Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Testing failure in Debian/Sparc #1009

Closed
olebole opened this issue Apr 23, 2013 · 7 comments
Closed

Testing failure in Debian/Sparc #1009

olebole opened this issue Apr 23, 2013 · 7 comments
Labels

Comments

@olebole
Copy link
Member

olebole commented Apr 23, 2013

While this looks like #963 I guess this is something and not connected to the Bigendian issue discussed there:

On Debian/Sparc, the testing step fails with a

[...]
astropy/io/votable/tests/exception_test.py:10: test_reraise PASSED
astropy/io/votable/tests/exception_test.py:25: test_parse_vowarning PASSED
make[1]: *** [test-python2.7] Error 246
astropy/io/votable/tests/table_test.py:31: test_table make[1]: Leaving directory `/build/buildd-python-astropy_0.2.1-1~exp5-sparc-ydjg4f/python-astropy-0.2.1'

https://buildd.debian.org/status/fetch.php?pkg=python-astropy&arch=sparc&ver=0.2.1-1%7Eexp5&stamp=1366710710
This "Error 246" was not introduced by the Bigendian fixes in #1003 but appeared also in the original 0.2.1 compilation.

@mdboom
Copy link
Contributor

mdboom commented Apr 25, 2013

Yeah -- I don't have much idea about this one. I haven't been having any luck getting a Sparc qemu VM up and running, so without access to the hardware, I'm not sure what we can do...

@embray
Copy link
Member

embray commented Apr 26, 2013

I asked IT a while back about getting access to either a Sparc VM or a physical machine for testing but that was a few weeks ago and they have yet to respond. So I'm going to harass them about that again....

@mdboom
Copy link
Contributor

mdboom commented Apr 26, 2013

It's not inconceivable that #1010 might fix this if sparc has the same nan floating point exceptions that MIPS does.

@olebole
Copy link
Member Author

olebole commented Apr 28, 2013

@mdboom I tested it with a new upload to Debian experimental, and it seems that this is different. The failure is still there. I am a bit curious why there is no explicite failure in the individual tests?
@iguananaut Unfortunately, the Sparc32 emulation (sun4m) on qemu is not capable to run new Linux distributions. There is a sparc64 emulation as well, which may work.

@olebole
Copy link
Member Author

olebole commented May 1, 2013

On a sparc64 port, the tests succeed:
http://buildd.debian-ports.org/status/fetch.php?pkg=python-astropy&arch=sparc64&ver=0.2.1-1&stamp=1367369364
One wild guess: can this be a pointer misalignment problem? Sparc requires, that pointer are well-alligned; otherwise the access ends up in a bus error.

@olebole
Copy link
Member Author

olebole commented Jun 24, 2013

With version 0.2.3, this problem seems to be silently fixed :-)
See https://buildd.debian.org/status/fetch.php?pkg=python-astropy&arch=sparc&ver=0.2.3-1&stamp=1371269136
for the log file.

@mdboom
Copy link
Contributor

mdboom commented Jun 24, 2013

Ok -- I'm going to close this then. Feel free to reopen if it silently breaks again ;)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

3 participants