[cups] Only able to print from CUPS web interface ("filter failed" error otherwise)

[Mario Sanchez Prada]

Hi,

On 06/11/14 09:19, Helge Blischke wrote:
> [...]
> from this log file it seems that the filter „bannertopdf“ writes garbage to the output
> so that Ghostscript (called from the pdftops filter?) crashes. 
> Unfortunately the error messages that ghostscript writes (I know gs well enough
> to know that gs *does* write error messages) are not written to the error_log (a quirk
> of the filter, I think), so that further diagnosis is not possible.
> 
> I tried to reproduce the issue on an Ubuntu based Linux installation (Peppermint 3) but
> on that system both the cups web interface and the Gnome printer utility produced
> identical results – no error whatsoever.
> 
> Could you please try the following:
> –	stop the printer
> –	do a test print
> –	save the file
> 	/var/spool/cups/d0xxxx-001
> 	(xxxx the jog ID)
> _	release the printer (so that the job files or completes)
> for both origins (web interface and system-config-printer)
> and post the saved files.

Thanks again for your feedback. I just run the tests you proposed and
I'm attaching the two files as requested:

  - d00058-001: from gnome-control-center (FAIL)
  - d00059-001: from web interface (OK)

Still, the two files are identical, so I'm not sure how helpful those
will be. However, I've grepped again in the journal for 'bin/gs' and
found this line for each execution:

>From gnome-control-center (FAIL):

nov 06 16:26:35 localmachine cupsd[29088]: [Job 58] pstocanonij:
/usr/bin/gs -sstdout=%stderr -r600 -g4958x7016 -q -dNOPROMPT -dSAFER
-sDEVICE=ppmraw -sOutputFile=%stdout -| /usr/bin/cifmg4100 --imageres
600 --papersize a4 --media plain --paperload front2 --quality 3
--renderintent photo --gamma 1.8 --balance_c 0 --balance_m 0 --balance_y
0 --density 0 --contrast 0 --inkcartridgesettings bkcolor --bbox
9,14,586,834 --fit


>From web interface (OK)

nov 06 16:27:56 localmachine cupsd[29088]: [Job 59] pstocanonij:
/usr/bin/gs -sstdout=%stderr -r600 -g4958x7016 -q -dNOPROMPT -dSAFER
-sDEVICE=ppmraw -sOutputFile=%stdout -| /usr/bin/cifmg4100 --imageres
600 --papersize a4 --media plain --paperload front2 --quality 3
--renderintent photo --gamma 1.8 --balance_c 0 --balance_m 0 --balance_y
0 --density 0 --contrast 0 --inkcartridgesettings bkcolor --bbox
9,14,586,834 --fit  --copies 1


Besides the fact that there's an additional parameter (--copies 1) when
printing from the web interface, I can also see the Canon specific
binary cifmg4100 binary in there, so perhaps the problem is there somehow?

> As the filter configuration in both cases is the same, the issue must be due
> to the contents of the input file.

Yes, or perhaps even in how this mysterious binary treats the data
passed from gs. Not sure, though

What do you think?

Thanks a lot,
Mario
-------------- next part --------------
#PDF-BANNER
Template default-testpage.pdf
Show printer-name printer-info printer-location printer-make-and-model printer-driver-name printer-driver-version paper-size imageable-area job-id options time-at-creation time-at-processing

-------------- next part --------------
#PDF-BANNER
Template default-testpage.pdf
Show printer-name printer-info printer-location printer-make-and-model printer-driver-name printer-driver-version paper-size imageable-area job-id options time-at-creation time-at-processing