| Age | Commit message (Collapse) | Author | Lines |
|---|
|
|
|
enable retpoline-related target features
|
|
|
|
|
|
Rollup of 7 pull requests
Successful merges:
- rust-lang/rust#129121 (Stabilize `tcp_quickack`)
- rust-lang/rust#142192 (De-duplicate f16 & f128 doctest attributes)
- rust-lang/rust#142193 (add tests for pattern binding drop order edge cases)
- rust-lang/rust#142222 (Dont make `ObligationCtxt`s with diagnostics unnecessarily)
- rust-lang/rust#142228 (rustc-dev-guide subtree update)
- rust-lang/rust#142231 (Run `calculate_matrix` job on `master` to cache citool builds)
- rust-lang/rust#142232 (add `Cargo.lock` to CI-rustc allowed list for non-CI env)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
add `Cargo.lock` to CI-rustc allowed list for non-CI env
Changes to dependencies usually require modifying `Cargo.toml`, which would already invalidate the CI-rustc cache if done in non-allowed paths. On non-CI environment, it should be safe to add `Cargo.lock` to the list of allowed paths as there is no real risk aside from a very rare false positive in cases like minor bumps to non-allowed path dependencies without modifying the `Cargo.toml` files.
Fixes rust-lang/rust#141986
|
|
Run `calculate_matrix` job on `master` to cache citool builds
As discussed in https://rust-lang.zulipchat.com/#narrow/channel/242791-t-infra/topic/PR.20ci.20seems.20much.20to.20slow/with/523028903, the current `rust-cache` solution for `citool` doesn't work, because we don't ever write to the cache from `master`, so the cache is empty on PR CI jobs.
This PR runs the `calculate_matrix` job on `master`, with the only motivation to actually prime the cache.
r? `@marcoieni`
|
|
rustc-dev-guide subtree update
r? `@ghost`
|
|
|
|
Signed-off-by: onur-ozkan <work@onurozkan.dev>
|
|
CI: rfl: move job forward to Linux v6.16-rc1
Another hopefully routine upgrade to Linux v6.16-rc1, just released.
r? `@lqd` `@Kobzol`
try-job: x86_64-rust-for-linux
`@rustbot` label A-rust-for-linux
`@bors` try
|
|
Do not checkout GCC submodule for the tidy job
This is not a fully general solution, but the GCC submodule checkout is so slow that I think it's worth it to special-case it. This brings down the time required to checkout submodules from ~1.5 minute to ~0.5 minute.
|
|
Only allow `bootstrap` cfg in rustc & related
Fixes https://github.com/rust-lang/rust/issues/142150
r? bootstrap
|
|
Remap compiler vs non-compiler sources differently (bootstrap side)
See [#t-compiler/help > Span pointing to wrong file location (`rustc-dev` component)](https://rust-lang.zulipchat.com/#narrow/channel/182449-t-compiler.2Fhelp/topic/Span.20pointing.20to.20wrong.20file.20location.20.28.60rustc-dev.60.20component.29/with/521087083).
The path remapping and unremapping for compiler sources (distributed via `rustc-dev` dist component) is broken because bootstrap currently remaps all sources unconditionally (if remapping is enabled) to the `/rustc/{hash}` form. However, the `rustc-dev` dist component (compiler sources) and `rust-src` dist component (library sources) unpacks differently:
- `rust-src` unpacks sources to a path like `$sysroot/lib/rustlib/src/rust`, whereas
- `rustc-dev` unpacks sources to a path like `$sysroot/lib/rustlib/rustc-src/rust`[^note],
meaning that the compiler need to unremap them differently. But the same remapping means that the compiler has no way to distinguish between compiler and non-compiler (esp. standard library) sources. To remedy this, this PR adopts the approach of:
- remapping compiler sources (corresponding to `rustc-dev` dist component) with `/rustc-dev/{hash}` (this is `RemapScheme::Compiler`), and
- remapping non-compiler sources (corresponding to `rust-src` dist component or other non-compiler sources) with `/rustc/{hash}` (this is `RemapScheme::NonCompiler`).
A different remapping allows the compiler to reverse the remapping differently.
This PR implements the bootstrap side. A follow-up compiler-side change is needed to implement the unremapping change to address the reported issue completely.
This PR introduces another env var `CFG_VIRTUAL_RUSTC_DEV_SOURCE_BASE_DIR` that is made available to the compiler when building compiler sources to know what the remap scheme for `rustc-dev` (`RemapScheme::Compiler`) is. Compiler sources are built with the compiler remapping scheme.
As far as I know, this change should not introduce new regressions, because the compiler source unremapping (through `rustc-dev`) is already broken.
[^note]: (Notice the `src` vs `rustc-src` difference.)
|
|
|
|
This allows us to reuse its cache on PR CI jobs.
|
|
So that we can make sure that they are reproducible locally.
|
|
|
|
|
|
|
|
|
|
Rollup of 12 pull requests
Successful merges:
- rust-lang/rust#141803 (Remove rustc's notion of "preferred" alignment AKA `__alignof`)
- rust-lang/rust#142053 (Add new Tier-3 targets: `loongarch32-unknown-none*`)
- rust-lang/rust#142089 (Replace all uses of sysroot_candidates with get_or_default_sysroot)
- rust-lang/rust#142108 (compiler: Add track_caller to AbiMapping::unwrap)
- rust-lang/rust#142132 (`tests/ui`: A New Order [6/N])
- rust-lang/rust#142162 (UnsafePinned: update get() docs and signature to allow shared mutation)
- rust-lang/rust#142171 (`tests/ui`: A New Order [7/N])
- rust-lang/rust#142179 (store `target.min_global_align` as an `Align`)
- rust-lang/rust#142183 (Added test for 30904)
- rust-lang/rust#142194 (Remove all unused feature gates from the compiler)
- rust-lang/rust#142199 (Do not free disk space in the `mingw-check-tidy` job)
- rust-lang/rust#142210 (Run `mingw-check-tidy` on auto builds)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Run `mingw-check-tidy` on auto builds
This has two advantages:
- It moves `auto` builds closer to being a superset of PR CI builds
- It allows us to reuse the Docker cache for the job in PR CI, thus speeding up the job in PR CI considerably
Discussed [here](https://rust-lang.zulipchat.com/#narrow/channel/242791-t-infra/topic/PR.20ci.20seems.20much.20to.20slow).
r? ``@Mark-Simulacrum``
|
|
Do not free disk space in the `mingw-check-tidy` job
It's not needed an it slows down the job considerably. It took ~2 minutes out of the total 8-9 minutes of running `mingw-check-tidy`.
|
|
Add new Tier-3 targets: `loongarch32-unknown-none*`
MCP: https://github.com/rust-lang/compiler-team/issues/865
NOTE: LoongArch32 ELF object support is available starting with object v0.37.0.
|
|
const-eval error: always say in which item the error occurred
I don't see why "is this generic" should make a difference. It may be reasonable to key this on whether the error occurs in a `const fn` that was invoked by a const (making it non-obvious which constant it is) vs inside the body of the const.
r? `@oli-obk`
|
|
|
|
Atomic intrinsics : use const generic ordering, part 2
This completes what got started in https://github.com/rust-lang/rust/pull/141507 by using a const generic for the ordering for all intrinsics. It is based on that PR; only the last commit is new.
Blocked on:
- https://github.com/rust-lang/rust/pull/141507
- https://github.com/rust-lang/rust/pull/141687
- https://github.com/rust-lang/stdarch/pull/1811
- https://github.com/rust-lang/rust/pull/141964
r? `@bjorn3`
|
|
|
|
Remove the comment on top as well, since that issue is now fixed in this
new tag.
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
|
|
Signed-off-by: Jakub Beránek <berykubik@gmail.com>
|
|
It's not needed an it slows down the job considerably.
Signed-off-by: Jakub Beránek <berykubik@gmail.com>
|
|
private-field diagnostic (previously it was only on no-such-field)
The difference between the diagnostics is that no-such-field is for record struct construction, while private-field is for dot syntax.
I tried to unify them, but there is a bit of uniqueness in each. This is possible but maybe not worth it.
Also, fix the quickfix when there is already a visibility to the field (replace it instead of appending to it).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Many of `std`'s dependency have a dependency on the crates.io
`compiler-builtins` when used with the feature
`rustc-std-workspace-core`. Use a Cargo patch to select the in-tree
version instead.
`compiler-builtins` is also added as a dependency of
`rustc-std-workspace-core` so these crates can remove their crates.io
dependency in the future.
|
|
|
|
`#[cfg_attr]` is a builtin attr, but it may still contain a macro.
|
|
Fix bootstrap tracing imports
|
|
paolobarbolini:stabilize-const_eq_ignore_ascii_case, r=Mark-Simulacrum
Stabilize `const_eq_ignore_ascii_case`
Tracking issue: rust-lang/rust#131719
Closes rust-lang/rust#131719
FCP Completed: https://github.com/rust-lang/rust/issues/131719#issuecomment-2941829167
|
|
Allow `#![doc(test(attr(..)))]` everywhere
This PR adds the ability to specify [`#![doc(test(attr(..)))]`](https://doc.rust-lang.org/nightly/rustdoc/write-documentation/the-doc-attribute.html#testattr) ~~at module level~~ everywhere in addition to allowing it at crate-root.
This is motivated by a recent PR #140323 (by ````@tgross35)```` where we have to duplicate 2 attributes to every single `f16` and `f128` doctests, by allowing `#![doc(test(attr(..)))]` at module level (and everywhere else) we can omit them entirely and just have (in both module):
```rust
#![doc(test(attr(feature(cfg_target_has_reliable_f16_f128))))]
#![doc(test(attr(expect(internal_features))))]
```
Those new attributes are appended to the one found at crate-root or at a previous module. Those "global" attributes are compatible with merged doctests (they already were before).
Given the small addition that this is, I'm proposing to insta-stabilize it, but I can feature-gate it if preferred.
Best reviewed commit by commit.
r? ````@GuillaumeGomez````
|
|
|
|
|
|
rustdoc: Further improve chapters and sections on testing
|
|
|
|
|
