Dylan DPC [Wed, 18 May 2022 06:41:15 +0000 (08:41 +0200)]
Rollup merge of #96378 - compiler-errors:trait-upcast-error, r=nagisa
Mention traits and types involved in unstable trait upcasting
Fixes #95972 by printing the traits being upcasted and the types being coerced that cause that upcasting...
---
the poor span mentioned in the original issue has nothing to do with trait upcasting diagnostic here...
> The original example I had that made me run into this issue had an even longer expression there (multiple chained
iterator methods) which just got all highlighted as one big block saying "somewhere here trait coercion is used and it's not allowed".
I don't think I can solve that issue in general without fixing the ObligationCauseCode and span that gets passed into Coerce.
Dylan DPC [Wed, 18 May 2022 06:41:13 +0000 (08:41 +0200)]
Rollup merge of #94639 - compiler-errors:rval-mutref, r=wesleywiser
Suggest dereferencing non-lval mutable reference on assignment
1. Adds deref suggestions for LHS of assignment (or assign-binop) when it implements `DerefMut`
2. Fixes missing deref suggestions for LHS when it isn't a place expr
Yuki Okushi [Tue, 17 May 2022 22:41:01 +0000 (07:41 +0900)]
Rollup merge of #97116 - RalfJung:ref-validity, r=oli-obk
interpret/validity: reject references to uninhabited types
According to https://doc.rust-lang.org/reference/behavior-considered-undefined.html, this is definitely UB. And we can check this without actually looking up anything in memory, we just need the reference value and its type, making this a great candidate for a validity invariant IMO and my favorite resolution of https://github.com/rust-lang/unsafe-code-guidelines/issues/77.
With this PR, Miri with `-Zmiri-check-number-validity` implements all my preferred options for what the validity invariants of our types could be. :)
CTFE has been doing recursive checking anyway, so this is backwards compatible but might change the error output. I will submit a PR with the new Miri tests soon.
Yuki Okushi [Tue, 17 May 2022 22:40:59 +0000 (07:40 +0900)]
Rollup merge of #97097 - chorman0773:add_tmm_clobers, r=joshtriplett
Add tmm_reg clobbers
This adds support for naming the 8 tile registers from intel AMX as clobbers from `asm!` invocations on x86_64 (only). It does not add the registers as input or output operands.
Yuki Okushi [Tue, 17 May 2022 22:40:55 +0000 (07:40 +0900)]
Rollup merge of #96647 - Enselic:fix-hrtb-for-wherepredicate, r=CraftSpider
rustdoc-json: Fix HRTBs for WherePredicate::BoundPredicate
Information about HRTBs are already present for `GenericBound:: TraitBound` and `FunctionPointer`. This PR adds HRTB info also to `WherePredicate::BoundPredicate`.
Use the same field name and type as for the other ones (`generic_params: Vec<GenericParamDef>`). I have verified that this gives rustdoc JSON clients the data they need and in a format that is easy to work with (see https://github.com/Enselic/public-api/pull/92).
I will be happy to add tests for this change (and bump `FORMAT_VERSION` which I just realized I forgot), but it is always nice to get one round of feedback first, so that I don't put a lot of effort into tests that then have to be discarded.
bors [Tue, 17 May 2022 19:50:44 +0000 (19:50 +0000)]
Auto merge of #92570 - jsha:rustdoc-search-refactor, r=GuillaumeGomez
Simplify rustdoc search test
Previously, rustdoc search attempted to parse search.js and extract out only certain methods and variables.
This change makes search.js and search-index.js loadable as [CommonJS modules](https://nodejs.org/api/modules.html#modules-commonjs-modules), so they can be loaded directly.
As part of that change, I had to separate execSearch from interacting with the DOM. This wound up being a nice cleanup that made more explicit what inputs it was taking.
I removed search.js' dependency on storage.js by moving hasOwnPropertyRustdoc directly into search.js, and replacing onEach with forEach in a path that is called by the tester.
Previously, search.js relied on the DOM and the `window` object. It can now be
loaded in the absence of the DOM, for instance by Node. The same is true of
search-index.js.
This allows removing a lot of code from src/tools/rustdoc-js/tester.js that
tried to parse search.js and extract specific functions that were needed for
testing.
bors [Tue, 17 May 2022 15:04:50 +0000 (15:04 +0000)]
Auto merge of #96959 - nbdd0121:unwind, r=Amanieu
Prevent unwinding when `-C panic=abort` is used regardless declared ABI
Ensures that Rust code will abort with `-C panic=abort` regardless ABI used.
```rust
extern "C-unwind" {
fn may_unwind();
}
// Will be nounwind with `-C panic=abort`, despite `C-unwind` ABI.
pub unsafe extern "C-unwind" fn rust_item_that_can_unwind() {
may_unwind();
}
```
Current behaviour is that unwind will propagate through. While the current behaviour won't cause unsoundness it is inconsistent with the text reading of [RFC2945](https://rust-lang.github.io/rfcs/2945-c-unwind-abi.html).
I tweaked `fn_can_unwind` instead of tweaking `AbortUnwindingCalls` because this approach would allow Rust (non-direct) callers to also see that this function is nounwind, so it can prevent excessive landing pads generation.
For more discussions: https://rust-lang.zulipchat.com/#narrow/stream/210922-project-ffi-unwind/topic/soundness.20in.20mixed.20panic.20mode.
Yuki Okushi [Tue, 17 May 2022 10:01:32 +0000 (19:01 +0900)]
Rollup merge of #97079 - SparrowLii:successors, r=lcnr
Change `Successors` to `impl Iterator<Item = BasicBlock>`
This PR fixes the FIXME in `compiler\rustc_middle\src\mir\mod.rs`.
This can omit several `&`, `*` or `cloned` operations on Successros' generated elements
Yuki Okushi [Tue, 17 May 2022 10:01:30 +0000 (19:01 +0900)]
Rollup merge of #97009 - fortanix:raoul/unused_macro_sgx, r=thomcc
Allow `unused_macro_rules` in path tests
PR #96150 adds a new lint to warn about unused macro rules (arms/matchers). This causes errors in `library/std/src/path/tests.rs` on the `x86_64-fortanix-unknown-sgx` platform. This PR fixes compilation errors on that platform by allowing unused macro rules.
This parenthetical is problematic, because it suggests that the following is sound:
```rust
let layout = Layout::new::<[u8; 32]>();
let p1 = alloc(layout);
let p2 = realloc(p1, layout, 32);
if p1 == p2 {
p1.write([0; 32]);
dealloc(p1, layout);
} else {
dealloc(p2, layout);
}
```
At the very least, this isn't the case for [ANSI `realloc`](https://en.cppreference.com/w/c/memory/realloc)
> The original pointer `ptr` is invalidated and any access to it is undefined behavior (even if reallocation was in-place).
and [Windows `HeapReAlloc`](https://docs.microsoft.com/en-us/windows/win32/api/heapapi/nf-heapapi-heaprealloc) is unclear at best (`HEAP_REALLOC_IN_PLACE_ONLY`'s description may imply that the old pointer may be used if `HEAP_REALLOC_IN_PLACE_ONLY` is provided).
The conservative position is to just remove the parenthetical.
cc `@rust-lang/wg-unsafe-code-guidelines` `@rust-lang/wg-allocators`
bors [Sun, 15 May 2022 21:17:46 +0000 (21:17 +0000)]
Auto merge of #97057 - bjorn3:sync_cg_clif-2022-05-15, r=bjorn3
Sync rustc_codegen_cranelift
Since the last sync there have mostly been fixes of various sorts. I also changed cg_clif from using a custom driver to `-Zcodegen-backend` when built as separate project. When built as part of rust it was already using `-Zcodegen-backend`.
bors [Sun, 15 May 2022 18:52:28 +0000 (18:52 +0000)]
Auto merge of #97063 - Dylan-DPC:rollup-u5el7hb, r=Dylan-DPC
Rollup of 4 pull requests
Successful merges:
- #96947 (Add rustc_nonnull_optimization_guaranteed to Owned/Borrowed Fd/Socket)
- #97021 (Added note in documentation)
- #97042 (Add new eslint rule about brace style)
- #97060 (Fix use of SetHandleInformation on UWP)
Dylan DPC [Sun, 15 May 2022 16:41:27 +0000 (18:41 +0200)]
Rollup merge of #97060 - bdbai:fix/uwphandle, r=ChrisDenton
Fix use of SetHandleInformation on UWP
The use of `SetHandleInformation` (introduced in #96441 to make `HANDLE` inheritable) breaks UWP builds because it is not available for UWP targets.
Proposed workaround: duplicate the `HANDLE` with `inherit = true` and immediately close the old one. Traditional Windows Desktop programs are not affected.
Dylan DPC [Sun, 15 May 2022 16:41:25 +0000 (18:41 +0200)]
Rollup merge of #96947 - sunfishcode:sunfishcode/rustc-nonnull-optimization-guaranteed, r=joshtriplett
Add rustc_nonnull_optimization_guaranteed to Owned/Borrowed Fd/Socket
PR #94586 added support for using
`rustc_nonnull_optimization_guaranteed` on values where the "null" value
is the all-ones bitpattern.
Now that #94586 has made it to the stage0 compiler, add
`rustc_nonnull_optimization_guaranteed` to `OwnedFd`, `BorrowedFd`,
`OwnedSocket`, and `BorrowedSocket`, since these types all exclude
all-ones bitpatterns.
This allows `Option<OwnedFd>`, `Option<BorrowedFd>`, `Option<OwnedSocket>`,
and `Option<BorrowedSocket>` to be used in FFI declarations, as described
in the [I/O safety RFC].
bors [Sun, 15 May 2022 11:19:27 +0000 (11:19 +0000)]
Auto merge of #96885 - petrochenkov:linkstrict2, r=cjgillot,luqmana
rustc: Stricter checking for #[link] attributes
A subset of https://github.com/rust-lang/rust/pull/94962 that doesn't touch library renaming/reordering/deduplication.
`#[link]` attributes are checked for all kinds of unexpected arguments inside them.
I also tried to make wording for these errors more consistent, that's why some existing errors are changed, including errors for command line `-l` options.
Spans are also made more precise where possible.
Matthias Krüger [Sun, 15 May 2022 06:10:44 +0000 (08:10 +0200)]
Rollup merge of #97041 - eddyb:nixos-llvm-ci-patchelf, r=Mark-Simulacrum
Fix `download-ci-llvm` NixOS patching for `.so`s.
See https://github.com/rust-lang/rust/pull/95170#discussion_r872960686 - in short, `Path::ends_with` doesn't do the same thing as `str::ends_with`, and can only be used to check for whole file names, not extensions.
With this PR, I get the full suite of:
```
extracting /home/eddy/Projects/rust-A/build/cache/llvm-ebb80ec4e90f8622440f3e33562db0d6e6c66555-true/rust-dev-nightly-x86_64-unknown-linux-gnu.tar.xz to /home/eddy/Projects/rust-A/build/x86_64-unknown-linux-gnu/ci-llvm
info: you seem to be using Nix. Attempting to patch /home/eddy/Projects/rust-A/build/x86_64-unknown-linux-gnu/ci-llvm/bin/llvm-config
/nix/store/r4bzq2xilvv8fmqjg626hzwi22ah3hf4-rust-stage0-dependencies
info: you seem to be using Nix. Attempting to patch /home/eddy/Projects/rust-A/build/x86_64-unknown-linux-gnu/ci-llvm/bin/FileCheck
info: you seem to be using Nix. Attempting to patch /home/eddy/Projects/rust-A/build/x86_64-unknown-linux-gnu/ci-llvm/lib/libLLVM-14-rust-1.62.0-nightly.so
```
(that `libLLVM-14-rust-1.62.0-nightly.so` at the end having been missing before)
Matthias Krüger [Sun, 15 May 2022 06:10:42 +0000 (08:10 +0200)]
Rollup merge of #97032 - est31:unused_macro_rules, r=petrochenkov
Allow the unused_macro_rules lint for now
It was newly added by #96150 with warn by default, which is great as it gave exposure to the community, and their feedback gave me ideas for improvements.
Allowing the lint is good for two reasons:
* It makes the transition easier as e.g. allow directives won't fire the unknown lint warning once it is turned to warn by default in the future. The [commit that allowed the lint in fuchsia](https://fuchsia.googlesource.com/fuchsia/+/9d8f96517c3963de2f0e25598fd36061914524cd%5E%21/) had to allow unknown lints for example.
This is especially important compared to other lints in the unused group,
because the _ prefix trick doesn't exist for macro rules, allowing is the
only option (either of unused_macro_rules, or of the entire unused group,
but that is not as informative to readers). Allowing the lint also makes it
possible to work on possible heuristics for disabling the macro in specific
cases.
* It gives time for implementing heuristics for when to suppress the lint, e.g.
when `compile_error!` is invoked by that arm (so it's only there to yield an error).
Matthias Krüger [Sun, 15 May 2022 06:10:42 +0000 (08:10 +0200)]
Rollup merge of #96958 - GuillaumeGomez:settings-menu-display, r=jsha
Improve settings menu display and remove theme menu
We talked about improving the settings menu and we mentioned that firefox pocket was a nice inspiration so I implemented it. The result looks like this:
bors [Sun, 15 May 2022 02:11:35 +0000 (02:11 +0000)]
Auto merge of #97016 - Mark-Simulacrum:bump-version, r=Mark-Simulacrum
Bump to 1.63
r? `@Mark-Simulacrum`
Posting this now, but will only approve later today / early tomorrow to give a little more time for not-yet-approved PRs to land on master (e.g., https://github.com/rust-lang/rust/pull/96970).
bors [Sat, 14 May 2022 23:53:11 +0000 (23:53 +0000)]
Auto merge of #96883 - jackh726:early-binder-2, r=oli-obk
Add EarlyBinder
Chalk has no concept of `Param` (https://github.com/rust-lang/chalk/blob/e0ade19d139bc784384acc6736cd960c91dd55a1/chalk-ir/src/lib.rs#L579) or `ReEarlyBound` (https://github.com/rust-lang/chalk/blob/e0ade19d139bc784384acc6736cd960c91dd55a1/chalk-ir/src/lib.rs#L1308). Everything is just "bound" - the equivalent of rustc's late-bound. It's not completely clear yet whether to move everything to the same time of binder in rustc or add `Param` and `ReEarlyBound` in Chalk.
Either way, tracking when we have or haven't already substituted out these in rustc can be helpful.
As a first step, I'm just adding a `EarlyBinder` newtype that is required to call `subst`. I also add a couple "transparent" `bound_*` wrappers around a couple query that are often immediately substituted.