@WG-prioritization/alerts issue #79103 has been requested for prioritization.
so I'm torn on this one, because on the one hand it's causing giant problems for docs.rs and stm32
but on the other hand, it's been like this for ages
for context, I've been talking with @seeplusplus about https://github.com/rust-lang/rust/issues/78761 and I'd guess they're probably 2-3 weeks from having working code
tl;dr rustdoc is a mess and I feel like I'm the only one cleaning it up :(
we appreciate it :)
these "janitorial" tasks are precisely what needs funding instead of volunteer work
I'm looking for jobs
but yeah realistically this is probably not something I can fix on my own, at least any sooner than ~a year from now
This may be ridiculous but maybe there’s some way of doing a joint PR among many people?
Then it wouldn’t fall all on one person.
Well getting rid of doctree can be split in smaller PRs, that's not the issue
It's just that no one else is working on this :/
@Camelid are you volunteering? :P
It also needs a fair amount of design work, mentoring someone to do it would be more work than doing it myself I think
Yeah, I'm not sure if I know enough about rustdoc and its interaction with rustc to work on this.
Is doctree the primary issue here?
Like the html-diff PR came out of the existing tools not being good enough to debug the doctree failures, I wouldn't expect someone not familiar with rustdoc/bootstrap to take that up
But I could try to help a bit :)
No, doctree is just blocking the actual changes I want to do
Which are things like getting rid of clean::Item and getting proper ParamEnvs for normalization
Really DocContext is what's bad, not clean::Item
It means rustdoc is holding all of the info about all of the crate through the life of the program
It's as bad as calling Box::leak
clean even do?
I don't have time right now but ping me tomorrow and I can help
or #rustdoc will know
Issue #79103's prioritization request has been removed.
P-medium. See issue for reasoning.
@Camelid curious to hear your reasoning for P-medium?
P-medium: it's annoying that it takes so much memory, but it's not like you can't document it. And it's being worked on.