Update comrak to 0.2.3, use ext_header_ids #1129
Conversation
|
It occurs to me this won't quite fix #1023; the anchors still won't work because we prefix I'll come up with an appropriate solution shortly; on GitHub we use JavaScript which fires on hash change to search for and scroll to the item with the |
|
So, to prevent ID collisions, we'll also need to sanitize the IDs not coming from comrak-outputted header tags; I've proposed https://github.com/notriddle/ammonia/pull/66 (now merged!) to give us the machinery to do that. Finally, we'll need either the JavaScript portion to scroll the appropriate tag into view (I could throw this together easily based on what we're using here), or we'd need to change all relative |
This is so fragments are scrolled to on load.
|
The JavaScript side of the solution is now in this PR; ready for your thoughts/comments! |
| @@ -53,8 +53,7 @@ impl<'a> MarkdownRenderer<'a> { | |||
| .cloned() | |||
| .collect(); | |||
| let tag_attributes = [ | |||
| ("a", ["href", "target"].iter().cloned().collect()), | |||
| ("code", ["class"].iter().cloned().collect()), | |||
There was a problem hiding this comment.
hm, does allowed_classes make this line unnecessary...? I guess so since the tests pass!
There was a problem hiding this comment.
Indeed, it even became required in the newer ammonia: https://github.com/notriddle/ammonia/blob/2581eec04ef98419fd202e0dd4d8997c9dfdb06e/src/lib.rs#L1030-L1037 (you're allowed to use allowed_classes or set class as an allowed tag, but not both)
|
Looks great, thank you!!! I'm going to review your other readme-related PRs then deploy and regenerate the readmes all at once, that'll probably take me a few days :) Thank you so much for taking care of all these things!!! bors: r+ |
Build succeeded |
|
Thanks so much! |
Fixes #1037. Use latest comrak, and use the header IDs extension to generate GitHub-compatible anchors.
/cc @kejadlen @treiff