-
Notifications
You must be signed in to change notification settings - Fork 12.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
ci: finish the migration to azure #62247
Conversation
Unless you want a review specifically by Alex, r=me |
@bors r=Mark-Simulacrum p=1000 |
📌 Commit effffe792da67a2a49f8819bb0fc1c0935044a34 has been approved by |
⌛ Testing commit effffe792da67a2a49f8819bb0fc1c0935044a34 with merge cf7c4892a8319d8d4f3fb3c8a42e612adc756a79... |
Fwiw I don't think we can delete Travis and appveyor yet since bors is still listening for those services to complete, so it may be best to just add one null job on both and we can later change bors after this lands |
effffe7
to
fffac31
Compare
Uh good point @alexcrichton, added a dummy job on each of them. @bors r=Mark-Simulacrum retry |
📌 Commit fffac31 has been approved by |
⌛ Testing commit fffac31 with merge 28bffc3f492cccedccc22608bf33fcb0b466e912... |
💔 Test failed - checks-azure |
@bors retry |
…mulacrum ci: finish the migration to azure This moves to azure the last builders as we discussed on Discord last night. r? @alexcrichton
💔 Test failed - checks-azure |
Now only Miri failed. @RalfJung is it spurious? Can we retry it or does it need an update? |
This doesn't look spurious. I have never seen this failure before:
And only on Windows? Things work fine on our own AppVeyor CI. What changed compared to the previous AppVeyor tools builder? |
Our configuration scripts are the same between AppVeyor and Azure, but the CI image is of course different. I don't know what are the exact changes though. |
Looks like somehow ether the encoding of the process arguments into a Windows command string, or the decoding (inside the interpreted program) into a list again, fails now where it worked before. No idea how that can happen; differences in the host platform shouldn't "leak" into the interpreted program. This is using the same architecture / target triple, right?
Alternatively, if this is an Azure vs AppVeyor thing, it could help to have Azure CI for Miri. We probably eventually want that anyway. But I won't have time to set that up any time soon, I'm afraid. |
Looking at this again... wait a second. "normalized" and "expected" stdout look the same, and still compiletest complains they are different? And I just now noticed that all tests with reference output are failing in the same way. Is this a |
Clippy had similar CRLF vs LF issue when adding Windows based Travis job (Appveyor was working just fine). Possibly Windows Azure jobs have misconfigured |
Thanks! Let's try a |
The Miri PR made it. :) So if you include that in here, maybe that helps with the line-ending problems? |
@bors r+ p=2000 |
📌 Commit 3159741fe0e5a752ea96f752c32a72ade3afbfbd has been approved by |
@pietroalbini I think you have conflicts with master now? There was a Miri update there since this branch started. |
Since switching CI to Azure Pipelines it seems that this test seems to fail more consistently, so let's disable that for now. It helps that we have less than a week before release - we disallow PRs that break the tools to land in this period, so this makes landing critical PRs smoother now. r? @alexcrichton
3159741
to
dc9d2b3
Compare
📌 Commit dc9d2b3 has been approved by |
…mulacrum ci: finish the migration to azure This moves to azure the last builders as we discussed on Discord last night. r? @alexcrichton
☀️ Test successful - checks-azure, checks-travis, status-appveyor |
This moves to azure the last builders as we discussed on Discord last night.
r? @alexcrichton