[Pkg-cups-devel] Bug#663712: cups: printing fails with ERROR: undefined (HP LaserJet 4600)

[Brian Potkin]

On Mon 27 Aug 2012 at 09:02:04 -0400, Sanjoy Mahajan wrote:

> > try printing again with the latest package versions and let us know
> > how you go on.
> 
> I retried the three methods of printing, each of which had problems before.
> Here's what happened with each one:
> 
> 1. lp -o fitplot file.pdf
> 
>    This method now works fine.  Progress!

Good. If you had stopped here we would all be happy bunnies and could go
home!

The process is:

   PDF ---> pdftopdf ---> pdftops ---> PRINTER

> 2. pdftops -paper match file.pdf - | lp -o fitplot
> 
>    This one is improved: Now all pages print, and the pink boxes are
>    pink (rather than green).  But the pages are still too small, taking
>    up roughly the upper left one-quarter of the page.  And the second
>    page comes out mangled; I've attached a b/w scan (the colors are
>    okay, but the placement of part of the text block is wrong).

Normally the input PostScript file would be converted to a PDF with
pstopdf from cups-filters and then be filtered as in 1. But your PPD
file hasn't a *CupsFilters line so, as a special concession, the input
goes through this chain:

   PostScript ---> pstops ---> PRINTER

I do not know enough about pstops to say whether it could be the cause
of your small or mangled pages. But, please see below.

> 3. pdftops -paper letter file.pdf - | lp -o raw
> 
>    This one got better in that the pink box on page 4 is now pink.  But
>    only the first 4 pages printed (out of 10).  But there was no error in
>    the cups error_log or any error reported by the printer.

As you say, the filtering system is by-passed. And the input file looks
ok before it gets to the printer.
 
> The pdftops in methods 2 and 3 is /usr/bin/pdftops from this package:
> 
>   poppler-utils  0.18.4-3
> 
> Maybe what remains of this bug should be reassigned to poppler-utils?
> Actually, that's not right.  The ps file produced in method 2, via
> "pdftops -paper match file.pdf -", looks fine in gs.  But when that ps
> file is sent to the printer, then the second page comes out mangled, and
> the pages are too small.  So that still seems like a cups-filter
> problem.
> 
> In method 3, I'm not sure what the failure point is, since it sends the
> postscript directly to the printer with lp -o raw.  The failure to print
> pages 5-10 then seems like a printer problem (but there's no error
> reported anywhere, so it's hard to find out more).

Once the file has gone to the printer cups has finished with it. There
is nothing to report.

The cups-filters README has a POSTSCRIPT PRINTING DEBUG MODE section.
Try printing into a file in /tmp with your three methods. I'll pre-empt
you by saying I have done this and a very quick look at the output files
with gv shows nothing out of the ordinary. These are the files which are
sent to the printer, so . . . .

Regards,

Brian.