Commit Graph

229291 Commits

Author SHA1 Message Date
Oli Scherer
8ac1a67d11 Name the destructure_mir_constant query appropriately 2023-07-05 15:54:43 +00:00
Oli Scherer
46cce98134 Use options instead of errors if the errors are never needed 2023-07-05 15:54:43 +00:00
Oli Scherer
09b89efa70 Remove a function argument that is always passed with the same value. 2023-07-05 15:54:43 +00:00
Oli Scherer
a0eb348d38 Specialize DestructuredConstant to its one user (pretty printing) 2023-07-05 15:54:43 +00:00
Oli Scherer
4dcf988360 Specialize try_destructure_mir_constant for its sole user 2023-07-05 15:54:43 +00:00
Michael Goulet
560136f15d
Rollup merge of #113356 - he32:netbsd-riscv64, r=oli-obk
Add support for NetBSD/riscv64 aka. riscv64gc-unknown-netbsd.
2023-07-05 08:45:46 -07:00
Michael Goulet
6f9addf6ed
Rollup merge of #113355 - Zalathar:ssa, r=oli-obk
Move most coverage code out of `rustc_codegen_ssa`

*This is one step in my larger coverage refactoring ambitions described at <https://github.com/rust-lang/compiler-team/issues/645>.*

The backend implementation of coverage instrumentation was originally split between SSA and LLVM, perhaps in the hopes that it could be used by other backends.

In practice, this split mostly just makes the coverage implementation harder to navigate and harder to modify. It seems unlikely that any backend will actually implement coverage instrumentation in the foreseeable future, especially since many parts of the existing implementation (outside the LLVM backend) are heavily tied to the specific details of LLVM's coverage instrumentation features.

The current shared implementation of `codegen_coverage` is heavily tied to the details of `StatementKind::Coverage`, which makes those details difficult to change. I have reason to want to change those details as part of future fixes/improvements, so this will reduce the amount of interface churn caused by those later changes.

---

This is intended to be a pure refactoring change, with no changes to actual behaviour. All of the “added” code has really just been moved from other files.
2023-07-05 08:45:46 -07:00
Michael Goulet
c31fe41453
Rollup merge of #113337 - compiler-errors:next-solver-winnow-specializing, r=lcnr
Winnow specialized impls during selection in new solver

We need to be able to winnow impls that are specialized by more specific impls in order for codegen to be able to proceed.

r? ``@lcnr``
2023-07-05 08:45:45 -07:00
Michael Goulet
b2b1a50751
Rollup merge of #113321 - BoxyUwU:move_constkind_to_typeir, r=oli-obk
Move `ty::ConstKind` to `rustc_type_ir`

Needed this in another PR for custom debug impls, and this will also be required to move the new solver into a separate crate that does not use `TyCtxt` so that r-a and friends can depend on the trait solver.

Rebased on top of #113325, only the second and third commits needs reviewing
2023-07-05 08:45:45 -07:00
Michael Goulet
0334b64cbb
Rollup merge of #113320 - oli-obk:eval_obligation_query, r=petrochenkov,BoxyUwU
Add some extra information to opaque type cycle errors

Plus a bunch of cleanups.

This should help users debug query cycles due to auto trait checking. We'll probably want to fix cycle errors in most (or all?) cases by looking at the current item's hidden types (new solver does this), and by delaying the auto trait checks to after typeck.
2023-07-05 08:45:44 -07:00
Michael Goulet
5c7a7d9ed4
Rollup merge of #113319 - lcnr:type-param-def-def-id, r=compiler-errors
`TypeParameterDefinition` always require a `DefId`

the `None` case never actually reaches diagnostics so it feels better for diagnostics to be able to rely on the `DefId` being there, cc #113310
2023-07-05 08:45:43 -07:00
Michael Goulet
a1f8edb5d5
Rollup merge of #113317 - lcnr:sketchy-new-select, r=oli-obk
-Ztrait-solver=next: stop depending on old solver

removes the final dependencies on the old solver when `-Ztrait-solver=next` is enabled.
2023-07-05 08:45:43 -07:00
Michael Goulet
1cb958a225
Rollup merge of #113010 - klensy:ri-rls-fmt, r=ozkanonur
rust-installer & rls: remove exclusion from rustfmt & tidy

<strike>based on #112884</strike>

`rust-installer` and `rls` no longer submodules, but not removed from exclude list for rustfmt and tidy, preventing running fmt and lints on them.
2023-07-05 08:45:42 -07:00
jyn
fd7f53112a Update compiler-builtins to 0.1.95
This pulls in the new `outline-atomics` intrinsics.
2023-07-05 10:41:41 -05:00
Havard Eidnes
6cc37bbee0 Add support for NetBSD/riscv64 aka. riscv64gc-unknown-netbsd. 2023-07-05 13:49:01 +00:00
bors
e4cd161006 Auto merge of #113210 - fee1-dead-contrib:effects-mvp, r=oli-obk
Effects/keyword generics MVP

This adds `feature(effects)`, which adds `const host: bool` to the generics of const functions, const traits and const impls. This will be used to replace the current logic around const traits.

r? `@oli-obk`
2023-07-05 13:42:00 +00:00
León Orell Valerian Liehr
22fd6a6abf
Add regression test 2023-07-05 13:45:12 +02:00
León Orell Valerian Liehr
3788b7ab32
Revert "use new c literals instead of cstr! macro"
This reverts commit a17561ffc9.
2023-07-05 13:11:27 +02:00
León Orell Valerian Liehr
9dbe67fc8c
Revert "use c literals in library"
This reverts commit f212ba6d6d.
2023-07-05 13:11:26 +02:00
León Orell Valerian Liehr
5b25f9d8bd
Revert "fix ptr cast"
This reverts commit 2f459f7f14.
2023-07-05 13:11:26 +02:00
León Orell Valerian Liehr
c6643b50ea
Revert the lexing of c_str_literals 2023-07-05 13:11:17 +02:00
bors
dfe0683138 Auto merge of #112594 - ChrisDenton:process=-kill, r=Amanieu
Return `Ok` on kill if process has already exited

This will require an FCP from `@rust-lang/libs-api.`

Fixes #112423. See that issue for more details.
2023-07-05 11:04:17 +00:00
Vadim Petrochenkov
9f3fba8da8 resolve: Add comments explaining use of Interned 2023-07-05 13:47:37 +03:00
Vadim Petrochenkov
c1f412f9a9 resolve: Use Interned for Module 2023-07-05 13:47:36 +03:00
Vadim Petrochenkov
4abdaeb67e resolve: Use Interned for Import 2023-07-05 13:47:36 +03:00
Vadim Petrochenkov
8efd9cc30d resolve: Use Interned for NameBinding 2023-07-05 13:47:36 +03:00
Zalathar
cb570d6bc1 Move coverageinfo::ffi and coverageinfo::map out of SSA 2023-07-05 20:40:40 +10:00
Zalathar
9c430d38cf Remove trait CoverageInfoMethods, since non-LLVM backends don't need it
These methods are only ever called from within `rustc_codegen_llvm`, so they
can just be declared there as well.
2023-07-05 20:40:40 +10:00
Zalathar
4169d0f756 Narrow trait CoverageInfoBuilderMethods down to just one method
This effectively inlines most of `FunctionCx::codegen_coverage` into the LLVM
implementation of `CoverageInfoBuilderMethods`.
2023-07-05 20:40:39 +10:00
Chris Denton
9e5f61fcdd
Workaround for old android not having echo 2023-07-05 09:54:16 +01:00
bors
99f7d368c0 Auto merge of #112319 - oli-obk:assoc_ty_sized_bound_for_object_safety2, r=compiler-errors
Don't require associated types with Self: Sized bounds in `dyn Trait` objects

Trait objects require *all* associated types to be specified, even if the associated type has an explicit `where Self: Sized` bound. The following snippet does not compile on master, but does with this PR.

```rust
fn _assert_is_object_safe(_: &dyn Foo) {}

pub trait Foo {
    type Bar where Self: Sized;
}
```

In contrast, if a `Self: Sized` bound is added to a method, the methodjust isn't callable on trait objects, but the trait can be made object safe just fine.

```rust
fn _assert_is_object_safe(_: &dyn Foo) {}

pub trait Foo {
    fn foo() where Self: Sized;
}
```

This PR closes this inconsistency (though it still exists for associated constants).

Additionally this PR adds a new lint that informs users they can remove associated type bounds from their trait objects if those associated type bounds have a `where Self: Sized` bound, and are thus useless.

r? `@compiler-errors`
2023-07-05 08:48:04 +00:00
Boxy
62174bfe72 Deal with fallout 2023-07-05 09:46:30 +01:00
Boxy
8ac3ffe834 move ConstKind to typeir and move inherent impls to Const 2023-07-05 09:46:30 +01:00
yukang
f25463e848 Fix the issue of wrong diagnosis for extern pub fn 2023-07-05 16:25:46 +08:00
Oli Scherer
f80aec7429 Test that you can't circumvent the Sized bound check 2023-07-05 07:46:05 +00:00
Oli Scherer
3219993fa8 Only use a single loop over the associated types 2023-07-05 07:46:05 +00:00
Oli Scherer
243687a37c Prefer retain over hand-rolling an inefficient version of it 2023-07-05 07:46:05 +00:00
Oli Scherer
25e3785b86 Make unused_associated_type_bounds's lint level changeable 2023-07-05 07:46:05 +00:00
Oli Scherer
307b5ffff3 Make all generics_require_sized_self go through the query to get caching. 2023-07-05 07:46:05 +00:00
Oli Scherer
ce3cff47e0 Add more tests 2023-07-05 07:46:05 +00:00
Oli Scherer
62fbbac2d9 tidy: move a large function out of an even larger file 2023-07-05 07:46:04 +00:00
Oli Scherer
9e98feb84c Add some extra information to opaque type cycle errors 2023-07-05 07:43:35 +00:00
Oli Scherer
66ae9998d5 Remove redundant delay_span_bug 2023-07-05 07:43:35 +00:00
Oli Scherer
9cacfae3e9 Fix some tests to *only* test their main issue and not have secondary failures 2023-07-05 07:43:35 +00:00
Oli Scherer
a49b736568 Lint now-unnecessary associated type bounds 2023-07-05 07:42:53 +00:00
Oli Scherer
ca581f9161 Don't require associated types with Self: Sized bounds in dyn Trait objects 2023-07-05 07:42:53 +00:00
Oli Scherer
97d831d008 Show which type was not specialized on query cycle misuse 2023-07-05 07:30:28 +00:00
Oli Scherer
d6b82ff761 Remove a redundant argument 2023-07-05 07:30:28 +00:00
Michael Goulet
d2a1803d6f Winnow specializing impls 2023-07-05 06:18:48 +00:00
bors
6dab6dc5fc Auto merge of #112697 - tgross35:explain-markdown, r=oli-obk
Add simple markdown formatting to `rustc --explain` output

This is a second attempt at #104540, which is #63128 without dependencies.

This PR adds basic markdown formatting to `rustc --explain` output when available. Currently, the output just displays raw markdown: this works of course, but it really doesn't look very elegant. (output is `rustc --explain E0038`)

<img width="583" alt="image" src="https://github.com/rust-lang/rust/assets/13724985/ea418117-47af-455b-83c0-6fc59276efee">

After this patch, sample output from the same file:

<img width="693" alt="image" src="https://github.com/rust-lang/rust/assets/13724985/12f7bf9b-a3fe-4104-b74b-c3e5227f3de9">

This also obeys the `--color always/auto/never` command option. Behavior:

- If pager is available and supports color, print with formatting to the pager
- If pager is not available or fails print with formatting to stdout - otherwise without formatting
- Follow `--color always/never` if suppied
- If everything fails, just print plain text to stdout

r? `@oli-obk`
cc `@estebank`
(since the two of you were involved in the previous discussion)
2023-07-05 06:18:46 +00:00