[TEST] Test a hashbrown branch with backward shift deletion - #162039
[TEST] Test a hashbrown branch with backward shift deletion#162039Zoxc wants to merge 2 commits into
hashbrown branch with backward shift deletion#162039Conversation
4316467 to
ed7546e
Compare
|
@bors try @rust-timer queue |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
[TEST] Test a `hashbrown` branch with backward shift deletion
This comment has been minimized.
This comment has been minimized.
|
Local results:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
💔 Test for b64e664 failed: CI. Failed job:
|
This comment has been minimized.
This comment has been minimized.
`cargo vendor` refuses to vendor two packages that share a name and version but
come from different sources, and the `linear-probes` `hashbrown` branch is
version `0.17.1`, the same as the crates.io `hashbrown` that `indexmap`,
`gimli`, `object`, `rkyv` and `wasmparser` keep using. Both end up in the
compiler workspace's `Cargo.lock`, so any `cargo vendor` run over it fails with
found duplicate version of package `hashbrown v0.17.1` vendored from two sources
which broke `dist` in CI: `dist::Rustc` generates the HTML copyright files, and
that vendors the whole tree first.
Skip the two places `dist` vendors the compiler workspace - the copyright
generation and the vendoring of the plain source tarball - behind
`VENDORING_BROKEN_BY_HASHBROWN_BRANCH` so it is a single flag to flip back when
the branch is dropped or gets a version of its own. The resulting `rustc` image
has no `COPYRIGHT.html` and the source tarball is not offline-buildable, which
is fine for the perf and try builds this branch exists for. `x vendor` itself
still fails.
|
@bors try |
[TEST] Test a `hashbrown` branch with backward shift deletion
This comment has been minimized.
This comment has been minimized.
|
The job Click to see the possible cause of the failure (guessed by this bot) |
This comment has been minimized.
This comment has been minimized.
|
Finished benchmarking commit (1cc5cfc): comparison URL. Overall result: ❌✅ regressions and improvements - please read:Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf. Next, please: If you can, justify the regressions found in this try perf run in writing along with @bors rollup=never rustc-perf Instruction countOur most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.
Max RSS (memory usage)Results (primary -1.0%, secondary -0.5%)A less reliable metric. May be of interest, but not used to determine the overall result above.
CyclesResults (primary -1.8%, secondary 0.1%)A less reliable metric. May be of interest, but not used to determine the overall result above.
Binary sizeResults (primary 0.0%, secondary 0.0%)A less reliable metric. May be of interest, but not used to determine the overall result above.
Bootstrap: 476.296s -> 470.082s (-1.30%) |
|
☔ The latest upstream changes (presumably #160524) made this pull request unmergeable. Please resolve the merge conflicts by rebasing. |
This tests a
hashbrownbranch with linear probing and backward shift deletion on the compiler.This branch and the
hashbrownbranch is LLM written.