From: Noel O'B. <bao...@gm...> - 2006-11-21 13:26:17
|
Yes - I've caught the Jaguar one now too. All unit tests and regression tests now pass (the latter also needed editing to remove a reference to Jaguar). I can't believe I missed all of this. I will send a prerelease to the user as you suggest. But I need to take some more time to update the Changelog, as I think I merged another bug fix for G03 since 0.6. The new openlogfile() command uses the extension of the log file to determine whether it is a .zip, .bz, .gz or none of the above. I'm not sure why the code failed though...even if a filename is passed which doesn't have an extension it should still work. What does this give on your system? >>> import os.path >>> os.path.splitext("base.ext") ('base', '.ext') >>> os.path.splitext("base") ('base', '') >>> "base.ext".rfind(".") 4 >>> "base".rfind(".") -1 Was there something strange about the filename you used? It says "AttributeError: rfind" in your message below. This implies that a string was not passed to splitext (I've just been looking at the source code...it's only a four line function or so), since all strings should have the .rfind() method. How did you call this method? Noel On 21/11/06, Adam Tenderholt <a-t...@st...> wrote: > > Looks like I messed up making the 0.6 release...I forgot to remove two > > references to the molproparser. > > Did you get the references to jaguarparser as well? I ran into that > problem when I removed cclib and tried reinstalling 0.6. > > Also, I'm having problems with trunk. Looks like it has to do with > utils.py and your new extension framework for detecting zipped files, > although I'm not positive since ccget seems to be working just fine. > > File "/Library/Frameworks/Python.framework/Versions/2.4/lib/ > python2.4/site-packages/pymolyze/pymolyze.py", line 128, in open > parser=ccopen(filename,progress,logging.ERROR) > File "/Library/Frameworks/Python.framework/Versions/2.4/lib/ > python2.4/site-packages/cclib/parser/utils.py", line 50, in ccopen > inputfile = openlogfile(filename) > File "/Library/Frameworks/Python.framework/Versions/2.4/lib/ > python2.4/site-packages/cclib/parser/utils.py", line 29, in openlogfile > extension = os.path.splitext(filename)[1] > File "/Library/Frameworks/Python.framework/Versions/2.4/lib/ > python2.4/posixpath.py", line 92, in splitext > i = p.rfind('.') > AttributeError: rfind > > > Of course, it worked for me because molparser.py *is* present on my > > system -- I need to uninstall old versions of cclib in future before > > trusting that they're fixed. I'd like to get the fix out as soon as > > possible, but I'm running short on time today, and I know that I won't > > have much time tomorrow. As well as this, I want to make *sure* this > > time that everything is fine, so I'm not going to rush out a version. > > Sounds good to me, although you probably should push a pre-released > version to the user who reported this bug as soon as possible so that > he can continue working with cclib---I'm guessing this is a > relatively minor problem. > > Adam > > > |