We use the Python shelve library which uses the Python anydbm library under the covers to select the "best" available dbm variant available on the system to store settings. The tag line of the module says:
anydbm is a generic interface to variants of the DBM database -- dbhash (requires bsddb), gdbm, or dbm. If none of these modules is installed, the slow-but-simple implementation in module dumbdbm will be used.
Since I don't know which dbm system is being used on your system, I'll need a bit of help to come up with a good fix. Could you do the following for me?
1) Run "python" in the ~/.lsr/profiles directory. 2) Type "import whichdb" 3) Type "print whichdb.whichdb('login.profile')" 4) Send me the resulting output.
Thanks, Pete
>Hi David, > >I think I see the bug now. The filenames of your profiles do not match >the filenames that are generated on any of our test machines. The >library that is creating these config files changes what simple >database backend it uses depending on your platform. > >What distribution name and version are you using? >
I'm using my own custom Linux based on LFS. Heavily modified of course. I can check it out of CVS and make sure it works before you release if you'd like. What package was it looking for that your test machines have anyway? I could very easily have not gotten around to installing something.
--- Linux Screen Reader Homepage: http://live.gnome.org/LSR Mailing list: lsr@xxxxxxxxxxxxx
To unsubscribe, mail lsr-request@xxxxxxxxxxxxx with 'unsubscribe' in the subject.
--- Linux Screen Reader Homepage: http://live.gnome.org/LSR Mailing list: lsr@xxxxxxxxxxxxx