Description of problem: Whoever wrote the foomatic XML parsing code in eggcups did so with very little thought. It just reads nodes and acts on them without any regard to their context. So for example, whichever 'driver' is listed last in the 'drivers' section (i.e. printer/drivers/driver) might take precendence over the actual recommended driver (i.e. printer/driver). But the most important mistake is that any old 'model' tag will be used as the for-display model name. Even this: <autodetect> <parallel> <commandset>MLC,PCL,PML</commandset> <description>Hewlett-Packard DeskJet 990C</description> <manufacturer>HEWLETT-PACKARD</manufacturer> <model>DESKJET 990C</model> </parallel> </autodetect> which should *only* ever be used for matching devices to database entries, and *never* be shown to the user. These device IDs are really rough-and-ready things that the manufacturers put in their devices. There is no consistency in them -- unlike the foomatic names which *are* consistent. Version-Release number of selected component (if applicable): desktop-printing-0.19-18 How reproducible: 100% Steps to Reproduce: 1.Provoke the eggcups driver prompt dialog
Created attachment 143415 [details] eggcups-xml-parsing.patch Here's a minimal fix (tested). Now it can actually create queues based on the user selection.