New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bump versions to 0.8.3 (and fix some clippy complaints) #645
Conversation
Maybe stating the obvious, but there's a bunch of breaking changes in I've slightly lost track of everything, but I did have a branch set up for this which I think is still good to use: The changes should fix both the docs.rs build, and using freetype via crates.io (also backports #624 which updated the internally used |
@thomcc For a big project like |
Hmm, not sure, sorry about that! I think I may just manually copied the change over to the branch rather than properly cherry-picking the particular commit (which preserves the authored-by metadata)
We do run lots of checks on each push/PR, but there's a few complications which means they sometimes miss issues:
😓 Of those I think that:
|
That's a lot of work, more so than fetching from the remote (it is already merged) or even running
That explains it. I didn't remember these
That's fair, depending on the project I usually allow the contributor to fix it or - to keep every commit its own isolated change - open a separate PR to address new lints as maintainer. It can be a little hassle at times but it's a way of staying up-to-date with lints. You can also schedule a cron job in the CI which
Fair enough, docs are built on
Always a good idea to test this, it happened on a few occasions (to me) that a crate built fine but missed some source files, readme, licenses when packaged and uploaded to
Perhaps bump it more often though? Unless
For one you're setting
Using an artifact to let people download it too, should be pretty trivial to add? With the artifact you can make sure the crate is tested on a completely clean image. |
Thanks. I didn't realize that. |
(nevermind, already discussed) |
Regarding pinned clippy, I'm opposed to it, and didn't realize we were doing it. It seems kind of to defeat the point. We could also just stop using clippy and use Regarding running against a |
Might be worth splitting this discussion out elsewhere, but.. with the clippy CI thing it seems like there is a few different levels of "progressiveness":
I guess there's also a forth option of running latest clippy it in CI with the As mentioned, personally I think the pinned version is a good middleground - while clippy is indeed very noisy, I think there a few lints which are reasonably worthwhile.. while equally I don't think they are too vital as to require constnatly updating clippy (especially as there was a time in the last year or so where lots of new lints were added without sufficient ecosystem-wide testing - there has since been work towards reducing this, rust-lang/rust-clippy#6623 rust-lang/rust-clippy#7666 ) I'd also be in favour of just ditching it from CI - although given the pinned clippy doesn't really add any extra work (and we already run builds with the pinned version to ensure we don't break MSRV), I also don't see much benefit in removing it
Ah, I hadn't really thought much into how to actually do this - I kind of assumed you could just specify it like |
Oh, actually, I just read the
..and, importantly, you can specify Edit: Made PR to test this theory, #646 |
Should verify the .crate file sent to crates.io can at least be built, which should catch things like the docs.rs build failure, as discussed in imgui-rs#645 and imgui-rs#631
Shouldn't the
Personal recommendation: why not name that branch |
Yeah, good question. I'm inclined to just release 0.9.0, since I don't think the approach that has worked for the backports in the past has really... worked well at all. |
@thomcc Thanks, that'd work for me. If you do that, please also delete stale branches (i.e. those associated with closed PRs) to prevent confusion: https://github.com/imgui-rs/imgui-rs/branches. The relevant commits are tagged anyway. |
I'd agree with skipping any further 0.8.x releases - best to make v0.9 and settle on a good system for future patch releases |
Closing this in favour of releasing v0.9 |
I'll cut the releases after this lands. @dbr is there any reason to think this might have the same issues as #631?