[SciPy-Dev] Random cancellations of Mac OS tests in the github workflow tests.

[Warren Weckesser]

On 5/2/20, Warren Weckesser <warren.weckesser at gmail.com> wrote:
> On 5/2/20, Andrew Nelson <andyfaff at gmail.com> wrote:
>> On Sun, 3 May 2020, 08:43 Andrew Nelson, <andyfaff at gmail.com> wrote:
>>
>>> I've not observed that. It does fail fast though, if one of 3.6, 3.7, or
>>> 3.8 fails then the other two stop.
>>>
>>
>> Closing and reopening a PR might stop runs from completing.
>>
>
> I've seen it in PRs,


For example, as I write this,
https://github.com/scipy/scipy/pull/12004 shows the first macOS test
as having been cancelled.  The other two macOS tests have passed.  The
remaining tests have either passed or not completed yet.

Warren


 and earlier today it happened in the tests that
> ran on the most recent commit to master.  I manually restarted that
> one, and the tests then all passed.
>
> Warren
>
>>>
>>
>