Reflect #9471, #9692, and #9695 in 3.1.1 document

website/versioned_docs/version-3.1.1/guides/markdown-features/markdown-features-react.mdx

@@ -25,6 +25,12 @@ Use the **[MDX playground](https://mdxjs.com/playground/)** to debug them and ma
 
 :::
 
+:::info
+
+Prettier, the most popular formatter, [supports only the legacy MDX v1](https://github.com/prettier/prettier/issues/12209). If you get an unintentional formatting result, you may want to add `{/* prettier-ignore */}` before the problematic area, or add `*.mdx` to your `.prettierignore`, until Prettier has proper support for MDX v3. [One of the main authors of MDX recommends `remark-cli` with `remark-mdx`](https://github.com/orgs/mdx-js/discussions/2067).
+
+:::
+
 ### Exporting components {#exporting-components}
 
 To define any custom component within an MDX file, you have to export it: only paragraphs that start with `export` will be parsed as components instead of prose.

website/versioned_docs/version-3.1.1/migration/v3.mdx

@@ -426,6 +426,79 @@ console.log('hello');
 
 :::
 
+### Other Markdown incompatibilities
+
+#### Emphasis starting or ending with a space or a punctuation
+
+New MDX parser now strictly complies with the CommonMark spec. CommonMark spec has introduced rules for emphasis around spaces and punctuation, which are incompatible especially with languages that do not use a space to split words, since v0.14.
+
+Japanese and Chinese are most affected by this, but there are some other languages that can be affected (e.g. Thai and Khmer), for example when you try to emphasize an inline code or a link. Languages that use a space to split words are much less affected.
+
+`**` (other than `` `**` ``) in the following example were parsed as intended in Docusaurus 2, but are not now in Docusaurus 3.
+
+{/* For Chinese translators: you can translate the following Japanese into Chinese. */}
+
+{/* prettier-ignore */}
+```md title="example.md"
+**Do not end a range of emphasis with a space. **Or `**` will not work as intended.
+
+<!-- Japanese -->
+**「。」の後に文を続けると`**`が意図した動作をしません。**また、**[リンク](https://docusaurus.io/)**や**`コード`**のすぐ外側に`**`、そのさらに外側に句読点以外がある場合も同様です。
+```
+
+<details>
+<summary>See the detailed conditions and how to upgrade</summary>
+
+If `*` or `**` matches either of the following conditions, it will not work as the beginning of an emphasis mark anymore:
+
+- The next character is a space (e.g. `word* word`)
+- The previous character is a punctuation character and the next character is a letter (not a space or punctuation character) (e.g. `文**（文）`)
+
+On the contrary, if `*` or `**` matches either of the following conditions, it will not work as the end of an emphasis mark anymore:
+
+- The previous character is a space (e.g. `word *word`)
+- The next character is a punctuation character and the previous character is a letter (e.g. `文。**文`)
+
+“A punctuation character” includes non-ASCII ones, brackets, quotation marks and some symbols including `%` and `@`. More strictly speaking, a character whose 2-letters Unicode category starts with `P` is treated as a punctuation character here.
+
+:::tip How to upgrade
+
+If the offending emphasis mark is next to a space, move the space out of the range of emphasis:
+
+```md title="english.md"
+**Do not end a range of emphasis with a space.** Or `**` will not work.
+```
+
+If the offending emphasis mark is surrounded by both a punctuation character and a letter, you can fix it without modifying the content by:
+
+1. Convert the document to MDX if it is a vanilla Markdown.
+2. replace the offending emphasis mark with a raw HTML tag (`<em>` or `<strong>`) instead:
+
+{/* prettier-ignore */}
+```mdx title="japanese.mdx"
+<strong>「。」の後に文を続けると`**`が意図した動作をしません。</strong>また、<strong>[リンク](https://docusaurus.io/)</strong>や<strong>`コード`</strong>のすぐ外側に`**`、そのさらに外側に句読点以外がある場合も同様です。
+```
+
+While not an ideal solution, you can also either of the following without converting the document to MDX:
+
+- Move the most outside punctuation character out of the emphasis mark.
+
+  {/* prettier-ignore */}
+  ```md title="japanese.md"
+  **「。」の後に文を続けると`**`が意図した動作をしません**。また、[**リンク**](https://docusaurus.io/)や・・・
+  ```
+
+- Put a space just outside of the offending `*` or `**`. This solution does not force you to convert the document to MDX.
+
+  {/* prettier-ignore */}
+  ```md title="japanese.md"
+  **「。」の後に文を続けると`**`が意図した動作をしません。** また、**[リンク](https://docusaurus.io/)** や **`コード`** のすぐ外側に`**`、そのさらに外側に句読点以外がある場合も同様です。
+  ```
+
+:::
+
+</details>
+
 ### MDX plugins
 
 All the official packages (Unified, Remark, Rehype...) in the MDX ecosystem are now [**ES Modules only**](https://gist.github.com/sindresorhus/a39789f98801d908bbc7ff3ecc99d99c) and do not support [CommonJS](https://nodejs.org/api/modules.html#modules-commonjs-modules) anymore.
@@ -464,6 +537,21 @@ If you created custom Remark or Rehype plugins, you may need to refactor those,
 
 :::
 
+### Formatters
+
+Prettier, the most common formatter, supports only the legacy MDX v1, not v3 yet as of Docusaurus v3.0.0. You can add `{/* prettier-ignore */}` before the incompatible parts of your code to make it work with Prettier.
+
+```mdx
+{/* prettier-ignore */}
+<SomeComponent>Some long text in the component</SomeComponent>
+```
+
+If you get tired of too many `{/* prettier-ignore */}` insertions, you can consider disabling MDX formatting by Prettier by adding the following to your `.prettierignore` file, until it starts supporting MDX v3:
+
+```txt title=".prettierignore"
+*.mdx
+```
+
 ## Other Breaking Changes
 
 Apart the MDX v3 upgrade, here is an exhaustive list of breaking changes coming with Docusaurus v3.
@@ -689,6 +777,8 @@ If you previously used the undocumented `:::warning` admonition, make sure to ve
 
 Docusaurus v3 also [deprecated the `:::caution`](https://github.com/facebook/docusaurus/pull/9308) admonition. Please refactor `:::caution` (yellow) to either `:::warning` (yellow) or `:::danger` (red).
 
+If you want to keep the title “caution”, you might want to refactor it to `:::warning[caution]` (yellow).
+
 :::
 
 ### Versioned Sidebars