chore: improve CJK emphasis handling in markdown processing

[3w36zj6]

In typst-docs, we use pulldown-cmark to parse Markdown. This library faithfully implements the CommonMark specification along with several dialects.

However, since the CommonMark spec is designed for languages that use word separation, its rules for emphasis are quite odd when applied to CJK markup.

Please add the following markup to an appropriate page:

**この強調は認識されません。**この文のせいで。

これは*強調*です。

- 非分かち書きでは**「鍵括弧」**を強調できない
- 非分かち書きでは**（丸括弧）**を強調できない
- 非分かち書きでは**“ダブルクオーテーション”**を強調できない
- 非分かち書きでは**句読点まで、**強調できない
- 非分かち書きでは**とても難しい？**強調

diff --git a/docs/overview.md b/docs/overview.md
index 9afe778c9..87e7d6c97 100644
--- a/docs/overview.md
+++ b/docs/overview.md
@@ -5,6 +5,16 @@ description: |
 
 # Typstについて
 
+**この強調は認識されません。**この文のせいで。
+
+これは*強調*です。
+
+- 非分かち書きでは**「鍵括弧」**を強調できない
+- 非分かち書きでは**（丸括弧）**を強調できない
+- 非分かち書きでは**“ダブルクオーテーション”**を強調できない
+- 非分かち書きでは**句読点まで、**強調できない
+- 非分かち書きでは**とても難しい？**強調
+
 <div class="info-box">
 
 **はじめに: Typst Japanese Communityより**

<div>
<h1>Typstについて</h1>
<p>**この強調は認識されません。**この文のせいで。</p>
<p>これは<em>強調</em>です。</p>
<ul>
<li>非分かち書きでは**「鍵括弧」**を強調できない</li>
<li>非分かち書きでは**（丸括弧）**を強調できない</li>
<li>非分かち書きでは**“ダブルクオーテーション”**を強調できない</li>
<li>非分かち書きでは**句読点まで、**強調できない</li>
<li>非分かち書きでは**とても難しい？**強調</li>
</ul>
</div>

In such cases, to apply emphasis, you need workarounds like inserting spaces or replacing with HTML tags.

To address this, we replaced pulldown-cmark with a custom version which supports markdown-cjk-friendly, so emphasis can be applied naturally without such workarounds.

References

• Markdown における強調の分かち書き #CommonMark - Qiita
• Markdownの太字（強調）記号**がたまに認識されないので緩和用プラグイン作った #JavaScript - Qiita

[Copilot]

Pull Request Overview

This PR improves CJK (Chinese, Japanese, Korean) emphasis handling in markdown processing by inserting zero-width space HTML entities around emphasis markers when they are adjacent to CJK characters without spaces.

• Adds preprocessing to detect CJK characters and insert HTML entities around emphasis markers
• Implements post-processing to remove these entities from final HTML output
• Includes comprehensive handling of code blocks and inline code spans to avoid affecting literal content

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[3w36zj6]

Hi @YDX-2147483647

What do you think about the idea in this PR? Could you please review it?

[YDX-2147483647]

Hmm, I don't think I'm capable of reviewing this PR. I haven't written any parser before, so I can't offer advices better than LLM.

I have indeed been tortured by this markdown parsing problem, but my usual solution is changing the parser (or installing an extension to the parser) rather than patching it on my own…

[3w36zj6]

Thank you for your comment.

I also believe that improving the parser logic is the smart way to address this. Since this is mainly a workaround, I don't feel confident enough to propose it upstream.

At the very least, if we can confirm that it doesn't break the documentation, it would be best to first operate this experimentally as a workaround within the Japanese community.

[kimushun1101]

I understood the situation and was able to verify the behavior.
However, I’m not as good as Copilot at reviewing my own source code.

I measured the time using the following command:
time mise run generate
The build time hasn’t increased noticeably compared to before.

[gomazarashi]

I’m not very familiar with the code itself, but I’ve confirmed that the documentation was generated correctly.
Thank you for addressing this issue.

[ultimatile]

Looks nice, thank you!

[3w36zj6]

pulldown-cmarkをmarkdown-cjk-friendlyのサポートが追加されたカスタムバージョンに変更し、既存のドキュメントのCJK強調の回避策を削除しました。お手数ですが再度レビューをお願いします。

[Copilot]

Pull request overview

Copilot reviewed 2 out of 30 changed files in this pull request and generated 1 comment.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.