Natalie Klestrup Röijezon
Natalie Klestrup Röijezon
That could be fair. Thinking about it a bit more, `discovery.resolve_typemeta(&object)` would have access to everything we care about + enable custom dynamic types.
Oh, I missed that actually (in general, I'd argue that `From`/`TryFrom` tend to surface very poorly in docs).
> And the dispatcher there currently has the setup of not passing up the delete events. Sometimes this is right (watches/owns relation typically don't care about deletes), sometimes it is...
Another piece of prior art would be Pandoc's [`attributes` extension](https://pandoc.org/MANUAL.html#extension-attributes), which reuses the header attribute syntax that pulldown-cmark already supports (`# asdf {.foo}`), so `baz` would be `[baz](foo){.bar}`.
Looks like we never actually deprecated this properly back in 24.11. I would propose rescheduling it for deprecation in 25.3 and removal in 25.3+1.
Marking this as "done" since #813 is merged, but we shouldn't close it before it's removed as well.
Pure Rust would, of course, be nice, but I'm still not _massively_ sold on taking on the new dependency from authors that I'm not really familiar with. In general, I'd...
Makes sense to me. Even if neither library is *that* heavy, it'd be an awkward impedence mismatch for users to deal with.