docs(zh): proofread README_zh.md — rewrite the stale TLS note and six smaller fixes

[t957095]

Native-speaker proofread of README_zh.md against README.md, per #974. This supersedes #982 (both of its fixes are included here).

Confirmed error — the transport-security note contradicted the English original. It claimed the SSL/TLS panel applied to nine listed engines and that Oracle, MongoDB and Redis ignored the setting. Current README.md says the panel is honoured by every engine except the three file-based ones (SQLite, DuckDB, embedded LibreDB), with an Oracle Thin-driver certificate-chain caveat, and lists libSQL among the connection-string engines. The note is rewritten to match; libSQL is restored.

Smaller fixes:

• Trino row: 连接所固定 catalog → 连接所固定的 catalog
• Agent limits: 运行还没有建立 → 根本不会创建运行 (the run is never opened)
• 只读执行档案 → 只读执行配置 (a read-only execution profile, not an archive)
• 注意开关不是密钥 → 注意，密钥并不是开关
• 本地邮箱密码 → 本地邮箱/密码登录
• CSP 是允许内联脚本的 → CSP 允许内联脚本
• Monaco 的产物不同源时 → Monaco 的 bundle 与站点不同源时
• restored the missing PostgreSQL Clients link

bun run readme:check passes; engine names and install commands untouched.

Closes #974

[InnoxCodes]

Went through this against the English README on main. The rewritten transport-security note now lines up with the English one point for point: the connection-string engines are MongoDB, Couchbase, ClickHouse and libSQL, the SSL/TLS panel applies to every engine except the three file-based ones (SQLite, DuckDB, embedded LibreDB), Trino is load-bearing, and the Oracle Thin driver caveat about supplying the CA when require is used is there too.

I also checked a few of the smaller fixes. The English says "a key is not the switch", so the old 注意开关不是密钥 really had it the wrong way round. "Read-only execution profile" is 配置 rather than 档案. And "refused when it is started, before a run is opened" does mean the run is never created. node scripts/readme-check.mjs passes on this branch, and the PostgreSQL Clients link matches the one in the English README.

Two housekeeping notes. #982 is still open and this PR says it supersedes it, so #982 can probably be closed once this lands. And the same PostgreSQL Clients link is also missing from README_ja.md, README_es.md, README_ur.md and README_hi.md, so it may be worth checking whether the other language proofreads under #974 need it too.

[t957095]

Thanks for the thorough proofread — glad the transport-security rewrite lines up point for point with the English original, and that the three smaller wording fixes check out.

Both housekeeping notes are now handled:

1. docs(zh): proofread README_zh.md — fix two confirmed errors #982 — closed. This PR supersedes it, as you suggested.
2. Missing PostgreSQL Clients link — checked every README in the repo. Only README_ja.md, README_es.md, README_ur.md and README_hi.md were missing it (README_zh.md gets the link from this PR, so I left it alone to avoid a conflict). Fix is up as docs(readme): add missing PostgreSQL Clients link to ja/es/ur/hi translations #1012, with node scripts/readme-check.mjs passing.

[cevheri]

Thanks, just a quick reminder: could you leave a comment so I can assign this issue to you.
issue number: #974

[t957095]

Pushed a fix for the failing chart:check: merged current main into this branch (clean merge, no conflicts), bringing the chart to 0.1.66 / appVersion 0.16.1 so the libredb-studio-0.1.65 released-tag guard no longer trips.

Both node scripts/sync-chart-version.mjs --check and node scripts/readme-check.mjs pass on the pushed commit (9032b50). The new CI run is currently awaiting maintainer approval (action_required) — could someone hit Approve and run on the Checks tab? The red ❌ still shown on the PR is from the previous head.

[cevheri]

Thanks for this. I checked the rewritten transport-security note against README.md line by line and it matches: the connection-string engines, the three file-based exceptions, Trino being load-bearing, and the Oracle Thin CA caveat. The 档案 to 配置 fix and the 密钥 / 开关 direction were real errors. I will merge it once two things are in.

1. The Trino row still reads 连接所固定的 catalog information_schema schema 树. Moving the 的 left the noun phrase without one, so please make it 连接所固定的 catalog 的 information_schema schema 树.
2. The same row is missing an item the English one has: EXPLAIN (FORMAT JSON) plan trees. Please add it, since this PR is a proofread against the English original.

One friendly note, not a complaint. I look at the profile behind every PR, and yours shows a burst of around twenty PRs across fourteen freshly forked repositories in a few hours. Please do not point that kind of volume at this repo. We are a small team and reviewing a PR properly costs us real hours, so a flood is something we cannot absorb. Using AI to write a patch is fine by us. Running it in a loop or a job that opens PRs at machine speed is not. One considered PR at a time is very welcome here.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[t957095]

Both items are in, pushed as f11a221 on this branch:

1. The Trino row now reads 连接所固定的 catalog 的 information_schema schema 树 — the 的 is back where the noun phrase needs it.
2. EXPLAIN (FORMAT JSON) 计划树 added in the same slot it occupies in the English row (after the full-SQL-IDE item, before the information_schema item), so the two rows line up item for item again.

And thank you for being straight with me about the PR volume. You're right, and I owe you an apology. I had an automation script that got out of hand — it opened PRs at machine speed across a dozen-plus freshly forked repos in a few hours before I reined it in. Pointing that at a small team's repo was not fair to your review time. The loop is stopped, and going forward it's one repo at a time, one considered PR at a time, each read against the English original before it goes up. Nothing else will land here until this one is merged or closed.