Conversation
rustbot
added
the
A-build-execution
0xPoe
changed the title
edcff6f to
930da71
Compare
There was a problem hiding this comment.
🔢 Self-check (PR reviewed by myself and ready for feedback)
-
Code compiles successfully
-
Unit tests added
-
No AI-generated elegant nonsense in PR.
-
Comments added where necessary
-
PR title and description updated
-
Documentation PR created (or confirmed not needed)
-
PR size is reasonable
r? @epage
| } | ||
|
|
||
| let bcx = create_bcx(ws, options, &interner, logger.as_ref())?; | ||
| crate::diagnostics::rules::mostly_unused_hint::diagnose(&bcx)?; |
There was a problem hiding this comment.
Note: The change moved diagnostics before compilation, so the same check now reports warnings even without rebuilding.
There was a problem hiding this comment.
I believe we already have some warnings reported even without rebuilding, so I guess this seems fine.
| if gctx.cli_unstable().profile_hint_mostly_unused { | ||
| continue; | ||
| } | ||
| let profile_hint = unit.profile.hint_mostly_unused; |
There was a problem hiding this comment.
I tried to find its source, but doing so here is too complex. It would almost require us to reimplement the precedence resolution.
rustbot
added
the
S-waiting-on-review
0xPoe
changed the title
| // Profile options come from the workspace itself, so they are reported even for | ||
| // non-local packages. | ||
| let title = format!( | ||
| "ignoring `hint-mostly-unused` profile option for `{}@{}`", |
There was a problem hiding this comment.
Why are we showing a package spec, rather than having an Origin element?
There was a problem hiding this comment.
See #17457 (comment)
If we want to find the exact original element, we have to mirror its precedence resolution here. I tried it, but it seemed too complex, so I abandoned it. So I decided to simply hint that this package has a profile option that has been ignored.
There was a problem hiding this comment.
But I’ll try to do it in the parse pass. I guess that would make it easier.
| } | ||
|
|
||
| let bcx = create_bcx(ws, options, &interner, logger.as_ref())?; | ||
| crate::diagnostics::rules::mostly_unused_hint::diagnose(&bcx)?; |
There was a problem hiding this comment.
Should these be in the parse pass?
| mod manual_readme; | ||
| mod missing_lints_features; | ||
| mod missing_lints_inheritance; | ||
| pub mod mostly_unused_hint; |
There was a problem hiding this comment.
This is a bit weird being publibc. Should we give it a lint pass instead, or should we make a new lint pass for it?
There was a problem hiding this comment.
Another diagnostic does this as we still work out what the diagnostic pass should look like. We only created a formal pass for parsing once we understood it.
930da71 to
74fda82
Compare
0xPoe
changed the title
74fda82 to
be89b18
Compare
be89b18 to
9c72948
Compare
rustbot
removed
the
S-waiting-on-review
4 participants
What does this PR try to resolve?
Try to use
annotate-snippetsto report mostly unused warnings. Inspired by #17368 (comment).How to test and review this PR?
Check the unit tests and review each commit individually.
r?@ghost
LLM disclosure: The unit tests were generated by AI.