99 Commits

Author SHA1 Message Date
Michael Goulet
bb99cdc7cd vars are ? 2023-04-25 19:53:09 +00:00
bors
1f5768bc67 Auto merge of - cjgillot:query-feed, r=oli-obk
Allow to feed a value in another query's cache and remove `WithOptConstParam`

I used it to remove `WithOptConstParam` queries, as an example.

The idea is that a query (here `typeck(function)`) can write into another query's cache (here `type_of(anon const)`). The dependency node for `type_of` would depend on all the current dependencies of `typeck`.

There is still an issue with cycles: if `type_of(anon const)` is accessed before `typeck(function)`, we will still have the usual cycle.  The way around this issue is to `ensure` that `typeck(function)` is called before accessing `type_of(anon const)`.

When replayed, we may the following cases:
- `typeck` is green, in that case `type_of` is green too, and all is right;
- `type_of` is green, `typeck` may still be marked as red (it depends on strictly more things than `type_of`) -> we verify that the saved value and the re-computed value of `type_of` have the same hash;
- `type_of` is red, then `typeck` is red -> it's the caller responsibility to ensure `typeck` is recomputed *before* `type_of`.

As `anon consts` have their own `DefPathData`, it's not possible to have the def-id of the anon-const point to something outside the original function, but the general case may have to be resolved before using this device more broadly.

There is an open question about loading from the on-disk cache.  If `typeck` is loaded from the on-disk cache, the side-effect does not happen. The regular `type_of` implementation can go and fetch the correct value from the decoded `typeck` results, and the dep-graph will check that the hashes match, but I'm not sure we want to rely on this behaviour.

I specifically allowed to feed the value to `type_of` from inside a call to `type_of`.  In that case, the dep-graph will check that the fingerprints of both values match.

This implementation is still very sensitive to cycles, and requires that we call `typeck(function)` before `typeck(anon const)`.  The reason is that `typeck(anon const)` calls `type_of(anon const)`, which calls `typeck(function)`, which feeds `type_of(anon const)`, and needs to build the MIR so needs `typeck(anon const)`.  The latter call would not cycle, since `type_of(anon const)` has been set, but I'd rather not remove the cycle check.
2023-04-21 08:04:58 +00:00
bors
8d09b8e206 Auto merge of - c410-f3r:dqewdas, r=petrochenkov
Move test files

r? `@petrochenkov`
2023-04-21 01:26:57 +00:00
Caio
4adc5f9281 Move test files 2023-04-20 15:06:17 -03:00
Camille GILLOT
a395d2a5de Give more descriptive names to queries. 2023-04-20 18:01:07 +00:00
Camille GILLOT
8fc1d68413 Account for variance in outlives verification. 2023-04-16 17:39:52 +00:00
Camille GILLOT
4e7edf3684 Add tests. 2023-04-16 09:58:08 +00:00
Esteban Küber
9fadcc143a Special-case item attributes in the suggestion output 2023-04-12 22:50:10 +00:00
Esteban Küber
5b40aa5eb4 Tweak output for 'add line' suggestion 2023-04-12 22:50:10 +00:00
Oli Scherer
f263f88bea Split out a separate feature gate for impl trait in associated types 2023-04-12 16:17:31 +00:00
Jack Huey
f0edcc8a6f Remove index from BrAnon 2023-04-06 23:01:40 -04:00
Camille GILLOT
05082f57af Perform match checking on THIR. 2023-04-03 15:59:21 +00:00
Yuki Okushi
2981d7781c
Rollup merge of - ehuss:overlapping-tests, r=Mark-Simulacrum
compiletest: Don't allow tests with overlapping prefix names

Some tests will delete their output directory before starting. The output directory is based on the test names. If one test is the prefix of another test, then when that test starts, it could try to delete the output directory of the other test with the longer path, or otherwise clash with it while the two tests are trying to create/delete/modify the same directory.

In practice, this manifested as a random error on macOS where two tests were trying to create/delete/create `rustdoc/primitive` and `rustdoc/primitive/no_std`, which resulted in an EINVAL (InvalidInput) error.

This renames some of the offending tests, adds `compiletest-ignore-dir` to prevent compiletest from processing some files, and adds a check to prevent this from happening in the future.

Fixes 
2023-03-30 21:07:00 +09:00
Santiago Pastorino
2ca350c776
Bless some extra working tests under -Zlower-impl-trait-in-trait-to-assoc-ty 2023-03-29 11:27:26 -03:00
Santiago Pastorino
66714658d0
Properly skip RPITITs from ModChild and give a name in AssocItem 2023-03-29 11:19:49 -03:00
Michael Goulet
76b0cf812b
Walk return-position impl trait in trait deeply in associated_item_def_ids 2023-03-29 11:19:48 -03:00
Matthias Krüger
3b49ad3814
Rollup merge of - compiler-errors:rpitit-wf, r=eholk
Deeply check well-formedness of return-position `impl Trait` in trait

Walk the bounds of RPITITs to see if we find any more RPITITs 😸
2023-03-25 03:37:12 +01:00
Michael Goulet
b605d0740f Deeply check WF for RPITITs 2023-03-23 23:37:22 +00:00
Michael Goulet
e5189cc7e4 Nested impl traits trigger opaque_hidden_inferred_bound too much 2023-03-23 06:19:52 +00:00
Eric Huss
2da2ade0f7 Rename tests to ensure they don't have overlapping names.
Some tests will delete their output directory before starting.
The output directory is based on the test names.
If one test is the prefix of another test, then when that test
starts, it could try to delete the output directory of the other
test with the longer path.
2023-03-22 21:12:40 -07:00
Dylan DPC
031640ccd2
Rollup merge of - spastorino:new-rpitit-16, r=compiler-errors
Do not consider synthesized RPITITs on missing items checks

Without this patch for `tests/ui/impl-trait/in-trait/dont-project-to-rpitit-with-no-value.rs` we get ...

```
warning: the feature `return_position_impl_trait_in_trait` is incomplete and may not be safe to use and/or cause compiler crashes
 --> tests/ui/impl-trait/in-trait/dont-project-to-rpitit-with-no-value.rs:4:12
  |
4 | #![feature(return_position_impl_trait_in_trait)]
  |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  |
  = note: see issue  <https://github.com/rust-lang/rust/issues/91611> for more information
  = note: `#[warn(incomplete_features)]` on by default

error[E0046]: not all trait items implemented, missing: `foo`, ``
  --> tests/ui/impl-trait/in-trait/dont-project-to-rpitit-with-no-value.rs:12:1
   |
8  |     fn foo(&self) -> impl Sized;
   |     ----------------------------
   |     |                |
   |     |                `` from trait
   |     `foo` from trait
...
12 | impl MyTrait for i32 {
   | ^^^^^^^^^^^^^^^^^^^^ missing `foo`, `` in implementation

error: aborting due to previous error; 1 warning emitted

For more information about this error, try `rustc --explain E0046`.
```

instead of ...

```
warning: the feature `return_position_impl_trait_in_trait` is incomplete and may not be safe to use and/or cause compiler crashes
  --> $DIR/dont-project-to-rpitit-with-no-value.rs:4:12
   |
LL | #![feature(return_position_impl_trait_in_trait)]
   |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
   |
   = note: see issue  <https://github.com/rust-lang/rust/issues/91611> for more information
   = note: `#[warn(incomplete_features)]` on by default

error[E0046]: not all trait items implemented, missing: `foo`
  --> $DIR/dont-project-to-rpitit-with-no-value.rs:12:1
   |
LL |     fn foo(&self) -> impl Sized;
   |     ---------------------------- `foo` from trait
...
LL | impl MyTrait for i32 {
   | ^^^^^^^^^^^^^^^^^^^^ missing `foo` in implementation

error: aborting due to previous error; 1 warning emitted

For more information about this error, try `rustc --explain E0046`.
```

r? `@compiler-errors`
2023-03-23 00:00:35 +05:30
Santiago Pastorino
c1f3529c91 Always encode RPITITs 2023-03-21 23:35:46 +00:00
Santiago Pastorino
364a5d4b54
Do not consider synthesized RPITITs on missing items checks 2023-03-21 15:44:12 -03:00
nils
66ba60a445
Rollup merge of - compiler-errors:dont-binder-twice, r=oli-obk
Walk un-shifted nested `impl Trait` in trait when setting up default trait method assumptions

Fixes a double subtraction in some binder math in return-position `impl Trait` in trait handling code.

Fixes 
2023-03-21 13:00:23 +01:00
Matthias Krüger
d86fd83ef6
Rollup merge of - spastorino:new-rpitit-14, r=compiler-errors
Fix generics_of for impl's RPITIT synthesized associated type

The only useful commit is the last one.

This makes `generics_of` for the impl side RPITIT copy from the trait's associated type and avoid the fn on the impl side which was previously wrongly used.
This solution is better but we still need to fix resolution of the generated generics.

r? ``@compiler-errors``
2023-03-20 09:46:53 +01:00
Michael Goulet
9f80c75703 Walk un-shifted nested impl Trait in trait when setting up default trait method assumptions 2023-03-20 04:50:02 +00:00
Matthias Krüger
8417c93092
Rollup merge of - spastorino:new-rpitit-12, r=compiler-errors
Fix generics mismatch errors for RPITITs on -Zlower-impl-trait-in-trait-to-assoc-ty

This PR stops reporting errors due to different count of generics on the new synthesized associated types for RPITITs. Those were already reported when we compare the function on the triat with the function on the impl.

r? ``@compiler-errors``
2023-03-18 12:04:23 +01:00
Santiago Pastorino
640c20272e
Fix generics_of for impl's RPITIT synthesized associated type 2023-03-17 20:01:57 -03:00
Santiago Pastorino
9139ed076d
Fix impl_trait_ty_to_ty substs 2023-03-17 16:28:00 -03:00
Santiago Pastorino
e0302bbc3b
Add revisions for -Zlower-impl-trait-in-trait-to-assoc-ty fixed tests 2023-03-17 16:01:53 -03:00
Santiago Pastorino
c7cc1c7442
Fix generics mismatch errors for RPITITs on -Zlower-impl-trait-in-trait-to-assoc-ty 2023-03-17 15:36:02 -03:00
Michael Goulet
ff7c3b854d Don't install default opaque projection predicates in RPITIT associated type's param-env 2023-03-16 01:59:41 +00:00
Michael Goulet
0949da8f4e Install projection from RPITIT to default trait method opaque correctly 2023-03-16 01:56:49 +00:00
Santiago Pastorino
c5c4340760
Add revisions to fixed tests in -Zlower-impl-trait-in-trait-to-assoc-ty 2023-03-15 16:58:37 -03:00
Matthias Krüger
21d15db1df
Rollup merge of - spastorino:new-rpitit-9, r=compiler-errors
Make fns from other crates with RPITIT work for -Zlower-impl-trait-in-trait-to-assoc-ty

Only the last two commits are meaningful.

r? `@compiler-errors`
2023-03-14 17:40:05 +01:00
Matthias Krüger
48934c48c6
Rollup merge of - spastorino:new-rpitit-7, r=compiler-errors
Fix object safety checks for new RPITITs

This one goes on top of 

r? `@compiler-errors`
2023-03-14 17:40:04 +01:00
Matthias Krüger
5037836daa
Rollup merge of - spastorino:new-rpitit-6, r=compiler-errors
Remove tests/ui/impl-trait/in-trait/new-lowering-strategy in favor of using revisions on existing tests

r? `@compiler-errors`

This one again sits on top of existing approved PRs and it still needs to add revisions to tests in `tests/ui/impl-trait/in-trait` as it only does so for async in traits.
2023-03-14 17:40:03 +01:00
Santiago Pastorino
a4e40370d0
Make fns from other crates with RPITIT work 2023-03-14 10:20:35 -03:00
bors
669e751639 Auto merge of - Swatinem:async-identity-future, r=compiler-errors
Remove `identity_future` indirection

This was previously needed because the indirection used to hide some unexplained lifetime errors, which it turned out were related to the `min_choice` algorithm.

Removing the indirection also solves a couple of cycle errors, large moves and makes async blocks support the `#[track_caller]`annotation.

Fixes https://github.com/rust-lang/rust/issues/104826.
2023-03-14 10:12:58 +00:00
Santiago Pastorino
8b9344acc3
Fix object safety checks for new RPITITs 2023-03-12 10:51:20 -03:00
Santiago Pastorino
b2f3198608
Filter out RPITITs in astconv when checking for missing associated types 2023-03-12 10:51:19 -03:00
Santiago Pastorino
b82d6a2e9e
Run existing impl trait in traits tests using -Zlower-impl-trait-in-trait-to-assoc-ty 2023-03-12 10:50:32 -03:00
Santiago Pastorino
8a5574bf62
Remove tests/ui/impl-trait/in-trait/new-lowering-strategy in favor of using revisions on existing tests 2023-03-12 10:50:31 -03:00
bors
9455a5591b Auto merge of - spastorino:new-rpitit-impl-side-2, r=compiler-errors
Make RPITITs simple cases work when using lower_impl_trait_in_trait_to_assoc_ty

r? `@compiler-errors`

It's probably best reviewed commit by commit.
2023-03-12 00:49:35 +00:00
Arpad Borsos
9f03cfc207
Remove identity_future indirection
This was previously needed because the indirection used to hide some unexplained lifetime errors, which it turned out were related to the `min_choice` algorithm.

Removing the indirection also solves a couple of cycle errors, large moves and makes async blocks support the `#[track_caller]` annotation.
2023-03-08 15:37:14 +01:00
Matthias Krüger
63635880f6
Rollup merge of - compiler-errors:rpitit-default-method-with-nested-rpitits, r=spastorino
Account for binders correctly when adding default RPITIT method assumption

As of , we install extra projection predicates into the param-env of a default trait method when it has return-position `impl Trait` (or is async).

The implementation didn't account for the fact that it's walking into and out of binders, so we just need to shift all the debruijn indices accordingly when constructing the projection predicates.

Fixes 

r? types
2023-03-07 19:57:44 +01:00
Santiago Pastorino
5daa01e4a6
Add simple impl trait test for RPITIT 2023-03-06 14:49:50 -03:00
Matthias Krüger
1c2f641e15
Rollup merge of - cjgillot:dpm-adapt, r=compiler-errors
Tweaks to -Zdrop-tracking-mir

Split from https://github.com/rust-lang/rust/pull/107421

3 commits: 1 diagnostic improvement and 2 ICEs.
2023-03-05 20:57:21 +01:00
Camille GILLOT
0916616fec drop_tracking_mir: diagnose recursive generator. 2023-03-05 08:29:35 +00:00
Michael Goulet
7634c5916a Don't project to RPITIT that has no default value 2023-03-04 18:36:02 +00:00