FreeBSD 12 Failures - Py_Initialize: can't initialize sys standard streams

[Gedare Bloom]

On Fri, Apr 17, 2020 at 10:17 AM Joel Sherrill <joel at rtems.org> wrote:
>
> After some Google searching and talking to Chris, I modified my nohup command that starts the script to add "</dev/null". It is still running now and reporting results.
>
> I don't know how to capture this info and the fact that removing gcc from the FreeBSD 12 test machine to improve the results.
>
Technically, it is "not adding gcc"? I think by default freebsd uses
clang now without shipping gcc?

> --joel
>
> On Wed, Apr 15, 2020 at 9:04 AM Joel Sherrill <joel at rtems.org> wrote:
>>
>> Hi
>>
>> I see this type of failure on rtems-tools and rtems-tester. I run nohup'ed and I suspect these happen after the connection is dropped. I see what I think is discussion of similar issues here: https://bugs.python.org/issue32849
>>
>> I don't see this on Linux although in that issue report, it was reproducible there.
>>
>>
>> ====================
>> + cd rtems-tools-ca7b290f94a19238ed068aad229854cf67dee9b5
>> + ./waf distclean configure '--prefix=/home/joel/rtems-cron-5/tools/5'
>> Fatal Python error: Py_Initialize: can't initialize sys standard streams
>> OSError: [Errno 9] Bad file descriptor
>>
>> Current thread 0x00000008009e2000 (most recent call first):
>> Abort trap (core dumped)
>> shell cmd failed: /bin/sh -ex  /usr/home/joel/rtems-cron-5/rtems-source-builder/
>> rtems/build/rtems-tools-ca7b290f94a19238ed068aad229854cf67dee9b5-1/do-build
>> error: building rtems-tools-ca7b290f94a19238ed068aad229854cf67dee9b5-1
>> ====================
>>
>> I also see this in my script that builds a bsp individually and then uses rtems-tester:
>>
>> =======================
>> Fatal Python error: Py_Initialize: can't initialize sys standard streams
>> OSError: [Errno 9] Bad file descriptor
>>
>> Current thread 0x00000008009e2000 (most recent call first):
>> time: command terminated abnormally
>> =======================
>>
>> I have everything nohup'ed and redirected to files which shouldn't be a big surprise.
>>
>> --joel
>
> _______________________________________________
> devel mailing list
> devel at rtems.org
> http://lists.rtems.org/mailman/listinfo/devel