Bug#916267: libgtk-3-0: Cannot print to IPP printers from the GTK dialog

[Brian Potkin]

On Thu 02 Apr 2020 at 16:42:08 +0100, Simon McVittie wrote:

> Version: 3.24.13-1
> 
> On Wed, 12 Dec 2018 at 11:20:04 +0000, Brian Potkin wrote:
> > An IPP printer is allocated the destination "print" in the dialog and
> > is tagged as "Rejecting Jobs". This is irrespective of whether its TXT
> > record pdl key has application/PDF in it or not.
> 
> This appears to have been fixed in GTK3 version 3.24.13 upstream, so it
> should now work in testing/unstable. Debian 10 'buster' is still affected.

With libgtk-3-0:i386 version 3.24.16-1, my printer is now identified as
ENVY4500 on Firefox, Evince and gedit. That's a change but the Status is
"Getting printer information... . There is no progress with being able
to print to the device.
 
> After removing cups-browsed and restarting cups, my printer still shows
> up in the Print dialog for gedit and Firefox.

I have stopped cups-browsed. This prevents any auto-setup of network
queues and printers as *local* queues. Local queues are handled through
CUPS on the client.

I have also stopped the CUPS service; network queues and printers are
dealt with by the GTK print dialog without CUPS and cups-filters. In
the case of queues the job (a PDF) is processed on the remote server.

The same PDF would be sent (or would be attempted to be sent) to an IPP
printer. My ENVY4500 will not accept a PDF as application/pdf is not a
PDL it is capable of processing. Your experience may be different, but
then we would have to know more about your printer. In the case of
successful printing, the output from

 avahi-browse -rt _ipp._tcp

would be of interest.

> > Furthermore, the dialog
> > seems incapable of displaying more than one entry for IPP printers on
> > the network.
> 
> Hopefully this was a side-effect of them all having incomplete information
> (all being named "print" is another symptom of that). If you have access
> to more than one IPP printer, please could you try with a testing/unstable
> system *without* cups-browsed, then retry *with* cups-browsed? I only have
> one IPP printer, so I can't tell what happens with more than one.
> 
> The ideal result is that you get one entry for each IPP printer on your
> network, with or without cups-browsed installed.
> 
> The next best thing is that you get one entry for each IPP printer
> (without cups-browsed) or two entries for each IPP printer (with
> cups-browsed). If you get duplicates with cups-browsed installed, then
> that would indicate that the code that is meant to eliminate duplicates
> isn't working.

I have only one IPP printer too. However, I set up a simulated IPP
printer on a buster machine with ippserver. The dialog displayed it
under the name I gave it.

Cheers,