1070 Commits

Author SHA1 Message Date
Lukas Wirth
dd41cda70e
Merge pull request #19678 from Veykril/push-mkznvpsktnnz
Arena allocate `LifetimeRef`s
2025-04-24 12:56:22 +00:00
Lukas Wirth
5ff4ba347d
Merge pull request #19675 from Veykril/push-uuluymsosttr
fix: Fix type argument mismatch incorrectly triggering on inferred trait args
2025-04-24 08:45:30 +00:00
Lukas Wirth
61df8ec4b8 Arena allocate LifetimeRefs 2025-04-24 09:55:06 +02:00
Chayim Refael Friedman
e9fc3b6f20 Correctly set infer_args = true in more places
Previously this being incorrect wasn't a problem, it just meant we put an error type that then changed to infer type, so exactly what rustc does at the end. But now there is a diagnostic.
2025-04-24 09:16:13 +02:00
Lukas Wirth
dce59ad8fb fix: Fix type argument mismatch incorrectly triggering on inferred trait args 2025-04-24 09:16:01 +02:00
Lukas Wirth
4a2eaddc77 refactor: Remove WherePredicateTypeTarget 2025-04-24 09:13:56 +02:00
Chayim Refael Friedman
adcf699ea3 Properly handle lifetimes when checking generic arguments len
And also, prepare for correct lowering of lifetime. We still don't handle most lifetimes correctly, but a bit more of the foundation to lifetime elision is now implemented.
2025-04-24 08:35:20 +03:00
Lukas Wirth
d3b6d88386 fix: Fix incorrect diagnostic for lifetime parameter count mismatch 2025-04-23 16:36:13 +02:00
Chayim Refael Friedman
547c124c05 Fix variance
This one does need fixpoint.
2025-04-22 15:20:07 +03:00
Chayim Refael Friedman
57c019a3c5 Adapt for new cycle handling changing in Salsa 2025-04-22 15:20:05 +03:00
Chayim Refael Friedman
db72e2ff41 Adjust for salsa::Id::from_u32() being unsafe
This impacts our manual `salsa::Id` wrappers. I refactored them a bit to improve safety.
2025-04-22 15:19:35 +03:00
Chayim Refael Friedman
99ce53b1d7 Add two new diagnostics: one for mismatch in generic arguments count, and another for mismatch in their kind
Also known as E0747 and E0107.

And by the way, rewrite how we lower generic arguments and deduplicate it between paths and method calls. The new version is taken almost straight from rustc.

This commit also changes the binders of `generic_defaults()`, to only include the binders of the arguments up to (and not including) the current argument. This make it easier to handle it in the rewritten lowering of generic args. It's also how rustc does it.
2025-04-22 14:55:43 +03:00
Lukas Wirth
2b4b483af1
Merge pull request #19662 from Veykril/push-lqqvmnonlwrx
minor: Fix outdated comment in hir-ty/generics.rs
2025-04-22 11:26:56 +00:00
Lukas Wirth
a6b24f47c1 minor: Fix outdated comment in hir-ty/generics.rs 2025-04-22 13:15:56 +02:00
Lukas Wirth
9a9f4e7f63
Merge pull request #19624 from jackh726/chalk-update
Update chalk
2025-04-22 10:58:31 +00:00
Lukas Wirth
34e7d60e30
Merge pull request #19644 from ChayimFriedman2/const-syms
internal: Make predefined symbols `const` instead of `static`
2025-04-21 12:34:59 +00:00
Lukas Wirth
d16b862ea5 fix: Support unstable UnsafePinned struct in type layout calc 2025-04-21 09:35:30 +02:00
Chayim Refael Friedman
0f325c7ff8 Remove unnecessary predefined symbol clones
Now that they're const it's no longer needed.

Nothing manual was performed: only a regexp search of `sym::([\w][\w\d]*)\.clone\(\)` and replace by `sym::$1`.
2025-04-21 03:10:13 +03:00
Chayim Refael Friedman
9477e46bec Get rid of static predefined symbols
Make them all `const`.
2025-04-21 03:07:35 +03:00
jackh726
6daa791fab Update chalk 2025-04-20 16:20:08 +00:00
Chayim Refael Friedman
c58ddafe90 Make HirFileId, EditionedFileId and macro files Salsa struct
And make more queries non-interned.

Also flip the default for queries, now the default is to not intern and to intern a query you need to say `invoke_interned`.
2025-04-19 22:10:52 +03:00
Lukas Wirth
32052e5054
Merge pull request #19616 from ChayimFriedman2/transparent-diags
internal: Make hir-ty non-diagnostic queries transparent
2025-04-18 08:53:53 +00:00
Chayim Refael Friedman
5e3a914dc2 Make hir-ty non-diagnostic queries transparent
That is, all queries that have a `with_diagnostic` variant.

Them being tracked was (maybe) needed before #19462, because back then diagnostics could refer `AstId`s (for macro types), but now they are no longer needed.
2025-04-18 07:04:13 +03:00
Chayim Refael Friedman
e1962d2e22 Fix a bug with predicate lowering of associated items
The essence of the bug is that bounds on `impl Trait` got lowered with incorrect bound vars, as if we were lowering the parent.
2025-04-17 00:19:54 +03:00
Chayim Refael Friedman
cdc5ba30fc Prevent panics when there is a cyclic dependency between closures
We didn't include them in the sorted closures list, therefore didn't analyze them, then failed to find them.
2025-04-14 07:12:51 +03:00
Florian Diebold
1bf840bb50
Merge pull request #19566 from flodiebold/push-vzpyzvpkwkyt
Fix dyn compatibility code bypassing callable_item_signature query
2025-04-10 15:11:34 +00:00
Florian Diebold
6e06115e6d Fix dyn compatibility code bypassing callable_item_signature query 2025-04-10 16:56:17 +02:00
Lukas Wirth
8df812f14d fix: Walk const block expressions for unsafety checking 2025-04-10 16:10:30 +02:00
Lukas Wirth
4fdc2507c6 fix: Fix invalid signature bitflags 2025-04-10 15:39:27 +02:00
Lukas Wirth
f880acd18c
Merge pull request #19461 from Hmikihiro/shadow_by_module
fix: shadow type by module
2025-04-10 12:20:14 +00:00
Chayim Refael Friedman
8a9a1e3345 Remove all upcasts!
It turns out there were a lot redundant too.
2025-04-10 11:08:38 +03:00
Hayashi Mikihiro
1b471ebfd5 TypeNs contain module
Signed-off-by: Hayashi Mikihiro <34ttrweoewiwe28@gmail.com>
2025-04-10 16:29:10 +09:00
Hayashi Mikihiro
a7dc60a2c4 return single option
Signed-off-by: Hayashi Mikihiro <34ttrweoewiwe28@gmail.com>
2025-04-10 16:29:10 +09:00
Hayashi Mikihiro
2854ad9a41 fix: shadow type by module
Signed-off-by: Hayashi Mikihiro <34ttrweoewiwe28@gmail.com>
2025-04-10 16:29:10 +09:00
Lukas Wirth
07212140db refactor: Turn LifetimeRef into an enum
This makes things more structured
2025-04-09 15:58:06 +02:00
Lukas Wirth
dc363f7f77
Merge pull request #19462 from Veykril/push-ypvprvvwkyll
refactor: Lower type-refs before type inference
2025-04-09 08:54:28 +00:00
Lukas Wirth
1fd9520c92 refactor: Lower type-refs before type inference
This refactors how we deal with items in hir-def lowering.

- It now lowers all of them through an "ExpressionStore" (kind of a misnomer as this point) as their so called *Signatures.
- We now uniformly lower type AST into TypeRefs before type inference.
- Likewise, this moves macro expansion out of type inference, resulting in a single place where we do non-defmap macro expansion.
- Finally, this PR removes a lot of information from ItemTree, making the DefMap a lot less likely to be recomputed and have it only depend on actual early name resolution related information (not 100% true, we still have ADT fields in there but thats a follow up removal).
2025-04-09 10:43:23 +02:00
David Barsky
f3f6b5f4f1
Merge pull request #19536 from jackh726/closure-infer
Port closure inference from rustc
2025-04-08 14:17:30 +00:00
jackh726
bc3e9d9fcb Port closure inference from rustc 2025-04-07 21:48:01 +00:00
Lukas Wirth
a2783704bf
Merge pull request #19432 from ShoyuVanilla/issue-19431
fix: Yet another false positive invalid cast diagnostic
2025-04-07 10:29:54 +00:00
Lukas Wirth
e7ce86ddea fix: Fix format_args lowering for >=1.87 2025-04-06 10:05:56 +02:00
Lukas Wirth
bcc8e3a0b6 chore: Replace some invokes with invoke_actual in hir-ty 2025-04-01 13:13:20 +02:00
Chayim Refael Friedman
cdcadb4265
Merge pull request #19492 from ChayimFriedman2/experiment
Avoid relying on `block_def_map()` needlessly
2025-04-01 04:08:26 +00:00
Chayim Refael Friedman
2d99b5ead5 Avoid relying on block_def_map() needlessly
We can compute the wanted information from the block's interned data.
2025-04-01 06:58:21 +03:00
Lukas Wirth
1e1571e1c8 fix: Fix new nightly lints 2025-03-31 13:27:54 +02:00
Lukas Wirth
2e1ff255ae
Merge pull request #19375 from ChayimFriedman2/do-not-complete
feat: Allow crate authors to control completion of their things
2025-03-28 09:06:19 +00:00
Chayim Refael Friedman
246d678d77 Fix a bug in orphan rules calculation
Where a fundamental type applied twice wasn't considered local.
2025-03-28 05:37:29 +03:00
Chayim Refael Friedman
7b584ef9bf Allow crate authors to control completion of their things
Via the new `#[rust_analyzer::completions(...)]` attribute.

Also fix a bug with existing settings for that where the paths wouldn't resolve correctly.
2025-03-27 14:01:26 +02:00
Lukas Wirth
7a1e9c4b0c ci: Require miri step 2025-03-25 08:10:27 +01:00
Lukas Wirth
dcd742301e refactor: Replace custom ThinVec with thin-vec crate 2025-03-24 13:44:13 +01:00