Commit graph

116 commits

Author SHA1 Message Date
lcnr
675c99f4d5 more eagerly discard constraints on overflow 2024-09-12 14:32:44 +02:00
Jubilee Young
d243c8fbc4 compiler: Inform the solver of concurrency
Parallel compilation of a program can cause unexpected event sequencing.
Inform the solver when this is true so it can skip invalid asserts, then
assert replaced solutions are equal if Some
2024-09-09 13:07:48 -07:00
Michael Goulet
5df13af56f Use the right type when coercing fn items to pointers 2024-08-13 16:23:20 -04:00
Pavel Grigorenko
168096f663 rustc_type_ir: derivative -> derive-where 2024-07-12 21:52:04 +03:00
Michael Goulet
fb8d5f1e13 Actually just make can_eq process obligations (almost) everywhere 2024-07-05 11:59:54 -04:00
Michael Goulet
fdde66acee Process alias-relate obligations when proving receiver_is_valid 2024-07-05 11:59:52 -04:00
Michael Goulet
ecdaff240d Actually report normalization-based type errors correctly for alias-relate obligations in new solver 2024-07-02 16:39:57 -04:00
Michael Goulet
789ee88bd0 Tighten spans for async blocks 2024-06-27 15:19:08 -04:00
Deadbeef
81da6a6d40 Make effects an incomplete feature 2024-06-22 14:11:11 +00:00
León Orell Valerian Liehr
2126c1d446
rustc_type_ir: Omit some struct fields from Debug output 2024-06-19 03:08:34 +02:00
Jacob Pratt
936d76009b
Rollup merge of #126127 - Alexendoo:other-trait-diag, r=pnkfelix
Spell out other trait diagnostic

I recently saw somebody confused about the diagnostic thinking it was suggesting to add an `as` cast. This change is longer but I think it's clearer
2024-06-16 03:41:57 -04:00
Michael Goulet
93ee07c756 Check that alias-relate terms are WF if reporting an error in alias-relate 2024-06-13 08:52:35 -04:00
Jubilee
25c55c51cb
Rollup merge of #126142 - compiler-errors:trait-ref-split, r=jackh726
Harmonize using root or leaf obligation in trait error reporting

When #121826 changed the error reporting to use root obligation and not the leafmost obligation, it didn't actually make sure that all the other diagnostics helper functions used the right obligation.

Specifically, when reporting similar impl candidates we are looking for impls of the root obligation, but trying to match them against the trait ref of the leaf obligation.

This does a few other miscellaneous changes. There's a lot more clean-up that could be done here, but working with this code is really grief-inducing due to how messy it has become over the years. Someone really needs to show it love. 😓

r? ``@estebank``

Fixes #126129
2024-06-12 20:03:19 -07:00
Michael Goulet
c453c82de4 Harmonize use of leaf and root obligation in trait error reporting 2024-06-12 20:57:23 -04:00
Michael Goulet
44040a0670 Also passthrough for projection clauses 2024-06-12 19:10:02 -04:00
Michael Goulet
b0c1474381 better error message for normalizes-to ambiguities 2024-06-12 19:03:37 -04:00
Michael Goulet
52b2c88bdf Walk into alias-eq nested goals even if normalization fails 2024-06-12 19:03:37 -04:00
Alex Macleod
d0112c6849 Spell out other trait diagnostic 2024-06-12 12:34:47 +00:00
Michael Goulet
0fc18e3a17 Remove DebugWithInfcx 2024-06-11 22:13:04 -04:00
lcnr
a8e091de4a bivariant alias: set has_unconstrained_ty_var 2024-06-05 11:48:03 +02:00
bors
1d52972dd8 Auto merge of #125778 - estebank:issue-67100, r=compiler-errors
Use parenthetical notation for `Fn` traits

Always use the `Fn(T) -> R` format when printing closure traits instead of `Fn<(T,), Output = R>`.

Address #67100:

```
error[E0277]: expected a `Fn()` closure, found `F`
 --> file.rs:6:13
  |
6 |     call_fn(f)
  |     ------- ^ expected an `Fn()` closure, found `F`
  |     |
  |     required by a bound introduced by this call
  |
  = note: wrap the `F` in a closure with no arguments: `|| { /* code */ }`
note: required by a bound in `call_fn`
 --> file.rs:1:15
  |
1 | fn call_fn<F: Fn() -> ()>(f: &F) {
  |               ^^^^^^^^^^ required by this bound in `call_fn`
help: consider further restricting this bound
  |
5 | fn call_any<F: std::any::Any + Fn()>(f: &F) {
  |                              ++++++
```
2024-06-03 08:14:03 +00:00
Matthias Krüger
ab55d42b74
Rollup merge of #125786 - compiler-errors:fold-item-bounds, r=lcnr
Fold item bounds before proving them in `check_type_bounds` in new solver

Vaguely confident that this is sufficient to prevent rust-lang/trait-system-refactor-initiative#46 and rust-lang/trait-system-refactor-initiative#62.

This is not the "correct" solution, but will probably suffice until coinduction, at which point we implement the right solution (`check_type_bounds` must prove `Assoc<...> alias-eq ConcreteType`, normalizing requires proving item bounds).

r? lcnr
2024-05-31 08:50:23 +02:00
Michael Goulet
2f4b7dc047 Fold item bound before checking that they hold 2024-05-30 15:52:29 -04:00
lcnr
f7d14b741e update UI tests 2024-05-30 15:26:48 +02:00
Esteban Küber
e6bd6c2044 Use parenthetical notation for Fn traits
Always use the `Fn(T) -> R` format when printing closure traits instead of `Fn<(T,), Output = R>`.

Fix #67100:

```
error[E0277]: expected a `Fn()` closure, found `F`
 --> file.rs:6:13
  |
6 |     call_fn(f)
  |     ------- ^ expected an `Fn()` closure, found `F`
  |     |
  |     required by a bound introduced by this call
  |
  = note: wrap the `F` in a closure with no arguments: `|| { /* code */ }`
note: required by a bound in `call_fn`
 --> file.rs:1:15
  |
1 | fn call_fn<F: Fn() -> ()>(f: &F) {
  |               ^^^^^^^^^^ required by this bound in `call_fn`
help: consider further restricting this bound
  |
5 | fn call_any<F: std::any::Any + Fn()>(f: &F) {
  |                              ++++++
```
2024-05-29 22:26:54 +00:00
lcnr
98bfd54b0a eagerly normalize when adding goals 2024-05-28 04:54:05 +00:00
lcnr
4d5a9bcb86 change selection test to run-pass 2024-05-28 04:44:45 +00:00
bors
b0f8618938 Auto merge of #125413 - lcnr:ambig-drop-region-constraints, r=compiler-errors
drop region constraints for ambiguous goals

See the comment in `compute_external_query_constraints`. While the underlying issue is preexisting, this fixes a bug introduced by #125343.

It slightly weakens the leak chec, even if we didn't have any test which was affected. I want to write such a test before merging this PR.

r? `@compiler-errors`
2024-05-27 15:28:51 +00:00
lcnr
24b5466892 drop region constraints for ambiguous goals 2024-05-24 20:32:35 +00:00
Michael Goulet
61aac551b8 Structurally resolve before builtin_index in EUV 2024-05-24 16:20:48 -04:00
lcnr
f99c9ffd88 track cycle participants per entry 2024-05-20 20:57:14 +00:00
Michael Goulet
119c7bbef7 Report better WF obligation leaf obligations in new solver 2024-05-16 21:08:42 -04:00
Michael Goulet
d3e510eb9d Don't ICE because recomputing overflow goals during find_best_leaf_obligation causes inference side-effects 2024-05-16 10:00:11 -04:00
Michael Goulet
fb298e80c3 Apply nits 2024-05-12 12:11:25 -04:00
Michael Goulet
5ab6dca6d3 Try structurally resolve 2024-05-12 12:11:25 -04:00
lcnr
83e6da0be5 analyse visitor: build proof tree in probe 2024-05-09 17:29:53 +00:00
Matthias Krüger
067f6327a5
Rollup merge of #124846 - compiler-errors:const-eval, r=lcnr
Don't ICE when we cannot eval a const to a valtree in the new solver

Use `const_eval_resolve` instead of `try_const_eval_resolve` because naming aside, the former doesn't ICE when a value can't be evaluated to a valtree.

r? lcnr
2024-05-07 18:12:56 +02:00
Matthias Krüger
4a5bf7b06e
Rollup merge of #124827 - lcnr:generalize-incomplete, r=compiler-errors
generalize hr alias: avoid unconstrainable infer vars

fixes https://github.com/rust-lang/trait-system-refactor-initiative/issues/108

see inline comments for more details

r? `@compiler-errors` cc `@BoxyUwU`
2024-05-07 18:12:55 +02:00
lcnr
690d5aa417 generalize hr alias: avoid unconstrainable infer vars 2024-05-07 15:58:06 +00:00
Michael Goulet
b58f5a7800 Don't ICE when we cannot eval a const to a valtree in the new solver 2024-05-07 11:14:25 -04:00
Matthias Krüger
f76c8f7f77
Rollup merge of #124759 - compiler-errors:impl-args, r=lcnr
Record impl args in the proof tree in new solver

Rather than rematching them during select.

Also use `ImplSource::Param` instead of `ImplSource::Builtin` for alias-bound candidates, so we don't ICE in `Instance::resolve`.

r? lcnr
2024-05-06 21:46:05 +02:00
Michael Goulet
e34723997a Use correct ImplSource for alias bounds 2024-05-06 14:38:35 -04:00
Matthias Krüger
ad73b1623d
Rollup merge of #124808 - compiler-errors:super, r=lcnr
Use `super_fold` in `RegionsToStatic` visitor

so as to avoid an infinite stack cycle

fixes #124805
r? lcnr
2024-05-06 18:50:36 +02:00
Michael Goulet
116f95bb46 Use super_fold in RegionsToStatic visitor 2024-05-06 12:22:15 -04:00
Michael Goulet
4e3350d43b Don't consider candidates with no failing where clauses 2024-05-06 11:32:50 -04:00
Matthias Krüger
07dc4aa837
Rollup merge of #124718 - compiler-errors:record-impl-args, r=lcnr
Record impl args in the proof tree

Weren't recording these since they went through a different infcx method

r? lcnr
2024-05-04 22:27:33 +02:00
Michael Goulet
50338aa59a Record impl args in the proof tree 2024-05-04 12:57:01 -04:00
Michael Goulet
6714216eaa Only consider ambiguous goals when finding best obligation for ambiguities 2024-05-04 12:05:36 -04:00
Michael Goulet
34e91ece90 Higher ranked goal source, do overflow handling less badly 2024-05-02 21:56:14 -04:00
Michael Goulet
3e03b1b190 Use a proof tree visitor to refine the Obligation for error reporting 2024-05-02 21:56:14 -04:00