Skip to content

fix(i18n): word four two-sentence messages as one catalog message - #31383

Merged
RohanK6 merged 1 commit into
mainfrom
crew/conductor/chat-whole-message-joins
Nov 6, 2026
Merged

RohanK6 merged 1 commit into
mainfrom
crew/conductor/chat-whole-message-joins

Conversation

@RohanK6

@RohanK6 RohanK6 commented Nov 6, 2026

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Four UI messages that joined two translated sentences in code are each one catalog message in every shipped language.

The source-links stop note, the browser permission explainer, the usage chart slider name and the crew memory-binding notice each glued two catalog sentences with an English space (the chart with . ). No language could word them as one message, and Japanese and Simplified Chinese got an ASCII space or full stop between full-width sentences.

Why it matters

These lines explain a stopped action, a browser permission, a chart a screen reader announces, and why a memory store cannot be used. A foreign separator between sentences reads as a rendering slip.

Not a goal

What changed (motivation → approach → change)

Site Before After
components/SourceLinksSubmenu.tsx stop note hide_all_source_links_stopped + space + hide_all_source_links_retry hide_all_source_links_stopped_retry
pages/chat/BrowserLiveCardViews.tsx explainer allow_explainer + space + allow_asked_once allow_explainer_asked_once
pages/overview/TokenStackedAreaChart.tsx slider name aria_label + . + keyboard_hint aria_label_keyboard_hint
components/crew/CrewEditorNextDialog.tsx memory notice mismatch or unavailable + space + memory_binding_diagnostic memory_binding_mismatch_diagnostic, memory_binding_unavailable_diagnostic

Each locale value reuses that locale's existing sentences: ja and zh-CN join with no space (the chart label gets a full-width 。), every other language keeps the space or . (Hindi takes a danda). Where a new value inherited wording the style guides reject, it is corrected: informal address in Bengali, Hindi and Russian, and U+202F before the French colon. Fragment keys with no other reader are removed from all 12 catalogs; aria_label and memory_binding_unavailable keep their other readers. The translator note in en.context.json moves to the new chart key. en-XA is regenerated. English is byte-identical.

Backwards compatibility

Compatible: only displayed text and one accessible name change, and the English text is byte-identical. No API, route, setting or stored value is touched.

Tests

website/src/test/ChatWholeMessageJoins.i18n.test.ts checks the English messages are byte-identical, that ja and zh-CN carry no ASCII space or full stop between sentences, that every shipped language has each message, and that none of the four files still glues the parts. All 7 checks fail on main and pass here. The per-language style suites pass with I18N_BASE_REF=origin/main.

Proof: not provable -- frontend test (prove.py runs Python only)

Manual verification

Rendered the memory-binding notice text in Russian and French on main and on this branch (screenshots below, from this branch at f55b8efb9b; the PNGs are not in the final commit). This host has no CJK font, so the Japanese and Chinese shapes are pinned by the test instead. Local reviews: GPT on gpt-5.6-sol and Opus; the one note (stale translator context) is fixed here.

Screenshots / video

Before (main):

before

After:

after

Related Issues

Fixes #31379

Sibling sweep

Swept: a line scanner over website/src for two i18nT calls joined by an English space, colon or period that form two sentences
Sibling: website/src/pages/ChannelPage.tsx:1205 -- not affected -- open PRs #30515 and #29402 touch this file
Sibling: website/src/pages/developer/AgentBackendTab.tsx:1101 -- not affected -- open PR #29509 touches this file

Pattern harvest

Rule candidate: lint
Pattern: a translated string built as ${i18nT(a)} ${i18nT(b)} where both are full sentences

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

The source-links stop note, the browser permission explainer, the
usage chart slider label and the crew memory-binding notice each joined
two translated sentences in code with an English space or full stop.
In Japanese and Simplified Chinese that put ASCII punctuation between
full-width sentences. Each is now one catalog message in all shipped
catalogs. English reads the same.
@RohanK6
RohanK6 requested review from a team and KiroCrewCaptain as code owners November 6, 2026 11:29
@RohanK6
RohanK6 requested a review from Zedmor November 6, 2026 11:29
@RohanK6
RohanK6 enabled auto-merge (squash) November 6, 2026 11:29
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 6, 2026
@github-actions

github-actions Bot commented Nov 6, 2026

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 9225f734f57938a20836d0c84896d6ef5ba02ab3 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks: the diff only rewords strings in existing UI, and no remaining code reads the catalog keys it removes.

Notes

  • Bengali aria_label_keyboard_hint joins its two sentences with an ASCII . rather than the language's own sentence mark. That is the same kind of separator problem this PR fixes for ja and zh-CN, and the new translator note asks for the language's own mark.

[DESIGN-REVIEWED] 9225f73

Description read

[DESCRIPTION-READ] 0fc4b092c8c585a8cb4eb7a9bb68e60fabe7d546d18841423667aff92e00b31d

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Nov 6, 2026

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ✅ PASS

UX-level review of 9225f734f57938a20836d0c84896d6ef5ba02ab3 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

UX-Verdict: PASS

Nothing blocks: only non-English copy changes and no control changes. English is byte-identical, so a screenshot or blind read would show no visible difference.

Notes

  • In bn.json, aria_label_keyboard_hint still joins its two sentences with an ASCII . ("…চার্ট. দিন…"). Use the danda "।" as Hindi does, so Bengali screen-reader users don't get the foreign separator this PR removes everywhere else.

[UX-REVIEWED] 9225f73

Description read

[DESCRIPTION-READ] 323eeaf96d620e5f93b1a9eab8e181e0e58f9881e6e35aa16180649780bd5590

That sha256 names everything this verdict read: the PR description, plus the 3 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Nov 6, 2026

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 9225f734f57938a20836d0c84896d6ef5ba02ab3 — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

First-Principles-Verdict: PASS

This is a fix, and it changes the four messages at the cause: the call sites that glued two catalog sentences together. The two siblings left unfixed are named and deferred because open PRs touch those files.

Notes

  • Item 5 — point patch, siblings deferred: the "Sibling sweep" names website/src/pages/ChannelPage.tsx:1205 and website/src/pages/developer/AgentBackendTab.tsx:1101 as the same glued-sentence pattern, left for later because "open PRs … touch this file". That makes 2 known sites still unfixed, with no pinning test.

What this change ships

Inventory (5 items) — 4 justified

Intent: in every language, these four UI messages read as one naturally worded message, with no English space or full stop forced between sentences. FIX.

  1. The note after "hide all chips" stops partway is now one translated message — justified
  2. The browser remote-debugging permission explainer is now one translated message — justified
  3. The usage chart's screen-reader name is now one message, using the language's own sentence mark — justified
  4. The crew editor's memory-store notice (mismatch or unavailable) is now one message per case — justified
  5. The same glued pattern remains in ChannelPage and AgentBackendTab (declared, deferred) — symptom-level

[FIRST-PRINCIPLES-REVIEWED] 9225f73

Description read

[DESCRIPTION-READ] 44ac6ee7e0c6eb7757fece4545e99179b8c8e7f2c69c0084f2e6cf68bdf38356

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions

github-actions Bot commented Nov 6, 2026

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 9225f734f57938a20836d0c84896d6ef5ba02ab3 — this comment is updated in place on each push.

Review details

No findings.

Notes:

  • The Bengali aria_label_keyboard_hint (bn.json:17163) still separates the two sentences with an ASCII . where Bengali normally uses the dari ।. This only changes how a screen reader pauses in the label, and it reads the same as before this PR. It's a wording point, not a blocking defect.

[OPUS-REVIEWED] 9225f73

Verdict parsed from the review's SHA-scoped output markers for commit 9225f734f57938a20836d0c84896d6ef5ba02ab3.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 9225f734f57938a20836d0c84896d6ef5ba02ab3 false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Nov 6, 2026

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 9225f734f57938a20836d0c84896d6ef5ba02ab3 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 9225f73

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 9225f734f57938a20836d0c84896d6ef5ba02ab3 false-positive: <reason, 20 to 500 characters>

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Nov 6, 2026

@KiroCrewCaptain KiroCrewCaptain left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Tier 1 auto-approve: fix (20 files). Criteria: no conflict, no requested changes, security path denylist clean, design-doc gate clean, security checklist all-NO, no reviewer-lane findings, readiness signal fresh for this head SHA, declared category matches the diff. Category: four UI messages built from two joined translations become one catalog message each.

@RohanK6
RohanK6 merged commit 7cc1a84 into main Nov 6, 2026
97 checks passed
@RohanK6
RohanK6 deleted the crew/conductor/chat-whole-message-joins branch November 6, 2026 13:35
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Nov 6, 2026
buluoray pushed a commit that referenced this pull request Nov 8, 2026
…1383)

The source-links stop note, the browser permission explainer, the
usage chart slider label and the crew memory-binding notice each joined
two translated sentences in code with an English space or full stop.
In Japanese and Simplified Chinese that put ASCII punctuation between
full-width sentences. Each is now one catalog message in all shipped
catalogs. English reads the same.

(cherry picked from commit 7cc1a84)
buluoray pushed a commit that referenced this pull request Nov 8, 2026
…1383)

The source-links stop note, the browser permission explainer, the
usage chart slider label and the crew memory-binding notice each joined
two translated sentences in code with an English space or full stop.
In Japanese and Simplified Chinese that put ASCII punctuation between
full-width sentences. Each is now one catalog message in all shipped
catalogs. English reads the same.

(cherry picked from commit 7cc1a84)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Four UI messages join two translated sentences in code with an English space or full stop

2 participants