msg318812 - (view) |
Author: STINNER Victor (vstinner) *  |
Date: 2018-06-06 11:12 |
On my PR https://github.com/python/cpython/pull/7447 "VSTS: Windows-PR" failed: "Windows-PR_20180606.13 failed". https://python.visualstudio.com/cpython/_build?buildId=6469 "The request was failed due to an internal service error. Please try again." |
|
|
msg318813 - (view) |
Author: STINNER Victor (vstinner) *  |
Date: 2018-06-06 11:15 |
By the way, it seems like tests were not run on the Linux-PR job: https://python.visualstudio.com/cpython/_build?buildId=6467&tab=summary&_a=summary Linux-PR / Build Linux-PR_20180606.13: "No test runs are available for this build." "Retained state Build not retained" What does it mean? Is it a bug? |
|
|
msg318818 - (view) |
Author: Steve Dower (steve.dower) *  |
Date: 2018-06-06 13:15 |
I'd already started discussing this with the VSTS team, it seems to be a bug on their side. The "build not retained" state just means it'll get cleaned up after seven days. People with login access can mark a build to be kept permanently, which changes that state. |
|
|
msg318823 - (view) |
Author: STINNER Victor (vstinner) *  |
Date: 2018-06-06 13:50 |
> The "build not retained" state just means it'll get cleaned up after seven days. Oh ok. But I'm confused, I don't see where are the logs. I clicked on Timeline, Code coverage, Tests: all are empty. In the logs ZIP, I cannot see any command related to Python (like building the C code or run tests). The build is failed but didn't run anything? |
|
|
msg318825 - (view) |
Author: Steve Dower (steve.dower) *  |
Date: 2018-06-06 14:15 |
Right. This build failed to start, which is why there's only one section on the left hand side (normally you'll see all the build steps there, and clicking on them will show their logs). I think it'll be an internal race condition. I'll pass back whatever info I can as I find out more. |
|
|
msg320049 - (view) |
Author: STINNER Victor (vstinner) *  |
Date: 2018-06-20 10:09 |
Steve: hi! Any update on the Microsoft side? |
|
|
msg320083 - (view) |
Author: Steve Dower (steve.dower) *  |
Date: 2018-06-20 15:37 |
It appears to be a known issue when a PR is retriggered (due to new commits) early enough in the CI run. I haven't got any info about when it'll be fixed. |
|
|
msg322979 - (view) |
Author: Terry J. Reedy (terry.reedy) *  |
Date: 2018-08-02 16:25 |
A quick follow-up commit can happen if one uses the web editor to edit more than one file. Editing a news file in Misc/News/next should not trigger any of the normal tests, only a blurb check if there is such. |
|
|
msg323018 - (view) |
Author: Steve Dower (steve.dower) *  |
Date: 2018-08-03 04:35 |
> A quick follow-up commit can happen if one uses the web editor to edit more than one file I hadn't thought of that one. Hopefully they'll just fix their bug though :) I spent a bit of time at the EuroPython sprints making the changes to quickly finish the build if nothing relevant has changed, but ironically it was so backed up with testing doc changes that I couldn't check it! Hopefully sometime next week I can finish that off. |
|
|