Split x86_64-msvc-ext into two jobs #130072
Conversation
|
rustbot has assigned @Mark-Simulacrum. Use |
rustbot
added
A-testsuite
|
@bors try |
Split x86_64-msvc-ext into two jobs This is an attempt to mitigate (but not resolve) the high failure rate of the x86_64-msvc-ext builder. The theory being that doing less makes it less likely to fail. But this may not work as having an extra job that may fail might be worse. try-job: x86_64-msvc-ext try-job: x86_64-msvc-ext2
|
☀️ Try build successful - checks-actions |
|
I'll try this a few times to see what the error rate is like @bors try |
Split x86_64-msvc-ext into two jobs This is an attempt to mitigate (but not resolve) the high failure rate of the x86_64-msvc-ext builder. The theory being that doing less makes it less likely to fail. But this may not work as having an extra job that may fail might be worse. try-job: x86_64-msvc-ext try-job: x86_64-msvc-ext2
This comment has been minimized.
This comment has been minimized.
|
Oops, I messed up the bash syntax. Fixed. |
|
The job Click to see the possible cause of the failure (guessed by this bot) |
|
That one didn't count because of the messed up bash but at least msvc-ext2 passed! Let's try again @bors try |
Split x86_64-msvc-ext into two jobs This is an attempt to mitigate (but not resolve) the high failure rate of the x86_64-msvc-ext builder. The theory being that doing less makes it less likely to fail. But this may not work as having an extra job that may fail might be worse. try-job: x86_64-msvc-ext try-job: x86_64-msvc-ext2
|
☀️ Try build successful - checks-actions |
|
@bors try |
Split x86_64-msvc-ext into two jobs This is an attempt to mitigate (but not resolve) the high failure rate of the x86_64-msvc-ext builder. The theory being that doing less makes it less likely to fail. But this may not work as having an extra job that may fail might be worse. try-job: x86_64-msvc-ext try-job: x86_64-msvc-ext2
|
☀️ Try build successful - checks-actions |
|
Succes again. I would have expected an access denied error by now so this does seem to be an improvement. I can keep running try jobs but my question to the reviewer is if adding a new job would be acceptable (hopefully as a temporary measure). I'd reiterate that failures are still likely but hopefully much rarer. @bors try |
Split x86_64-msvc-ext into two jobs This is an attempt to mitigate (but not resolve) the high failure rate of the x86_64-msvc-ext builder. The theory being that doing less makes it less likely to fail. But this may not work as having an extra job that may fail might be worse. try-job: x86_64-msvc-ext try-job: x86_64-msvc-ext2
|
☀️ Try build successful - checks-actions |
bors
added
S-waiting-on-review
|
Ok, let's try again. @bors r+ |
bors
added
S-waiting-on-bors
|
@bors p=1 |
|
...it's already p=1, right. |
|
☀️ Test successful - checks-actions |
|
Finished benchmarking commit (8f93a10): comparison URL. Overall result: no relevant changes - no action needed@rustbot label: -perf-regression Instruction countThis benchmark run did not return any relevant results for this metric. Max RSS (memory usage)Results (secondary -2.6%)This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.
CyclesThis benchmark run did not return any relevant results for this metric. Binary sizeThis benchmark run did not return any relevant results for this metric. Bootstrap: 756.515s -> 756.659s (0.02%) |
…try> Revert MSVC CI changes - rust-lang#130151 - rust-lang#130072 try-job: x86_64-msvc try-job: x86_64-msvc-ext
cuviper
added
the
beta-accepted
[beta] backports - Only add an automatic SONAME for Rust dylibs rust-lang#130960 - Reject leading unsafe in `cfg!(...)` and `--check-cfg` rust-lang#131057, resolving rust-lang#131055 - Disable jump threading `UnOp::Not` for non-bool rust-lang#131201 - Update LLVM submodule rust-lang#131448 - Split x86_64-msvc-ext into two jobs rust-lang#130072 - Use a small runner for msvc-ext2 job rust-lang#130151 r? ghost
| @@ -58,8 +58,9 @@ case $HOST_TARGET in | |||
| # Strangely, Linux targets do not work here. cargo always says | |||
| # "error: cannot produce cdylib for ... as the target ... does not support these crate types". | |||
| # Only run "pass" tests, which is quite a bit faster. | |||
| python3 "$X_PY" test --stage 2 src/tools/miri --target aarch64-apple-darwin --test-args pass | |||
| python3 "$X_PY" test --stage 2 src/tools/miri --target i686-pc-windows-gnu --test-args pass | |||
| #FIXME: Re-enable this once CI issues are fixed | |||
There was a problem hiding this comment.
Is there an issue that tracks this? Would be good to reference it here so that it's easy to check when the FIXME can be resolved.
There was a problem hiding this comment.
Also, would have been nice to ping @rust-lang/miri here; this change could mean a breakage on those targets only gets detected by Miri CI, and it would have been very confusing to figure out what happened when we are sure that Rust CI already covers these targets.
There was a problem hiding this comment.
Ah, I found the issue: #127883
Added it to the PR description, too.
There was a problem hiding this comment.
Unless I messed something up, we should still be testing everything that we were testing. It's just being tested in a separate job (x86_64-msvc-ext2).
I only intended this to be very temporary but it seems the issues is persisting longer than I'd hoped.
There was a problem hiding this comment.
Oh, now I see, this got moved to src/ci/github-actions/jobs.yml... which is not where our tools testing logic usually lives so I missed that.
…youxu checktools.sh: add link to issue for more context about disabled Miri tests Adds some context for the changes made in rust-lang#130072.
…isDenton checktools.sh: add link to issue for more context about disabled Miri tests Adds some context for the changes made in rust-lang#130072.
…isDenton checktools.sh: add link to issue for more context about disabled Miri tests Adds some context for the changes made in rust-lang#130072.
Rollup merge of rust-lang#131877 - RalfJung:checktools-comment, r=ChrisDenton checktools.sh: add link to issue for more context about disabled Miri tests Adds some context for the changes made in rust-lang#130072.
[experiment] Try reverting the MSVC workarounds Get a fresh set of runs with the following workarounds reverted: - rust-lang#130151 - rust-lang#130072 - rust-lang#133248 try-job: x86_64-msvc try-job: x86_64-msvc-ext
This is an attempt to mitigate (but not resolve) the high failure rate of the x86_64-msvc-ext builder. The theory being that doing less makes it less likely to fail. But this may not work as having an extra job that may fail might be worse.
Related to #127883
try-job: x86_64-msvc-ext
try-job: x86_64-msvc-ext2