On 04Jun2018 0932, Victor Stinner wrote:
2018-06-04 18:18 GMT+02:00 Brett Cannon <br...@python.org>:
I'm currently not in the mood to argue about VSTS' stability so I don't feel
comfortable flipping that on as a requirement quite yet.
I don't suggest to make it mandatory right now.
I will try to keep on eye on VSTS ;-)
FWIW I had a quick look through the history of Windows commit builds on
VSTS. There last three failures were:
* test_subprocess getting wedged and the build timed out (surprisingly
gracefully though, it sent a Ctrl+C to the app so we still got all the
output)
* Tcl/tk and/or OpenSSL build got stuck (this was on 3.6 - we don't
build these every time as of 3.7)
* test_asyncio issues
Since the last of those issue, there have been 50-ish successful builds.
(Note that I looked at commit builds, not PR builds, so I wouldn't have
to wade through every single PR to see if it was a genuine bug.)
You can review the history here:
Windows commits:
https://python.visualstudio.com/cpython/_build/index?definitionId=4&_a=history&path=%5C
Windows PRs:
https://python.visualstudio.com/cpython/_build/index?definitionId=9&_a=history&path=%5C
macOS commits:
https://python.visualstudio.com/cpython/_build/index?definitionId=5&_a=history&path=%5C
macOS PRs:
https://python.visualstudio.com/cpython/_build/index?definitionId=8&_a=history&path=%5C
Linux commits:
https://python.visualstudio.com/cpython/_build/index?definitionId=6&_a=history&path=%5C
Linux PRs:
https://python.visualstudio.com/cpython/_build/index?definitionId=7&_a=history&path=%5C
Linux with code coverage:
https://python.visualstudio.com/cpython/_build/index?definitionId=13&_a=history&path=%5C
(both commits and PRs right now, still undecided how to best make use of
this run, as we don't seem to report coverage stats anywhere?)
Cheers,
Steve
_______________________________________________
python-committers mailing list
python-committers@python.org
https://mail.python.org/mailman/listinfo/python-committers
Code of Conduct: https://www.python.org/psf/codeofconduct/