Warnings for issue #32330 #33137
Warnings for issue #32330 #33137
Conversation
|
☔ The latest upstream changes (presumably #33179) made this pull request unmergeable. Please resolve the merge conflicts. |
nikomatsakis
force-pushed
the
issue-32330-lbr-in-return-type-warning-2
branch
from
a254c26 to
4fe2845
Compare
| @@ -626,3 +635,39 @@ impl<'tcx> TypeVisitor<'tcx> for HasTypeFlagsVisitor { | |||
| false | |||
| } | |||
| } | |||
|
|
|||
| /// Collects all the late-bound regions it finds into a hash set. | |||
There was a problem hiding this comment.
This comment might want to live on the public function instead.
|
☔ The latest upstream changes (presumably #32756) made this pull request unmergeable. Please resolve the merge conflicts. |
nikomatsakis
force-pushed
the
issue-32330-lbr-in-return-type-warning-2
branch
from
4fe2845 to
8380fe9
Compare
|
I remembered I want to update the issue in the lint to point at a tracking issue. I'll do a crater run tonight and fix it up in the morning. |
nikomatsakis
force-pushed
the
issue-32330-lbr-in-return-type-warning-2
branch
2 times, most recently
from
0a98537 to
cfa3471
Compare
|
Hmm. Did a crater run without the improvements to subtyping and was started by the results. My memory is that (with the lint set to deny) we got 4 root regressions before, but now I see 73: https://gist.github.com/nikomatsakis/2b324ddc0d68c90ccbafb8f71e812344 |
|
I am running a new crater run using the more correct subtyping rules to reconfirm (ti's possible that crates.io changed, or perhaps that those are timeouts actually). |
|
OK I looked a bit more closely. The results are not timeouts (though there was one). They are a mix of expected lint errors but also some unexpected type errors -- most of them are in fact in the nom crate, despite crater reporting them as a "root regression". I will investigate a bit. |
We ought not to be affecting inference state when assembling candidates, so invoke select inside of a probe.
nikomatsakis
force-pushed
the
issue-32330-lbr-in-return-type-warning-2
branch
2 times, most recently
from
8debd69 to
90ff7dd
Compare
|
In an effort to just land something dang it, I've pared this branch way back. It is now just warning at the declaration site for types and where-clauses that are clearly invalid. Doing a make check to make sure I didn't mess something up, but this is a subset of the previous code. |
|
Result from the crater run with the changes to subtyping: https://gist.github.com/nikomatsakis/4edf5ef8f7bf65bbb2029c05b46f7818 10 root regressions. More than before; so it's definitely important to start landing something. |
|
But a good number of those are these errors that I don't fully understand, so I'd rather not land said changes now... |
This is a step towards fixing rust-lang#32330. The full fix would be a breaking change, so we begin by issuing warnings for scenarios that will break.
nikomatsakis
force-pushed
the
issue-32330-lbr-in-return-type-warning-2
branch
from
90ff7dd to
639890d
Compare
|
@bors r=aturon |
|
📌 Commit 639890d has been approved by |
…rning-2, r=aturon Warnings for issue #32330 This is an extension of the previous PR that issues warnings in more situations than before. It does not handle *all* cases of #32330 but I believe it issues warnings for all cases I've seen in practice. Before merging I'd like to address: - open a good issue explaining the problem and how to fix it (I have a [draft writeup][]) - work on the error message, which I think is not as clear as it could/should be (suggestions welcome) r? @aturon [draft writeup]: https://gist.github.com/nikomatsakis/631ec8b4af9a18b5d062d9d9b7d3d967
This is an extension of the previous PR that issues warnings in more situations than before. It does not handle all cases of #32330 but I believe it issues warnings for all cases I've seen in practice.
Before merging I'd like to address:
r? @aturon