Track labels in the HIR #7021
Track labels in the HIR #7021
Conversation
crates/hir_def/src/body.rs
Outdated
| pub fn label_syntax(&self, label: LabelId) -> Result<LabelSource, SyntheticSyntax> { | ||
| self.label_map_back[label].clone() | ||
| } |
There was a problem hiding this comment.
I think labels can't be synthetic (ie, we don't produce labels during desugaring), so this should return a value rather than a result. And, more generally, we shouldn't store a Result in the source map in the frist palce.
There was a problem hiding this comment.
Right, I partially based this off of patterns hence why I used this here but I think you are right, we aren't producing new labels.
matklad
left a comment
matklad
left a comment
There was a problem hiding this comment.
LGTM, although it maybe makes sense to fold basic goto definition support here as well, just to be able to implement a smoke test.
|
I already have everything done for the the entire rename and goto scheme so its all tested in a sense, I just split the work up to get smaller PRs hence this only contains the changes to HIR. |
|
bors r+ Splitting is generally good, but I personally also strongly prefer at least a hello-world test. There's no crisp and clear solutions here, ideally, it should be possible to spike the main functionality such that it has smoke tests, and then add extra bells&whistles in follow ups |
|
Fair point, I'll keep that in mind 👍 |
2 participants
Groundwork for #6966