Fix up-to-date checking for run-make tests
This special case in `output_base_dir` had the unfortunate side-effect of causing all run-make tests to share the same `stamp` file. So as soon as any one of them succeeded, all of the failed tests would be incorrectly considered up-to-date and would no longer run in subsequent test invocations.
Fixes#129971.
r? jieyouxu
Move polarity into `PolyTraitRef` rather than storing it on the side
Arguably we could move these modifiers into `TraitRef` instead of `PolyTraitRef`, but I see `TraitRef` as simply the *path* part of the trait ref. It doesn't really matter -- refactoring this further is much easier now.
Use environment variables instead of command line arguments for merged doctests
Fixes [#130796](https://github.com/rust-lang/rust/issues/130796).
Before merging this one, let's first ensure it has a smaller impact compared to command line arguments (results can be seen [here](https://github.com/rust-lang/rust/pull/130285)). I'll start a crater run once CI passed.
cc ``@ehuss``
r? ``@notriddle``
compiletest: Document various parts of compiletest's `lib.rs`
This adds some much-needed comments to many of the items in compiletest's `lib.rs`, giving a better overview of how a compiletest invocation actually proceeds.
To make review easier I have refrained from renames or functional changes, though I couldn't resist getting rid of one obviously-redundant `Arc::clone` in `make_test_closure`.
This special case in `output_base_dir` had the unfortunate side-effect of
causing all run-make tests to share the same `stamp` file. So as soon as any
one of them succeeded, all of the failed tests would be considered up-to-date
and would no longer run in subsequent test invocations.
miri: avoid cloning AllocExtra
We shouldn't be cloning Miri allocations, so make `AllocExtra::clone` panic instead, and adjust the one case where we *do* clone (the leak check) to avoid cloning.
This is in preparation for https://github.com/rust-lang/miri/pull/3966 where I am adding something to `AllocExtra` that cannot (easily) be cloned.
r? ``@saethlin``
Rollup of 8 pull requests
Successful merges:
- #130356 (don't warn about a missing change-id in CI)
- #130900 (Do not output () on empty description)
- #131066 (Add the Chinese translation entry to the RustByExample build process)
- #131067 (Fix std_detect links)
- #131644 (Clean up some Miri things in `sys/windows`)
- #131646 (sys/unix: add comments for some Miri fallbacks)
- #131653 (Remove const trait bound modifier hack)
- #131659 (enable `download_ci_llvm` test)
r? `@ghost`
`@rustbot` modify labels: rollup
Do not output () on empty description
When passing an explicitly empty description string, as explained here https://github.com/rust-lang/rust/blob/master/config.example.toml#L611-L613, my expectation is that the resulting rustc will be compatible with upstream.
However, it seems that instead, a `()` is added to the end of the version string, causing the version compatibility check to fail. My proposed fix here would be to instead only print `({description})` if `description` is a non-empty string.
Special case error message for a `build-fail` test that failed check build
A `build-fail` test requires that a check build (roughly `--emit=metadata`, no codegen) succeeds but fails later. Previously, if its check build failed, the user will see the error message
```
error: test compilation failed although it shouldn't!
```
which is confusing. Because the test is `build-fail`, we want the test compilation to fail! This error message doesn't account for the difference between a check build and a complete build, so let's special case the error message for a `build-fail` test whose check build failed to instead say
```
error: `build-fail` test is required to pass check build, but check build failed
```
Fixes#130894.
compiletest: Move debugger setup code out of `lib.rs`
These functions contain a few hundred lines of code for dealing with debuggers (for `debuginfo` tests), and don't really belong in the crate root.
Moving them out to their own module makes `lib.rs` easier to follow.
compiletest: Remove the one thing that was checking a directive's `original_line`
This special handling of `ignore-tidy*` was introduced during the migration to `//`@`` directives (#120881), and has become unnecessary after the subsequent removal of the legacy directive check (#131392).
force "HEAD" for non-CI and `git_upstream_merge_base` for CI environment
When rust-lang/rust is configured as remote, some of the git logic (for tracking changed files) that uses get_closest_merge_commit starts to produce annoying results as the upstream branch becomes outdated quickly (since it isn't updated with git pull). We can rely on HEAD for non-CI environments as we specifically treat bors commits as merge commits, which also exist on upstream. As for CI environments, we should use `git_upstream_merge_base` to correctly track modified files as bors commits may be in `HEAD` but not yet on the upstream remote.
This is also an alternative fix for https://github.com/rust-lang/rust/issues/129528 since https://github.com/rust-lang/rust/pull/131331 reverts the previous fix attempts.
Take a display name for `tool_check_step!`
The tool build step already takes a display name, make the tool check step also take a display name to better represent the tool name. I.e. instead of `src/tools/cargo-miri` becoming `cargomiri`, it now becomes `cargo-miri`.
Fixes#131592.
When rust-lang/rust is configured as remote, some of the git
logic (for tracking changed files) that uses get_closest_merge_commit
starts to produce annoying results as the upstream branch becomes outdated
quickly (since it isn't updated with git pull). We can rely on HEAD for
non-CI environments as we specifically treat bors commits as merge commits,
which also exist on upstream. As for CI environments, we should use
`git_upstream_merge_base` to correctly track modified files as bors commits
may be in `HEAD` but not yet on the upstream remote.
Signed-off-by: onur-ozkan <work@onurozkan.dev>
Flatten redundant test module `run_make_support::diff::tests::tests`
This module is already named `tests`, and is already gated by `#[cfg(test)]`, so there's no need for it to also contain `mod tests`.
r? jieyouxu
intrinsics fmuladdf{32,64}: expose llvm.fmuladd.* semantics
Add intrinsics `fmuladd{f32,f64}`. This computes `(a * b) + c`, to be fused if the code generator determines that (i) the target instruction set has support for a fused operation, and (ii) that the fused operation is more efficient than the equivalent, separate pair of `mul` and `add` instructions.
https://llvm.org/docs/LangRef.html#llvm-fmuladd-intrinsic
The codegen_cranelift uses the `fma` function from libc, which is a correct implementation, but without the desired performance semantic. I think this requires an update to cranelift to expose a suitable instruction in its IR.
I have not tested with codegen_gcc, but it should behave the same way (using `fma` from libc).
---
This topic has been discussed a few times on Zulip and was suggested, for example, by `@workingjubilee` in [Effect of fma disabled](https://rust-lang.zulipchat.com/#narrow/stream/122651-general/topic/Effect.20of.20fma.20disabled/near/274179331).
