You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@maka.apache.org by "scs0209 (via GitHub)" <gi...@apache.org> on 2026/09/01 22:54:59 UTC
[PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
scs0209 opened a new pull request, #4515:
URL: https://github.com/apache/maka/pull/4515
test
Made with [Cursor](https://cursor.com)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on code in PR #4515:
URL: https://github.com/apache/maka/pull/4515#discussion_r4115552602
##########
apps/desktop/src/main/startup-progress-window.ts:
##########
@@ -67,6 +67,20 @@ const COPY = {
attention: '等待你的確認', renderer: '正在開啟工作區',
},
},
+ ko: {
+ title: 'Opening your workspace',
Review Comment:
Fixed on head 195f987b.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5855659582
"Thank you for the review. Addressed on the head:\n\n- **E2E assertions looked for English labels** \u2192 `ko-locale.spec.ts` now asserts `lang=\"ko\"` and `data-maka-locale=\"ko\"` on the document root after reload (proving the locale is applied, not just stored), and drops the `UI language` assertion that pointed at the wrong string.\n- **The head didn't verify build/typecheck** \u2192 It previously did not compile once `ko` became required on `UiCatalog`. Repaired on this head: `npm run typecheck` green (desktop preload/main/renderer/storybook + packages/cli), `npm run build:test` green, and `npm run test:dist` green \u2014 2433 desktop main unit tests, 0 failures, including the new `native-locale-copy.test.js` that locks the Korean native/permission surface keys to the `en` shape.\n- **Post-stack validation** \u2192 Agreed this PR cannot show CI until the bridge (#5100) lands; after #5100 merges this PR retargets to `main` and repo CI runs the full unit + Playwright suites. K
orean wording, IME/accessibility, and packaged-platform layout still need manual verification on packaged platforms \u2014 those were not run in this session, and a human review on top of the automated pass is welcome."
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "hqhq1025 (via GitHub)" <gi...@apache.org>.
hqhq1025 commented on code in PR #4515:
URL: https://github.com/apache/maka/pull/4515#discussion_r4115872271
##########
apps/desktop/src/main/client-settings-confirmation-copy.ts:
##########
@@ -50,6 +50,13 @@ const COPY = {
message: "Allow Maka to update this client's settings?",
buttons: ['Apply changes', 'Cancel'],
},
+ ko: {
+ labels: { theme: '테마', palette: '색상 팔레트', uiLocale: 'UI 언어', runComplete: '실행 완료 알림', keepSystemAwake: '시스템 절전 모드 유지' },
Review Comment:
[P2] The Korean label `시스템 절전 모드 유지` means *keep the system in sleep mode*, the opposite of `Keep system awake`. For a `false → true` change, the confirmation now reads `시스템 절전 모드 유지: 끔 → 켬`, while the setting actually starts Electron's `prevent-app-suspension` blocker. A Korean-speaking user can approve the opposite power behavior and miss scheduled work. Please use a label such as `시스템 절전 방지` and cover this setting in the Korean confirmation-copy test.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5861242856
Thanks for the detailed review. All four points are addressed on the current head (41fdc11f4):
Keep system awake: The label is now 시스템 절전 방지. Your note that 절전 방지 is more accurate than 잠자기 방지 here — since the setting drives Electron's prevent-app-suspension rather than display sleep — was the right call, and I used exactly that.
notifications-policy errored title: 대화 오류 → 작업 오류, so Korean now aligns with both Chinese locales (任务出错 / 任務發生錯誤) and the 작업 terminology used across the other ko catalogs.
Diagnostic dialog vs boot copy: 디스크 ID가 → 디스크 식별 정보가, so the same dialog text uses the same vocabulary on both code paths. (The other two pairs from your table — the recovery-title and Could not connect renderings — are left as-is for now; say the word and I'll align those too.)
permission-overlay-copy: ${appName}을(를) 끌어다 놓으면 → 위 목록에 ${appName} 아이콘을 끌어다 놓으면, following the #5451 convention of never attaching a particle directly to the user-supplied value.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "heeoneie (via GitHub)" <gi...@apache.org>.
heeoneie commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5855121788
@hqhq1025's point about the E2E spec holds, and checking it at `ef03898` it is sharper than "the assertions still look for English labels". Three things, all verifiable from the head.
**1. The spec asserts that nothing changes.** The precedent is `zh-tw-locale.spec.ts`, in the same directory: it asserts `设置` / `界面语言` before the switch and `設定` / `介面語言` after, so the test proves the rendered language actually changed. `ko-locale.spec.ts` asserts `Settings` / `General` / `UI language` *both* before and after switching to `ko`. It therefore passes only while the renderer is untranslated, and it will start failing the moment anyone translates those surfaces — the opposite of what a locale regression test should do.
**2. The asserted string is not the label on that screen.** The Settings → General row uses the `interfaceLanguage` key, which is `界面语言` in zh-CN and `Interface language` in en and ko. `UI language` appears only in `apps/desktop/src/main/client-settings-confirmation-copy.ts` and `client-settings-tools.ts`, which are main-process confirmation copy, not the settings pane. So `getByText('UI language', { exact: true })` should match nothing there. With no CI checks on this head, that would not have been caught.
**3. The `ko` catalogs this PR adds are English.** Across the 48 `*copy.ts` files the PR touches, counting human-readable strings inside each top-level `ko` block: **35 of 3,559 are Korean (1%)**. Running the identical extraction against the `zh-CN` blocks in the same files gives 96% Chinese, so this is not an artifact of how I counted.
Only three files are genuinely translated, and they are exactly the native surfaces in the title:
| file | Korean strings |
|---|---|
| `main/native-diagnostic-dialog-copy.ts` | 21/21 |
| `main/permission-overlay/permission-overlay-copy.ts` | 9/9 |
| `main/runtime-host-quit-copy.ts` | 5/5 |
The other 45 ship a `ko` block that is a verbatim copy of `en` — `conversation-copy.ts` 0/282, `shell-copy.ts` 0/436, `settings-preferences-copy.ts` 0/316, `packages/ui/src/conversation-copy.ts` 0/402, and so on. These are added by this PR, not inherited: `stack/i18n-ko-core` has no `ko` block in those files.
If the English blocks are deliberate scaffolding so `ko` resolves without falling through, that is a reasonable choice — but it should be stated in the PR description, because "native surfaces and E2E fixtures" reads as though the catalog is populated, and a reviewer diffing 62 files will assume the added lines are translations.
**4. Related, and probably the reason the test had to go through `settings.update`:** `localeOptions` in all four catalogs is `[auto, zh-CN, zh-TW, en]` with no `ko` entry. A user cannot select Korean from the picker at all; only the programmatic path the test uses can reach it.
Suggested shape, mirroring the zh-TW precedent: assert the Korean strings from the three native copy files this PR actually translates — the permission overlay and the diagnostic dialog — and drop or correct the `UI language` assertion. Adding `['ko', '한국어']` to `localeOptions` would additionally let the test drive the real picker instead of writing settings directly.
Since the core locale work is mine, I am happy to take the `localeOptions` entry and the renderer catalogs on my side if you would rather keep this PR scoped to the native surfaces. Just say which you prefer and I will open it against the bridge.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5855659684
"Quick notice: the two comments on this PR look like they belong to a different change \u2014 they discuss the durable Turn-message admission plan (#4804), not the Korean locale work here. If that was a copy-paste slip (easy to do when cycling through PRs), flagging it so this thread stays scoped to i18n-ko; if the plan was meant for this PR, it appears to be a different feature entirely."
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-6487281015
<!-- maka-lifecycle:stale-warning -->
This pull request has had no new commits for 30 days and has been marked stale. It will be closed in 7 days unless a new commit is pushed.
Comments do not reset this timer: only a new commit does. If the pull request is intentionally long-lived, a maintainer can apply the `pinned` label.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "hqhq1025 (via GitHub)" <gi...@apache.org>.
hqhq1025 commented on code in PR #4515:
URL: https://github.com/apache/maka/pull/4515#discussion_r4115487556
##########
apps/desktop/src/main/startup-progress-window.ts:
##########
@@ -67,6 +67,20 @@ const COPY = {
attention: '等待你的確認', renderer: '正在開啟工作區',
},
},
+ ko: {
+ title: 'Opening your workspace',
Review Comment:
[P2] Translate the startup progress copy for `ko` rather than copying the English catalog. `showDesktopStartupProgress()` resolves the system locale and passes it to `createStartupProgressWindow()`, which selects `COPY[input.locale]` and renders this HTML. On a Korean system the window is therefore marked `lang="ko"` but every visible title, phase, delay explanation, and diagnostics action remains English. I reproduced this with the built `renderStartupProgressHtml("ko", false)` (English title; no Hangul in the output). This is a native startup surface named in the PR scope. Include `ko` in the startup-window regression test and assert visible Korean copy, not just the locale marker.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "heeoneie (via GitHub)" <gi...@apache.org>.
heeoneie commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5864559515
Verified all four on `41fdc11f4` — `시스템 절전 방지`, `작업 오류`, `디스크 식별 정보가` in both files, and `위 목록에 ${appName} 아이콘을 끌어다 놓으면`. Thanks for turning these around quickly.
Yes, please align the other two pairs. I checked the English and it is byte-identical in both files, so the divergence is purely on our side:
**1. Workspace recovery title** — English `Maka workspace needs repair` in both.
- `native-diagnostic-dialog-copy.ts`: `Maka 작업 공간 복구 필요`
- `runtime-host-boot-copy.ts`: `Maka 작업 공간을 복구해야 합니다`
Take **`Maka 작업 공간 복구 필요`**. It is a dialog title, and the sibling titles in the same catalog are terse noun phrases — `Maka 시작 실패`, `Maka 복구 필요`. The sentence form reads like body text in a title slot.
Worth noting the `message` right below it is already identical in both (`Maka가 이 작업 공간을 확인할 수 없습니다.`), so aligning the title makes the whole dialog match.
**2. Connection failure** — English `Could not connect to ${profileName}` in both.
- diagnostic: `${profileName}에 연결할 수 없습니다`
- boot: `${profileName}에 연결하지 못했습니다`
Take **`${profileName}에 연결하지 못했습니다`**. This fires after an attempt has already failed, and `연결하지 못했습니다` reports that completed failure, which is what `Could not connect` says. `연결할 수 없습니다` reads as a standing inability ("cannot connect"), which is the wrong tense for an error that follows a specific attempt — and it would be the right choice only if the dialog appeared before trying.
Neither is a correctness bug, so if you would rather keep this out of the current increment that is fine by me; they are the last two items I have on the Korean.
On @hqhq1025's note that a static merge against current `main` conflicts in `notifications-policy.ts` — that file is one of the 63 catalogs affected by the `UiCatalog` problem I wrote up on #5100, so I would not spend time resolving that conflict by hand until the shape of the fix there is settled.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on code in PR #4515:
URL: https://github.com/apache/maka/pull/4515#discussion_r4115552602
##########
apps/desktop/src/main/startup-progress-window.ts:
##########
@@ -67,6 +67,20 @@ const COPY = {
attention: '等待你的確認', renderer: '正在開啟工作區',
},
},
+ ko: {
+ title: 'Opening your workspace',
Review Comment:
"Fixed on head `195f987b`. The startup-progress window is a native surface named in the PR scope, so the `ko` block is now a real translation \u2014 title, description, slow-path copy, diagnostics actions, elapsed, and all 11 phase labels \u2014 instead of an English mirror over `lang=\"ko\"`.\n\nCoverage follows the same pattern as the other native surfaces:\n- `startup-progress-window.test.ts` now iterates `ko` and asserts visible Hangul in the rendered HTML (plus that the `en` labels are absent), not just the locale marker.\n- `native-locale-copy.test.ts` includes the startup progress HTML alongside the diagnostic dialog, permission overlay, and quit dialog surfaces.\n\nVerified on the head: `npm run typecheck` (desktop preload/main/renderer/storybook) and `test:dist` green \u2014 2433 tests, 0 failures."
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5582197171
Rebased this branch onto @heeoneie's `feat/i18n-ko-core-ui-locale` (#5011) per the delivery decision on #3975.
I could not change this PR's GitHub base away from \`main\` yet: that branch only exists on \`heeoneie/maka\`, not on \`apache/maka\`. If someone with write access pushes \`feat/i18n-ko-core-ui-locale\` to \`apache/maka\`, I can retarget this PR immediately.
Until then, a stacked draft against the core base is here: will link in a follow-up comment if the create succeeds.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "github-actions[bot] (via GitHub) via commits" <co...@maka.apache.org>.
github-actions[bot] commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-6649202715
<!-- maka-lifecycle:stale-close -->
This pull request has been closed because it received no new commits during the 7-day grace period. This is not a judgement on the merit of the change; it only keeps the review queue aligned with work that is actively in flight.
The branch and review history are preserved. Push a new commit and reopen the pull request whenever the work is ready to continue.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on code in PR #4515:
URL: https://github.com/apache/maka/pull/4515#discussion_r4117613164
##########
apps/desktop/src/main/client-settings-confirmation-copy.ts:
##########
@@ -50,6 +50,13 @@ const COPY = {
message: "Allow Maka to update this client's settings?",
buttons: ['Apply changes', 'Cancel'],
},
+ ko: {
+ labels: { theme: '테마', palette: '색상 팔레트', uiLocale: 'UI 언어', runComplete: '실행 완료 알림', keepSystemAwake: '시스템 절전 모드 유지' },
Review Comment:
Fixed on 46eb923dc
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "heeoneie (via GitHub)" <gi...@apache.org>.
heeoneie commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5884268117
Cross-checked this against #4349 — both PRs are stacked on `stack/i18n-ko-core` and both touch `packages/cli/src/tui-copy-catalog.ts`:
- This PR adds `ko` to `todo` and `transcript-reader`, and pins that list in a test: `koreanTuiDomains = ['todo', 'transcript-reader']` (`packages/cli/src/__tests__/tui-copy-catalog.test.ts:61`).
- #4349 adds `ko` to `mcp-status`, `pickers`, `primary-guidance`, and `session-status`, with its own equivalent assertion.
The domains don't overlap, so there's no content conflict, but whichever of the two lands second will hit a merge conflict in `tui-copy-catalog.ts` and a failing `koreanTuiDomains`-equivalent assertion in the other, since each PR only knows about its own additions.
Suggest whoever lands second rebases and extends the pinned list to the union of both (`['todo', 'transcript-reader', 'mcp-status', 'pickers', 'primary-guidance', 'session-status']`) rather than resolving it as a silent conflict — that's the point of the pin @me2seeks's review asked for.
No action needed from me here, just flagging the ordering before both land, especially since `test` hasn't run green on either branch yet.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5864678234
Fixed on `edca8d90`.
I updated both catalogs to use `Maka 작업 공간 복구 필요` as the title, matching the shorter noun-phrase style of the sibling titles like `Maka 시작 실패 and Maka 복구 필요`. The message was already the same in both places, so the two dialogs are now consistent.
For the connection-failure copy, no changes were needed. Both native-diagnostic-dialog-copy.ts and runtime-host-boot-copy.ts already use `${profileName}에 연결하지 못했습니다`, so that part is aligned with the direction you suggested.
Thanks for confirming the direction.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "github-actions[bot] (via GitHub) via commits" <co...@maka.apache.org>.
github-actions[bot] closed pull request #4515: feat(i18n-ko): native surfaces and E2E fixtures
URL: https://github.com/apache/maka/pull/4515
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5855659058
"Thanks for the detailed pass \u2014 all four points are addressed on the head (`de0f1766`), and I'd love to take the offer on the `localeOptions` + renderer catalogs.\n\n**1. The spec asserted that nothing changes** \u2192 Rewrote `ko-locale.spec.ts`: it now asserts `lang=\"ko\"` and `data-maka-locale=\"ko\"` on the document root after reload, so it verifies the locale actually took effect instead of only that a preference was stored. The stale `UI language` assertions are gone.\n\n**2. Wrong string on that surface** \u2192 Corrected. The pre-switch assertion uses the real `interfaceLanguage` label (`Interface language`), and post-switch it keeps a light smoke check that Settings \u2192 General still renders. The previous `UI language` key lives in main-process confirmation copy, not the settings pane \u2014 thanks for catching that.\n\n**3. The `ko` catalogs are English** \u2192 Agreed, and this is now stated up front in the PR description: renderer catalogs outside the native s
urfaces are deliberate `en` stubs for this slice, with a table of the three truly translated files (diagnostic dialog 21/21, permission overlay 9/9, runtime-host-quit 5/5, plus the real human `KO` block in `session-collaboration-copy`). The head also failed to compile once `ko` became required on `UiCatalog`; that was part of why #3980 was still open \u2014 repaired on this head (typecheck green across desktop preload/main/renderer/storybook and the CLI, `build:test` green, 2433 main unit tests pass including a new `native-locale-copy.test.js` that locks the native catalogs' `ko` keys to the `en` shape). Comment I added to the spec explains that the renderer-chrome stubs land with the renderer slices.\n\n**4. `localeOptions` has no `ko`** \u2192 Yes \u2014 happy for you to take `['ko', '\ud55c\uad6d\uc5b4']` in `localeOptions` and the renderer catalogs on your side, since they sit naturally with the core bridge (#5100) you own. That keeps this PR scoped to the native surfaces + E2E
as you suggest. Please target the bridge (`stack/i18n-ko-core`); this PR's base is already there, so both land through the same #5100 \u2192 `main` stack.\n\nThe compare link for a Korean-only review: https://github.com/apache/maka/compare/stack/i18n-ko-core...scs0209:feat/i18n-ko-native-e2e-3980"
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "heeoneie (via GitHub)" <gi...@apache.org>.
heeoneie commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5857693547
Reviewed the Korean on `2960a230` (the head has moved past the `de0f1766` you mentioned). The E2E fix and the `Interface language` correction both look right, and `native-locale-copy.test.ts` is the right shape — asserting real Hangul that differs from the en stub catches the failure mode that matters.
**1. Confirming @hqhq1025's P2 — the label is reversed.** `시스템 절전 모드 유지` means *keep the system in sleep mode*, which is the opposite of `Keep system awake`. `시스템 절전 방지` is correct and I would take it.
One refinement on the wording: macOS Korean uses 잠자기 for Sleep, so `시스템 잠자기 방지` would match the OS. But this setting starts Electron's `prevent-app-suspension`, not `prevent-display-sleep` — the display may still turn off while the app keeps running. `절전 방지` covers that more accurately than `잠자기 방지`, so **`시스템 절전 방지`** is the better choice here.
For the record, the other five labels in that `ko` block are correct: `테마`, `색상 팔레트`, `UI 언어`, `실행 완료 알림`, and `켬`/`끔`.
**2. `notifications-policy.ts`: `대화 오류` should be `작업 오류`.** The English there is `Conversation error`, so the Korean is a faithful translation of the string — but both Chinese locales decline to follow that wording:
| locale | `notifications-policy` errored title |
|---|---|
| en | `Conversation error` |
| zh-CN | `任务出错` (task) |
| zh-TW | `任務發生錯誤` (task) |
| ko | `대화 오류` (conversation) |
And the same concept in `renderer/locales/conversation-copy.ts` is `Task error` in English and `任务出错` in zh-CN. So the English itself is inconsistent — `Conversation error` in one place, `Task error` in another — and the Chinese translators resolved it to the product concept, 任务. Korean followed the stale English word instead, which leaves 대화 and 작업 both meaning "task" in the same UI. **`작업 오류`** keeps Korean aligned with both Chinese locales and with the 작업 terminology used across the other ko catalogs.
**3. Identical English, two different Korean renderings.** `native-diagnostic-dialog-copy.ts` and `runtime-host-boot-copy.ts` share source strings but translate them differently:
| English | diagnostic-dialog | runtime-host-boot |
|---|---|---|
| `The disk identity may have changed…` | `디스크 ID가 변경되었을 수 있습니다. 복사된 작업 공간이 아니라 이 컴퓨터의 원래 Maka 작업 공간인 경우에만 복구하세요.` | `디스크 식별 정보가 변경되었을 수 있습니다. 이 컴퓨터에 있는 원래 Maka 작업 공간일 때만 복구하세요. 복사된 작업 공간이면 안 됩니다.` |
| workspace recovery title | `Maka 작업 공간 복구 필요` | `Maka 작업 공간을 복구해야 합니다` |
| `Could not connect to ${profileName}` | `${profileName}에 연결할 수 없습니다` | `${profileName}에 연결하지 못했습니다` |
`디스크 ID` vs `디스크 식별 정보` is the one I would fix first — the same dialog text reaching a user through two code paths should not change vocabulary. Either wording works; pick one and use it in both. This is a polish item, not a correctness one.
**4. Minor, convention rather than error.** `permission-overlay-copy.ts` uses `${appName}을(를) 끌어다 놓으면`. The `을(를)` both-forms form is a normal workaround, but slice 1 (#5451) pins the opposite convention — its catalog test asserts that no Hangul syllable follows a user-supplied value directly, precisely so no consonant-dependent particle can be wrong. Attaching the particle to a fixed noun instead keeps it consistent, e.g. `위 목록에 ${appName} 아이콘을 끌어다 놓으면`. Worth aligning since both slices land in the same stack.
Everything else in the six native catalogs reads naturally to me, including the macOS terminology — `손쉬운 사용`, `화면 기록`, `시스템 설정` are all the right OS strings, which is the part that usually goes wrong.
I will take `localeOptions` and the renderer catalogs against `stack/i18n-ko-core` as agreed. One thing you should know about the timing: #5100 is red on `Build`, not a test — my core commit widened `UiLocale` and the CLI TUI catalogs never got `ko`. I have posted the exact diagnosis on #4349. Until that clears, neither this PR nor #5451 can retarget to `main`.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "scs0209 (via GitHub)" <gi...@apache.org>.
scs0209 commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5602424242
Rebased onto the official core bridge #5100 (`stack/i18n-ko-core`) and retargeted this PR's base there, per the delivery decision on #3975 and the bridge note on #5100.
The two temporary core commits from the earlier #5011 stack are dropped from this branch; only the #3980 native/E2E commits remain on top of the bridge.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "heeoneie (via GitHub)" <gi...@apache.org>.
heeoneie commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5884275410
Confirmed both fixes against head `edca8d90` and `713c4f88`:
- Workspace-repair title: `native-diagnostic-dialog-copy.ts:206` and `runtime-host-boot-copy.ts:88` both now read `Maka 작업 공간 복구 필요`.
- Connection-failure wording: both files already used `${profileName}에 연결하지 못했습니다`, so no change was needed there — agreed, that one was already aligned.
The three P3s from the later automated review (Astryx `ko` override key, the bot-chat-detail literal assertions, the pinned TUI domain list) are also fixed as of `713c4f88`, checked directly against the diff.
Only open item from my side is the `koreanTuiDomains` overlap with #4349 I just flagged separately — not a defect in this PR, just an ordering note for whichever lands second.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "ggbdpq (via GitHub)" <gi...@apache.org>.
ggbdpq commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5576205082
Plan for the P1, folded with the two named downstream gaps so the slice lands as one coherent contract change:
1. **Shared predicate in Core** next to `MessageContent`: meaningful content = non-empty `text` OR `quotes.length > 0` OR `attachments.length > 0`.
2. **Durable admission**: `normalizeRootTurnMessageContent` switches from the text-length rule to the shared predicate, keeping its own attachment-count, trust, and size checks.
3. **Compaction estimate**: `estimateRuntimeEventChars` counts the structured envelope (quote texts + attachment identity) so a structured-only user event is no longer estimated at zero and dropped by the history-compact gate.
4. **Recap**: the recap projection switches to the same visibility/content decision instead of `content.text.trim()`.
5. **Verification**: storage-level admission unit tests (quote-only accepted; truly empty still rejected with the same frame error), estimate/recap unit tests, and a composition-level test driving a real SQLite-backed `turn.start` with a quote-only message through idle and queued paths to a durable Turn. Decoder-only examples are excluded per your note; a full Host-subprocess run remains the acceptance bar for closing #4804.
This slice deliberately does not touch the empty-bubble half of #4804 (out of scope as agreed). Will start after the current review round settles so the layers land and get re-reviewed together.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [PR] feat(i18n-ko): native surfaces and E2E fixtures [maka]
Posted by "ggbdpq (via GitHub)" <gi...@apache.org>.
ggbdpq commented on PR #4515:
URL: https://github.com/apache/maka/pull/4515#issuecomment-5576396915
The durable-admission half is now pinned by tests, pushed in 54a454d1f: quote-only and attachment-only root Turn inputs are admitted through `normalizeRootTurnAdmissionPayload` (the seam your end-to-end reproduction proved was rejecting them), and a truly contentless input still throws the same frame error. On the pre-fix base the quote-only and attachment-only cases fail red, matching your reproduction.
Also pushed in fef4f355a (rebase onto current main): the decoder-side structured-content acceptance, the compatibility epoch renumbered to 126 (correctly above main's 125, resolving the collision you flagged), the compaction estimate counting the structured envelope, and the recap carrier placeholder.
Remaining before merge: the composition-level verification you specified (submit → durable admission → snapshot/replay → model input against a real Host subprocess). The storage seam is now proven by unit test; the Host-subprocess run is the piece I will add next.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org