Sync with cmark-gfm-0.29.0.gfm.11

[digitalmoksha]

This syncs with cmark-gfm-0.29.0.gfm.11 release

[digitalmoksha]

@kivikakk I haven't updated fuzz/quadratic.rs or created fuzz/quadratic_brackets.rs yet.

I also have a small doubt on render.c in github/cmark-gfm@0.29.0.gfm.10...0.29.0.gfm.11. I don't think we need it, or at least I can't find where we would incorporate the changes 🤔

[digitalmoksha]

Woohoo, green pipeline!

[digitalmoksha]

So I think I've made the necessary changes to fuzz/quadratic.rs.

Looking at how you're using Markdown::Markdown, Markdown::Sandwich, and Markdown::Tree, I'm guessing we would want a Markdown::Brackets, rather than creating an entirely separate quadratic_brackets.rs file.

Though I admit my eyes are glazing over right now, and I'm not sure how it should be coded. I'm quite fuzzy on the fuzzer 😬

[kivikakk]

Amazing! Thank you so much!

Looking at how you're using Markdown::Markdown, Markdown::Sandwich, and Markdown::Tree, I'm guessing we would want a Markdown::Brackets, rather than creating an entirely separate quadratic_brackets.rs file.

I'm not sure myself—I didn't write it either. It looks to me like Markdown::Sandwich might actually be that? fuzz_quadratic_brackets.c seems to do something very similar—there's a section repeated some number of times before and after.

I also have a small doubt on render.c in github/[email protected]. I don't think we need it, or at least I can't find where we would incorporate the changes 🤔

This is actually making it clear that there's another difference between Comrak and cmark-gfm, perhaps since that change. See:

$ printf '3. x\n5. z' | cargo run -- --to commonmark
    Finished dev [unoptimized + debuginfo] target(s) in 0.09s
     Running `target/debug/comrak --to commonmark`
3.  x
5.  z

$ printf '3. x\n5. z' | build/src/cmark-gfm -t commonmark
3.  x
4.  z

This difference isn't really able to be captured in the spec, since it only checks HTML output. You could consider Comrak's roundtripping here more information-preserving, and it's not clear that our approach is worse.

You've done really well. Thank you. I'm happy to merge this now if you are.

[kivikakk]

(Also, don't worry about the conflict, I'll resolve it when I merge.)

[digitalmoksha]

You could consider Comrak's roundtripping here more information-preserving, and it's not clear that our approach is worse.

Hmm, you're right. But I think the main reason that I would use the commonmark output option would be to cleanup some text. I think in that case it would be preferable to have the items renumbered.

But it certainly could be looked at either way. I went ahead and opened #323, because I found a problem with the numbering when using task lists. Neither of these are a big deal since the resulting commonmark is valid.

It looks to me like Markdown::Sandwich might actually be that?

I think you're probably right. 😌

I'm happy to merge this now if you are.

Let's do it! 🚀

[kivikakk]

🎉 🎉 🎉 Amazing! Thank you so much! We're finally CI green again! 🍏

[digitalmoksha]

Hey thanks for working through these with me so fast!