RE: [PATCH v3] build: support z/OS (OS/390).

From: rsbecker@nexbridge.com <rsbecker@nexbridge.com>
Date: Mar 1, 2024, 18:25 UTC
Message-ID: <04df01da6c05$caaa5dd0$5fff1970$@nexbridge.com>
In-Reply-To: <xmqqbk7x6cox.fsf@gitster.g>

On Friday, March 1, 2024 1:15 PM, Junio C Hamano wrote:

Show 7 quoted lines

>Haritha D <Harithamma.D@ibm.com> writes:
>
>> The win test(7) test case failed stating the reason as "The Operation
>> cancelled". I saw that it failed after 5 hours 59 minutes of running
>> the test cases(build). How do I handle this?
>
>Those "win test (n)", at least some of them, seem to have been somewhat

flaky [*1*].  If you are certain you did not break them with

>your change, you do not have to fix them yourself.
>
>I am a wrong person to ask how the test failure that may [*2*] block GGG

submission can be circumvented, as I am not involved in that

Show 14 quoted lines

>machinery at all.
>
>Thanks.
>
>
>[Footnotes]
>
>*1* Also I've seen osx-clang job time-out from time to time, without
>    failing any specific test.  Re-running failed jobs from the menu
>    often make them pass, which is why I said "somewhat flaky".
>
>*2* I do not even know if GGG refuses to submit a series with a test
>    failure, let alone if it allows to override such a safety if
>    exists.

Which tests have been hanging on S390? We have occasional hangs on NonStop that end up being attributed to our CI build system not supplying pipes properly to git. It would be interesting if the same tests are having issues on different platforms. --Randall