Issue1352
Created on 2009-05-21.05:44:13 by pjenvey, last changed 2009-05-23.20:02:14 by pjenvey.
msg4703 (view) |
Author: Philip Jenvey (pjenvey) |
Date: 2009-05-21.05:43:59 |
|
When running jython.bat via subprocess (e.g. test_subprocess running
sys.executable), the return code is always 0, even though the .bat does
an exit /b errorcode. Oddly enough that error code seems to register
when testing on the command line
According to https://issues.apache.org/jira/browse/GERONIMO-4525
doing cmd /c exit /b errorcode instead "enforces the process exit code".
I have no idea what this means but it helps test_subprocess
Oti could you review this change? Does it look evil, can we always rely
on cmd.exe?
|
msg4709 (view) |
Author: Philip Jenvey (pjenvey) |
Date: 2009-05-22.22:30:08 |
|
I suppose we could make it %COMSPEC% /c
|
msg4710 (view) |
Author: Oti Humbel (otmarhumbel) |
Date: 2009-05-23.06:16:30 |
|
In all 3 variants, i get exactly one failing test in test_subprocess.py:
FAIL: test_communicate_pipe_buf (__main__.ProcessTestCase)
I'm in favour of %COMSPEC% /c, the installer autotests pass with this.
|
msg4711 (view) |
Author: Philip Jenvey (pjenvey) |
Date: 2009-05-23.06:22:05 |
|
That one's a separate issue. I didn't totally diagnose it but I found
that the child process of that test doesn't get the full -c command line
arg passed to it.
subprocess seemed to be doing its job there so I have a suspicion that
the .bat is cutting it off somehow. Though I could be wrong
|
msg4712 (view) |
Author: Philip Jenvey (pjenvey) |
Date: 2009-05-23.20:02:14 |
|
.bat change applied in r6370
|
|
Date |
User |
Action |
Args |
2009-05-23 20:02:14 | pjenvey | set | status: open -> closed resolution: fixed messages:
+ msg4712 |
2009-05-23 06:22:06 | pjenvey | set | messages:
+ msg4711 |
2009-05-23 06:16:40 | otmarhumbel | set | messages:
+ msg4710 |
2009-05-22 22:30:08 | pjenvey | set | messages:
+ msg4709 |
2009-05-22 16:13:08 | zyasoft | set | nosy:
+ fwierzbicki, zyasoft |
2009-05-21 05:44:14 | pjenvey | create | |
|