Mirrors/rust-analyzer
2
0
Fork 0
mirror of https://github.com/rust-lang/rust-analyzer synced 2025-01-12 13:18:47 +00:00
Code Issues Projects Releases Packages Wiki Activity

Rollup merge of #134528 - jieyouxu:fix-rustc-bootstrap-test, r=Kobzol

Browse source 
opt-dist: propagate channel info to bootstrap

Fixes #133503.

Previously, `tests/ui/bootstrap/rustc_bootstap.rs` [sic] failed during [beta bump](https://github.com/rust-lang/rust/pull/133447#issuecomment-2501298794) in opt-dist tests. This is because:

- `opt-dist` tried to run `./x test` against beta-channel dist `rustc` through `bootstrap`.
- The dist build produced during the beta bump produces a `rustc` which correctly thinks that it is a beta compiler based on `src/ci/channel` info.
- `opt-dist` tries to run `./x test` on the beta `rustc` from the dist build, but without specifying channel through a synthetic `config.toml`, so `bootstrap` tells `compiletest` that we're on the `nightly` channel (by default).
- Now there's a channel mismatch: `compiletest` believes the `rustc` under test is a *nightly* rustc, but the `rustc` under test actually considers itself a *beta* rustc. This means that `//@ only-nightly` will be satisfied yet the test will fail as the *beta* rustc is not a *nightly* rustc.

This PR:

- Fixes the test failure during beta bump (i.e. #133503) by having `opt-dist` faithfully report the channel of the dist `rustc` being tested (i.e. "beta" in a beta bump PR). This will properly make the test be ignored during beta bump as the `rustc` under test is not a *nightly* rustc.
- Fixes the test name `rustc_bootstap.rs` -> `rustc_bootstrap.rs`. No more stapping.
- Slightly adjusts the doc comment in the test to make it more clear.

I ran a try-job against the beta branch (explicitly running the opt-dist tests by modifying the job definition) with these changes in #134131, and it appears that the try-job was [successful](https://github.com/rust-lang/rust/pull/134131#issuecomment-2555492215). The two commits in this PR are cherry-picked from #134131, with the test commit slightly modified (to also adjust the test comments).

r? `@Kobzol` (or compiler or bootstrap or infra I guess?)
This commit is contained in:
Matthias Krüger 2024-12-23 14:44:21 +01:00 committed by GitHub
commit 50d2a384ba
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
Show stats Download patch file Download diff file Expand all files Collapse all files

Diff content is not available