I added this back in 111999, but I no longer think it's a good idea
- It had to get scaled back to only power-of-two things to not break a bunch of targets
- LLVM seems to be getting better at memcpy removal anyway
- Introducing vector instructions has seemed to sometimes (115515) make autovectorization worse
So this removes it from the codegen crates entirely, and instead just tries to use <https://doc.rust-lang.org/nightly/nightly-rustc/rustc_codegen_ssa/traits/builder/trait.BuilderMethods.html#method.typed_place_copy> instead of direct `memcpy` so things will still use load/store for immediates.
Rollup of 8 pull requests
Successful merges:
- #123254 (Do not allocate for ZST ThinBox (attempt 2 using const_allocate))
- #123626 (Add MC/DC support to coverage test tools)
- #123638 (rustdoc: synthetic auto: filter out clauses from the implementor's ParamEnv)
- #123653 (Split `non_local_definitions` lint tests in separate test files)
- #123658 (Stop making any assumption about the projections applied to the upvars in the `ByMoveBody` pass)
- #123662 (Don't rely on upvars being assigned just because coroutine-closure kind is assigned)
- #123665 (Fix typo in `Future::poll()` docs)
- #123672 (compiletest: unset `RUSTC_LOG_COLOR`)
r? `@ghost`
`@rustbot` modify labels: rollup
compiletest: unset `RUSTC_LOG_COLOR`
If this leaks in from the environment then it can make tests fail when they deliberately trigger `WARN` or `ERROR` logging, currently this stops these tests from failing if you set `RUSTC_LOG_COLOR=always` in the parent environment:
- `tests/ui/coherence/occurs-check/associated-type.rs#next`
- `tests/ui/coherence/occurs-check/associated-type.rs#old`
- `tests/ui/higher-ranked/structually-relate-aliases.rs`
- `tests/ui/self/arbitrary-self-from-method-substs.rs#default`
- `tests/ui/traits/next-solver/issue-118950-root-region.rs`
Don't rely on upvars being assigned just because coroutine-closure kind is assigned
Previously, code relied on the implicit assumption that if a coroutine-closure's kind variable was constrained, then its upvars were also constrained. This is because we assign all of them at once at the end up upvar analysis.
However, there's another way that a coroutine-closure's kind can be constrained: from a signature hint in closure signature deduction. After #123350, we use these hints, which means the implicit assumption above no longer holds.
This PR adds the necessary checks so that we don't ICE.
r? oli-obk
Stop making any assumption about the projections applied to the upvars in the `ByMoveBody` pass
So it turns out that because of subtle optimizations like [`truncate_capture_for_optimization`](ab5bda1aa7/compiler/rustc_hir_typeck/src/upvar.rs (L2351)), we simply cannot make any assumptions about the shape of the projections applied to the upvar locals in a coroutine body.
So stop doing that -- the code is resilient to such projections, so the assertion really existed only to "protect against the unknown".
r? oli-obk
Fixes#123650
Split `non_local_definitions` lint tests in separate test files
This PR splits the giant `non_local_definitions` lint UI test in separate test files.
This change is extracted from #123594 (where it was requested https://github.com/rust-lang/rust/pull/123594#discussion_r1555261772), to ease the review of the other PR and to reduce the size of the other PR.
r? ``@compiler-errors``
rustdoc: synthetic auto: filter out clauses from the implementor's ParamEnv
... not just the elaborated clauses.
Fixes another regression introduced by me in #123340, oops!
Fixes https://github.com/rust-lang/rust/pull/123340#issuecomment-2034195786, cc ``@tamird.``
An earlier local iteration of branch `rustdoc-simplify-auto-trait-impl-synth` (PR #123340) contained a fix for issue #111101 before I decided to limit the scope. I must've introduced this bug when manually reverting that part of the code.
r? ``@GuillaumeGomez`` or rustdoc
Add MC/DC support to coverage test tools
Extracted and squashed from #123409 by `@ZhuUx.`
These updates to the coverage test tools can land ahead of the main changes, slightly reducing the size and complexity of that PR.
---
The `coverage-dump` changes aren't directly tested in this PR, but the tests in #123409 demonstrate that they do work on real MC/DC coverage output.
`@rustbot` label +A-code-coverage
Do not allocate for ZST ThinBox (attempt 2 using const_allocate)
There's PR https://github.com/rust-lang/rust/pull/123184 which avoids allocation for ZST ThinBox.
That PR has an issue with unsoundness with padding in `MaybeUninit` (see comments in that PR). Also that PR relies on `Freeze` trait.
This PR is much simpler implementation which does not have this problem, but it uses `const_allocate` feature.
`@oli-obk` suggested that `const_allocate` should not be used for that feature. But I like how easy it to do this feature with `const_allocate`. Maybe it's OK to use `const_allocate` while `ThinBox` is unstable? Or, well, we can abandon this PR.
r? `@oli-obk`
Ensure we do not accidentally insert new early aborts in the analysis passes
pulling the infallible part out into a separate function makes sure that someone needs to change the signature in order to regress this.
We only want to stop compilation in the presence of errors after all analyses are done, but before we start running lints.
per-item we can still stop doing work if previous queries returned errors, but that's a separate story.
If this leaks in from the environment then it can make tests fail when
they deliberately trigger `WARN` or `ERROR` logging.
Signed-off-by: David Wood <david@davidtw.co>
Fix UI tests with dist-vendored dependencies
There is already a workaround in `compiletest` to deal with custom
`CARGO_HOME` using `-Zignore-directory-in-diagnostics-source-blocks={}`.
A similar need exists when dependencies come from the local `vendor`
directory, which distro builds often use, so now we ignore that too.
Also, `issue-21763.rs` was normalizing `hashbrown-` paths, presumably
expecting a version suffix, but the vendored path doesn't include the
version. Now that matches `[\\/]hashbrown` instead.
KCFI: Use legal charset in shim encoding
To separate `ReifyReason::FnPtr` from `ReifyReason::VTable`, we hyphenated the shims. Hyphens are not actually legal, but underscores are, so use those instead.
r? `@compiler-errors`
rustdoc: slightly clean up the synthesis of blanket impls
Small follow-up to #123340 as promised in https://github.com/rust-lang/rust/pull/123340#discussion_r1546918604. No functional changes whatsoever.
* inline the over-engineered “type namespace” (struct) `BlanketImplFinder` just like I did with `AutoTraitFinder` in #123340
* use the new `synthesize_*` terminology over the old nondescript / misleading `get_*` one
* inline a `use super::*;` (not super modular, lead to `clean/mod.rs` (!) accumulating cruft)
* use `tracing` properly
r? GuillaumeGomez or rustdoc
Update books
## rust-lang/book
23 commits in 19c40bfd2d57641d962f3119a1c343355f1b3c5e..3131aa4642c627a24f523c82566b94a7d920f68c
2024-04-05 22:09:59 UTC to 2024-03-27 18:14:05 UTC
- Correct the description of Listing 5-6 (rust-lang/book#3878)
- Minor text improvements (rust-lang/book#3790)
- Update full code reference main.rs in ch20-03 (rust-lang/book#3857)
- Make Note text in ch. 20 consistent with other notes (rust-lang/book#3876)
- Consistency fix: start note sentence with a capital (rust-lang/book#3652)
- Update README.md (rust-lang/book#3656)
- Update panic! formatting style for Guess example (rust-lang/book#3767)
- Small correction in ch13-01 (rust-lang/book#3780)
- Fix a wording typo in ch07 (rust-lang/book#3694)
- Fix a typo: remove an extra 's' from ch. 18.01 (rust-lang/book#3874)
- Remove redundant words (rust-lang/book#3672)
- Fix#3703 (rust-lang/book#3704)
- Update loop to Result: 10 (rust-lang/book#3711)
- Added "--" between run and args on ch12-03-l12 (rust-lang/book#3726)
- Add maintenance time section in appendix-07-nightly-rust.md (rust-lang/book#3859)
- Update the link to the farsi translation repository (rust-lang/book#3839)
- Update README to use --locked for installing mdbook (rust-lang/book#3830)
- Reword "`union`s" to "a `union`" (rust-lang/book#3738)
- Fix two typos in 04-03 and 07-03 (rust-lang/book#3849)
- Fix typo in Chapter 7 Section 3 (rust-lang/book#3743)
- Improve ch03-05-control-flow collection looping wording (rust-lang/book#3758)
- Fix missing column separator (rust-lang/book#3855)
- Update compiler message (rust-lang/book#3856)
## rust-lang/edition-guide
3 commits in 98b33e9a441457b0a491fe1be90e7de64eafc3e5..eb3eb80e106d03250c1fb7c5666b1c8c59672862
2024-03-27 20:44:27 UTC to 2024-03-26 19:26:15 UTC
- some mdbook conveniences (rust-lang/edition-guide#297)
- typo (rust-lang/edition-guide#296)
- Clean up the `editions/index.md` page (rust-lang/edition-guide#294)
## rust-embedded/book
1 commits in 2e95fc2fd31d669947e993aa07ef10dc9828bee7..aa7d4b0b4653ddb47cb1de2036d090ec2ba9dab1
2024-04-05 07:42:54 UTC to 2024-04-05 07:42:54 UTC
- Dependencies: changed "qemu-arch-extra" to "qemu-system-arm" on arch section (rust-embedded/book#368)
## rust-lang/nomicon
2 commits in 6bc2415218d4dd0cb01433d8320f5ccf79c343a1..0d5f88475fe285affa6dbbc806e9e44d730797c0
2024-04-06 13:51:11 UTC to 2024-04-03 02:23:07 UTC
- chore: fix typo (rust-lang/nomicon#448)
- add link to reference about undefined behavior (rust-lang/nomicon#447)
## rust-lang/reference
3 commits in 984b36eca4b9293df04d5ba4eb5c4f77db0f51dc..55694913b1301cc809f9bf4a1ad1b3d6920efbd9
2024-04-03 21:31:14 UTC to 2024-04-01 19:56:13 UTC
- Add the `#[diagnostic]` attribute namespace and the `#[diagnostic::on_unimplemented]` feature to the reference (rust-lang/reference#1449)
- type-layout: be more specific about 32-bit alignments (rust-lang/reference#1393)
- Fix clippy warning in procedural macro example (rust-lang/reference#1488)
## rust-lang/rust-by-example
1 commits in 7601e0c5ad29d5bd3b518700ea63fddfff5915a7..60d34b5fd33db1346f9aabfc0c9d0bda6c8e42be
2024-04-07 13:00:53 UTC to 2024-04-07 13:00:53 UTC
- chore: fix some typos (rust-lang/rust-by-example#1833)
## rust-lang/rustc-dev-guide
11 commits in ffa246b7fd95a96e1cd54883e613aed42c32547d..b77a34bd46399687b4ce6a17198e9f316c988794
2024-04-06 20:41:09 UTC to 2024-03-27 08:49:05 UTC
- Explicitly mention compiletest directives are supported in rmake.rs (rust-lang/rustc-dev-guide#1949)
- Add docs for sharded descriptions (rust-lang/rustc-dev-guide#1959)
- Add basic docs for the new `aux-bin` header (rust-lang/rustc-dev-guide#1942)
- Add needs-threads header command (rust-lang/rustc-dev-guide#1943)
- Fix some broken links under bootstrapping. (rust-lang/rustc-dev-guide#1958)
- Replace -Zno-parallel-llvm with -Zno-parallel-backend (rust-lang/rustc-dev-guide#1957)
- Rewrite the `Parameter Environments` chapter (rust-lang/rustc-dev-guide#1953)
- Add quickstart for how to build and run the compiler (rust-lang/rustc-dev-guide#1951)
- Delete length check (rust-lang/rustc-dev-guide#1952)
- Fix some comments (rust-lang/rustc-dev-guide#1950)
- add opaque-types-region-inference-restrictions (rust-lang/rustc-dev-guide#1948)
[rustdoc] [GUI tests] Make theme switching closer to reality
Better to actually perform actions user do rather than only testing the change through local storage.
As for `browser-ui-test` update: I updated `puppeteer` version (to `0.19.4`) and fixed a bug when displaying the file if it came from an `include`.
r? `@notriddle`
sanitizers: Create the rustc_sanitizers crate
Create the `rustc_sanitizers` crate and move the source code for the CFI and KCFI sanitizers to it. The tracking issue for reviewing and moving sanitizers into a compiler crate is #123619. This is part of our work to organize and stabilize support for the sanitizers. (See our roadmap at https://hackmd.io/`@rcvalle/S1Ou9K6H6.)`
Replace some `CrateStore` trait methods with hooks.
Just like with the `CrateStore` trait, this avoids the cyclic definition issues with `CStore` being
defined after TyCtxt, but needing to be used in TyCtxt.
Re-enable the early otherwise branch optimization
Closes#95162. Fixes#119014.
This is the first part of #121397.
An invalid enum discriminant can come from anywhere. We have to check to see if all successors contain the discriminant statement. This should have a pass to hoist instructions.
r? cjgillot
Pass list of defineable opaque types into canonical queries
This eliminates `DefiningAnchor::Bubble` for good and brings the old solver closer to the new one wrt cycles and nested obligations. At that point the difference between `DefiningAnchor::Bind([])` and `DefiningAnchor::Error` was academic. We only used the difference for some sanity checks, which actually had to be worked around in places, so I just removed `DefiningAnchor` entirely and just stored the list of opaques that may be defined.
fixes#108498
fixes https://github.com/rust-lang/rust/issues/116877
* [x] run crater
- https://github.com/rust-lang/rust/pull/122077#issuecomment-2013293931
There is already a workaround in `compiletest` to deal with custom
`CARGO_HOME` using `-Zignore-directory-in-diagnostics-source-blocks={}`.
A similar need exists when dependencies come from the local `vendor`
directory, which distro builds often use, so now we ignore that too.
Also, `issue-21763.rs` was normalizing `hashbrown-` paths, presumably
expecting a version suffix, but the vendored path doesn't include the
version. Now that matches `[\\/]hashbrown` instead.
To separate `ReifyReason::FnPtr` from `ReifyReason::VTable`, we
hyphenated the shims. Hyphens are not actually legal, but underscores
are, so use those instead.
CFI: Fix ICE in KCFI non-associated function pointers
We oddly weren't testing the more usual case of casting non-methods to function pointers. The KCFI shim insertion logic would ICE on these due to asking for an irrefutable associated item if we cast a function to a function pointer without needing a traditional shim.
r? `@compiler-errors`
parser: reduce visibility of unnecessary public `UnmatchedDelim`
`lexer::UnmatchedDelim` struct in `rustc_parse` is unnecessary public outside of the crate. This commit reduces the visibility to `pub(crate)`.
Beside, this removes unnecessary field `expected_delim` that causes warnings after changing the visibility.
Remove unnecessary cast from `LLVMRustGetInstrProfIncrementIntrinsic`
(Noticed while reviewing #123409.)
This particular cast appears to have been copied over from clang, but there are plenty of other call sites in clang that don't bother with a cast here, and it works fine without one.
For context, `llvm::Intrinsic::ID` is a typedef for `unsigned`, and `llvm::Intrinsic::instrprof_increment` is a member of `enum IndependentIntrinsics : unsigned`.
---
The formatting change in `unwrap(M)` is the result of manually running `clang-format` on this file, and then reverting all changes other than the ones affecting these lines.
Don't emit divide-by-zero panic paths in `StepBy::len`
I happened to notice today that there's actually two such calls emitted in the assembly: <https://rust.godbolt.org/z/1Wbbd3Ts6>
Since they're impossible, hopefully telling LLVM that will also help optimizations elsewhere.
Fix `ByMove` coroutine-closure shim (for 2021 precise closure capturing behavior)
This PR reworks the way that we perform the `ByMove` coroutine-closure shim to account for the fact that the upvars of the outer coroutine-closure and the inner coroutine might not line up due to edition-2021 closure capture rules changes.
Specifically, the number of upvars may differ *and/or* the inner coroutine may have additional projections applied to an upvar. This PR reworks the information we pass into the `ByMoveBody` MIR visitor to account for both of these facts.
I tried to leave comments explaining exactly what everything is doing, but let me know if you have questions.
r? oli-obk
Safe Transmute: Compute transmutability from `rustc_target::abi::Layout`
In its first step of computing transmutability, `rustc_transmutability` constructs a byte-level representation of type layout (`Tree`). Previously, this representation was computed for ADTs by inspecting the ADT definition and performing our own layout computations. This process was error-prone, verbose, and limited our ability to analyze many types (particularly default-repr types).
In this PR, we instead construct `Tree`s from `rustc_target::abi::Layout`s. This helps ensure that layout optimizations are reflected our analyses, and increases the kinds of types we can now analyze, including:
- default repr ADTs
- transparent unions
- `UnsafeCell`-containing types
Overall, this PR expands the expressvity of `rustc_transmutability` to be much closer to the transmutability analysis performed by miri. Future PRs will work to close the remaining gaps (e.g., support for `Box`, raw pointers, `NonZero*`, coroutines, etc.).
r? `@compiler-errors`
Fix argument ABI for overaligned structs on ppc64le
When passing a 16 (or higher) aligned struct by value on ppc64le, it needs to be passed as an array of `i128` rather than an array of `i64`. This will force the use of an even starting doubleword.
For the case of a 16 byte struct with alignment 16 it is important that `[1 x i128]` is used instead of `i128` -- apparently, the latter will get treated similarly to `[2 x i64]`, not exhibiting the correct ABI. Add a `force_array` flag to `Uniform` to support this.
The relevant clang code can be found here:
fe2119a7b0/clang/lib/CodeGen/Targets/PPC.cpp (L878-L884)fe2119a7b0/clang/lib/CodeGen/Targets/PPC.cpp (L780-L784)
I think the corresponding psABI wording is this:
> Fixed size aggregates and unions passed by value are mapped to as
> many doublewords of the parameter save area as the value uses in
> memory. Aggregrates and unions are aligned according to their
> alignment requirements. This may result in doublewords being
> skipped for alignment.
In particular the last sentence. Though I didn't find any wording for Clang's behavior of clamping the alignment to 16.
Fixes https://github.com/rust-lang/rust/issues/122767.
r? `@cuviper`
Create the rustc_sanitizers crate and move the source code for the CFI
and KCFI sanitizers to it.
Co-authored-by: David Wood <agile.lion3441@fuligin.ink>
