Hi all,
as you likely have not noticed :), hedge's regression gave it a clean
bill of health today (yes, there were still some failures, but those
were in PyMetis and Pyrticle).
Therefore, today's staging branch just became the new master branch.
Andreas
Hi Mark,
please find attached a patch against silo-4.7.2-smalltest. It fixes an
issue in DBCreateReal, where the output of stat() (especially device and
inode number) is used to check for whether a file is already registered
even if that stat() failed.
This caused spurious errors of the sort
-------------------------------------------------------------------------------
DBCreate: File multiply opened and all not read-only.: fld-0000.silo
Traceback (most recent call last):
File "wave.py", line 158, in <module>
main()
File "wave.py", line 134, in main
visf = vis.make_file("fld-%04d" % step)
File
"/Library/Python/2.6/site-packages/hedge-0.91-py2.6-macosx-10.6-
i386.egg/hedge/visualization.py",
line 457, in make_file
self.pcontext.rank, self.pcontext.ranks)
File "/Library/Python/2.6/site-packages/pylo/__init__.py", line 203,
in __init__
self.master_file = SiloFile(head_pathname, *args, **kwargs)
File "/Library/Python/2.6/site-packages/pylo/__init__.py", line 94,
in __init__
fileinfo, filetype)
RuntimeError: DBCreate failed
-------------------------------------------------------------------------------
for a user of my code.
Andreas
On Mittwoch 10 März 2010, Christoph Statz wrote:
> Hello Andreas;
>
> I am using Silo 4.7.2 with HDF5 1.8.4.
> -> libsiloh5
I've reproduced your problem with these versions. Still figuring out
what's going on...
Andreas