Hi Jonas,
> > Attached is a patch that fixes the problem in _mysql.c. It was made > against > > 1.2.3, but applies to 1.2.2 with some fuzz. > > the bugreport contains a different patch than you attached to your mail. > Yes, it's a similar report, but the submitter tries to fix the problem in python-land instead of the .c file. As he says himself ; + # Similar to _mysql_Exception. Not perfect. > > Patch has also been submitted to python-mysqldb at sourceforge. > > unfortunately i cannot find either of the patches (the one from upstream > mysql-python sourceforge bug tracking system, and the one you attached > to this debian bugreport) applied against upstream mysql-python sources. > I attached the patch to a report in sourceforge ; http://sourceforge.net/tracker/?func=detail&aid=3008217&group_id=22307&atid=374934 > neither in svn, nor in hg repository. please point me to the upstream > source where one of the patches has been applied. > We just realised this problem a few days ago, after 14 months of mysterious "why is the cursor.description None in python-land ? who silenced the exception ?". After a while, we finally managed to reproduce the problem with a simple script, and a after a bit of digging, we found the missing check in _mysql.c. However, this was just last week, so no upstream versions have applied the patch (yet..). Otoh, the project doesn't seem to be that regularly maintained these days, hence I decided to also submit patches to debian and ubuntu directly. cheers, eskil ---
