Red Hat Bugzilla – Bug 162033
system-config-printer crashes with a traceback when driver==Postscript
Last modified: 2007-11-30 17:11:09 EST
I had some problems with my printer so I decided to reset all my printer
configfiles to default, start system-config-printer and add my printer.
My printer is a HP Laserjet 2100, this one contains in the foomatic xml file:
-Selecting this printer (/dev/lp0)
-finish adding it
The printer HP LaserJet 2100 (ID: HP-LaserJet_2100) is not supported by the
Couldn't load driver information.
Traceback (most recent call last):
File "/usr/share/printconf/util/queueTree.py", line 600, in edit_button_clicked
if self.editQueue.editQueueDialog (iter):
File "/usr/share/printconf/util/editQueue.py", line 427, in editQueueDialog
File "/usr/share/printconf/util/editQueue.py", line 511, in select_printer_iter
File "/usr/share/printconf/util/editQueue.py", line 841, in printer_model_selected
driver_index = self.recommended_driver
AttributeError: editQueue instance has no attribute 'recommended_driver'
Selecting 2100M works fine.
Also the entry for the 2100 is plain wrong since this one can't do Postscript,
it should use the plxmono or lj5gray driver. See the entry for the 2100M which
can do postscript but defaults to plxmono because the 2100M identifies itself as
2100 which makes it impossible to distuingish the two for hw autodetect.
I don't want to be a PITA, but my reason for reporting this seperatly and
against system-config-printer, is that I believe that an error in the foomatic
database should not lead to a traceback and a crash but to a nice userfriendly
This particular issue is that the database is not self-consistent. There's not
much friendly you can say to the user at that point. :-/
Hmm, perhaps foomatic ought to have a consistency check in its %install step.
Yes that would be an excellent idea and solution.
Now all thats left is coding it :)
Turns out there are several such inconsistencies: