popups.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* chore(i18n): translate Add all / Clear all toolbar shortcuts (#4014)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(annotator): size selection popup to visible tool count (#4014)

With the customizable toolbar a fixed-width popup looked sparse for a 2-3
tool toolbar (buttons spread to the corners). Size the popup to the number
of visible tools (responsive) capped at the previous max; annotated
selections keep the max width since they show highlight options / notes.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(settings): reword empty-toolbar hint to 'No tools, drag one here' (#4014)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* test(annotator): render default tools (not Share) in popup layout screenshot (#4014)

The visual regression test rendered every annotationToolButtons entry, so
adding the Share tool shifted the toolbar to 9 buttons and broke the
baselines. Share is hidden by default (added via Customize Toolbar), so the
popup screenshot should mirror the default-enabled set — filter to
DEFAULT_ANNOTATION_TOOLBAR_ITEMS, keeping the existing baselines valid.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-13 14:11:59 +02:00
Huang Xin b6937f43f1 chore(agent): stage memories (#4569) 2026-06-13 08:57:21 +02:00
Huang Xin a56cc6c61a feat(tts): word-by-word highlighting for Edge TTS, closes #4017 (#4566) 
Highlight each word as it is spoken (Edge TTS only) instead of keeping the
whole sentence highlighted, and keep the view tracking the spoken word
across page boundaries.

Word boundaries
- Capture Edge's audio.metadata WordBoundary frames (offset/duration in
  100ns ticks plus the verbatim input-text span) in the Tauri, browser, and
  Cloudflare-Workers WebSocket transports.
- Carry boundaries through the authenticated HTTPS proxy route via an
  X-TTS-Word-Boundaries response header (percent-encoded JSON, ASCII-safe),
  so word highlighting works on the web where the browser cannot open the
  wss connection directly. Cache them alongside the audio blob URL.

Highlighting
- Sync a requestAnimationFrame loop to audio.currentTime against the
  boundary table and highlight the word sub-range within the spoken
  sentence. Synthesis stays sentence-level (natural prosody); only the
  visual highlight is word-level.
- Suppress the sentence highlight when the active client reports word
  boundaries and draw the first word immediately, so the whole sentence
  never flashes before the first word. Fall back to the sentence highlight
  when a chunk has no boundaries (other engines, empty metadata).
- Re-apply the current word (not the sentence) when the view relocates.

Page following
- Turn the page as soon as the spoken word crosses a page boundary (a
  tts-highlight-word event scrolls only when the word is outside the visible
  range), instead of waiting for the next sentence.
- Check the word's position for the "back to TTS location" badge so it no
  longer appears while the view follows the word onto the next page.

Also fixes a pre-existing bug where the browser WebSocket was constructed
with an options object (valid only for the Node ws package), which threw in
browsers and made the wss path unusable on the web.

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-06-13 08:29:34 +02:00
Huang Xin 763b579c8f fix(android): launch installed dictionary for system lookup, closes #4559 (#4568) 
On targetSdk 36, ACTION_PROCESS_TEXT handlers were hidden by Android 11+
package-visibility filtering — only auto-visible web browsers resolved the
intent, so system-dictionary lookups landed in the OEM browser (VIVO/iQOO)
even with a dictionary like Eudic installed. Add a <queries> declaration so
dictionary apps are visible, and filter web browsers out of the handler set
so an OEM browser that registers PROCESS_TEXT can't swallow the lookup:

- no browser among handlers → unchanged implicit dispatch (keeps native Always)
- browser + one dictionary → launch it directly (explicit component)
- browser + several dictionaries → chooser excluding browsers, remembering the
  pick via EXTRA_CHOSEN_COMPONENT so later lookups go straight through
- only a browser installed → report unavailable instead of opening it

Routing is a pure, JUnit-tested decideLookupDispatch(). Adds get/clear
lookup-dictionary commands + an Android-only reset row in the dictionary
settings to switch the remembered app.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-13 04:36:31 +02:00
Huang Xin c72afe269a fix(tts): keep voice list stable across region variants of a language, closes #4033 (#4565) 
The voice panel filtered voices by the full locale of the currently
speaking text (v.lang.startsWith(locale)), so a book mixing region
variants of one language flip-flopped its voice list: Standard Ebooks
tag their boilerplate front matter en-US (17 Edge voices) while the
body text is en-GB (5 Edge voices).

Filter by primary language instead (isSameLang) in all three TTS
clients so every English variant yields the same voice set, and sort
voices matching the requested locale first
(TTSUtils.sortVoicesPreferLocaleFunc) so default-voice resolution via
getVoiceIdFromLang still picks an exact-locale voice. This also fixes
languages whose tags never matched a voice locale prefix at all (e.g.
zh-Hans books previously got an empty Edge voice list).

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-06-13 04:32:27 +02:00
Huang Xin 852d0ae3e9 fix(reader): keep dark-mode page body transparent so the bg texture shows, closes #4446 (#4564) 
The body.theme-dark catch-all from #4392 painted every section iframe's
body with the opaque theme bg in dark mode, occluding the host
background texture and poisoning foliate's docBackground capture (so
paginated segments and scrolled view backgrounds resolved opaque too).
Force transparent instead: the dark page fill already comes from the
paginator container / reader grid cell, and book-forced light page
backgrounds stay neutralized since the theme-dark fill shows through.
Unconditional rather than texture-gated because docBackground is
captured once per section load and a gated rule would go stale on live
texture toggling.

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-06-12 19:37:23 +02:00
loveheaven 7f57af8f90 perf(cfi): bucket booknotes per chapter and batch-collapse location matcher (#4561) 
* perf(cfi): bucket booknotes per chapter and batch-collapse location matcher

When iterating a list of CFIs against the same currentLocation (Annotator
on every page turn, useSearchNav, useBooknotesNav), the standalone
isCfiInLocation collapses the location twice per CFI. With 1000+
booknotes -- which a heavy user reported -- that's 2000 CFI parses
per page turn. The foliate epubcfi.js chunk showed up as ~15% of
self time in Bottom-Up profiles of the release Android build.

Fix:
- createCfiLocationMatcher(location) collapses once and returns a
  matches(cfi) predicate that reuses the cached bounds. O(N) calls
  become 1 collapse + N compares.
- getCfiSpinePrefix(cfi) extracts the spine path via pure string ops
  (no CFI.parse round-trip) for use as a chapter bucket key.
- Annotator builds annotationIndex = { bySection, globals } via
  useMemo([config.booknotes]) once when booknotes change, not per
  page turn. The progress-driven effect then only scans the current
  chapter's bucket -- ~50 CFIs in a typical book instead of all 1000.
  globals are pre-filtered too.
- useSearchNav / useBooknotesNav switch to the batched matcher for
  the same reason.

Includes parity tests covering empty/malformed inputs, equality
shortcut, prefix shortcut, in-range, and out-of-range cases.

* fix(annotator): keep note-only annotations in the per-chapter bucket

The booknote bucketing gated entries on `item.style`, which dropped
note-only annotations (a `note` with no highlight style/color, created
via the Notebook flow) from the per-relocate re-apply path. Their note
bubble was no longer redrawn on relocate or when booknotes changed while
a section stayed rendered.

Restore the original two-list semantics: bucket on style OR note, then
classify per location (annotations need a style, notes need a note).

Extract the logic into a dedicated, unit-tested `annotationIndex` module
(buildAnnotationIndex + selectLocationAnnotations) instead of inlining it
in Annotator, matching the reader/utils domain-named convention.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Huang Xin <chrox.huang@gmail.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-12 18:06:13 +02:00
loveheaven 7cba22ab31 perf(reader): coalesce relocate events and memoize BookCell to stop per-swipe storm (#4562) 
* perf(reader): coalesce relocate events and memoize BookCell to stop per-swipe storm

FoliateViewer
-------------
foliate fires `relocate` multiple times during a swipe burst (snap
steps + intermediate stabilize). Each one ended up in setProgress,
which writes to readerProgressStore + bookDataStore. Coalesce them to
a single commit per animation frame so only the final viewport state
is persisted.

Earlier this used requestIdleCallback, but profiling on Android showed
"Fire Idle Callback" ballooning to 2.0+ s of total time per ~28 s
session: rIC backed up under sustained pressure and dumped the whole
queue into the post-swipe pause, producing exactly the "feels sluggish
right after I let go" jank we were trying to fix. rAF runs once per
frame, gets scheduled by the browser's normal vsync loop, and doesn't
accumulate when the page is busy.

BooksGrid -> BookCell
---------------------
Previously BooksGrid subscribed to the entire progresses map and
rendered every book inline. The map changes on every page turn, so the
whole bookKeys.map(...) body re-ran for every swipe. On top of that
inset-related objects (gridInsets, contentInsets) were rebuilt every
render and threaded as fresh references into 7+ children, so even
unchanged children couldn't bail out. That accounted for ~27% of
main-thread time in the Bottom-Up profile ("Animation Frame Fired"
2.6s / 27%).

Extract BookCell as its own React.memo'd component:
- Each cell subscribes only to its own book's progress via
  useBookProgress(bookKey). A page turn re-renders one BookCell, not
  the grid.
- viewInsets / contentInsets are memoized off their numeric inputs so
  children get stable prop references across renders.
- BookCell uses per-field selectors internally for the same reason
  spelled out in store/readerProgressStore.ts header.
- Dropdown handlers are wrapped in useCallback so HeaderBar's props
  object stays stable.

* fix(reader): subscribe BookCell to its own viewState so settings/ribbon toggles apply live

BookCell subscribed reactively only to useBookProgress and read
viewState/viewSettings imperatively. Settings that save with
applyStyles=false (Show Header/Footer, Double Border, Border Color) and
the bookmark ribbon toggle write no progress, so the cell didn't
re-render and the chrome it gates (SectionInfo, ProgressBar, DoubleBorder,
Ribbon) only updated on the next page turn.

Subscribe to the per-book viewStates[key] slice. This is safe now that
progress lives in its own store — viewStates[key] only bumps on
low-frequency events (settings toggles, ribbon, init, sync), never on
the per-swipe relocate path — so it does not reintroduce the commit
storm the progress-store split removed.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Huang Xin <chrox.huang@gmail.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-12 17:59:34 +02:00
Huang Xin ee01fcd123 fix(reader): texture the scrolled-mode top inset mask, closes #4486 (#4563) 
In scrolled mode the notch-area masks the top safe-area inset with
opaque bg-base-100 so content scrolling under the status bar is hidden,
but it painted over the background texture (.foliate-viewer::before at
the z-0 layer), leaving a flat untextured strip across the unsafe
header area.

Give the mask its own texture ::before (.notch-masked in textures.ts)
and make the element span the grid cell, clipped down to the inset
strip with clip-path — background-size cover/contain resolves against
the element box, so the full-cell box is what keeps the mask's tiles
aligned with the viewer's at the seam. clip-path also clips
hit-testing, so the click target stays the inset strip only.

Verified on a Xiaomi 13: the strip now renders the texture with a
pixel-continuous seam (row-to-row MAE at the boundary dropped from
11913 to 230, the level of ordinary texture rows), and
elementsFromPoint confirms the notch is hit-testable only inside the
strip.

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-06-12 17:39:17 +02:00
loveheaven 59d4f0aa33 perf(reader): split progress into its own store to cut React commit storm (#4557) 
setProgress was called multiple times per swipe burst, each call writing
into readerStore.viewStates[key].progress. ~65 places in the reader
subtree subscribed to useReaderStore() without a selector, so every
setProgress fan-out re-rendered all of them -- even the 51 that didn't
care about progress. On Android release builds this showed up as
Layout = 9.8% and Function Call = 9.6% of main-thread self time in
Chrome DevTools' Bottom-Up profile during a reading session.

Fix:
- New tiny store store/readerProgressStore.ts holds the per-book
  BookProgress map. setBookProgress only fires its own subscribers.
- readerStore.setProgress now writes progress to the new store and only
  touches bookDataStore for the primary view (secondary parallel views
  shouldn't overwrite the shared config).
- readerStore.getProgress is kept as a delegating facade so existing
  imperative call sites don't break.
- Components / hooks that genuinely need to react to progress changes
  subscribe via the new useBookProgress(bookKey) hook. The handful of
  call sites that just want a one-shot read use getBookProgress(key) so
  they don't subscribe at all.
- readerStore.clearViewState calls clearBookProgress so the map doesn't
  grow unbounded across book opens/closes.

See store/readerProgressStore.ts header for the full rationale.
 2026-06-12 17:14:49 +02:00
loveheaven 1c392de0fa perf(reader): throttle library.json writes and cache known dirs to cut IPC (#4556) 
useProgressAutoSave fires saveConfig ~once per second of reading. Two
write-time sources were doubling its IPC cost:

1. saveConfig wrote the WHOLE library.json (+ backup) on every call, so a
   user with N books paid 2*JSON.stringify(N) per save. Chrome DevTools'
   Bottom-Up profile on a release Android build showed processIpcMessage
   chewing ~25% of main-thread time during a reading session.
2. nativeFileSystem.writeFile / copyFile defensively called plugin:fs|exists
   before every write to ensure the parent dir existed. Same dir gets
   probed once per save -- ~50% of IPC time per save was just exists()
   round-trips against directories that have been there since the book
   was opened.

Fix:
- LIBRARY_SAVE_THROTTLE_MS=30s coalesces a swipe burst into a single
  library.json write. Per-book config.json is still written eagerly --
  it's the sync source-of-truth and is small. flushPendingLibrarySave()
  is called on hook unmount + window blur so closing the book always
  flushes.
- In-process knownExistingDirs Set caches verified directories per app
  session. createDir adds, removeDir (incl. recursive) clears. Cold
  start still does the original exists+createDir dance once per dir.
 2026-06-12 16:32:21 +02:00
loveheaven 1ce79d9abf perf(reader): reduce open-book TBT by batching layout-thrashing reads/writes and deferring annotation page back-fill (#4554) 
* perf(reader): batch keepTextAlignment reads/writes to avoid layout thrashing

keepTextAlignment iterates every <div>, <p>, <blockquote>, <dd> in a
freshly-loaded section and tags each with an aligned-{center,left,
right,justify} class based on its computed text-align. The previous
implementation read getComputedStyle and wrote classList.add inside
the SAME forEach pass, which is the textbook layout-thrashing
anti-pattern: classList.add invalidates the document's style cache
(class-based selectors can affect descendants), so the next
getComputedStyle call forces the browser to recompute style for the
whole document.

For a long chapter (~hundreds of p/div/blockquote/dd elements — a
typical Harry Potter section), that turned the loop into N x layout
recalcs. On a release Android build it surfaced as:

  - Browser console violation: 'Forced reflow while executing
    JavaScript took 1210ms'
  - The dominant chunk of the open-book Bottom-Up profile's
    Layout = 32.8% / Recalculate Style = 17.5% of TBT (2503ms total)
  - The 'load' handler also tripped a 1249ms violation, dominated by
    keepTextAlignment running inside it

Fix: split into a read pass (O(N) getComputedStyle into an array) +
a write pass (O(N) classList.add). The browser computes style once
for the document at the start of the read pass and reuses that
result for every subsequent getComputedStyle call; the write pass
then batches all class mutations together so style invalidation
happens at most once at the end.

* perf(reader): back-fill annotation pages off the open-book hot window

Each call to view.getCFIProgress(cfi) synchronously decompresses the matching section's XHTML from the EPUB zip and walks its text nodes (foliate-js progress.js #getCache), costing 100-300ms per cold section on a release Android build. For users with annotations spread across many chapters that's seconds of zip-IPC + main-thread work that was happening inside the open-book TBT window.

First attempt scheduled the back-fill via requestIdleCallback. On Android Tauri the WebView fires rIC aggressively while the main thread is still doing layout/style work for the freshly-opened book — the Bottom-Up profile after that change still showed 1.5s+ of sendIpcMessage -> readData -> loadDocument -> getCFIProgress chains nested under "Fire Idle Callback" inside the same hot window.

New strategy:

  - Hard gate on the renderer's first 'stabilized' event so the back-fill can't possibly start before the open-book paint settles.

  - Add a 5s grace timer after stabilized so the user's first page-turns and paginator's adjacent-section preload can finish without contention.

  - Process annotations one at a time with a 250ms setTimeout gap between each, instead of chained idle callbacks. Each getCFIProgress shows up as its own short task with input-handling slots in between.

  - 10s safety-net fallback if 'stabilized' never arrives, plus full cleanup on unmount.

  - Batch the saveConfig write at the end (one IPC instead of N).

  - Skip entirely when there are no annotations missing a page.

The page field still only feeds the secondary 'p NN ·' label in the sidebar BooknoteItem, so the on-screen highlight rendering paths (progress-driven addAnnotation in the [progress] effect, plus onCreateOverlay on section load) are completely independent and unaffected by this change.
 2026-06-12 16:22:35 +02:00
Huang Xin 61d804a54f fix(dict): resolve Android content-URI filenames via native basename (#4553) 
On some Android devices the SAF picker returns an opaque, extension-less
content:// document URI (e.g. .../downloads.documents/document/msf%3A20).
Dictionary bundle grouping derived each filename from getFilename() — a pure
string-parse of the URI — so no .ifo/.idx/.dict marker was found, every file
was orphaned, and the user saw "Skipped incomplete bundles" even though the
bundle was complete. Devices whose URI happens to embed the name (e.g.
primary%3ADictionaries%3A21cen.dict.dz) worked, which is why it reproduced
only on some Android devices. The same string-parse also wrote the bundle
files (and synced metadata / contentId) under the mangled URI-segment names,
so a re-import elsewhere did not dedupe.

tauri's Android path.file_name (basename) special-cases content:// / file://
URIs and queries the content resolver for the real DISPLAY_NAME — the same
call AppService.openFile already relies on. Resolve the display name once at
selection time, store it on SelectedFile.name, and have bundle grouping
classify by that name instead of re-parsing the URI. The old extension filter
already used basename but discarded the resolved name; threading it through
removes that divergence.

Also fix the Settings -> Dictionaries "+" badges (Import Dictionary / Add Web
Search) collapsing to a black spot in e-ink mode by adding eink-inverted,
mirroring the font import button (#4454).

Fixes #4489
Fixes #4472

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-12 12:01:29 +02:00
Huang Xin 12ac7ae6c0 fix(reader): draw annotation highlights over bullet lists (#4552) 
Highlights spanning paragraphs and a bullet list painted the
paragraphs but not the list items: the overlayer split ranges with a
hard-coded 'p, h1, h2, h3, h4' selector before collecting client
rects, so li/blockquote/td text fell into no sub-range and produced
no SVG rects. Bump foliate-js to split by text nodes (plus img/svg)
instead, which covers every block type while still excluding the
block border boxes that over-highlight blank space.

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
 2026-06-12 10:34:34 +02:00
Huang Xin 28767fecd9 docs(readme): add a Documentation section linking to readest.com/docs (#4551) 
Add a Documentation section to the README pointing to the official docs at https://readest.com/docs, with a matching entry in the top navigation and a reference-style link.

Also bundle accumulated agent memory updates under apps/readest-app/.claude/memory/.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-12 09:54:37 +02:00
Huang Xin da6f45f69c ci: single, workspace-aware rust-cache for build_tauri_app (#4550) 
* ci(pull-request): cache the vendored tauri workspace crates

build_tauri_app rebuilt the whole tauri stack every run. The fork is wired via
[patch.crates-io] to path crates (packages/tauri, packages/tauri-plugins) plus
local src-tauri/plugins/*, all workspace members that Swatinem/rust-cache prunes
by default (cache-workspace-crates: false). Every crates.io plugin depending on
the patched `tauri` then rebuilt transitively, while unrelated deps stayed cached.

Set cache-workspace-crates: true so those sporadically-updated submodule crates
are cached, and bump the cache key (tauri-cargo -> tauri-cargo-ws) so the old
workspace-crate-less cache is invalidated and repopulated (rust-cache won't
re-save on a full key match). The first run after this is a full rebuild;
subsequent runs reuse the cached tauri stack.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* ci(pull-request): keep a single rust-cache for build_tauri_app

build_tauri_app ran two rust-cache actions: actions-rust-lang/setup-rust-toolchain's
built-in one (cache-workspace-crates: false) plus the explicit Swatinem/rust-cache.
They doubled cache storage and competed over the shared target/. Set cache: false on
setup-rust-toolchain so the explicit cache — the one configured with
cache-workspace-crates for the vendored tauri fork — is the only one.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-12 09:38:19 +02:00