loongarch: Fix ELF header flags
This patch changes the ELF header flags so that the ABI matches the floating-point features. It also updates the link to the new official documentation.
Test benchmarks with `-Z panic-abort-tests`
During test execution, when a `#[bench]` benchmark is encountered it's executed once to check whether it works. Unfortunately that was not compatible with `-Z panic-abort-tests`: the feature works by spawning a subprocess for each test, which prevents the use of dynamic tests as we cannot pass closures to child processes, and before this PR the conversion from benchmark to test was done by turning benchmarks into dynamic tests whose closures execute the benchmark once.
The approach this PR took was to add two new kinds of `TestFn`s: `StaticBenchAsTestFn` and `DynBenchAsTestFn` (⚠️ **this is a breaking change** ⚠️). With that change, a `StaticBenchFn` can be converted into a `StaticBenchAsTestFn` without creating dynamic tests, and making it possible to test `#[bench]` functions with `-Z panic-abort-tests`. The subprocess test runner also had to be updated to perform the conversion from benchmark to test when appropriate.
Along with the bug fix, in the first commit I refactored how tests are executed: rather than executing the test function in multiple places across `libtest`, there is now a private `TestFn::into_runnable()` method, which returns either a `RunnableTest` or `RunnableBench`, on which you can call the `run()` method. This simplified the rest of the changes in the PR.
This PR is best reviewed commit-by-commit.
Fixes https://github.com/rust-lang/rust/issues/73509
By reversing the logic I felt that the code became a clearer. Also,
added a comment to make it clear that we need to take the trailing
semicolon for the `let-else` statement into account.
This allows users to configure the maximum length of a single line
`let-else` statements. `let-else` statements that otherwise meet the
requirements to be formatted on a single line will have their divergent
`else` block formatted over multiple lines if they exceed this length.
**Note**: `single_line_let_else_max_widt` will be introduced as a stable
configuration option.
Mark wrapped intrinsics as inline(always)
This should mitigate having the inliner decide not to inline when the architecture is lacking an implementation of
TargetTransformInfo::areInlineCompatible aware of the target features (e.g. PowerPC as today).
See https://github.com/rust-lang/stdarch/pull/1443#issuecomment-1613788080
Rollup of 8 pull requests
Successful merges:
- #113072 (str docs: remove "Basic usage" text where not useful)
- #113153 (make HashMap::or_insert_with example more simple)
- #113185 (Set `channel = nightly` in dist profile)
- #113186 (document that the panic in collect_intra_doc_links is load-bearing)
- #113187 (No need to distinguish `LocalTy` from `Ty`)
- #113189 (compiletest: Only trim the end of process output)
- #113191 (Update browser-ui-test version and improve GUI test)
- #113206 (User may want to skip tidy check sometimes)
r? `@ghost`
`@rustbot` modify labels: rollup
compiletest: Only trim the end of process output
As of #94196, compiletest automatically trims process stderr/stdout output before printing it, to make failure info more compact.
This causes the first line of `run-coverage` output to be displayed incorrectly, because it uses leading whitespace to align line numbers.
Trimming only the end of the output string should still have the intended effect (e.g. removing trailing newlines), without causing problems for output that deliberately uses leading whitespace on the first line.
## Before
```
--- stdout -------------------------------
1| 1|fn main() { //
2| 1| let num = 9;
3| 1| while num >= 10 {
4| 0| }
5| 1|}
------------------------------------------
stderr: none
```
## After
```
--- stdout -------------------------------
1| 1|fn main() { //
2| 1| let num = 9;
3| 1| while num >= 10 {
4| 0| }
5| 1|}
------------------------------------------
stderr: none
```
No need to distinguish `LocalTy` from `Ty`
I think the distinction between `decl_ty` and `revealed_ty` was from when you were allowed to put `impl Trait` in let bindings... I don't think we need that anymore, and it makes typeck that much more confusing 😆
Side-note: I don't know why we store this in a separate field [`locals`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_hir_typeck/struct.Inherited.html#structfield.locals) in `Inherited`, rather than just the `TypeckResults`... Might look into changing that next.
Set `channel = nightly` in dist profile
This avoids some channel-specific defaults leaking into local installs. It also makes it easier to set options for compiler/library/codegen profiles in the future, since they can be gated off `channel` instead of being duplicated between all three files.
Here are the exact things `channel` controls today:
68d458bb40/src/bootstrap/llvm.rs (L466-L470)85c4ea0138/src/bootstrap/config.rs (L1374-L1375)85c4ea0138/src/bootstrap/config.rs (L1464-L1465)
``@cuviper`` i expect you don't want any of those to be set in distro builds, right?
linker flavors
- only the stable values for `-Clink-self-contained` can be used on stable until we
have more feedback on the interface
- `-Zunstable-options` is required to use unstable linker flavors
Fix loading target specs in compiletest not working with custom targets
In https://github.com/rust-lang/rust/pull/112454#issuecomment-1611351168 it was pointed out that the PR broke blessing mir-opt tests. Since #112418, blessing mir-opt tests generates "synthetic targets", which are custom target specs. Those specs are not included in `--print=all-target-specs-json`, and #112454 required that the current target was returned by that flag.
This PR fixes the breakage by loading the target spec for the current target explicitly, if a custom target is detected.
r? `@oli-obk`
