Do not lint use_self in proc macro expansion
#9454
Conversation
|
r? @flip1995 (rust-highfive has picked a reviewer for you, use r? to override) |
rust-highfive
added
the
S-waiting-on-review
| @@ -103,6 +103,7 @@ impl<'tcx> LateLintPass<'tcx> for UseSelf { | |||
| if parameters.as_ref().map_or(true, |params| { | |||
| !params.parenthesized && !params.args.iter().any(|arg| matches!(arg, GenericArg::Lifetime(_))) | |||
| }); | |||
| if !is_from_proc_macro(cx, item); // expensive, should be last check | |||
There was a problem hiding this comment.
Would item_path.span.from_expansion() be enough here?
There was a problem hiding this comment.
Just tried testing it, and from_expansion still produces the false positive.
There was a problem hiding this comment.
huh, what about for item.span?
There was a problem hiding this comment.
The issue is the proc macro using a real span instead of making a new one. Basically every lint can be triggered from a proc macro, but with the span of the item being linted coming from a token outside the proc macro.
There was a problem hiding this comment.
The path span happened to do that but the whole item doesn't, so checking its span should be fine
There was a problem hiding this comment.
It depends what span the proc macro sets. If it doesn't set a span with expansion info, but just uses the span given from the item/tokens it looks to clippy as if it is unexpanded code.
|
@bors r+ Thanks! |
|
☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test |
|
In which version will this patch be released ? |
fix #9440
fix #8910
fix #6902
changelog: [
use_self]: Do not lint in proc macro expansion