Valid fragments w/ URL-encoded emoji false positive

[adamlui]

like https://github.com/adamlui/js-utils/actions/runs/23705290073/job/69056053612

[katrinafyi]

Investigating and there seems to be a problem with the "🏛️ MIT 许可证" header link at https://github.com/adamlui/js-utils/blob/main/minify.js/node.js/docs/zh-cn/README.md#%EF%B8%8F-mit-%E8%AE%B8%E5%8F%AF%E8%AF%81. It works in browser but not in lychee.

[a](https://github.com/adamlui/js-utils/blob/main/minify.js/node.js/docs/zh-cn/README.md#%EF%B8%8F-mit-%E8%AE%B8%E5%8F%AF%E8%AF%81)

$ cargo run b.md --include-fragments -vvvv     
 [DEBUG] Applied quirk 'fetch raw GitHub Markdown files' to https://github.com/adamlui/js-utils/blob/main/minify.js/node.js/docs/zh-cn/README.md#%EF%B8%8F-mit-%E8%AE%B8%E5%8F%AF%E8%AF%81
 [ERROR] https://github.com/adamlui/js-utils/blob/main/minify.js/node.js/docs/zh-cn/README.md#%EF%B8%8F-mit-%E8%AE%B8%E5%8F%AF%E8%AF%81 (at 2:1) | Cannot find fragment
  [WARN] There were issues with GitHub URLs. You could try setting a GitHub token and running lychee again.
Issues found in 1 input. Find details below.

[b.md]:
 [ERROR] https://github.com/adamlui/js-utils/blob/main/minify.js/node.js/docs/zh-cn/README.md#%EF%B8%8F-mit-%E8%AE%B8%E5%8F%AF%E8%AF%81 (at 2:1) | Cannot find fragment

The percent encoded string is really odd:

from urllib.parse import *
unquote('%EF%B8%8F-mit-%E8%AE%B8%E5%8F%AF%E8%AF%81').encode('unicode_escape')
# b'\\ufe0f-mit-\\u8bb8\\u53ef\\u8bc1'

The Chinese characters after mit are fine but the first character is FE0F VARIATION SELECTOR-16. This is the emoji-presentation selector which is appended after emojis to request emoji display (see also). It seems that the emoji is stripped but the variation selector is kept.

Is there some specification about how to derive fragment IDs from markdown? Edit: found github/cmark-gfm#361

I've also found other emoji which are kept in the fragment... https://gist.github.com/katrinafyi/c5f767d2952ae2ac2a3059b267325334#%F0%9F%85%B0%EF%B8%8F-d