Update pulldown-cmark dependency to 0.10

[max-heller]

Version 0.10 of pulldown-cmark was just released with a number of breaking changes. Unfortunately, it doesn't seem straightforward to update pulldown-cmark-to-cmark's dependency on it. This is a start, but there are some breaking changes that will have to be worked around (compilation currently errors)

[liningpan]

Do you have any plans for fixing those breaking changes? It seems like upstream removed some information from the end tags to save some memory allocations pulldown-cmark/pulldown-cmark#517. I suppose it means that we have to push start tags onto a stack as our internal state.

[max-heller]

Do you have any plans for fixing those breaking changes? It seems like upstream removed some information from the end tags to save some memory allocations pulldown-cmark/pulldown-cmark#517. I suppose it means that we have to push start tags onto a stack as our internal state.

pushed a commit that introduces some additional stacks to keep track of links/images, and the current heading. need to fix compilation of the tests to see if it works

[max-heller]

updated most of the tests but tests/display.rs is tricky because it attempts to format end tags on their own, which now do not contain the necessary information

[TerakomariGandesblood]

Any updates here?

[max-heller]

Any updates here?

Mainly not sure how to update some of the tests, @Byron do you have any ideas for the end tag tests?

[Byron]

It looks like only a single test is failing, which lets me hope this PR is close to merging actually.

updated most of the tests but tests/display.rs is tricky because it attempts to format end tags on their own, which now do not contain the necessary information

Is the intent of the failing test clear? If it is, can it be adjusted to deal with the changes made here? Could the same thing be tested differently?
If the intent is not clear, it could possibly be removed and replaced with something that makes more sense. I still hope that the test is protecting something valuable though, and that it can be salvaged.

[max-heller]

Adjusted the end tag display tests and fixed the failing test, this needs some cleanup to avoid tons of extra allocations but it's passing the tests!

[max-heller]

Is there a reason State isn't allowed to borrow from events? It's constrained to State<'static> in many APIs, but this doesn't seem necessary, and we could avoid allocating a ton of strings if we let it borrow. Taking in Iterator<Borrow<Event>> is also somewhat restrictive compared to Iterator<Event>, since we can't reuse owned strings from events

[Byron]

Thanks a lot, great work!

The spec-tests improved conformity by 4 as well, and I appreciated the added test to check support for ids and attributes as well.

A new release is now available: https://github.com/Byron/pulldown-cmark-to-cmark/releases/tag/v12.0.0