This allows these test modules to pass on builds/targets without
unwinding support, where `panic = "abort"` - the ignored tests are for
functionality that's not supported on those targets.
Revert "Use workspace lints for crates in `compiler/` #138084"
Revert <https://github.com/rust-lang/rust/pull/138084> to buy time to consider options that avoids breaking downstream usages of cargo on distributed `rustc-src` artifacts, where such cargo invocations fail due to inability to inherit `lints` from workspace root manifest's `workspace.lints` (this is only valid for the source rust-lang/rust workspace, but not really the distributed `rustc-src` artifacts). The problem is that the `rustc-src` component doesn't include the root `Cargo.toml` manifest.
This breakage was reported in https://github.com/rust-lang/rust/issues/138304.
This reverts commit 48caf81484, reversing changes made to c6662879b2.
cc `@RalfJung`
r? `@nnethercote` (sorry, I didn't consider this being a thing 💀)
Fix O(tests) stack usage in edition 2024 mergeable doctests
Fixes https://github.com/rust-lang/rust/issues/138248
The important change here is that we are not passing a potentially-large array by value. Between the fact that `TestFn` cannot be `Clone` and `test_main` takes a `Vec<TestDescAndFn>`, the only way to call `test::test_main` without O(tests) stack use is to call `Vec::push` many times.
The normal test harness does not have this problem because it calls `test_main_static` or `test_main_static_abort`, which take `&[TestDescAndFn]`. Changing `test::test_main` to take a slice is not a simple change, so I'm avoiding doing it here.
Fix post-merge workflow
I forgot that `actions/checkout` only checks out a single commit by default. I also forgot to set the environment variable required for the `gh` CLI commands.
I did a few more tests on my fork and hopefully now it should work properly. I also tested it with fake rollup PRs and the comment was sent only to the merged rollup, as it should be.
r? `@marcoieni`
Pass `InferCtxt` to `InlineAsmCtxt` to properly taint on error
Pass an `InferCtxt` into the `InlineAsmCtxt` so that we can use the right dcx to taint the infcx properly, so that we don't try to build invalid MIR for inline asm blocks.
Fixes#138130
Shares the first commit with #137502.
r? oli-obk
