fix: resolve extern prelude for local mods in block modules #17251
Conversation
rustbot
added
the
S-waiting-on-review
|
I'm not sure how to add unittests for it. 🤔 |
| if self.block.is_some() { | ||
| // Don't resolve extern prelude in block `DefMap`s, defer it to the crate def map so | ||
| // that blocks can properly shadow them | ||
| if matches!(self[module].origin, ModuleOrigin::BlockExpr { .. }) { |
There was a problem hiding this comment.
| if matches!(self[module].origin, ModuleOrigin::BlockExpr { .. }) { | |
| if self.block.is_some() && module == DefMap::ROOT { |
If I understand the code here correctly, likjewise below. Though I am unsure whether this is the correct fix here
There was a problem hiding this comment.
Though it should be more correct than the status quo
There was a problem hiding this comment.
They are equivalent. If a module is a BlockExpr, its local_id should be equivalent to ROOT, and it should be within a block; conversely, this is also true. Should I change it to the latter?
There was a problem hiding this comment.
Yes, sorry I wasn't clear. With unsure whether this is the correct fix I meant the general approach here. My code suggestion is the more correct way to do your check here (they are equivalent, it's just that the origin shouldn't be used to check this property).
| if matches!(self[module].origin, ModuleOrigin::BlockExpr { .. }) { | ||
| // Don't resolve extern prelude in pseudo-module of a block. |
|
Thanks! |
|
@bors r+ |
|
☀️ Test successful - checks-actions |
fix #17057, #17032.
We should use
ModuleOriginto check if the current module is a pseudo-module introduced by blocks (where names might be shadowed), rather than checkingblock_def_map.