You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
jobEnvironment dropping content after = in value #33
That's very strange. From looking at where we handle setting DictAttribute values, I don't see anything that jumps out as wrong.
Could you try changing {'ALEX': 'alex=2'} to {'ALEX': '"alex=2"'} (with extra inner double-quotes around the value)? I'm wondering if the DRMAA C implementation you're using is doing something stupid on its end.
It's worth noting that users may be able to hack around this by using nativeSpecification. In any event, this is an issue with the particular DRMAA implementation itself and not drmaa-python. As such, this probably should be closed. Possibly could add a note in the docs about this issue and the (proposed) workaround.
Activity
cancan101 commentedon Jul 22, 2015
For reference, this works fine:
$ qsub -v ALEX=alex=2 -b y -j y python env.py.dan-blanchard commentedon Jul 23, 2015
That's very strange. From looking at where we handle setting
DictAttributevalues, I don't see anything that jumps out as wrong.Could you try changing
{'ALEX': 'alex=2'}to{'ALEX': '"alex=2"'}(with extra inner double-quotes around the value)? I'm wondering if the DRMAA C implementation you're using is doing something stupid on its end.cancan101 commentedon Jul 23, 2015
I believe I was able to get this bug reproduced using your travis setup:
c8a9c5e
results: https://travis-ci.org/cancan101/drmaa-python/jobs/72308774
cancan101 commentedon Jul 23, 2015
Wrapping with
"still leads to the same truncation after the=:'"alexnicoulaj commentedon Jun 16, 2017
FYI, this bug is in the C implementation of DRMAA, see http://gridengine.org/pipermail/users/2016-April/009042.html
jakirkham commentedon Mar 12, 2018
It's worth noting that users may be able to hack around this by using
nativeSpecification. In any event, this is an issue with the particular DRMAA implementation itself and not drmaa-python. As such, this probably should be closed. Possibly could add a note in the docs about this issue and the (proposed) workaround.