Repository navigation
Rustc pull update - #2246
Open
workflows-stdarch[bot] wants to merge 126 commits into
Open
Rustc pull update#2246workflows-stdarch[bot] wants to merge 126 commits into
workflows-stdarch[bot] wants to merge 126 commits into
Conversation
Pull recent changes from https://github.com/rust-lang/rust via Josh. Upstream ref: rust-lang/rust@1396514 Filtered ref: rust-lang/rustc-dev-guide@30addec Upstream diff: rust-lang/rust@44e34e1...1396514 This merge was created using https://github.com/rust-lang/josh-sync.
Rustc pull update
Pull recent changes from https://github.com/rust-lang/rust via Josh. Upstream ref: rust-lang/rust@c78a294 Filtered ref: rust-lang/rustc-dev-guide@2f61036 Upstream diff: rust-lang/rust@1396514...c78a294 This merge was created using https://github.com/rust-lang/josh-sync.
Rustc pull update
all but one test suite support revisions directive
Pull recent changes from https://github.com/rust-lang/rust via Josh. Upstream ref: rust-lang/rust@eda4fc7 Filtered ref: rust-lang/rustc-dev-guide@b8ceb2d Upstream diff: rust-lang/rust@e7d90c6...eda4fc7 This merge was created using https://github.com/rust-lang/josh-sync.
Rustc pull update
add autodiff download instructions
Pull recent changes from https://github.com/rust-lang/rust via Josh. Upstream ref: rust-lang/rust@91021cc Filtered ref: rust-lang/rustc-dev-guide@3ea5559 Upstream diff: rust-lang/rust@1e21831...91021cc This merge was created using https://github.com/rust-lang/josh-sync.
Rustc pull update
add chapter on projection vs trait split
projection vs trait bound split cleanup
Pull recent changes from https://github.com/rust-lang/rust via Josh. Upstream ref: rust-lang/rust@9602bda Filtered ref: rust-lang/rustc-dev-guide@17d8cbf Upstream diff: rust-lang/rust@562dee4...9602bda This merge was created using https://github.com/rust-lang/josh-sync.
Rustc pull update
new rustc_on_unimplemented match values
Document AST lowering implementation
…otif-grp-links Fix GPU target notification group links
Update neovim instructions for neovim 0.11+
Pull recent changes from https://github.com/rust-lang/rust via Josh. Upstream ref: rust-lang/rust@e22c616 Filtered ref: rust-lang/rustc-dev-guide@592b9fb Upstream diff: rust-lang/rust@17584a1...e22c616 This merge was created using https://github.com/rust-lang/josh-sync.
Rustc pull update
…r/new-features-params implementing-new-features.md: update parameters to `declare_features!`
Mentions rustc-docs for rustc_private
Pull recent changes from https://github.com/rust-lang/rust via Josh. Upstream ref: rust-lang/rust@045b177 Filtered ref: rust-lang/rustc-dev-guide@2622ac3 Upstream diff: rust-lang/rust@f53b654...045b177 This merge was created using https://github.com/rust-lang/josh-sync.
…c-pull Rustc pull update
Add CI testing instructions for offload
Pull recent changes from https://github.com/rust-lang/rust via Josh. Upstream ref: rust-lang/rust@f2b291d Filtered ref: rust-lang/rustc-dev-guide@f2fe946 Upstream diff: rust-lang/rust@045b177...f2b291d This merge was created using https://github.com/rust-lang/josh-sync.
Rustc pull update
…0505-stage Fix name of first compiled stage, as stage0 is downloaded
Add docs for non defining vs defining uses in opaque types
Shrink MIR basic-block statement vectors to fit before caching them Every MIR basic block allocates a vector for its statements. These statement vectors grow via the usual heuristics, and never get re-shrunk to the correct size for their contents. This holds on to a fair bit of memory. Shrink them to fit right before we cache them. For an aws-sdk-ec2 check, this saves ~80 MiB of peak max-RSS (~1.45%), at the cost of 0.3% in instructions. r? nnethercote
Don't rely on `Global` handing out dangling pointers This code is incorrect and will leak memory once the global allocator supports zero-sized allocations, since `Box` doesn't deallocate zero-sized layouts
Pull recent changes from https://github.com/rust-lang/rust via Josh. Previous upstream ref: rust-lang/rust@260f1ac New upstream ref: rust-lang/rust@5ec56e3 Filtered ref: rust-lang/rustc-dev-guide@5a5b852 Upstream diff: rust-lang/rust@260f1ac...5ec56e3 This merge was created using https://github.com/rust-lang/josh-sync.
Rustc pull update
thread::LocalKey: document meaning of 'inner' function argument rust-lang/rust#92123 gave this function an argument, but without documenting what the argument means or does. I hope I reverse engineered this correctly. Cc @m-ou-se @joboet
remove redundant unsafe Related Zulipchat topic link: [Safety requirements of ARM hint APIs](https://rust-lang.zulipchat.com/#narrow/channel/219381-t-libs/topic/Safety.20requirements.20of.20ARM.20hint.20APIs/with/629609101) This PR tries to remove the unsafe in `core_arch/arm_shared/hint.rs`. As the Zulip topic discussed, these APIs have `#[cfg(any(...))]` attribute to restrict that these can only be compiled when specific architecture is met. So there seems no need to mark these APIs unsafe. For `__nop`, I found that in `riscv_shared` there exist the same `nop` instruction, but it is marked safe. The concrete path is `core_arch/src/riscv_shared/mod.rs` ```rust /// Generates the `NOP` instruction /// /// The NOP instruction does not change any architecturally visible state, except for /// advancing the `pc` and incrementing any applicable performance counters. #[inline] #[unstable(feature = "riscv_ext_intrinsics", issue = "114544")] pub fn nop() { unsafe { asm!("nop", options(nomem, nostack, preserves_flags)); } } ``` But there is a remaining question: For `wfe` and `wfi`, they can be configurable to trap. So I think if needed, I can complement some discription about trap. Hoping you can give me some suggestion.
Remove indirection for `Expr` in `Body` While this does not seem to have the measurable perf effect that I hoped, I still think this is nice to merge since it does theoretically remove one pointer dereference anytime a body is visited. I think this has no measurable effect because the `Expr` in the `Body` is allocated right next to the `Body` itself in the arena, so this should always be in L1 cache. This also works towards making the `Body` type simpler and possibly removing it, as after rust-lang/compiler-team#1044 it won't contain the Parameters anymore. I think we won't end up removing this type though because `BodyId` is a useful concept that is inherently linked to bodies. I might also take a look if there are other places in the compiler where a similar optimization might have measurable effect. r? @Kivooeo
rustc-dev-guide subtree update Subtree update of `rustc-dev-guide` to rust-lang/rustc-dev-guide@7a32071. Created using https://github.com/rust-lang/josh-sync. r? @ghost
…uwer Rollup of 3 pull requests Successful merges: - rust-lang/rust#163918 (remove redundant unsafe) - rust-lang/rust#164047 (Remove indirection for `Expr` in `Body`) - rust-lang/rust#164154 (rustc-dev-guide subtree update)
fast path for BTreeMap::Iter::next `BTreeMap::Iter` currently obtains the next kv through leaf navigation even when the next kv is in the same node. This PR adds fast paths for `BTreeMap::Iter` and `BTreeMap::Range`: when advancing does not cross a node boundary, the next edge and kv handle are constructed directly from their indexes. Result on aarch64: - map::iteration benchmark | baseline ns/iter | new ns/iter | change -- | -- | -- | -- map::iteration_20 | 362.74 ± 3.80 | 116.60 ± 1.46 | -67.9% map::iteration_1000 | 18743.31 ± 166.97 | 7181.88 ± 91.56 | -61.7% map::iteration_100000 | 2048028.00 ± 38646.50 | 893023.00 ± 6601.05 | -56.4% map::iteration_rev_20 | 398.15 ± 6.02 | 111.56 ± 1.24 | -72.0% map::iteration_rev_1000 | 19650.95 ± 120.38 | 6719.33 ± 33.38 | -65.8% map::iteration_rev_100000 | 2045171.75 ± 88630.03 | 826631.00 ± 7210.95 | -59.6% map::iteration_mut_20 | 332.97 ± 3.07 | 116.83 ± 9.19 | -64.9% map::iteration_mut_1000 | 17282.91 ± 142.66 | 7105.18 ± 64.92 | -58.9% map::iteration_mut_100000 | 1872307.50 ± 116359.10 | 930589.50 ± 11501.75 | -50.3% map::iteration_mut_rev_20 | 314.35 ± 5.69 | 99.88 ± 4.46 | -68.2% map::iteration_mut_rev_1000 | 16153.08 ± 77.67 | 6333.19 ± 9.81 | -60.8% map::iteration_mut_rev_100000 | 1863366.00 ± 90111.10 | 832836.00 ± 13949.30 | -55.3% - map::range_iteration Benchmark | Baseline (ns/iter) | New (ns/iter) | Change -- | -- | -- | -- map::range_iteration_20 | 380.62 ± 4.55 | 99.54 ± 1.49 | -73.8% map::range_iteration_1000 | 18,545.75 ± 193.17 | 5,870.29 ± 66.97 | -68.3% map::range_iteration_100000 | 1,886,226.50 ± 7,167.85 | 606,092.00 ± 5,516.90 | -67.9% map::range_iteration_rev_20 | 354.25 ± 3.59 | 98.63 ± 4.56 | -72.2% map::range_iteration_rev_1000 | 16,976.08 ± 186.23 | 5,604.28 ± 104.99 | -67.0% map::range_iteration_rev_100000 | 1,717,588.50 ± 6,813.65 | 579,875.00 ± 6,914.20 | -66.2% map::range_iteration_mut_20 | 336.01 ± 14.14 | 94.51 ± 2.88 | -71.9% map::range_iteration_mut_1000 | 15,517.95 ± 218.64 | 5,095.03 ± 44.77 | -67.2% map::range_iteration_mut_100000 | 1,604,837.00 ± 12,474.30 | 510,776.50 ± 1,607.21 | -68.2% map::range_iteration_mut_rev_20 | 293.22 ± 4.60 | 86.87 ± 0.85 | -70.4% map::range_iteration_mut_rev_1000 | 13,705.48 ± 126.54 | 5,035.34 ± 34.81 | -63.3% map::range_iteration_mut_rev_100000 | 1,393,440.00 ± 5,983.90 | 517,439.50 ± 4,770.85 | -62.9% - map::first_and_last Benchmark | Baseline (ns/iter) | New (ns/iter) | Change -- | -- | -- | -- map::first_and_last_0_nightly | 21.06 ± 0.44 | 21.05 ± 0.22 | -0.0% map::first_and_last_0_stable | 42.63 ± 1.34 | 42.53 ± 0.94 | -0.2% map::first_and_last_100_nightly | 47.77 ± 1.22 | 46.62 ± 0.60 | -2.4% map::first_and_last_100_stable | 302.27 ± 1.00 | 126.24 ± 4.89 | -58.2% map::first_and_last_10k_nightly | 73.27 ± 1.01 | 73.22 ± 0.92 | -0.1% map::first_and_last_10k_stable | 368.89 ± 4.71 | 172.73 ± 2.44 | -53.2% - set Benchmark | Baseline (ns/iter) | New (ns/iter) | Change -- | -- | -- | -- set::difference_random_100_vs_100 | 2,932.50 ± 5.08 | 1,377.14 ± 20.87 | -53.0% set::difference_random_100_vs_10k | 2,864.22 ± 9.71 | 2,294.26 ± 6.83 | -19.9% set::difference_random_10k_vs_100 | 193,483.00 ± 1,447.92 | 94,580.40 ± 1,345.45 | -51.1% set::difference_random_10k_vs_10k | 296,728.67 ± 2,560.70 | 145,264.67 ± 1,775.17 | -51.0% set::difference_staggered_100_vs_100 | 2,896.54 ± 10.72 | 1,477.26 ± 14.43 | -49.0% set::difference_staggered_100_vs_10k | 3,443.02 ± 72.13 | 2,715.43 ± 25.13 | -21.1% set::difference_staggered_10k_vs_10k | 293,658.33 ± 897.93 | 152,474.25 ± 2,005.09 | -48.1% set::intersection_100_neg_vs_100_pos | 14.39 ± 0.04 | 14.12 ± 0.14 | -1.9% set::intersection_100_neg_vs_10k_pos | 17.77 ± 0.06 | 17.63 ± 0.06 | -0.8% set::intersection_100_pos_vs_100_neg | 14.44 ± 0.06 | 13.99 ± 0.14 | -3.1% set::intersection_100_pos_vs_10k_neg | 17.72 ± 0.22 | 17.38 ± 0.03 | -1.9% set::intersection_10k_neg_vs_100_pos | 17.21 ± 0.07 | 17.16 ± 0.21 | -0.3% set::intersection_10k_neg_vs_10k_pos | 19.56 ± 0.11 | 19.38 ± 0.18 | -0.9% set::intersection_10k_pos_vs_100_neg | 17.08 ± 0.07 | 16.94 ± 0.17 | -0.8% set::intersection_10k_pos_vs_10k_neg | 19.42 ± 0.05 | 19.20 ± 0.22 | -1.1% set::intersection_random_100_vs_100 | 2,765.31 ± 19.46 | 1,250.69 ± 7.81 | -54.8% set::intersection_random_100_vs_10k | 3,149.91 ± 11.38 | 2,536.46 ± 43.26 | -19.5% set::intersection_random_10k_vs_100 | 3,140.24 ± 40.56 | 2,549.25 ± 46.99 | -18.8% set::intersection_random_10k_vs_10k | 280,063.67 ± 1,732.87 | 135,042.57 ± 1,463.94 | -51.8% set::intersection_staggered_100_vs_100 | 2,911.66 ± 21.58 | 1,556.65 ± 19.25 | -46.5% set::intersection_staggered_100_vs_10k | 3,082.33 ± 6.55 | 2,555.09 ± 18.25 | -17.1% set::intersection_staggered_10k_vs_10k | 294,074.33 ± 2,705.90 | 164,741.10 ± 650.57 | -44.0% set::is_subset_100_vs_100 | 2,804.09 ± 6.16 | 1,025.71 ± 9.23 | -63.4% set::is_subset_100_vs_10k | 2,343.67 ± 21.98 | 1,832.94 ± 30.82 | -21.8% set::is_subset_10k_vs_100 | 2.35 ± 0.01 | 2.46 ± 0.07 | +4.7% set::is_subset_10k_vs_10k | 274,705.33 ± 381.00 | 102,497.67 ± 1,043.56 | -62.7%
remove unnecessary char position file lookup for file!() What this PR does is remove unnecessary computation from the file!() macro. Each invocation performed a general lookup which calculated line and column for each invocation, when all it needs is the file name (the function itself discards the other information). My few tests (which I can provide if requested, not shown since the change is so trivial) showed on many (1000+ invocations) the compilation speedup is over 10x.
Lint default field values in types with type or const parameters as they won't be evaluated pre-mono
Default field values in types that have const generics are no longer being evaluated until mono. Emit a warn-by-default lint so that API designers are not caught of guard by this behavior.
```
warning: field `multiline_field` has a default value that is only checked when a value of `Z` is constructed
--> $DIR/field-references-param-accurate-span.rs:9:15
|
LL | struct Z<const X: usize> {
LL | multiline_field:
LL | ()
LL | = {
| _______________^
LL | | f::<X>();
LL | | panic!();
LL | | },
| |_____________^ this can't be const-evaluated until use
|
= help: structs and enums with type and const parameters only evaluate their default field values during construction, not eagerly when declared
= note: `#[warn(unevaluated_default_field_value)]` on by default
help: if this behavior is acceptable, allow the lint and preferably write a test relying on the default value
|
LL + #[expect(unevaluated_default_field_value)]
LL | struct Z<const X: usize> {
|
```
Support `Span` context in lints.
Fixes rust-lang/rust#146496.
Part of rust-lang/rust#132162.
Alternative to rust-lang/rust#163182.
CC @fmease @BoxyUwU
github-actions
Bot
force-pushed
the
rustc-pull
branch
from
October 12, 2026 04:22
5b9bc20 to
adb6972
Compare
…ercote
Avoid double-checking self.bound_vars' len in ValidateBoundVars' TypeVisitor impl
`ValidateBoundVars` has three instances of the following:
```rust
if list.len() <= idx {
panic!();
}
bound_ty.assert_eq(list.get(idx).unwrap());
```
The length check on `list` ends up being performed twice: once in the function itself and once in the implementation of the `get` method.
Replacing this by a `let-else` allows to ensure that the bound check is done exactly once. This results in IC improvements ranging from -0.26% to -0.08% except 5 benchmarks that end up being +0.05% to +0.08%.
r? nnethercote
@bors try @rust-timer queue
github-actions
Bot
force-pushed
the
rustc-pull
branch
from
October 15, 2026 04:13
adb6972 to
44a4ec5
Compare
compiler-builtins subtree update Subtree update of `compiler-builtins` to rust-lang/compiler-builtins@c8f734e. Created using https://github.com/rust-lang/josh-sync. r? @ghost
Add bootstrap step to build the TPDE plugin Adds TPDE steps to build the TPDE plugin on Linux targets and to optionally include it as part of the compiler sysroot for local development. Part of rust-lang/goals#802. This does not make any effort to be able to use TPDE in the compiler; all it does is stick a dylib in the place where it will eventually be needed. cc @ZuseZ4 r? Kobzol --------------- No LLMs were used to write this PR.
…ormat, r=petrochenkov Fix rmeta-link metadata not being read on AIX on XCOFF file format the payload goes to `.info` and the writer knows this from the file format on the other hand `rmeta_link::read` has no target so it always looks at the named section which on AIX returns nothing and returns `None` This PR is the fix so that it checks the format just like the writer does Note that there is no AIX in CI so I couldn't test it and don't have access to a machine so was not tested locally either
c-variadic: support ppcf128 tracking issue: rust-lang/rust#163323 cc rust-lang/rust#161787 On `powerpc64` this needs a change to lower the alignment of `ppcf128` to only 8 when read from a `VaList`. On `powerpc` this needs some additional changes to the logic to match GCC. Of course, in clang the ABI of this type is completely broken, we'll see if my fix for that gets merged llvm/llvm-project#225956.
…rtdev stdarch subtree update Subtree update of `stdarch` to eb9b302. Created using https://github.com/rust-lang/josh-sync. r? @ghost
…omez Generate redirect pages for hidden modules This fixes an intra-doc link, shown in the test case, where the private path gets used as the link target. Bug was reported [on the unofficial Discord](https://discord.com/channels/273534239310479360/273541522815713281/1561027634681348187).
…onthey Add more syscall doc aliases to std docs Some of these are quite trivial but we should optimize for the user not having to think about it. Blocked on rust-lang/std-dev-guide#64
breakpoint: warn against relying on SIGTRAP aborts This is mostly based on https://issues.chromium.org/issues/40625592 where Chromium devs report > Note: on Intel, the debugger breakpoint instruction (int3) is only one byte, while the fatal termination instruction (ud2) is two bytes. Unfortunately, crash reports seem to be indicating that int3 is non-fatal. Since it's important that TRAP_SEQUENCE() terminates, it cannot rely on int3. FWIW they seem to also be worried about debugger users > Make sure it crashes in case a developer steps past the breakpoint That was 7 years ago, so maybe there's an update we could get somehow -- does anyone have contacts with the Chrome team at Google? That said, looking at the code, they [still have the extra ud2 today](https://github.com/v8/v8/blob/5fe0aa3bc79c0a9d3ad546b79211f07105f09585/src/base/immediate-crash.h#L58). The code also has a comment saying ``` // TODO(https://crbug.com/958675): In theory, it should be possible to use just // int3. However, there are a number of crashes with SIGILL as the exception // code, so it seems likely that there's a signal handler that allows execution // to continue after SIGTRAP. ``` That link goes to the above bug. The code also has indications of an unfinished transition ``` // [...] In addition, the int3/bkpt/brk will // be removed in followups, so splitting it up like this now makes it easy to // land the followups. ``` That comment was written at least 5 years ago though. r? @tgross35 Cc @joshtriplett
…rouwer
improve diagnostic for literals in lint attributes
Improves the diagnostic for literals in lint attributes, like `#[allow("meow")]`.
Before:
```
error[E0452]: malformed lint attribute input
--> main.rs:5:9
|
5 | #[allow("meow")]
| ^^^^^^ bad attribute argument
error[E0452]: malformed lint attribute input
--> main.rs:8:8
|
8 | #[deny(1.0)]
| ^^^ bad attribute argument
```
After:
```
error: expected path to a lint, found literal
--> main.rs:5:9
|
5 | #[allow("meow")]
| ^^^^^^ not a lint
|
= help: try using `#[allow(meow)]`
error: expected path to a lint, found literal
--> main.rs:8:8
|
8 | #[deny(1.0)]
| ^^^ not a lint
|
= help: for example, write `#[deny(dead_code)]` for `dead_code`
```
Probably easier to review with whitespace changes turned off.
r? JonathanBrouwer
abort_on_unwind: clarify docs Tracking issue: rust-lang/rust#130338
…uwer Rollup of 6 pull requests Successful merges: - rust-lang/rust#164518 (stdarch subtree update) - rust-lang/rust#164497 (Generate redirect pages for hidden modules) - rust-lang/rust#113891 (Add more syscall doc aliases to std docs) - rust-lang/rust#163692 (breakpoint: warn against relying on SIGTRAP aborts) - rust-lang/rust#164486 (improve diagnostic for literals in lint attributes) - rust-lang/rust#164511 (abort_on_unwind: clarify docs)
This updates the rust-version file to b6eeb93c992027a03e1fd120b6652546396b142c.
Pull recent changes from https://github.com/rust-lang/rust via Josh. Previous upstream ref: rust-lang/rust@260f1ac New upstream ref: rust-lang/rust@b6eeb93 Filtered ref: 32f5516 Upstream diff: rust-lang/rust@260f1ac...b6eeb93 This merge was created using https://github.com/rust-lang/josh-sync.
github-actions
Bot
force-pushed
the
rustc-pull
branch
from
October 19, 2026 04:20
44a4ec5 to
2efcf9e
Compare
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
15 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Latest update from rustc.