Complete a plural category the target locale needs from the message's own
`other` branch during sync, so one missing branch no longer costs the whole
translation. Romanian arrived with `one` and `other` for all 156 of its
plurals and rendered English for every one of them. See docs/LOCALIZATION.md.
Fail the nightly export only when a catalog loses far more than churn
explains, rewrite a Crowdin em dash to the hyphen the source uses instead
of discarding the translation, and report a rejected message as a
rejection rather than a lost one. See docs/LOCALIZATION.md.
Two independent guards were failing the workflow on healthy input.
Retention read "exported value equals the English source" as a lost translation, but exports are
built with skipUntranslatedStrings, so Crowdin omits a key entirely once it stops carrying a
translation. Anything present in the export is by definition translated, and the equality check only
ever fired on correct work: Spanish "Error", "Feb" and "Cookie", French "Description", Portuguese
"Readwise". Retention now compares presence alone, which is what absence already meant. The same
inference lived in a legacyComplete branch that would have silently stopped protecting the 2,318
keys whose translation legitimately equals the source once any locale reached full translation.
Scheduled runs had no way out of that. The only escape hatch is allowed_translation_losses, a
workflow_dispatch input that is never persisted, so every schedule rediscovered the same losses and
failed until someone dispatched the run by hand.
The slot count rule then only runs for the keys it is given, and the sync passed none, so its
rejection filter was weaker than the locale validation running on its output one step later. An
invalid ja message survived the filter, reached the catalog, and failed the next step instead of
being dropped and reported. Move the source scan into locale-source-keys.mjs so the sync and
validate-locales.mjs derive the same keys.
Traditional Chinese cannot reuse the `%two_letters_code%` filename because it would collide with
Simplified under `zh.json`, so `crowdin.yml` maps `zh-TW` to a `zh-Hant` catalog and locale
matching resolves by script subtag before falling back to the base language. Every other new
locale keeps its bare two-letter ID.
Twenty-five entries outgrew the old menus. The mobile submenu inherited `overflow-hidden` with no
max-height, so the last nine languages were unreachable on a phone, and the list was ordered by
ISO code while displaying endonyms, which reads as unsorted. Both header menus now render one
LanguagePicker that searches the native name, the English name and the locale code, groups the
browser's own languages as suggestions, and orders the rest with a collator on the string it
actually shows. Mobile opens a sheet instead of a nested flyout, since a flyout is the wrong
shape for touch regardless of height. Settings gains the Appearance language section its copy
has anticipated since the keys were added.
Tighten the Crowdin progress chart, which the longer language list had stretched to 810px tall.
Both catalogs held nothing but English fallback values, so selecting either language produced a
fully English interface that reads as broken rather than untranslated. Crowdin Translate covers only
24 languages and includes neither, so neither locale had any machine translation route, and no human
translations existed yet.
Drop them from the shared locale list, the Crowdin export configuration, the pull request allowlist,
the catalogs, and the tests, leaving fifteen target languages. Remove both from the Crowdin project
as well, after confirming each export was byte-identical to the English source so no translation work
was discarded.
Re-adding either is cheap once a translation route exists, whether that is a machine translation
engine that supports them or a volunteer translator.
Give the locale tests the same fallback semantics the application runs with. Production always loads
the English catalog and points `fallbackLocale` at it, so a key a target catalog does not carry
renders English. The test instances omitted both, which resolves a missing key to its own name, and
several assertions compared a translated label against the English one without checking that the
language actually translates that key. Both patterns pass only while every catalog is complete, and
would fail confusingly the first time a language lags behind a new key.
Delete the Crowdin service branch automatically once its pull request closes or merges. Crowdin
appends every export to the same branch and never rebases it, so a branch that outlives its pull
request keeps its original merge base and each later export diverges further from main until every
catalog conflicts. Deleting it forces Crowdin to recreate the branch from the current main.
Ref #576
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
* 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
Replace pipe-delimited Vue I18n plural messages with ICU MessageFormat syntax across all locale catalogs and Vue components. Configure custom message formatting and validation to enforce ICU syntax and argument parity.
Closes#792
neon
Chris Mansell