On Wed 04 Oct 2017 at 18:38:32 +0100, Brian Potkin wrote: > Package: cups-browsed > Version: 1.17.8-1 > Severity: normal > Tags: upstream > > > > I am working on an unstable machine: cups-browsed 1.17.8-1 and cups > 2.2.4-7. There is a single local queue, pcl. On the network there is > a jessie server (cups 1.7.5-11+deb8u1) advertising two queues and > an IPP printer (an Envy 4500) with AirPrint on.
Still on unstable but with cups 2.2.8-3 and cups-browsed 1.20.3+b1. > With cups-browsed started 'lpstat -a' shows > > dotmatrix_desktop accepting requests since Wed 04 Oct 2017 17:50 BST > ENVY4500 accepting requests since Wed 04 Oct 2017 17:50 BST > LaserJet_300_desktop accepting requests since Wed 04 Oct 2017 17:50 BST > pcl accepting requests since Wed 04 Oct 2017 17:50 BST More or less the same as before. > This is what I would expect. (desktop is the remote print server). > > Stopping cups (which stops cups-browsed) and looking at a print dialog > in Firefox or gedit I see > > dotmatrix > LaserJet-300 > print > > Not quite what I expected (I get this on jessie and stretch, too). print > is the Envy 4500. In addition, it hangs at "Getting printer information" > when highlighted and shows "Rejecting jobs" when not selected. Two bugs > in GTK? I've altered my view on this. Firefox will want to send a PDF. The Envy 4500 does not accept PDF. Not a GTK bug. > Now for the cups-browsed part. I start it and a dialog has > > dotmatrix > dotmatrix_desktop > ENVY4500 > LaserJet-300 > LaserJet_300_desktop > pcl > print > > Is this duplication of GTK discovered printers/print queues supposed to > happen? I would think not; it does not happen on jessie. My mistake was to compare a far different version of cups-browsed with the present version and to neglect how the gtk print dialog behaves in the presence of cups-browsed. https://bugzilla.gnome.org/show_bug.cgi?id=786794 ackowledges that duplicates should be checked for and gets the UUIDs of printers installed by CUPS (by extracting them from the device-uri) and compares them with the UUIDs of printers being shared via DNSSD. This doesn't work for me and duplicates are still shown. I also note the advice in https://bugreports.qt.io/browse/QTBUG-28188 But important is to take care of duplicate listings then (the same printer listed both by cups-browsed and the dialog itself), preferably by the dialog automatically discovering that entries are duplicate, but at least by making Bonjour discovery optional, by a run-time switchable option. I do not have the energy to pursue this with Gnome upstream so am closing this report as not a cups-browsed bug. Regards, Brian.