collect doc alias as tips during resolution #127721
Conversation
rustbot
added
S-waiting-on-review
|
In general I'd prefer to maybe not do this at all for the local crate. |
petrochenkov
added
S-waiting-on-author
Agree and accept this suggestion. @rustbot ready |
rustbot
added
S-waiting-on-review
This comment was marked as resolved.
This comment was marked as resolved.
lolbinarycat
added
S-waiting-on-author
|
@bvanjoi triage here, any interest in rebasing this? |
|
@bvanjoi |
|
Rebased. @rustbot ready |
rustbot
added
S-waiting-on-review
This comment was marked as resolved.
This comment was marked as resolved.
|
Update: Fixes the incorrect submodule... |
This comment was marked as resolved.
This comment was marked as resolved.
|
Hi @estebank could you take a look at this when you get a chance? Thanks 🙂 |
|
Maybe r? @petrochenkov (or reroll)? |
| @@ -1285,3 +1285,29 @@ pub fn parse_confusables(attr: &Attribute) -> Option<Vec<Symbol>> { | |||
|
|
|||
| Some(candidates) | |||
| } | |||
|
|
|||
| pub fn collect_doc_alias_symbol_from_attrs<'tcx>( | |||
There was a problem hiding this comment.
Both calls to this function immediately use contains(sym) on the result, so it's probably better to pass sym to here and avoid the collection.
| format!( | ||
| "`{}` has a name defined in the doc alias attribute as `{}`", | ||
| self.r.tcx.item_name(did), | ||
| path.last().unwrap().ident.as_str() |
There was a problem hiding this comment.
| path.last().unwrap().ident.as_str() | |
| path.last().unwrap().ident |
| @@ -783,6 +796,50 @@ impl<'ast, 'ra: 'ast, 'tcx> LateResolutionVisitor<'_, 'ast, 'ra, 'tcx> { | |||
| (false, candidates) | |||
| } | |||
|
|
|||
| fn lookup_doc_alias_name(&mut self, path: &[Segment], ns: Namespace) -> Option<DefId> { | |||
| let item_str = path.last().unwrap().ident; | |||
There was a problem hiding this comment.
| let item_str = path.last().unwrap().ident; | |
| let item_name = path.last().unwrap().ident.name; |
| } else { | ||
| let mod_path = &path[..path.len() - 1]; | ||
| if let PathResult::Module(ModuleOrUniformRoot::Module(module)) = | ||
| self.resolve_path(mod_path, Some(TypeNS), None) |
There was a problem hiding this comment.
I don't think the logic is right, we are reporting an error for one specific unresolved segment in a path (and it's not necessarily the last one), not for all segments, and all segments before the one unresolved segment should be already resolved and recorded into partial_res_map.
So
- no need to search candidates for all segments, only for the currently unresolved one
- no need to use
resolve_pathfor the previous segments,partial_res_mapshould already be filled.
petrochenkov
added
S-waiting-on-author
Close #124273
Collect the symbol in the doc alias attributes and provide a tip when a match is found.
r? @estebank