feat: add Telugu (te) locale support

CONTRIBUTING.md

@@ -204,7 +204,7 @@ FloCafe runs on real business data that must survive software upgrades.
 
 ## Translations and i18n
 
-FloCafe currently provides translations for English (`en`), Spanish (`es`), French (`fr`), Brazilian Portuguese (`pt`), Russian (`ru`), Filipino (`fil`), Turkish (`tr`), Persian (`fa`, including RTL support), Arabic (`ar`, including RTL support), Urdu (`ur`, including RTL support), German (`de`), Italian (`it`), Japanese (`ja`), Simplified Chinese (`zh`), Traditional Chinese for Taiwan (`zh-tw`), Korean (`ko`), Bahasa Indonesia (`id`), Dutch (`nl`), Hindi (`hi`), Bengali/Bangla (`bn`), Albanian (`sq`), Vietnamese (`vi`), Thai (`th`), Nepali (`ne`), Hausa (`ha`), Javanese (`jv`), Marathi (`mr`), Swahili (`sw`), and Punjabi (`pa`).
+FloCafe currently provides translations for English (`en`), Spanish (`es`), French (`fr`), Brazilian Portuguese (`pt`), Russian (`ru`), Filipino (`fil`), Turkish (`tr`), Persian (`fa`, including RTL support), Arabic (`ar`, including RTL support), Urdu (`ur`, including RTL support), German (`de`), Italian (`it`), Japanese (`ja`), Simplified Chinese (`zh`), Traditional Chinese for Taiwan (`zh-tw`), Korean (`ko`), Bahasa Indonesia (`id`), Dutch (`nl`), Hindi (`hi`), Bengali/Bangla (`bn`), Albanian (`sq`), Vietnamese (`vi`), Thai (`th`), Nepali (`ne`), Hausa (`ha`), Javanese (`jv`), Marathi (`mr`), Swahili (`sw`), Punjabi (`pa`), and Telugu (`te`).
 
 - **Existing languages:** Narrowly scoped fixes and improvements to existing translation strings are always welcome. Verify changes with `npm run i18n:check`.
 - **New languages:** Adding an entirely new language requires maintainer coordination through an issue first.

README.md

@@ -194,7 +194,7 @@ Both scripts ask whether to keep application data. Do not choose their data-purg
 
 - **Offline-first architecture:** Local SQLite storage with auto-migrating schema and fail-closed safety.
 - **Granular roles & permissions:** Owner, Manager, Cashier, Server, and Chef roles with context-aware access policies.
-- **Multi-language support:** 29 localized languages with full bidirectional RTL rendering (Arabic, Persian, Urdu).
+- **Multi-language support:** 30 localized languages with full bidirectional RTL rendering (Arabic, Persian, Urdu).
 - **131 Country profiles:** 109 world currencies with localized phone parsing, number formatting, and timezone handling.
 - **Country tax packs:** Signed regional tax pack integration and manual rule engine.
 - **Automated database backups:** Pre-migration automatic snapshots, manual exports, and optional Google Drive sync.
@@ -248,6 +248,7 @@ FloCafe includes UI translations for:
 - Marathi
 - Swahili
 - Punjabi
+- Telugu
 
 UI language is independent of store country and regional settings, and tax calculation rules remain a separate concern. For details on contributing translations or adding languages, see the [Internationalization and translation guide](docs/architecture/internationalization.md).
 

docs/architecture/frontend.md

@@ -542,10 +542,10 @@ and the auth store keeps the languages that failed. Every print action calls
 a receipt warning saying English labels were used. A receipt that renders in the wrong language is
 a financial document, not a cosmetic bug.
 
-## Internationalization: 29 locales
+## Internationalization: 30 locales
 
 [`frontend/src/lib/i18n/languages.ts`](../../frontend/src/lib/i18n/languages.ts) is the renderer's
-registry. It holds 29 entries, and
+registry. It holds 30 entries, and
 [`frontend/src/lib/i18n/messages/`](../../frontend/src/lib/i18n/messages) holds one JSON file per
 entry. Each registry entry carries the BCP-47 locale tag, the native name, the text direction, the
 `selectable` flag, and a dynamic chunk loader.
@@ -667,7 +667,7 @@ frontend/src/
 │   └── updates/                Update download prompt and install guard dialog
 ├── hooks/                      Printer, KDS, theme, update, cash, approver, and formatting hooks
 ├── lib/
-│   ├── i18n/                   Language registry, loader, and the 29 message files
+│   ├── i18n/                   Language registry, loader, and the 30 message files
 │   ├── printer/                Encoders, document builders, warnings, WebUSB bridge,
 │   │                           and the delivery-slip encoder and browser renderer
 │   ├── updates/                Beta channel, auto-check, and restart-and-install helpers

docs/architecture/internationalization.md

@@ -1,6 +1,6 @@
 # Internationalization
 
-FloCafe supports 29 UI languages. This page describes how a language is declared, loaded, and
+FloCafe supports 30 UI languages. This page describes how a language is declared, loaded, and
 validated, and how the UI language stays separate from a store's regional settings.
 
 ## Source of truth and the parity rule
@@ -15,8 +15,8 @@ translation test fails. `npm run i18n:check` runs `tests/translations.test.ts`,
 registry and file consistency, key parity, ICU validity, placeholder and tag parity, per-language
 completeness, and frontend key safety.
 
-The 29 committed catalogues are: `ar`, `bn`, `de`, `en`, `es`, `fa`, `fil`, `fr`, `ha`, `hi`, `id`,
-`it`, `ja`, `jv`, `ko`, `mr`, `ne`, `nl`, `pa`, `pt`, `ru`, `sq`, `sw`, `th`, `tr`, `ur`, `vi`, `zh`, `zh-tw`.
+The 30 committed catalogues are: `ar`, `bn`, `de`, `en`, `es`, `fa`, `fil`, `fr`, `ha`, `hi`, `id`,
+`it`, `ja`, `jv`, `ko`, `mr`, `ne`, `nl`, `pa`, `pt`, `ru`, `sq`, `sw`, `te`, `th`, `tr`, `ur`, `vi`, `zh`, `zh-tw`.
 
 ## The registry
 

docs/architecture/printing.md

@@ -309,7 +309,7 @@ The drawer pulse is the ESC/POS sequence `ESC p 0 0x19 0xFA`. It is emitted only
 
 `shared/print/width.ts` measures in grapheme clusters and display cells, not UTF-16 code units. A
 combining mark attaches to its base, an Indic conjunct is never split at a virama
-(U+094D, U+09CD, U+0A4D), and a ZWNJ stays with the grapheme it follows. Measuring in code units silently
+(U+094D, U+09CD, U+0A4D, U+0C4D), and a ZWNJ stays with the grapheme it follows. Measuring in code units silently
 breaks every non-BMP and every combining-script receipt, which is why every renderer in the
 pipeline uses this helper rather than `String.length`.
 

docs/guides/adding-a-language.md

@@ -64,8 +64,8 @@ While translating:
   such as an email or IP address. Anything else renders as untranslated English to your users, and
   the check will reject it.
 - For a language with combining marks, the value must be in Unicode NFC. `tests/translations.test.ts`
-  enforces NFC for the languages it checks by name, which are `ha`, `jv`, `mr`, `ne`, `pa`, `ru`, `sw`, `th`, and `vi`.
-- Avoid invisible formatting characters. The Russian, Javanese, and Marathi validators in
+  enforces NFC for the languages it checks by name, which are `ha`, `jv`, `mr`, `ne`, `pa`, `ru`, `sw`, `te`, `th`, and `vi`.
+- Avoid invisible formatting characters. The Russian, Javanese, Marathi, and Telugu validators in
   `tests/translations.test.ts` reject U+200B, U+200C, U+200D, U+00AD, and U+FEFF, because they
   corrupt printed text silently. Other language validators do not check for them yet, so this is
   your responsibility to get right rather than something the test will catch.
@@ -74,7 +74,7 @@ While translating:
 ## 4. Register the intentional-identical keys
 
 Each language has a `<LANG>_INTENTIONAL_IDENTICAL` set in `tests/translations.test.ts` listing the
-keys allowed to be English-identical, most with a comment saying why. There are 24 such sets.
+keys allowed to be English-identical, most with a comment saying why. There are 25 such sets.
 
 Add an entry for every key you deliberately left identical, with a comment giving the reason. Without
 the entry the check fails, and the failure is what keeps an accidentally untranslated string from
@@ -129,13 +129,13 @@ language is not covered by them until you add it:
 
 | File | Array | Current coverage |
 | --- | --- | --- |
-| `tests/phase3-print-regressions.test.ts` | `languages` | 28 of 29; omits `ar`. |
-| `tests/issue-241-localized-errors.test.ts` | `LANGUAGES` | All 29 (`Object.keys(LANGUAGE_REGISTRY)`). |
-| `tests/rtl-setup-auth-settings.test.ts` | `languages` | 25 of 29; omits `ar`, `de`, `fil`, `tr`. |
-| `tests/browser-receipts.test.ts` | loop array | 25 of 29; omits `ar`, `de`, `fil`, `tr`. |
-| `tests/i18n-audit-remediations.test.ts` | two loop arrays | 25 of 29 each; omits `ar`, `de`, `fil`, `tr`. |
-| `tests/rtl-kds-server-whatsapp.test.ts` | four loop arrays | 25 of 29 for the all-language loops, 23 for the LTR-only loop. |
-| `tests/rtl-dashboard-pos-common.test.ts` | two LTR-only loop arrays | 23 of 29; omits the RTL languages plus `de`, `fil`, `tr`. |
+| `tests/phase3-print-regressions.test.ts` | `languages` | 29 of 30; omits `ar`. |
+| `tests/issue-241-localized-errors.test.ts` | `LANGUAGES` | All 30 (`Object.keys(LANGUAGE_REGISTRY)`). |
+| `tests/rtl-setup-auth-settings.test.ts` | `languages` | 26 of 30; omits `ar`, `de`, `fil`, `tr`. |
+| `tests/browser-receipts.test.ts` | loop array | 26 of 30; omits `ar`, `de`, `fil`, `tr`. |
+| `tests/i18n-audit-remediations.test.ts` | two loop arrays | 26 of 30 each; omits `ar`, `de`, `fil`, `tr`. |
+| `tests/rtl-kds-server-whatsapp.test.ts` | four loop arrays | 26 of 30 for the all-language loops, 24 for the LTR-only loop. |
+| `tests/rtl-dashboard-pos-common.test.ts` | two LTR-only loop arrays | 24 of 30; omits the RTL languages plus `de`, `fil`, `tr`. |
 
 This is a known baseline, not something this procedure introduces. Several of these arrays already
 omit registered locales, so they are not a reliable coverage signal. The arrays whose names include

docs/reference/translation-terminology.md

@@ -27,8 +27,8 @@ Pinned by `tests/print-kernel.test.ts`.
 
 ## Indic conjuncts are never split
 
-A Devanagari, Bengali, or Gurmukhi conjunct is a single grapheme cluster and must stay one. The
-virama characters that join them, U+094D in Devanagari, U+09CD in Bengali, and U+0A4D in Gurmukhi,
+A Devanagari, Bengali, Gurmukhi, or Telugu conjunct is a single grapheme cluster and must stay one. The
+virama characters that join them, U+094D in Devanagari, U+09CD in Bengali, U+0A4D in Gurmukhi, and U+0C4D in Telugu,
 are the ones a naive splitter would break.
 
 Pinned by `tests/print-kernel.test.ts`, which asserts that wrapping a two-cluster Indic string at one
@@ -55,7 +55,7 @@ is invisible on screen and corrupts comparison, searching, and print labels.
 identically and wrap identically after normalisation, which is what makes NFC a requirement rather
 than a preference. `tests/translations.test.ts` rejects a `vi.json` value that is not already NFC.
 
-The same NFC requirement is enforced for `ha.json`, `jv.json`, `mr.json`, `ne.json`, `pa.json`, `ru.json`, `sw.json`, and `th.json`.
+The same NFC requirement is enforced for `ha.json`, `jv.json`, `mr.json`, `ne.json`, `pa.json`, `ru.json`, `sw.json`, `te.json`, and `th.json`.
 
 ## Albanian and Vietnamese on a generic thermal printer
 
@@ -198,6 +198,12 @@ pay-out on both the dashboard and the Z-report, and takeaway is `ਟੇਕਅਵ
 `pos.*`. Gurmukhi virama conjuncts (`ਪ੍ਰ`, `ਸ੍ਰ`) and combining signs (`ੰ` tippi, `ਂ` bindi, `ੱ`
 addak, `਼` nukta) must remain NFC-normalized and attached to their base consonant.
 
+**Telugu.** Subtotal `ఉప మొత్తం` and grand total `మొత్తం` are distinct labels and must stay
+distinct on the receipt. The vegetarian tags are `శాకాహారం` and the non-vegetarian tags
+`మాంసాహారం`, identical between `pos.*` and `products.*`. Cash movement uses `నగదు జమ` for pay-in
+and `నగదు ఉపసంహరణ` for pay-out, identically on the dashboard and on the Z-report. Takeaway is
+`టేక్అవే` and delivery is `డెలివరీ`.
+
 ## Verification
 
 The contracts on this page are pinned by `tests/print-kernel.test.ts`,

frontend/e2e/phase7-setup-print-i18n.spec.ts

@@ -103,6 +103,7 @@ const LABELS: Record<string, { basic: string; web: string; whatsapp: string }> =
     web: 'ਵੈੱਬ ਪ੍ਰਿੰਟ (ਬ੍ਰਾਊਜ਼ਰ)',
     whatsapp: 'WhatsApp ਸ਼ੇਅਰ',
   },
+  te: { basic: 'ప్రాథమిక రసీదు (థర్మల్)', web: 'వెబ్ ప్రింట్ (బ్రౌజర్)', whatsapp: 'WhatsApp షేర్' },
 };
 
 async function loginAsOwner(page: Page): Promise<void> {