[cups.bugs] [HIGH] STR #3566: Printing PDF files generates Unable to write print data: Broken pipe errors

[Brent Goodrick]

DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

This occurs on Debian Linux, but I suspect this is not unique to Debian but
to CUPS.  I doubt this is specific to the printer (but I doubt it is the
printer because writing the file to PostScript and sending that straight
away to the same printer prints just fine).

The printer being used is a Xerox 6180MFP-D using the "Xerox Phaser 6180
Foomatic/Postscript (recommended)" printer driver (PC filename is
POSTSCRI.PPD). I am not using the vendor-supplied (Xerox) PPD file as I
found it unworkable (but I currently believe that is not germane to this
problem).

Steps to reproduce:

First, generate the PDF file (versus trying to attach the 4MB file to this
bug report):

1. Browse to http://en.wikipedia.org/wiki/File:Acacia_greggii_thorns.jpg
2. Click on the "Full Resolution" link and save it to a file.
3. Open the file in eog (Eye of GNOME 2.28.2
http://projectgs/gnome.org/eog/)
4. File --> Print and then select Print To File, and select PDF as the
output file format, change the filename (e.g., acacia.pdf), and click
Print.

Next, attempt to print the file: From a command prompt, type commands
similar to:

  cd <the_appropriate_directory>
  lpr -P  Xerox-6180MFP-D acacia.pdf

Then wait. And wait some more, and then finally we see the failure in the
web interface (http://localhost:631):

"Unable to write print data: Broken pipe" 

Attached is the contents of /var/log/cups/error_log for that specific job
failure.

I saw this originally when attempting to print from "eog" directly,
without first printing to a file. I do not believe this is "eog"
specifically.

Now, the data corresponding to the original .jpg file prints just fine if
instead one prints to a PostScript file and not a PDF file, and then print
the resulting .ps file directly via the similar lpr command.

I did a web searche and the closest report I could find was:

http://forums.linux-foundation.org/read.php?32,10479,10479

(not my post).

Is this a case of a bad pipe between the filters themselves?

Also attaching the .ppd file that I believe is in action in this context.

I can provide additional diagnostics or debugging data you need me to
provide, please request it and I will provide it if I can (but please be
specific because I'm fairly new to CUPS infrastructure).

Thanks!
bg

(Sigh. Apparently the bug submit page limits you to ONE file at a time, so
I will have to submit it and then follow up with the additional file).

Link: http://www.cups.org/str.php?L3566
Version: 1.4.2
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Xerox-6180MFP-D.ppd
Type: application/octet-stream
Size: 9331 bytes
Desc: not available
URL: <https://lists.cups.org/pipermail/cups/attachments/20100426/fbd5d7bb/attachment.obj>