Repository navigation
docs(i18n): add Turkish translation of README and quickstart - #3846
berkeagaya1597 wants to merge 2 commits into
Conversation
|
Welcome to the Agent Governance Toolkit! Thanks for your first pull request. |
|
Azure Pipelines: There may be pipelines that require an authorized user to comment /azp run to run. |
PR Review Summary
Verdict: AI review comments are untrusted advisory output. The summary reports workflow-generated completion status only, not model-authored pass/fail claims. |
|
@microsoft-github-policy-service agree |
在本 fork 實測重現:已註冊的 server 用 http:// 被 entry.require_tls 擋下,但同一個 URL 換成未註冊的名字,落到 fallback 路徑就完全沒有 TLS 檢查、直接放行。也就是說 allowlist 的鍵打錯一個字,傳輸層下限就等於關掉——治理工具本身的護欄失效。 落地照上游:McpAuthPolicy 新增 default_require_tls=True(fail-closed)、fallback 分支補上與已註冊條目相同的 scheme 檢查(只認 https/wss)、from_yaml 讀該欄位。 只在真的有給 URL 時才擋,與已註冊條目一致,所以不誤擋從不傳 URL 的呼叫端。 測試四條(含上游沒有的「未註冊+不給 URL 仍放行」),21 passed。 同輪另三筆: - microsoft#3848(OPA timeout 測試 30ms 太緊)採用:假 opa sleep 30 秒,30ms 預算可能在 行程還沒 spawn 完就到期,失敗的是 spawn 不是逾時,測到的不是這條要測的東西。 本 fork 的 ci.yml 與 policy-engine-ci.yml 都跑 cargo test --workspace,opa.rs:254 就是同一行。驗證限制照實說:本機沒裝 cargo,這一行沒實跑過,靠 CI。 - microsoft#3846(土耳其文翻譯)不引用:本 fork 沒有 docs/i18n/,維護一份讀不懂也無法審校的 翻譯只會變成長期漂移的死文件。 - microsoft#3850(清理 python-app.yml 的 flake8)不引用:本 fork 沒有那支 workflow。 水位 PR 3845 → 3850。順帶把 test_fork_overlay 釘字面日期與字面水位的斷言改成釘性質 (第五個 repo 出現同一條)。 驗證:compileall / ruff / fork tests / 連結檢查 / mcp_auth 測試全部 exit 0。 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
上游是 microsoft 的高速開發線,本輪處理完 microsoft#3846–microsoft#3850 之後同一天內又出現兩筆。 microsoft#3851(DecisionAssure Impact 治理變更影響分析引擎):OPEN、100 檔 +8742/-434 的 新功能。不引用——是提案不是上游狀態,且本 fork 沒有現在就在痛的缺陷需要它;提前 引用一個 100 檔的未採納功能等於扛下一條上游可能不會走的分支。 microsoft#3852(放寬 serde/serde_json/thiserror 的精確釘版):+4/-4。上游的理由成立,但那是 「發佈者」的理由——它說發佈到 crates.io 的 library crate 用精確釘版會鎖死下游的 Cargo.lock。本 fork 不是發佈方:publish.yml 有 9 處 official-repo guard、 publish-containers.yml 1 處,發佈流程全被閘門擋掉,沒有下游消費者受影響。而放寬釘版 對本線是負向的:精確釘版讓建置可重現,放寬後同一個 commit 在不同時間會拉到不同 patch 版,反而讓 dependency-freshness 的比對失去基準。 另記一句給後人:水位代表「到某個編號為止已逐筆看過」,不是「以後不會再有」。之後的 新項目由每週排程接手,不需要在同一輪裡追到底。 驗證:fork tests / 連結檢查 exit 0。 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Imran Siddique (imran-siddique)
left a comment
There was a problem hiding this comment.
The translation itself follows the house pattern correctly: frontmatter present, the
"this is a translation, check the English for the latest" banner at the top, the language switcher
updated inside the translated file, and product names, commands and API identifiers left in English
as docs/i18n/README.md asks for. That is more of the convention than most translation PRs get right
first time.
Three integration points are missing, and they are all in the checklist the repo already publishes.
docs/i18n/README.md is not updated. That file carries the Language Availability table and its
own contributing steps say, at step 3, to add the translated files to it. Turkish is absent, so the
i18n index will not list a translation the root README links to. Two rows, README and quickstart,
both "Available".
mkdocs.yml has no Turkish entry. Lines 115 to 124 carry one link: per language
(i18n/README.ja/, i18n/README.ko/, i18n/README.zh-CN/, i18n/README.zh-TW/). Without a matching
entry the Turkish pages ship in the built site but are unreachable from its navigation.
This conflicts with #3673. That PR adds a Spanish translation and edits the exact same two lines
of the root README.md, the language bar at line 1 and the "Also in:" line at 211. Whichever merges
second will conflict. Worth knowing now rather than at the merge button, and the resolution is
trivial in both directions: both languages belong on both lines.
On the spell-check ignore, which is the interesting difference between the two PRs. Your approach
and #3673's are not the same: you add **/docs/i18n/*.tr.md to ignorePaths in .cspell.json, and
#3673 adds a term to .cspell-repo-terms.txt. Nothing on main handles this today, because
.cspell.json has no i18n entries at all and the existing translations are all CJK, which cspell does
not flag. Yours is the first Latin-script translation to hit the problem, so whatever lands here sets
the pattern.
For what it is worth, I think ignoring the path is the right answer and per-word dictionary entries
will not scale, but the pattern should be **/docs/i18n/*.md rather than *.tr.md. Spanish is
already queued in #3673 and the index file names French, German, Portuguese and Hindi as wanted, so a
language-specific line means a new entry per language forever, and each one is a shared append point
that conflicts with the next translation PR. One path covers all of them and this PR would be the
natural place to establish it.
One pre-existing thing this PR sits next to, not yours to fix: Traditional Chinese is complete on
main (README.zh-TW.md, quickstart.zh-TW.md), is listed in the i18n index, and is in the mkdocs
nav, but it is missing from the root README language bar, which is the line you are editing. So the
bar is already one language behind. Adding 繁體中文 while you are in there would close it, if you
would rather not leave a known gap in a line you touched.
Nothing blocking on the translation.
MohammadHaroonAbuomar
left a comment
There was a problem hiding this comment.
- Commit 00c8a91 has no Signed-off-by trailer. The DCO check has not run yet because the fork workflows are gated, but it will fail once a maintainer approves them. Please
git rebase --signoffand force-push; everything else in this PR is already verified and this is the only remaining item.
Signed-off-by: Muhammet Berke Ağaya <berke.agaya@setur.local>
00c8a91 to
0200941
Compare
|
Added the missing Signed-off-by trailer and force-pushed the updated commit. Thanks for the review. |
|
Thanks for your contribution; we're not taking this change, so I'm closing it. |
7. PR açıklaması (kopyalayıp kullanın)