i18n(ja): restore English TiDB Cloud UI labels and remove stray quotes - #23554
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
|
Important Review skippedToo many files! This PR contains 267 files, which is 117 over the limit of 150. To get a review, reduce the PR to 150 files or fewer by splitting it into smaller PRs or changing its base branch. Upgrade to a paid plan to raise the limit. This review couldn't start because sufficient usage credits or metered capacity aren't available. Add credits or update usage-based reviews in the billing tab, then retry. ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: ASSERTIVE Plan: Pro Plus Run ID: 📒 Files selected for processing (267)
You can disable this status message by setting the 📝 WalkthroughWalkthroughThis PR updates Japanese TiDB Cloud documentation to use current English console labels. It covers backup, imports, integrations, monitoring, migrations, billing, security, and administration procedures. Documented workflows and configuration behavior remain unchanged. ChangesEnglish console terminology
Estimated code review effort: 2 (Simple) | ~15 minutes Merge Risk: 🟡 Moderate · up to This documentation-only change restores English console labels across many Japanese procedures, but the current head still includes incorrect labels and cloud/provider instructions plus malformed Markdown and emphasis in several workflows. These defects can misdirect users or make steps render ambiguously, so targeted corrections are needed before merge; the PR does not introduce runtime or availability risk. 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches 💡 1🛠️ Fix failing CI checks 💡
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Independent 6-way review of the 1-word UI-label restoration diff (PR pingcap#23554) found a systemic defect class: wherever the original Japanese bold span already had a trailing particle (を/の/に) or a straddling quote mark (「/」) bundled into its content, the mechanical word-swap replaced the whole bundle with just the bare English word, silently dropping the particle or leaving its quote partner orphaned. Fixed ~25 files worth of these regressions plus a few pre-existing straddling-bracket defects surfaced by the same corpus-wide scan (Destination/Pulsar, Bucket Overview, Description/Support Tickets patterns, Target Database/Target Table reconstruction in naming-conventions-for-data-import.md), a genuine wrong-word bug (Edit/Delete mixed up in data-service-custom-domain.md), and a compound-word-splitting error (一時+Pausing) that should not have been converted at all. Verified via a corpus-wide bracket-balance and duplicate-word scan against upstream: 0 remaining issues across the full diff.
6cc6d76 to
d37666a
Compare
d37666a to
d3a2b20
Compare
Corpus-wide sweep for the 1-word UI-label restoration campaign (the final and highest-false-positive-risk batch after the 4/3-word and 2-word batches). Regenerated the candidate list fresh (329 words, 2,039 occurrences) via EN/JA bold-span positional pairing, restricted to files where EN and JA bold-span counts match per file. Verification: each candidate word was checked against the actual TiDB Cloud console frontend source (dbaas-ui) for a genuine rendered UI string (label/title/name/placeholder/data-mp-event/tight JSX text), not a source comment or prose-embedded coincidental match. Only words with at least one such confirmation were restored, and only at occurrences where the specific sentence context is a genuine UI reference rather than generic descriptive prose (the same "category header vs UI reference" trap documented in earlier batches, e.g. a word can be a real dbaas-ui field name AND a common noun used to introduce a concept elsewhere). Fixed 1606 occurrences across 145 files via this method, plus several occurrences requiring manual reconstruction rather than a simple content swap: - Trapped particles/oversized spans where the original Japanese bold span included more than the bare term (a trailing particle, a wrapping quote plus particle, or a full verb phrase) - reconstructed to keep only the label bolded and the rest as plain text. - Several EN/JA clause-reordering mispairs (the same sentence bolds terms in a different order between EN and JA, causing positional extraction to pair the wrong span) - found via independent re-verification and targeted anomaly scans (duplicate-word-per-line, verb-ending-inside-bold, oversized-span-length checks), then fixed by hand against the EN source. - A few adjacent defects found while touching these lines: a false-friend mistranslation (Alpine misrendered as 高山), a missing menu-path fix (Settings > Networking rendered as Networking > 「ネットワーク」 in 3 files), and a couple of stray-quote/bracket cleanups on the same lines per the standing policy of removing decoration EN doesn't have. Verification: corpus-wide bracket/`**`-parity check across the full diff found 0 issues across 145 files; a same-line-duplicate-word scan and a verb-phrase-inside-bold scan (both used to catch the reordering/oversized-span defect classes) found 0 remaining issues after fixes were applied.
Reverted per feedback: unlike the adjacent Limitations/Permissions bullet headers being purely descriptive prose labels (not literal TiDB Cloud UI strings), keep this one in Japanese as a plain bold term rather than restoring the English word.
- Restore Pending (保留中) to English in 4 private-endpoint status lists, matching sibling statuses (Active/Deleting/Failed/Creating) already fixed in the same bullet lists in these files. - Resolve 3 occurrences of Enable that a verification pass had flagged as a positional-mismatch artifact and skipped without a manual follow-up: dual-layer-data-encryption-premium.md (2x, click Enable in the Dual-Layer Data Encryption section) and tidb-node-group-management.md (1x, click Enable in the Public Endpoint section, also restoring the Public Endpoint label).
Independent 6-way review of the 1-word UI-label restoration diff (PR bold span already had a trailing particle (を/の/に) or a straddling quote mark (「/」) bundled into its content, the mechanical word-swap replaced the whole bundle with just the bare English word, silently dropping the particle or leaving its quote partner orphaned. Fixed ~25 files worth of these regressions plus a few pre-existing straddling-bracket defects surfaced by the same corpus-wide scan (Destination/Pulsar, Bucket Overview, Description/Support Tickets patterns, Target Database/Target Table reconstruction in naming-conventions-for-data-import.md), a genuine wrong-word bug (Edit/Delete mixed up in data-service-custom-domain.md), and a compound-word-splitting error (一時+Pausing) that should not have been converted at all. Verified via a corpus-wide bracket-balance and duplicate-word scan against upstream: 0 remaining issues across the full diff.
d3a2b20 to
edecb6d
Compare
|
Caution CodeRabbit couldn't update its existing comment. The review summary may be out of date. Error details |
The English source never wraps a bold UI-label span (e.g. Click **Connect**) in quotes, but the ja translation had accumulated 「」 around many such spans (e.g. **「接続」**), across single-word button labels, symbols (**「+」**, **「...」**), and longer descriptive labels (**「クラウドストレージからデータをインポート」**). Verified this pattern holds against EN for a representative sample across short and long labels before doing a mechanical, corpus-wide strip: **「X」** -> **X**, applied only where the quote marks sit immediately inside the bold delimiters (1,577 occurrences across 233 files). This supersedes the narrower carve-outs used in PR pingcap#23551/pingcap#23552, which kept a handful of terms quoted for in-file convention (接続 タイプ, スキーマロード, 接続方法, 組織設定, etc.) - per the standing policy, in-file consistency is not a reason to keep quoting a term EN never quotes. Verified per-file: line count, ** count, [ count, and ] count are all unchanged (0 anomalies across all 233 changed files) - only the 「」 characters were removed. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…026-08-18 Resolved 118 file conflicts: took this branch's English-restoration fixes as the base, then re-applied the quote-stripping substitution (**「X」**->**X**) on top, so both fixes combine cleanly. Verified 0 remaining quote-wrapped bold spans and 0 bracket-count anomalies across all 118 files. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
test seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account. You have signed the CLA already but the status is still pending? Let us recheck it. |
|
The PR title mentions "restore 1-word TiDB Cloud UI labels to English", but these 1-word TiDB Cloud UI labels are still in Japanese in the current PR. |
Human reviewer feedback on PR pingcap#23554 (qiancai) noted that some 1-word TiDB Cloud UI labels claimed to be restored in that PR were still in Japanese. Investigation confirmed this: the original positional bold-span-pairing method restricted itself to "files where EN and JA bold-span counts match per file", silently skipping every file where counts didn't match - even though individual lines within those files still had clean, unambiguous EN/JA correspondences. Re-swept using a per-line EN/JA line-window match (each JA line checked against a +/-10 line window of the EN source for the literal English bold term, not file-level bold-count parity) across a curated list of ~30 common TiDB Cloud console button/label words. Every one of the 232 sites below was individually confirmed against the EN source at that exact position before fixing - no blind global find/replace. Biggest misses, all now fixed: - **接続** -> **Connect** (116 occurrences, 59 files) - the single biggest miss; "Connect" is one of the most common buttons across the entire TiDB Cloud console and appears in dozens of client integration guides. - **概要** -> **Overview** (31), **次へ** -> **Next** (14), **インポート** -> **Import** (11), **ダウンロード** -> **Download** (9), **作成** -> **Create** (6), **設定** -> **Settings** (6), **保存** -> **Save** (5), **追加** -> **Add** (4), **編集** -> **Edit** (4), **再起動** -> **Restart** (4), **続行** -> **Continue** (3), **エクスポート** -> **Export** (3), **検索** -> **Search** (2), **メンバー** -> **Members** (2), **招待** -> **Invite** (2), **接続する** -> **Connect** (2), and 6 single-occurrence terms (削除/Delete, 確認/Confirm, キャンセル/Cancel, 再試行/Retry, ダッシュボード/Dashboard, 開始/Start). - Bonus: develop/dev-guide-gui-navicat.md's "**接続**" (2 occurrences) actually corresponds to EN "**Connection**" (Navicat's own top-left button, not TiDB Cloud's "Connect" button) - fixed to the correct English word rather than "Connect". Explicitly NOT fixed (checked and excluded): a third-party-tool "**接続**" in the Airbyte integration guide corresponds to EN "**Connections**" (Airbyte's own sidebar tab name, not TiDB Cloud's dbaas-ui, and a different word besides), left as a separate, lower-priority candidate for a future pass if that vendor's UI is ever verified. This was itself found while double-checking the 接続->Connect sweep for false positives, not part of the original review comment. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
@qiancai Thank you for catching this — you were right, and investigating it turned up a real gap in the original sweep. What was actually wrong: the original method restricted itself to files where the EN and JA bold-span ( I re-swept using a per-line EN/JA match instead (checking each Japanese line against the English source around that exact position, not file-level bold-count parity), and found 232 additional genuine misses across 88 files, now fixed and pushed. The biggest one was One clarification, in case it helps when re-reviewing the diff: this branch also has an earlier quote-stripping commit merged in (removing stray |
Found via a corpus-wide scan for bold JA spans immediately followed by a UI-interaction verb (クリック/選択/オン/オフ/等), then checked against the EN source at each site. Covers both remaining 1-word terms (Public, Authorize, Publish, Previous, Resume/Resuming/Restart, default, Allow, Finish, Zone, Object, Container, Custom, Branching, Yes, Expand, Copy, Columns, Refresh, Database) and multi-word phrases (Reset Password, Generate Password, Test Connection, Connection test, Server and Port, SAVE CONNECTION, Add environment variable, Enable Feature, Open Settings, Successfully connected!) across 67 files.
…m UI labels
Fix the TiDB Cloud connection-dialog checklist template repeated across
33 develop/*.md and ai/*.md files:
- Connection Type and Branch had been translated to Japanese
(接続タイプ / ブランチ) while Connect With and other sibling fields in
the same list stayed in English, per EN source which bolds all four
field names literally (Connection Type / Branch / Connect With /
Operating System).
- Several files had "Connect With" with the topic particle は trapped
inside the bold span ("**Connect With は**") instead of outside it.
- Operating System was translated (オペレーティングシステム) in most
files; in a 5-file subset the は particle was also trapped inside the
bold span.
- dev-guide-wordpress.md additionally had Database's は trapped inside
the bold span.
All values verified against the EN source (upstream/release-8.5) at the
corresponding sentence position.
Fix dashboard-diagnostics-access.md and dashboard-metrics-relation.md: Range Start Time, Range Duration, Start, Compare by Baseline, Baseline Range Start Time, and Cluster Diagnostics had been translated to Japanese (some also with the topic particle trapped inside the bold span, or wrapped in brackets) instead of kept as the literal English TiDB Dashboard UI labels, verified against upstream/release-8.5.
Chrome DevTools panel and menu names (Application, Local Storage) stay in English regardless of the browser's display language, matching the EN source's literal bold UI strings. Restore them from アプリケーション /ローカルストレージ to English, consistent with the third-party-tool-UI scope of this campaign.
dashboard-log-search.md: the Download Selected button had been fully translated (選択項目をダウンロード) while its sibling buttons Cancel/ Retry stayed in English; and the Running status bold label had been translated (実行中) while its siblings Success/Failed stayed in English. Restore both to match the EN source and sibling convention.
dashboard-overview.md: four UI labels had been translated to Japanese or wrapped in brackets while sibling labels in the same lists/notes stayed in English (View Alerts, and the Note's own reference to "View Metrics"), creating self-contradictions: - **[概要]** -> **Overview** (left nav menu item) - **インスタンスの** (trapped particle) -> **Instance**のタイトル - **メトリクスの表示** -> **View Metrics** - **診断の実行** -> **Run Diagnostics**
…n-sso.md Full pass over the file against the EN source. Fixes: - Missing を particle before ご覧ください in the SYSTEM_VARIABLES_ADMIN note. - Bracket-wrapped Japanese translations of TiDB Dashboard UI labels (Single Sign-On, Authorize and Save, Update, Authorize Impersonation, Sign in via Company Account, Sign On, Settings) restored to literal English, matching the EN source's bold UI strings. - Okta/Auth0 dialog field names (Sign-in method, Application Type, Single-Page Application, Sign-in/Sign-out redirect URIs, Assignments, OpenID Connect ID Token, Basic Information, Display name) restored to English. - OIDC Client ID / OIDC Discovery URL / Client ID / Domain field names (TiDB Dashboard's own SSO settings fields, referenced repeatedly throughout the file) unified to literal English instead of a mix of katakana and English.
The Domain-field-value prefix/suffix instruction was missing its verb object particle, reading as an incomplete sentence. Rephrase using the established "先頭にX、末尾にYを追加して" pattern already used for the same kind of instruction in tidb-cloud/dedicated-external-storage.md.
dashboard-slow-query.md: ☰ menu tooltip (その他 -> More), the Export button (translated to エクスポート with the particle trapped inside bold right after being correctly called Export one sentence earlier), and two column headers (終了時刻 -> Finish Time, 説明 -> Description) restored to the literal English TiDB Dashboard UI strings.
…e UI labels - Fix dev-guide-bookshop-schema-design.md: Data Format, Folder URI, Bucket Access, Role ARN, and the Import Data from Amazon S3 page title had been translated to Japanese instead of kept as literal TiDB Cloud UI strings. - Fix dev-guide-gui-datagrip.md: bracket-wrapped Data Source and Drivers window name restored to English. - Fix dev-guide-gui-dbeaver.md: Driver properties tab, CA certificate field, and Server Host/Port connection-form field names restored to English (some had the topic particle trapped inside the bold span). - Fix the repeated "In addition to the Public connection type, ... supports Private Endpoint connections" / "... supports Private Endpoint and VPC Peering connection types" boilerplate across all 26 develop/dev-guide-*.md files that include it: Private Endpoint and VPC Peering had been translated to katakana (プライベート エンドポイント / VPC ピアリング) while the sibling term Public stayed in English in the same sentence. All values verified against upstream/release-8.5 at the corresponding sentence position.
- dev-guide-gui-dbeaver.md: "SSLを使用する" -> "Use SSL" (isolated occurrence; sibling files datagrip.md/navicat.md already used the correct English checkbox label). - dev-guide-gui-vscode-sqltools.md: full pass restoring every VS Code/SQLTools UI field and button name to literal English (Connection name/group, Connect using, Server Address, Port, Username, Password mode, SQLTools Driver Credentials, MySQL driver specific options, Install, Extensions, CONNECTIONS, Add New Connection, Settings, Networking, Public Endpoint, Enable, Add IP Address, Enabled, Disabled, SSL Options (node.TLSSocket), Certificate Authority (CA) Certificate File, TEST CONNECTION), most of which had been translated to Japanese or wrapped in brackets while sibling fields in the same lists stayed correctly in English. Verified the file now has full 1:1 parity of bold English UI terms against upstream/release-8.5.
…ple-app files Systematic pass: extracted every bold English UI term from the EN source for each file and diffed against the JA file's bold terms, fixing every gap found. - dev-guide-gui-datagrip.md: Settings, Networking, Advanced tab, DataGrip field table header, and Host/Port/User/Password (both connection-string mapping tables and the Self-Managed bullet list) restored to English; also fixed a second occurrence of the bracket-wrapped Data Source and Drivers window name. - dev-guide-gui-dbeaver.md: Settings > Networking restored to English. - dev-guide-gui-navicat.md: Connection Name, Host, Port, User Name, Password, Settings, Networking, Public Endpoint, Enable, and Add IP Address restored to English. - dev-guide-gui-mysql-workbench.md: Connection Name, Hostname, Port, Username, Password, Store in Keychain .../Store in Vault, Settings, and Workbench Preferences restored to English. - All 21 develop/dev-guide-sample-application-*.md files: the bracket-wrapped "Settings > Networking" left navigation instruction restored to English (identical boilerplate across all files). - dev-guide-sample-application-aws-lambda.md: Functions page link text, Configuration tab, and Environment variables restored to English. - dev-guide-sample-application-ruby-rails.md: Connect With and Connection Type restored to English (were bracket-wrapped and translated, inconsistent with the same file's other two occurrences which were already correct English). Confirmed each file now has full 1:1 parity of bold English UI terms against upstream/release-8.5 (excluding a handful of unrelated prose-emphasis bold terms like MUST/RECOMMENDED and code-example section headings, which are not UI labels and out of scope here).
…e-app guides The earlier fix for this shared boilerplate sentence removed a garbled bold span but dropped the trailing "to configure it" clause, leaving the sentence end abruptly after the markdown link. EN: "...or follow the steps in [Configure an IP Access List](url) to configure it before your first connection." Restored the missing clause across all 24 files sharing this exact template.
Extended the systematic EN-bold-term-parity check to ~77 previously unswept tidb-cloud/*.md files (connect/import/migrate/monitor/ serverless/premium guides). Fixes span: bracket-wrapped translated button/menu labels, particles trapped inside bold spans, changefeed Action-column and Kafka distribution-mode option labels translated where EN keeps them literal, stale "クラスター" page links updated to the renamed "My TiDB" page/URL, half-width vs full-width punctuation (? and >) mismatches against console UI, and several >-escaping inconsistencies. Also fixed the same Action-column bracket defect found as a sibling gap in essential-changefeed-overview.md while reviewing changefeed-overview.md.
|
Fifth round: extended the full EN-bold-term-parity check (per-file, not word-list-based) to ~77 previously unswept `tidb-cloud/*.md` connect/import/migrate/monitor/serverless/premium guides, plus completed a truncated sentence in the shared "IP access list" boilerplate that an earlier round's fix had accidentally cut short. ~130 more sites fixed across 33 files (commits 69ce13e, dc2bd89): bracket-wrapped translated button/menu labels, particles trapped inside bold spans, changefeed Action-column and Kafka distribution-mode option labels, stale "クラスター" page links updated to the renamed "My TiDB" page, half-width vs full-width punctuation mismatches (?, >) against literal console UI, and several `>`-escaping inconsistencies. Two genuine content gaps surfaced along the way that are NOT simple bold-term fixes (flagging rather than fixing here, since they need net-new translation, not a mechanical swap):
PR remains `MERGEABLE`. |
Right. Updated this pull request to use English label. not only 1-word, but every word. |
… label fixes) set-up-vpc-peering-connections.md: JA described an older "combo box in upper-left corner" project-switching flow; EN has since moved to "My TiDB page -> Project view tab -> project settings icon". Updated both occurrences (AWS and Google Cloud sections) to match, renumbering the surrounding steps, and corrected the now-narrower "combo box" tip (org-switching only, not project/cluster) to match EN. troubleshoot-import-access-denied-error.md: the "Check the trust entity" section was missing an entire paragraph (intro sentence + a 4-step list for retrieving the TiDB Cloud Account ID/External ID from the console) that EN has but JA never received. Translated and inserted it, matching established UI-label conventions already used in sibling import-guide files.
|
Closed out the two content gaps flagged in the last comment (commit 5369975):
PR remains `MERGEABLE`. |
Extends the campaign to the ~205 previously-unswept tidb-cloud files (including CLI reference and Terraform docs, which turned out to have near-zero defect density as expected -- pure flag/attribute tables with no bold UI-copy spans to compare). Fixes across 24 files include stray-bracket button/tab labels, a stale "My TiDB" page link update (4 more sites), a garbled Failover/Reprotect sentence with a missing を particle, a link-target/link-text mismatch where a bold link's visible text didn't match what its own href pointed to, and several >-escaping and ellipsis-glyph normalization fixes.
|
Sixth round: extended the EN-parity sweep to the remaining ~205 previously-unswept tidb-cloud/.md files, including CLI reference (`ticloud-.md`) and Terraform resource docs this time. As expected, those reference-doc categories came back almost entirely clean (pure flag/attribute tables with no bold UI-copy spans to compare at all). 24 files fixed (commit f6a34f8), including stray-bracket button/tab labels, 4 more stale "クラスター" → "My TiDB" page link updates, a garbled Failover/Reprotect sentence missing a を particle, and a link where the bold visible text didn't match its own href target (linked to the wrong anchor's label). PR remains `MERGEABLE`. |
[LGTM Timeline notifier]Timeline:
|
What is changed, added or deleted? (Required)
Corpus-wide sweep for the 1-word UI-label restoration campaign (the final and highest-false-positive-risk batch after the earlier 4/3-word and 2-word batches). Regenerated the candidate list fresh (329 words, 2,039 occurrences) via EN/JA bold-span positional pairing, restricted to files where EN and JA bold-span counts match per file.
Verification: each candidate word was checked against the actual TiDB Cloud console frontend source (dbaas-ui) for a genuine rendered UI string (
label:/title:/name:/placeholder:/data-mp-event/tight JSX text), not a source comment or prose-embedded coincidental match. Only words with at least one such confirmation were restored, and only at occurrences where the specific sentence context is a genuine UI reference rather than generic descriptive prose (the "category header vs UI reference" trap — a word can be a real dbaas-ui field name AND a common noun used to introduce a concept elsewhere).Fixed over 1600 occurrences across 145 files via this method, plus several occurrences requiring manual reconstruction rather than a simple content swap:
Verification
**-parity check across the full diff: 0 issues.Which TiDB version(s) do your changes apply to? (Required)
Tips for choosing the affected version(s):
By default, CHOOSE MASTER ONLY so your changes will be applied to the next TiDB major or minor releases. If your PR involves a product feature behavior change or a compatibility change, CHOOSE THE AFFECTED RELEASE BRANCH(ES) AND MASTER.
For details, see tips for choosing the affected versions.
What is the related PR or file link(s)?
AI agent involvement
Do your changes match any of the following descriptions?
Summary by CodeRabbit