Filter by Unicode General_Category #35
Closed
Conversation
This file contains 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
jablko
added a commit
to jablko/DefinitelyTyped
that referenced
this pull request
Jan 20, 2021
sandersn
pushed a commit
to DefinitelyTyped/DefinitelyTyped
that referenced
this pull request
Jan 20, 2021
* [README] Correct Unicode slugs * Don't cherry pick Flet/github-slugger#35
|
any chance of merging this @Flet? |
kaznovac
pushed a commit
to kaznovac/DefinitelyTyped
that referenced
this pull request
Mar 2, 2021
* [README] Correct Unicode slugs * Don't cherry pick Flet/github-slugger#35
wooorm
added a commit
that referenced
this pull request
Aug 22, 2021
I reverse engineered GitHub’s slugging algorithm. Somewhat based on #25 and #35. To do that, I created two scripts: * `generate-fixtures.mjs`, which generates a markdown file, in part from manual fixtures and in part on the Unicode General Categories, creates a gist, crawls the gist, removes it, and saves fixtures annotated with the expected result from GitHub * `generate-regex.mjs`, which generates the regex that GitHub uses for characters to ignore. The regex is about 2.5kb minzipped. This increases the file size of this project a bit. But matching GitHub is worth it in my opinion. I also investigated regex `\p{}` classes in `/u` regexes. They work mostly fine, with two caveats: a) they don’t work everywhere, so would be a major release, b) GitHub does not implement the same Unicode version as browsers. I tested with Unicode 13 and 14, and they include characters that GitHub handles differently. In the end, GitHub’s algorithm is mostly fine: strip non-alphanumericals, allow `-`, and turn ` ` (space) into `-`. Finally, I removed the trim functionality, because it is not implemented by GitHub. To assert this, make a heading like so in a readme: `#  `. This is a space encoded as a character reference, meaning that the markdown does not see it as the whitespace between the `#` and the content. In fact, this makes it the content. And GitHub creates a slug of `-` for it. Further work: I think it would be nice to release this as is. Then, afterwards, I’d like to modernize the project, add GH Actions to generate the build, add types, and move to ESM. /cc @Flet @jablkojablko Closes GH-22. Closes GH-25. Closes GH-35. Co-authored-by: Dan Flettre <flettre@gmail.com> Co-authored-by: Jack Bates <jack@nottheoilrig.com>
wooorm
added a commit
that referenced
this pull request
Aug 24, 2021
I reverse engineered GitHub’s slugging algorithm. Somewhat based on #25 and #35. To do that, I created two scripts: * `generate-fixtures.mjs`, which generates a markdown file, in part from manual fixtures and in part on the Unicode General Categories, creates a gist, crawls the gist, removes it, and saves fixtures annotated with the expected result from GitHub * `generate-regex.mjs`, which generates the regex that GitHub uses for characters to ignore. The regex is about 2.5kb minzipped. This increases the file size of this project a bit. But matching GitHub is worth it in my opinion. I also investigated regex `\p{}` classes in `/u` regexes. They work mostly fine, with two caveats: a) they don’t work everywhere, so would be a major release, b) GitHub does not implement the same Unicode version as browsers. I tested with Unicode 13 and 14, and they include characters that GitHub handles differently. In the end, GitHub’s algorithm is mostly fine: strip non-alphanumericals, allow `-`, and turn ` ` (space) into `-`. Finally, I removed the trim functionality, because it is not implemented by GitHub. To assert this, make a heading like so in a readme: `#  `. This is a space encoded as a character reference, meaning that the markdown does not see it as the whitespace between the `#` and the content. In fact, this makes it the content. And GitHub creates a slug of `-` for it. Closes GH-22. Closes GH-25. Closes GH-35. Closes GH-38. Co-authored-by: Dan Flettre <flettre@gmail.com> Co-authored-by: Jack Bates <jack@nottheoilrig.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Would you consider adding these changes to #25/custom-regex-filter? This PR isn't meant to supersede #25, it's a request to pull this PR into that PR. It builds on #25 but filters based on Unicode General_Category vs. Block/Sequence_Property.
What I did was:
test/General_Category/index.js) that loops over each category, as compiled by the unicode-12.1.0 package (excluding union categories, Cased_Letter, Letter, etc.) and writes Markdown files consisting of a heading containing all the code points in that category, and JSON test cases.I started with categories because there are fewer of them than blocks and I did find that most are either all stripped or all kept in the slug. Notable exceptions are:
Other_Symbol is all stripped except for code points with the Alphabetic property, e.g. U+00A6 (and emojis) are stripped but U+24D0 is kept, so in the second commit I reversed the logic in
script/generate-regex.js, now listing categories to keep vs. strip:The resulting regex conforms to all but four of the new test cases:
.trim()is more aggressive than GitHub but I'm not sure that's worth fixing? There's no regression compared to today's behavior and the difference from GitHub's behavior is purely academic?As mentioned in #25 (comment) the motivation for this is e.g. https://github.com/DefinitelyTyped/DefinitelyTyped/blob/master/README.ja.md#型定義ファイルとは何ですか-またどのように入手できますか
github-slugger (master and today's custom-regex-filter branch) doesn't match GitHub's slug because it doesn't strip U+FF1F (Halfwidth and Fullwidth Forms block):