zh-CN: complete the Day 1-4 translation - #3256
Open
TianfangChang wants to merge 10 commits into
Open
Conversation
…ng/lifetimes translations
Agenda lists and course-structure tables rendered in English because their
entries were either fuzzy, empty, or filled with the English source verbatim
(msgstr == msgid), all three of which mdbook-i18n-helpers skips.
Translate all 146 of them, reusing the chapter titles already translated for
SUMMARY.md and the existing "（5 分钟）" / "（1 小时 5 分钟）" duration format.
Also fixes stale translations these entries carried:
- [Lifetimes](../lifetimes.md) was translated as [引用](../references.md)
- several durations left over from an older schedule (20/55/1 小时)
- [Exercise: Geometry] rendered as 练习：面向 GUI 库的模块
- a few agenda slots holding an unrelated sentence about pattern matching
Other correctness fixes:
- SUMMARY.md "Copy Types" was 复合类型 ("compound types") -> Copy 类型
- Labels -> 标签, Matching Values -> 匹配值, Source -> 源代码
- memory-management/move.md svgbob diagram: the s1 "（无法访问）" row was
8 characters short, breaking the dashed box border
- borrowck.md: _借用检查器_ glued to CJK on both sides is not emphasis in
CommonMark; spaced out to match the rest of the file
msgids are unchanged.
…y 1-4)
Mechanical pass over the 120 Day 1-4 entries that needed no translation
judgement. Two shapes:
42x the translation was already identical to the source, but carried a
trailing \n the msgid does not have -- that injects a blank line into
the rendered code block. Strip it and drop the fuzzy marker.
72x code comments and string literals whose msgstr had drifted onto some
unrelated entry during past merges, e.g.
"Alice" -> 切片
"Optional box" -> 异常
"Calling function on {input}"-> 调用 Unsafe 函数
"Size of none: {}" -> "before: {a}"
// undefined behavior. -> Rust 没有运行时未定义行为：
Restored to the source text, matching the convention already used by
the other ~830 code-string entries in this file.
Six were prose that the code-string heuristic caught by mistake and are
translated instead:
Lifetimes 结构体生命周期 -> 生命周期 (matches the agenda entries)
Supertraits 更多 trait -> 超特征
Option 异常 -> Option
Testing. -> 测试。
`Iterator` implements -> `Iterator` 实现了
`PartialEq` and `Eq` / `PartialOrd` and `Ord` were already correct; only
the fuzzy marker was dropped.
Verified: every rewritten msgstr equals its msgid byte for byte apart from
those six, no entry left fuzzy, msgids unchanged.
Day 1-4 remaining: 366 -> 246 (all prose).
The remaining 13 Day 3 entries, all previously untranslated: 7 borrowing/interior-mutability.md speaker notes and body text 2 smart-pointers/box.md notes on the null pointer optimization 1 lifetimes/lifetime-elision.md speaker note 2 the "Including 10 minute breaks, ..." session timings 1 smart-pointers/trait-objects.md svgbob diagram Terminology follows the Day 3 translations already live in this file: 共享引用 / 独占引用 / 内部可变性 / 借用规则 / 所有权. The trait-objects diagram was fuzzy because the upstream art was redrawn; the stale msgstr was the previous English diagram. Redrawn from the new source with only the two floating labels translated (Stack/Heap -> 栈/堆, "Program text" -> "程序文本"), matching the sibling diagrams in move.md and box.md. Neither label sits on a box border, so every border row is still 81 characters wide. Emphasis markers are spaced away from adjacent CJK (_安全_, _始终_); glued to a Han character on both sides CommonMark does not treat `_` as emphasis. Verified: msgids unchanged, no line over 80 columns, po syntax clean. Day 3 remaining: 13 -> 0. Day 1-4 remaining: 246 -> 233.
… unsafe) The remaining 53 Day 4 entries: 37 stale fuzzy translations rediffed against the current English, 16 never translated. Notable corrections in the stale ones -- these were rendering as English, so the wrong text was not visible, but it was wrong: Welcome to Day 4 欢迎来到第一天 -> 欢迎来到第四天 Cargo Setup 设置 -> Cargo 设置 anyhow::Error is essentially a wrapper.. ”是“Result\<V, anyhow::Error>”的类型别名。 anyhow::Result<V> is a type alias.. “anyhow::Result Actual error type inside of it.. “anyhow::Error”本质上是“Box The three anyhow entries had their translations shifted one entry out of step, so each carried the neighbour's text. Retranslated from the current English. Also fixes four rendering defects in the FFI conversion list in unsafe-rust/exercise.md, which were live (not fuzzy) and so shipping broken: `\*const i8` backslash inside a code span renders literally -> `*const i8` `&\[u8\]` same -> `&[u8]` “&OsStr” curly quotes where the source has backticks -> `&OsStr` plus a stray space before the full-width colon in three items thiserror-and-anyhow.md and iterators/fromiterator.md were split/renamed upstream (into thiserror.md + anyhow.md, and collect.md); their msgids are still live in this po, and roughly half the sentences survived the move verbatim, so they are translated here rather than skipped. Verified: msgids unchanged, no line over 80 columns, po syntax clean. Day 4 remaining: 53 -> 0. Day 1-4 remaining: 233 -> 180.
The remaining 83 Day 2 entries: 59 stale fuzzy translations rediffed against
the current English, 24 never translated.
Several were shifted onto a neighbouring entry, so the Chinese described
something else entirely:
Associated Types 共享类型 -> 关联类型
Implementing Traits 实现 Unsafe Trait -> 实现特征
Generic Traits 泛型 -> 泛型特征
`let else` expressions `while let` 表达式 -> `let else` 表达式
`|` as an `or` 解释模式匹配中的绑定的原理…
`..` can expand as much as… 你可以展示如何匹配一个引用。
`1..=5` represents an inclusive range 现在是一个讲解不可反驳模式的好时机…
Show students the `Ord` trait… [“OsStr”](…)和[“OsString”](…)
Factual fixes:
`Result` is the standard type…as we will see on Day 4 第 3 天 -> 第 4 天
The `_` pattern…expressions _must_ be exhaustive 不可反驳 -> 穷尽的
The partial struct initialization… 结构体副本 -> 部分初始化
Unlike `match`, `if let` does not have to cover all
branches the old text claimed
if let lacks guard
clauses, unrelated
Also repairs std-traits/default.md, which was live rather than fuzzy and so
was rendering mangled backticks:
标准的 Rust 类型通常会以合理的值（例如 ` 0`\"\" \``等）实现`Default\``。
-> 标准的 Rust 类型通常会用合理的值实现 `Default`（例如 `0`、`\"\"` 等）。
and std-types/string.md, where `&\*s1` had a backslash inside a code span and
the surrounding code spans used curly quotes.
Verified: msgids unchanged, po syntax clean. Four new lines exceed 80 columns;
each is a single unbreakable markdown link, same as the 14 already in the file.
Day 2 remaining: 83 -> 0. Day 1-4 remaining: 180 -> 97 (Day 1 only).
The file mixed both second-person forms: 99 你 against 230 您, the latter mostly left over from the original machine translation. Every translation written on this branch uses 你, so normalise the rest to match. Mechanical replacement of every 您 with 你. 你 and 您 are both full-width, so no line needed rewrapping. 您 appears in no msgid, and the 41 occurrences outside live msgstr are in obsolete #~ entries, replaced for consistency in case msgmerge ever revives one. 271 occurrences, no other change. msgids unchanged, po syntax clean.
…d types)
The remaining 97 Day 1 entries: 77 stale fuzzy translations rediffed against
the current English, 20 never translated. Day 1-4 is now fully translated.
Entries whose translation had drifted onto an unrelated one:
The [`for` loop] iterates over ranges… [`Drop`] 用于定义析构函数。
The [`loop` statement] just loops… [`Drop`] 用于定义析构函数。
`println!(format, ..)` prints a line… 运算符重载是通过 [`std::ops`] 中的特征实现的：
`s` is a reference to a slice of `i32`s 切片会从另外一个对象中借用数据…
Slices always borrow from another object 关于修改 `a[3]` 的问题可能会引发…
What is the type of this variable? 每个循环中的“word”是什么类型？
Welcome Back 欢迎
Factual fixes:
think of `&str` as `std::string_view` old text said `const char*`
…wraps in a release build 封装 (wrapping/boxing) -> 回绕
`static` is similar to mutable global
variables in C++ old text said it resembles `const`
Be sure to note the difference between
`let mut x_coord` and `let x_coord: &mut` old text called the first one a
mutable reference
Rust is very much like other languages… old text added 「而非函数式」, not
in the source
Rust uses macros…[overloading](…) link pointed at a path that has not
existed since the 2023 restructure
Also makes the two timing strings consistent across all four days, which
previously both rendered as 本节:
"In this session:" / "this session should take about" -> 本场 (half-day)
"In this segment:" / "This segment should take about" -> 本节 (chapter)
That touches six entries outside Day 1 which were translated earlier on this
branch.
Verified: msgids unchanged, po syntax clean. Two new lines exceed 80 columns,
both a single unbreakable URL.
Day 1 remaining: 97 -> 0. Day 1-4 remaining: 97 -> 0.
|
Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA). View this failed invocation of the CLA check for more information. For the most up to date status, view the checks section at the bottom of the pull request. |
These 86 entries were neither fuzzy nor empty, so they never showed up as outstanding work — they were already "translated" and were being rendered, just rendered wrong. 75 entries / 128 occurrences: a code span in the source had become a pair of Chinese quotation marks in the translation, so it rendered as proportional text instead of monospace. EN: Only functions marked `const` can be called at compile time… ZH: 在编译时只能调用标记为“const”的函数… -> 标记为 `const` 的函数 Applied only where the quoted text appears verbatim as a code span in the msgid, which a genuine Chinese quotation never does; `“2”`, `“=>”` and similar were left alone because the msgid has no matching span. 5 entries: a markdown link had been pulled apart so nothing rendered as a link. ZH: 适用于 [`rust_static_library` gn 模板] 的选项 (https://…) -> [`rust_static_library` gn 模板](https://…)可用的选项 memory-management/review.md had lost the System Allocator URL entirely and kept only \[系统分配器\]. 3 entries: unbalanced backticks (`Cargo.toml``, `next() 及其, `LinesReader). 3 entries: `_x_` with a Han character on both sides, which CommonMark does not treat as emphasis, so the underscores rendered literally. async/runtimes.md was worse — both closing underscores were missing, so the surviving pair italicised everything between 反应器 and 执行器. Found by scanning all 3301 rendering entries for markdown that differs structurally from its msgid. Verified against a local zh-CN `mdbook build`: `标记为<code>const</code>的函数`, `<em>降级</em>`, and the System Allocator anchor all render correctly now. Of the 86, 72 correspond to strings still present in the source; the other 14 are in entries the source has since dropped. msgids unchanged, po syntax clean.
TianfangChang
marked this pull request as ready for review
August 1, 2026 05:00
TianfangChang
requested review from
AgainstEntropy,
candysonya,
superwhd and
wnghl
as code owners
August 1, 2026 05:00
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This is translation work only — no
msgmergewas run, and no msgid changed.Days 1–4 of the course (
welcome-day-1throughunsafe-rust) are now fullytranslated: the 612 entries that were fuzzy, empty, or filled with the English
source are down to 0. Android / Chromium / bare-metal / concurrency / async are
untouched and still have 482 to go.
896 entries changed in total.
Why so much was rendering as English
Three causes, all of which make
mdbook-i18n-helpersfall back to the msgid:Fuzzy entries whose English had been reworded. Many were not merely stale
but shifted — the msgstr held a neighbouring entry's text, so unfuzzying
them blindly would have shipped the wrong Chinese. Every one was rediffed
against the current English. A sample:
The [forloop] iterates over ranges…[Drop] 用于定义析构函数。Associated Types共享类型Welcome to Day 4欢迎来到第一天Show students theOrdtrait…[“OsStr”](…)和[“OsString”](…)Empty entries.
Entries whose msgstr was the English source verbatim. These render exactly
like an empty entry but count as translated, so the translation report
overstated coverage and
msgmergewould never flag them again.A separate class: entries that rendered, but rendered wrong
The last commit deals with 86 entries that were neither fuzzy nor empty, so they
never appeared as outstanding work:
“const”ratherthan
`const`), rendering as proportional textURL entirely
_x_flanked by Han characters, which CommonMark does not treat asemphasis; in
async/runtimes.mdboth closing underscores were missing, so thesurviving pair italicised a whole clause
Also fixed earlier in the branch:
`\*const i8`and`&\[u8\]`(abackslash inside a code span renders literally), a row of the
memory-management/move.mdsvgbob diagram that was 8 characters short and brokethe dashed border, and
`&str`described as C++'sconst char*rather thanstd::string_view.Consistency
你throughout, rather than a 99/230 split with您SUMMARY.mdtranslationsChecks
Verified with a local
MDBOOK_BOOK__LANGUAGE=zh-CN mdbook build(exit 0) —标记为<code>const</code>的函数,<em>降级</em>and the restored SystemAllocator anchor all render correctly.
msgfmtwas not available, so the filewas additionally validated by script: msgid set byte-for-byte unchanged, every
string literal correctly escaped, link targets in each msgstr matching its
msgid.
i18n-reportparses it cleanly and agrees on the counts.dprintno longer formats.pofiles — the exec command was dropped in #2381 —so nothing here was run through
dprint fmt. TRANSLATIONS.md still tellstranslators to do so; happy to send a separate PR fixing that.
Worth knowing: this PO file is two years behind
po/zh-CN.powas last synced on 2024-04-15. Regenerating the POT from thecurrent source with
mdbook-xgettextgives:po/zh-CN.poSo 353 of the 896 entries touched here are for strings the source has since
dropped, and they will become obsolete on the next
msgmerge. The largestsingle group is the agenda lists:
- [Loops](./control-flow-basics/loops.md) (5 minutes)no longer exists anywhere insrc/— those tables are generated by{{%segment outline}}now, and the duration column is emitted in English by theplugin rather than going through the PO file at all.
I have deliberately not run
msgmergehere, per TRANSLATIONS.md. But zh-CNwould benefit a lot from a sync, and the generated agenda tables may be worth a
look as a translation gap in their own right.
🤖 Generated with Claude Code