Re: Bookworm CUPS Printing Revisited

[Charles Curley]

On Wed, 17 May 2023 13:10:21 -0600
Charles Curley  wrote:

> Thanks to Brian  on the thread "Re: CUPS on
> Bullseye and Bookworm" I now have CUPS working on dragon, my bookworm
> i386 architecture laptop.

Dragon being an i386 arch machine, I booted ideapc (AMD 64) to the
Bookworm RC1 XFCE live CD.
(https://cdimage.debian.org/cdimage/bookworm_di_rc1-live/amd64/iso-hybrid/)
I installed nothing. I ran several tests on that.

Using the automagical instant printer: nothing worked, not even using
"lp -d ".

Using the IP address queue, Netscape, Libreoffice and "lp -d" all
worked. At one point the system even told me the printer was out of
paper.

-- 
Does anybody read signatures any more?

https://charlescurley.com
https://charlescurley.com/blog/

Re: Bookworm CUPS Printing Revisited

[Charles Curley]

On Fri, 19 May 2023 16:07:56 +0100
Brian  wrote:

> On Thu 18 May 2023 at 14:26:13 -0600, Charles Curley wrote:
> 
>  [...]  
> 
> [...]
> 
>  [...]  
>  [...]  
> 
> The PDF queue is not part of the printing system.

|snarky comment omitted.|

> 
> > Firefox also allows printing from the system printer dialog. When I
> > selected that, it showed the same three queues. The two printer
> > queues had error messages.
> > 
> > For HP_Laserjet_MFP_m234sdw_C0FB67, "No destination host name
> > supplied by cups-browserd for printer "HP_Las..."  
> 
> cups-browserd uses mDNS to get the destination host name. OK, it can
> come off the rails at times, but ideapc is happy enough with it.
> 
> I'll again mention that neiher cups-browserd nor a manual queue is
> required for printing; I gave a test of this in another post.
> CUPS should form a temporary queue.

I ran "apt purge cups-browsed", then:

root@dragon:~# lpstat -l -e
HP_LaserJet_MFP_M234sdw_C0FB67 network none 
ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/
M234ip permanent ipp://localhost/printers/M234ip ipp://192.168.100.134/ipp/print
root@dragon:~# lpadmin -x M234ip
root@dragon:~# lpadmin -x HP_LaserJet_MFP_M234sdw_C0FB67
root@dragon:~# lpstat -l -e
HP_LaserJet_MFP_M234sdw_C0FB67 network none 
ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/
root@dragon:~# lp /usr/share/cups/data/form_english.pdf
lp: Error - No default destination.
root@dragon:~# lp -d HP_LaserJet_MFP_M234sdw_C0FB67 
/usr/share/cups/data/form_english.pdf
request id is HP_LaserJet_MFP_M234sdw_C0FB67-24 (1 file(s))
root@dragon:~# 

And nothing happened on the printer.

I then opened a PDF document in evince, opened the print dialog, and
selected that printer. I got a short-lived message, "rejecting jobs". I
turned the printer off. The printer went away in the print dialog. I
turned it back on. The printer re-appeared, along with the same
short-lived "rejecting jobs" message.

I went to Netscape. That printed.

I begin to think the problem is with GTK's printer dialog.

> 
> > For HP_Laserjet_MFP_m234sdw_C0FB67@hpbm234ethernet.local, I got an
> > error message, which disappeared before I could write it down. There
> > is now no error message. I tried printing to that, and this time the
> > printer came up, showed its "busy" indicator, and did nothing.  
> 
> This is often an indication that the dialog is unable to query the
> printer. It shouldn't happen on bookworm.
>   
> > > An alternative URI for M234 is
> > > 
> > >   ipp://192.168.100.134/ipp/print
> > > 
> > > and the previous lpadmin command could be run with this. It's OK
> > > as long as the IP does not change and it also cuts out the moving
> > > part of DNS-SD resolution.  
> > 
> > That worked:
> > 
> > root@dragon:~# lpadmin -p M234ip -v
> > "ipp://192.168.100.134/ipp/print" -E -m everywhere root@dragon:~#
> > lp -d M234ip /usr/share/cups/data/form_english.pdf request id is
> > M234ip-20 (1 file(s)) root@dragon:~# 
> > 
> > It also showed up in Firefox's system printer dialog. I printed to
> > it. Nothing. I then printed from Firefox's internal printer
> > dialog. That worked.  
> 
> The difference between M234 and M234ip is that the latter does not
> need to use resolving to get the IP of the ptinter. The queue should
> work in Abiword too.

I re-installed the M234ip queue. It did not work from abiword.

> 
> > > The problem with the GTK dialog is that it never handled its
> > > relationship with CUUPS correctly before bullseye (APIs and all
> > > that). However, it *always* got along with manually set up queues,
> > > which M234 is.
> > >   
> > 
> > That isn't what I see here.  
> 
> Maybe not, but it is still the case that it was GTK's handling of
> printer discovery that was the issue.
> 
> > John Doe seems to think MDNS doesn't like my network
> > setup. https://lists.debian.org/debian-user/2023/05/msg00767.html  
> 
> I queried hpm234ethernet.localdomain but you seemed happy with it, so
> I let it go.
> 

You asked where it came from, and I replied my DHCP and DNS servers. Is
there a problem with it?

-- 
Does anybody read signatures any more?

https://charlescurley.com
https://charlescurley.com/blog/

Re: Bookworm CUPS Printing Revisited

[Brian]

On Thu 18 May 2023 at 14:26:13 -0600, Charles Curley wrote:

> On Thu, 18 May 2023 15:09:00 +0100
> Brian  wrote:

[...]

> > I haven't any idea why M234 does not work from the dialog. It might be
> > an issue with the GTK version. Personally, I would prefer that Firefox
> > was uused to confirm or not.
> 
> I installed xfce and firefox, and ran firefox on the computer (rather
> than over ssh to another computer). Firefox showed three queues: PDF,
> and two printer queues. The PDF one worked. The other two failed
> silently. The printer never woke up for either.

The PDF queue is not part of the printing system.

> Firefox also allows printing from the system printer dialog. When I
> selected that, it showed the same three queues. The two printer queues
> had error messages.
> 
> For HP_Laserjet_MFP_m234sdw_C0FB67, "No destination host name supplied
> by cups-browserd for printer "HP_Las..."

cups-browserd uses mDNS to get the destination host name. OK, it can
come off the rails at times, but ideapc is happy enough with it.

I'll again mention that neiher cups-browserd nor a manual queue is
required for printing; I gave a test of this in another post.
CUPS should form a temporary queue.

> For HP_Laserjet_MFP_m234sdw_C0FB67@hpbm234ethernet.local, I got an
> error message, which disappeared before I could write it down. There
> is now no error message. I tried printing to that, and this time the
> printer came up, showed its "busy" indicator, and did nothing.

This is often an indication that the dialog is unable to query the
printer. It shouldn't happen on bookworm.
  
> > An alternative URI for M234 is
> > 
> >   ipp://192.168.100.134/ipp/print
> > 
> > and the previous lpadmin command could be run with this. It's OK as
> > long as the IP does not change and it also cuts out the moving part
> > of DNS-SD resolution.
> 
> That worked:
> 
> root@dragon:~# lpadmin -p M234ip -v "ipp://192.168.100.134/ipp/print"
> -E -m everywhere root@dragon:~# lp -d M234ip
> /usr/share/cups/data/form_english.pdf request id is M234ip-20 (1
> file(s)) root@dragon:~# 
> 
> It also showed up in Firefox's system printer dialog. I printed to
> it. Nothing. I then printed from Firefox's internal printer
> dialog. That worked.

The difference between M234 and M234ip is that the latter does not
need to use resolving to get the IP of the ptinter. The queue should
work in Abiword too.

> > The problem with the GTK dialog is that it never handled its
> > relationship with CUUPS correctly before bullseye (APIs and all
> > that). However, it *always* got along with manually set up queues,
> > which M234 is.
> > 
> 
> That isn't what I see here.

Maybe not, but it is still the case that it was GTK's handling of
printer discovery that was the issue.

> John Doe seems to think MDNS doesn't like my network
> setup. https://lists.debian.org/debian-user/2023/05/msg00767.html

I queried hpm234ethernet.localdomain but you seemed happy with it, so
I let it go.

-- 
Brian.

Re: Bookworm CUPS Printing Revisited

[Charles Curley]

On Thu, 18 May 2023 15:09:00 +0100
Brian  wrote:

> On Wed 17 May 2023 at 13:10:21 -0600, Charles Curley wrote:
> 
> > Thanks to Brian  on the thread "Re: CUPS on
> > Bullseye and Bookworm" I now have CUPS working on dragon, my
> > bookworm i386 architecture laptop.
> > 
> > Or I should say, CUPS printing is more or less working but something
> > else (GTK printing?) isn't. I can find the driverless printer on the
> > network, create a print queue for it, and print to it.
> > 
> > root@dragon:~# driverless
> > ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/
> > root@dragon:~# lpadmin -p M234 -v
> > "ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/"
> > -E -m everywhere root@dragon:~# lp -d M234
> > /usr/share/cups/data/form_english.pdf request id is M234-3 (1
> > file(s)) root@dragon:~# lp -d M234 /etc/nsswitch.conf request id is
> > M234-4 (1 file(s)) root@dragon:~# 
> > 
> > So far, so good.
> > 
> > I have two problems:
> > 
> > 1) I cannot print from abiword. abiword finds three printer queues:
> > print to file (which works), M234 (as previously set up), and
> > HP_Laserjet_MFP_M234dsw_COFB67. The two printer queues do not work.
> > The former see the printer light up, the "busy" indicator light up
> > briefly, then nothing. The latter evinces no action on the printer
> > at all.  
> 
> I haven't any idea why M234 does not work from the dialog. It might be
> an issue with the GTK version. Personally, I would prefer that Firefox
> was uused to confirm or not.

I installed xfce and firefox, and ran firefox on the computer (rather
than over ssh to another computer). Firefox showed three queues: PDF,
and two printer queues. The PDF one worked. The other two failed
silently. The printer never woke up for either.

Firefox also allows printing from the system printer dialog. When I
selected that, it showed the same three queues. The two printer queues
had error messages.

For HP_Laserjet_MFP_m234sdw_C0FB67, "No destination host name supplied
by cups-browserd for printer "HP_Las..."

For HP_Laserjet_MFP_m234sdw_C0FB67@hpbm234ethernet.local, I got an
error message, which disappeared before I could write it down. There
is now no error message. I tried printing to that, and this time the
printer came up, showed its "busy" indicator, and did nothing.



> 
> An alternative URI for M234 is
> 
>   ipp://192.168.100.134/ipp/print
> 
> and the previous lpadmin command could be run with this. It's OK as
> long as the IP does not change and it also cuts out the moving part
> of DNS-SD resolution.

That worked:

root@dragon:~# lpadmin -p M234ip -v "ipp://192.168.100.134/ipp/print"
-E -m everywhere root@dragon:~# lp -d M234ip
/usr/share/cups/data/form_english.pdf request id is M234ip-20 (1
file(s)) root@dragon:~# 

It also showed up in Firefox's system printer dialog. I printed to
it. Nothing. I then printed from Firefox's internal printer
dialog. That worked.

> 
> The problem with the GTK dialog is that it never handled its
> relationship with CUUPS correctly before bullseye (APIs and all
> that). However, it *always* got along with manually set up queues,
> which M234 is.
> 

That isn't what I see here.

John Doe seems to think MDNS doesn't like my network
setup. https://lists.debian.org/debian-user/2023/05/msg00767.html

-- 
Does anybody read signatures any more?

https://charlescurley.com
https://charlescurley.com/blog/

Re: Bookworm CUPS Printing Revisited

[Brian]

On Wed 17 May 2023 at 13:10:21 -0600, Charles Curley wrote:

> Thanks to Brian  on the thread "Re: CUPS on
> Bullseye and Bookworm" I now have CUPS working on dragon, my bookworm
> i386 architecture laptop.
> 
> Or I should say, CUPS printing is more or less working but something
> else (GTK printing?) isn't. I can find the driverless printer on the
> network, create a print queue for it, and print to it.
> 
> root@dragon:~# driverless
> ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/
> root@dragon:~# lpadmin -p M234 -v 
> "ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/" -E -m 
> everywhere
> root@dragon:~# lp -d M234 /usr/share/cups/data/form_english.pdf
> request id is M234-3 (1 file(s))
> root@dragon:~# lp -d M234 /etc/nsswitch.conf
> request id is M234-4 (1 file(s))
> root@dragon:~# 
> 
> So far, so good.
> 
> I have two problems:
> 
> 1) I cannot print from abiword. abiword finds three printer queues:
> print to file (which works), M234 (as previously set up), and
> HP_Laserjet_MFP_M234dsw_COFB67. The two printer queues do not work. The
> former see the printer light up, the "busy" indicator light up briefly,
> then nothing. The latter evinces no action on the printer at all.

I haven't any idea why M234 does not work from the dialog. It might be
an issue with the GTK version. Personally, I would prefer that Firefox
was uused to confirm or not.

An alternative URI for M234 is

  ipp://192.168.100.134/ipp/print

and the previous lpadmin command could be run with this. It's OK as long
as the IP does not change and it also cuts out the moving part of DNS-SD
resolution.

> (I selected abiword rather than LibreOffice because LO is a bit of a
> resource hog.)

LibreOffice has its own dialog with its own issues.
 
> 2) When I print from the command line, I get results printing to M234.
> But when I try printing to the other queue (directly to the printer as
> I understand it), I see:

Strictly speaking CUPS never prints directly to a printer but via a
queue, whether it be manually set up, auto set up or temporary.

> charles@dragon:~$ lp -d HP_Laserjet_MFP_M234dsw_C0FB67 ~/test.document.pdf 
> lp: Error - The printer or class does not exist.
> charles@dragon:~$

Pass on this for the moment.

> I copied the printer name from abiword's list of available printers. Is
> that the correct way to specify a driverless printer? I also tried
> copying and pasting the ipps:// and dnssd:// printers shown by lpinfo
> -v. Same error.
> 
> According to https://wiki.debian.org/CUPSDriverlessPrinting, "the GTK
> print dialog on buster and before (firefox and evince, for example),
> has its own way of dealing with a network printer. Unfortunately,
> applications that print through this dialog do not make use of CUPS'
> temporary queue formation. To have a queue for an IPP printer visible
> and usable, users should manually set up a queue with lpadmin, the CUPS
> web interface or system-config-printer or rely on cups-browsed to do it
> for them. The situation on bullseye and later has improved." Regression?

The problem with the GTK dialog is that it never handled its relationship
with CUUPS correctly before bullseye (APIs and all that). However, it
*always* got along with manually set up queues, which M234 is.

-- 
Brian.

Re: Bookworm CUPS Printing Revisited

[gene heskett]

On 5/18/23 00:28, David Wright wrote:

On Wed 17 May 2023 at 16:15:11 (-0400), gene heskett wrote:


Couple folks rather pointedly asked if I ever read changelogs.  But
before they can be read, they have to be found.  I just spent 2 hours
with mc, punching f3 on changelog.gz's, trolling thru /usr/share/docs
w/o finding an entry for kernels.  So where do I find this famous
changelog I'm supposed to read?  Or are we becoming windoze, and its a
secret?


Google:   linux kernel changelog

Hit 1:https://www.kernel.org/

Select:   5.0.180 chagelog

Edit: Last two characters in address bar s/80/79/

Enjoy.

Cheers,
David.

.
While this might work, it depends on a photographic memory to remember 
all that. At my age of 88, I can't remember what if anything I had for 
breakfast.


What we need is a web script that works something like tail, but shows 
everything newer the $date.  That would be very helpfull if the code is 
well explained and the reader has coding experience which I do. However 
its not with recent 64 and 128 bit stuff. OpenSCAD and rs-274-D gcode 
these days.


Cheers, Gene Heskett.
--
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 

Re: Bookworm CUPS Printing Revisited

[David Wright]

On Wed 17 May 2023 at 16:15:11 (-0400), gene heskett wrote:

> Couple folks rather pointedly asked if I ever read changelogs.  But
> before they can be read, they have to be found.  I just spent 2 hours
> with mc, punching f3 on changelog.gz's, trolling thru /usr/share/docs
> w/o finding an entry for kernels.  So where do I find this famous
> changelog I'm supposed to read?  Or are we becoming windoze, and its a
> secret?

Google:   linux kernel changelog

Hit 1:https://www.kernel.org/

Select:   5.0.180 chagelog

Edit: Last two characters in address bar s/80/79/

Enjoy.

Cheers,
David.

Re: Bookworm CUPS Printing Revisited

[gene heskett]

On 5/17/23 15:11, Charles Curley wrote:

Thanks to Brian  on the thread "Re: CUPS on
Bullseye and Bookworm" I now have CUPS working on dragon, my bookworm
i386 architecture laptop.

Or I should say, CUPS printing is more or less working but something
else (GTK printing?) isn't. I can find the driverless printer on the
network, create a print queue for it, and print to it.

root@dragon:~# driverless
ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/
root@dragon:~# lpadmin -p M234 -v 
"ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/" -E -m 
everywhere
root@dragon:~# lp -d M234 /usr/share/cups/data/form_english.pdf
request id is M234-3 (1 file(s))
root@dragon:~# lp -d M234 /etc/nsswitch.conf
request id is M234-4 (1 file(s))
root@dragon:~#

So far, so good.

I have two problems:

1) I cannot print from abiword. abiword finds three printer queues:
print to file (which works), M234 (as previously set up), and
HP_Laserjet_MFP_M234dsw_COFB67. The two printer queues do not work. The
former see the printer light up, the "busy" indicator light up briefly,
then nothing. The latter evinces no action on the printer at all.

(I selected abiword rather than LibreOffice because LO is a bit of a
resource hog.)

2) When I print from the command line, I get results printing to M234.
But when I try printing to the other queue (directly to the printer as
I understand it), I see:

charles@dragon:~$ lp -d HP_Laserjet_MFP_M234dsw_C0FB67 ~/test.document.pdf
lp: Error - The printer or class does not exist.
charles@dragon:~$

I copied the printer name from abiword's list of available printers. Is
that the correct way to specify a driverless printer? I also tried
copying and pasting the ipps:// and dnssd:// printers shown by lpinfo
-v. Same error.

According to https://wiki.debian.org/CUPSDriverlessPrinting, "the GTK
print dialog on buster and before (firefox and evince, for example),
has its own way of dealing with a network printer. Unfortunately,
applications that print through this dialog do not make use of CUPS'
temporary queue formation. To have a queue for an IPP printer visible
and usable, users should manually set up a queue with lpadmin, the CUPS
web interface or system-config-printer or rely on cups-browsed to do it
for them. The situation on bullseye and later has improved." Regression?

I started this original thread because I could not get a bullseye client 
to work with this bullseye server but buster clients worked fine..  For 
me, the fix was something that can only be described as serendipity.


Apt had installed a newer kernel about a week back, probably a security 
fix, but since my whole system is behind the best guard dog ever, dd-wrt 
in my router, and I had just rebooted from an overnight freeze up, I 
ignored its blabbing that I needed to reboot until I had another similar 
freeze up Monday morning and had to reboot with the front panel reset 
button. New kernel now running, bullseye server to bullseye client now 
Just Works.


Not supposed to be snarky, but...

Couple folks rather pointedly asked if I ever read changelogs.  But 
before they can be read, they have to be found.  I just spent 2 hours 
with mc, punching f3 on changelog.gz's, trolling thru /usr/share/docs 
w/o finding an entry for kernels.  So where do I find this famous 
changelog I'm supposed to read?  Or are we becoming windoze, and its a 
secret?


Since 1992, a truely complete changelog is quite likely several 
terabytes so a URL link to a tail output would suffice.



Cheers, Gene Heskett.
--
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 

Bookworm CUPS Printing Revisited

[Charles Curley]

Thanks to Brian  on the thread "Re: CUPS on
Bullseye and Bookworm" I now have CUPS working on dragon, my bookworm
i386 architecture laptop.

Or I should say, CUPS printing is more or less working but something
else (GTK printing?) isn't. I can find the driverless printer on the
network, create a print queue for it, and print to it.

root@dragon:~# driverless
ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/
root@dragon:~# lpadmin -p M234 -v 
"ipps://HP%20LaserJet%20MFP%20M234sdw%20(C0FB67)._ipps._tcp.local/" -E -m 
everywhere
root@dragon:~# lp -d M234 /usr/share/cups/data/form_english.pdf
request id is M234-3 (1 file(s))
root@dragon:~# lp -d M234 /etc/nsswitch.conf
request id is M234-4 (1 file(s))
root@dragon:~# 

So far, so good.

I have two problems:

1) I cannot print from abiword. abiword finds three printer queues:
print to file (which works), M234 (as previously set up), and
HP_Laserjet_MFP_M234dsw_COFB67. The two printer queues do not work. The
former see the printer light up, the "busy" indicator light up briefly,
then nothing. The latter evinces no action on the printer at all.

(I selected abiword rather than LibreOffice because LO is a bit of a
resource hog.)

2) When I print from the command line, I get results printing to M234.
But when I try printing to the other queue (directly to the printer as
I understand it), I see:

charles@dragon:~$ lp -d HP_Laserjet_MFP_M234dsw_C0FB67 ~/test.document.pdf 
lp: Error - The printer or class does not exist.
charles@dragon:~$

I copied the printer name from abiword's list of available printers. Is
that the correct way to specify a driverless printer? I also tried
copying and pasting the ipps:// and dnssd:// printers shown by lpinfo
-v. Same error.

According to https://wiki.debian.org/CUPSDriverlessPrinting, "the GTK
print dialog on buster and before (firefox and evince, for example),
has its own way of dealing with a network printer. Unfortunately,
applications that print through this dialog do not make use of CUPS'
temporary queue formation. To have a queue for an IPP printer visible
and usable, users should manually set up a queue with lpadmin, the CUPS
web interface or system-config-printer or rely on cups-browsed to do it
for them. The situation on bullseye and later has improved." Regression?

-- 
Does anybody read signatures any more?

https://charlescurley.com
https://charlescurley.com/blog/