Convert some things from chalk_ir types to rustc_type_ir types #20454
Conversation
rustbot
added
the
S-waiting-on-review
rust-cloud-vms
bot
force-pushed
the
next-trait-solver-next
branch
2 times, most recently
from
e0fa509 to
d12e875
Compare
| hir_fmt_generic_arguments(f, parameters_to_write, self_)?; | ||
| write!(f, ">")?; | ||
| } | ||
| impl<'db> HirDisplay for crate::next_solver::Ty<'db> { |
There was a problem hiding this comment.
I think you should replace the chalk ty display impl to converting to new solver ty then displaying it, to reduce duplicate code and maintenance efforts.
There was a problem hiding this comment.
It will also help us better test the code.
There was a problem hiding this comment.
I forwarded these impls. Associated type printing changes (only in tests). Had to add some salsa::attach calls, but otherwise no problems.
| mut check_alias: impl FnMut(&Name, TraitRef<'db>, TypeAliasId) -> Option<R>, | ||
| ) -> Option<R> { | ||
| let db = interner.db; | ||
| let mut search = |t: TraitRef<'db>| -> Option<R> { |
There was a problem hiding this comment.
You basically inlined all_super_trait_refs()? Why? Isn't it better to extract it to its own function?
There was a problem hiding this comment.
Basically just moving things around here, not adding anything new.
IIRC, all_super_trait_refs does not work because of cycles, but I don't quite remember. Personally, I would not want to try to abstract this out in this PR since it's just moving things.
|
|
||
| // we're _in_ the impl -- the binders get added back later. Correct, | ||
| // but it would be nice to make this more explicit | ||
| search(trait_ref.skip_binder()) |
There was a problem hiding this comment.
Why did you not add the check for being inside an impl method?
There was a problem hiding this comment.
I don't remember, I actually ported this months ago. I'll add a FIXME to come back to it, but don't want to figure this out here.
| })) | ||
| } | ||
|
|
||
| pub(crate) fn associated_ty_item_bounds<'db>( |
There was a problem hiding this comment.
Why not querify this? Although I see you haven't put the interner's one a query either. But at least reuse the code from DbInterner::item_bounds()?
There was a problem hiding this comment.
The equivalent code from chalk_db was not querified either. I added a FIXME to reuse with explicit_item_bounds.
rust-cloud-vms
bot
force-pushed
the
next-trait-solver-next
branch
3 times, most recently
from
152571b to
77cd3dd
Compare
github-merge-queue
bot
removed this pull request from the merge queue due to failed status checks
…redicates_without_parent_query
rust-cloud-vms
bot
force-pushed
the
next-trait-solver-next
branch
from
77cd3dd to
596a6bf
Compare
|
Rebased - the test from #20462 had its output change (associated type printing in tests) |
rustbot
removed
the
S-waiting-on-review
Each commit compiles, so these could land in parts if desired.