On 12/04/2019 10:28, Michael Felt wrote:
> Michael Felt <aixtools@felt.demon.nl> added the comment:
>
> On 10/04/2019 18:49, STINNER Victor wrote:
>> STINNER Victor <vstinner@redhat.com> added the comment:
>>
>> "I am looking into this - but as it seems to have gone away again - is
>> there a simple way to get that code back, and/or see what the diff is,
>> before/badrun/after?"
>>
>> Maybe it's just a flacky test. It's hard to guess.
>>
>> You can get the Git revision of a build by looking at details:
>>
>> https://buildbot.python.org/all/#/builders/10/builds/2389
>> => Properties: "got_revision: 8702b67dad62a9084f6c1823dce10653743667c8"
>>
>> See also "Changes" tab.
>>
>> Sometimes when a bug disappears, I just close it as "outdated" after a few weeks.
>>
>> ----------
>>
>> _______________________________________
>> Python tracker <report@bugs.python.org>
>> <https://bugs.python.org/issue36579>
>> _______________________________________
>>
> I'll try a manual build on the gccfarm. Not exactly the same
> environment, but as close as I can get.
>
> ----------

The single test passes:

$ ./python -m test -v test_venv
== CPython 3.8.0a3+ (v3.8.0a3-109-g8702b67:8702b67, Apr 12 2019,
04:35:08) [GCC 7.2.0]
== AIX-2-00FA74164C00-powerpc-32bit big-endian
== cwd: /home/aixtools/python/test_venv/build/test_python_27132304
== CPU count: 64
== encodings: locale=ISO8859-1, FS=iso8859-1
Run tests sequentially
0:00:00 [1/1] test_venv
test_defaults (test.test_venv.BasicTest) ... ok
test_executable (test.test_venv.BasicTest) ... ok
test_executable_symlinks (test.test_venv.BasicTest) ... ok
test_isolation (test.test_venv.BasicTest) ... ok
test_multiprocessing (test.test_venv.BasicTest) ... ok
test_overwrite_existing (test.test_venv.BasicTest) ... ok
test_prefixes (test.test_venv.BasicTest) ... ok
test_prompt (test.test_venv.BasicTest) ... ok
test_symlinking (test.test_venv.BasicTest) ... ok
test_unicode_in_batch_file (test.test_venv.BasicTest) ... skipped 'only
relevant on Windows'
test_unoverwritable_fails (test.test_venv.BasicTest) ... ok
test_upgrade (test.test_venv.BasicTest) ... ok
test_devnull (test.test_venv.EnsurePipTest) ... ok
test_explicit_no_pip (test.test_venv.EnsurePipTest) ... ok
test_no_pip_by_default (test.test_venv.EnsurePipTest) ... ok
test_with_pip (test.test_venv.EnsurePipTest) ... ok

----------------------------------------------------------------------

Ran 16 tests in 172.244s

OK (skipped=1)
test_venv passed in 2 min 52 sec

== Tests result: SUCCESS ==

With "make test" - they also all pass (make buildbottest did not run as
expected, only one thread, make test runs with CPU: 64)

...

0:05:50 [420/420/1] test_venv passed (5 min 43 sec)

== Tests result: SUCCESS ==

401 tests OK.

1 test altered the execution environment:
    test_concurrent_futures

18 tests skipped:
    test_curses test_devpoll test_epoll test_gdb test_kqueue
    test_msilib test_ossaudiodev test_spwd test_startfile test_tix
    test_tk test_ttk_guionly test_unicode_file
    test_unicode_file_functions test_winconsoleio test_winreg
    test_winsound test_zipfile64

Total duration: 5 min 50 sec
Tests result: SUCCESS

So, I guess "flacky". Cannot reproduce in any case.

>
> _______________________________________
> Python tracker <report@bugs.python.org>
> <https://bugs.python.org/issue36579>
> _______________________________________
>