tests: delete `cat-and-grep-sanity-check`
Part of #121876.
All remaining `Makefile`s have open PRs that do not rely on platform `cat` or `grep` or the `cat-and-grep` script.
Add Kobzol on vacation
I will be mostly reviewing diaper contents in the upcoming weeks, so I'm (proactively, for now) removing myself from the auto review rotation. Feel free to CC me (explicit `r?` won't work) explicitly though, I will be around.
r? `````@ghost`````
Ignore `mermaid.min.js`
It is very long and tends to match a lot of search queries. It's not useful to show with ripgrep.
Note that this does not actually untrack the file; changes can still be committed (although I suspect it may not have been intentional to commit originally?). This just changes how it interacts with tools that use `.gitignore` as a default filter.
r? rustc-dev-guide
Add an example of using `carrying_mul_add` to write wider multiplication
Just the basic quadratic version that you wouldn't actually use for really-big integers, but it's nice and short so is useful as for a demonstration of why you might find `carrying_mul_add` useful :)
cc #85532 ``````@clarfonthey``````
CI: build FreeBSD artifacts on FreeBSD 13.4
13.2 is EoL, and 13.3 will be EoL too in about 2 months. Plus, both suffer from a bug in LLVM's libunwind. It causes a segfault inside of std::backtrace::Backtrace::capture().
Fixes#132185
cc ``````@ehuss`````` . before you can do the trybuild, you'll also have to download new FreeBSD 13.4 base.txz images and place them in https://ci-mirrors.rust-lang.org/rustc , then update this PR with the correct file names.
try-job: dist-x86_64-freebsd
try-job: dist-various-2
remove support for the (unstable) #[start] attribute
As explained by `@Noratrieb:`
`#[start]` should be deleted. It's nothing but an accidentally leaked implementation detail that's a not very useful mix between "portable" entrypoint logic and bad abstraction.
I think the way the stable user-facing entrypoint should work (and works today on stable) is pretty simple:
- `std`-using cross-platform programs should use `fn main()`. the compiler, together with `std`, will then ensure that code ends up at `main` (by having a platform-specific entrypoint that gets directed through `lang_start` in `std` to `main` - but that's just an implementation detail)
- `no_std` platform-specific programs should use `#![no_main]` and define their own platform-specific entrypoint symbol with `#[no_mangle]`, like `main`, `_start`, `WinMain` or `my_embedded_platform_wants_to_start_here`. most of them only support a single platform anyways, and need cfg for the different platform's ways of passing arguments or other things *anyways*
`#[start]` is in a super weird position of being neither of those two. It tries to pretend that it's cross-platform, but its signature is a total lie. Those arguments are just stubbed out to zero on ~~Windows~~ wasm, for example. It also only handles the platform-specific entrypoints for a few platforms that are supported by `std`, like Windows or Unix-likes. `my_embedded_platform_wants_to_start_here` can't use it, and neither could a libc-less Linux program.
So we have an attribute that only works in some cases anyways, that has a signature that's a total lie (and a signature that, as I might want to add, has changed recently, and that I definitely would not be comfortable giving *any* stability guarantees on), and where there's a pretty easy way to get things working without it in the first place.
Note that this feature has **not** been RFCed in the first place.
*This comment was posted [in May](https://github.com/rust-lang/rust/issues/29633#issuecomment-2088596042) and so far nobody spoke up in that issue with a usecase that would require keeping the attribute.*
Closes https://github.com/rust-lang/rust/issues/29633
try-job: x86_64-gnu-nopt
try-job: x86_64-msvc-1
try-job: x86_64-msvc-2
try-job: test-various
Rework dyn trait lowering to stop being so intertwined with trait alias expansion
This PR reworks the trait object lowering code to stop handling trait aliases so funky, and removes the `TraitAliasExpander` in favor of a much simpler design. This refactoring is important for making the code that I'm writing in https://github.com/rust-lang/rust/pull/133397 understandable and easy to maintain, so the diagnostics regressions are IMO inevitable.
In the old trait object lowering code, we used to be a bit sloppy with the lists of traits in their unexpanded and expanded forms. This PR largely rewrites this logic to expand the trait aliases *once* and handle them more responsibly throughout afterwards.
Please review this with whitespace disabled.
r? lcnr
codegen: store ScalarPair via memset when one side is undef and the other side can be memset
Basically since `undef` can be any byte, it can also be the byte(s) that are in the non-undef parts of a value. So we can just treat the `undef` at not being there and only look at the initialized bytes and memset over them
fixes#104290
based on https://github.com/rust-lang/rust/pull/135258
Enable `unreachable_pub` lint in core
This PR enables the [`unreachable_pub`](https://doc.rust-lang.org/rustc/lints/listing/allowed-by-default.html#unreachable-pub) as warn in `core`, `rtstartup` and `panic_unwind`.
The motivation is similar to the compiler [MCP: Enable deny(unreachable_pub) on `rustc_*` crates](https://github.com/rust-lang/compiler-team/issues/773#issue-2467219005) :
> "Where is this thing used?" is a question I ask all the time when reading unfamiliar code. Because of this, I generally find it annoying when things are marked with a more permissive visibility than necessary. "This thing marked pub, which other crates is it used in? Oh, it's not used in any other crates."
Another motivation is to help to lint by utilizing it in-tree and seeing it's limitation in more complex scenarios.
The diff was mostly generated with `./x.py fix --stage 1 library/core/ -- --broken-code`, as well as manual edits for code in macros, generated code and other targets.
r? libs
It is very long and tends to match a lot of search queries. It's not useful to show with ripgrep.
Note that this does not actually untrack the file; changes can still be committed (although I suspect it may not have been intentional to commit originally?). This just changes how it interacts with tools that use `.gitignore` as a default filter.
Subtree sync for rustc_codegen_cranelift
Nothing too exciting this time, but this includes a fix for a linker hang on Windows: https://github.com/rust-lang/rustc_codegen_cranelift/pull/1554
r? ``@ghost``
``@rustbot`` label +A-codegen +A-cranelift +T-compiler
Partial progress on #132735: Replace extern "rust-intrinsic" with #[rustc_intrinsic] across the codebase
Part of #132735: Replace `extern "rust-intrinsic"` with `#[rustc_intrinsic]` macro
- Updated all instances of `extern "rust-intrinsic"` to use the `#[rustc_intrinsic]` macro.
- Skipped `.md` files and test files to avoid unnecessary changes.
Reexport likely/unlikely in std::hint
Since `likely`/`unlikely` should be working now, we could reexport them in `std::hint`. I'm not sure if this is already approved or if it requires approval
Tracking issue: #26179
Rollup of 5 pull requests
Successful merges:
- #135433 (Add Profile Override for Non-Git Sources)
- #135626 (doc: Point to methods on `Command` as alternatives to `set/remove_var`)
- #135658 (Do not include GCC source code in source tarballs)
- #135676 (rustc_resolve: use structured fields in traces)
- #135762 (Correct counting to four in cell module docs)
r? `@ghost`
`@rustbot` modify labels: rollup
When LLVM's location discriminator value limit is exceeded, emit locations with dummy spans instead of dropping them entirely
Dropping them fails `-Zverify-llvm-ir`.
Fixes#135332.
r? `@jieyouxu`
Correct counting to four in cell module docs
It could also be argued that `OnceCell<T>` and `LazyCell<T>` don't really provide safe interior mutability in different ways. But it's a vague enough claim that I'm not sure it's worth being pedantic about.
rustc_resolve: use structured fields in traces
I think this crate was written before `tracing` was adopted, and was manually writing fields into trace logs instead of using structured fields.
I kept function names in the trace messages even though I added `#[instrument]` invocations so that the events will be in named spans, wasn't sure if spans are always printed.
Do not include GCC source code in source tarballs
The licensing story is unclear, it makes the archive much larger, and we should not need it for building anything in the tarballs (yet).
```
Before:
121s building the archive
1.3 GiB gzipped size
5.7 GiB extracted size
402519 extracted files
After:
64s building the archive
961 MiB gzipped size
4.5 GiB extracted size
257719 extracfed files
```
Fixes: https://github.com/rust-lang/rust/issues/135606
r? `@ehuss`
