Can't print from Flash application

[Chris Conway]

Binary package hint: flashplugin-nonfree

In hopes of a fix for Bug #530520, I upgraded to flashplugin-installer 10.1.82.76ubuntu0.10.04.2. Now Flash won't print at all. I've experienced this problem on kenken.com.

Steps to reproduce:
1) Go to http://kenken.com/playnow.html. Select a puzzle, then scroll down and click "Print a puzzle."
2) A print dialog appears, choose a printer and click "Print."

What happens:
A print job appears in the queue with "Document" as "(stdin)", "Size" as "0k", and "Status" as "Processing." The status never changes and the document never prints.

What should happen:
The document should print!

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: flashplugin-installer 10.1.82.76ubuntu0.10.04.2
ProcVersionSignature: Ubuntu 2.6.32-24.39-generic 2.6.32.15+drm33.5
Uname: Linux 2.6.32-24-generic x86_64
NonfreeKernelModules: wl
Architecture: amd64
Date: Sat Aug 28 12:39:31 2010
ProcEnviron:
 PATH=(custom, user)
 LANG=en_US.utf8
 SHELL=/bin/bash
SourcePackage: flashplugin-nonfree

[Chris Conway (cconway)]

I see the same behavior trying to print from Noteflight.com, only the size is non-zero.

[Chris Conway (cconway)]

If I look at http://localhost:631/jobs/, the job appears as "held" with status "Unable to write print data: Broken pipe," but this doesn't appear in the system-config-printer-applet unless I manually refresh. (Refreshing also updates the size to non-zero.)

[Till Kamppeter (till-kamppeter)]

Perhaps the job consumes too much memory/disk space. See also bug 598940.

[Chris Conway (cconway)]

Attaching a CUPS error log. I do have a rather large spool file (163 MB), so it's possible this duplicates #598940.

[Chris Conway (cconway)]

On second thought, I'm not so sure this is a duplicate. If I print a short core from Noteflight, I see the same behavior as above: a spool file of several hundred megabytes is generated, after a period of time a message about a broken pipe is generated, and the job transitions to held. If I print a larger score, I see the behavior described in Bug #598940: a multi-GB spool file is generated and a gs process starts to monopolize the CPU, showing no signs of terminating after several minutes; I have to kill the job before it transitions to held.

[Chris Conway (cconway)]

I'm able to print from Flash on a different printer (different model, different driver). I'm attaching a CUPS error_log for a successful print job (it's job 236 in the log).

The portion of the error_log in #6 that appears to go wrong is as follows:

D [02/Sep/2010:11:07:34 -0400] [Job 233] Starting renderer with command: gs -dFirstPage=1 -q -dBATCH -dPARANOIDSAFER -dNOPAUSE -sDEVICE=hl1250 -dEconoMode=0 -dDEVICEWIDTHPOINTS=612 -dDEVICEHEIGHTPOINTS=792 -r300x300 -dSourceTray=0 -dPaperType=0 -sOutputFile=- -f /var/spool/cups/tmp/foomatic-m3SNxh
D [02/Sep/2010:11:07:34 -0400] [Job 233] Starting process "kid3" (generation 1)
D [02/Sep/2010:11:07:34 -0400] [Job 233] Starting process "kid4" (generation 2)
D [02/Sep/2010:11:07:34 -0400] [Job 233] Starting process "renderer" (generation 2)
D [02/Sep/2010:11:07:34 -0400] [Job 233] JCL: %-12345X@PJL
D [02/Sep/2010:11:07:34 -0400] [Job 233] <job data>
D [02/Sep/2010:11:07:34 -0400] [Job 233]
D [02/Sep/2010:11:07:35 -0400] [Job 233] Read 8192 bytes of print data...
D [02/Sep/2010:11:07:35 -0400] [Job 233] Wrote 8192 bytes of print data...
D [02/Sep/2010:11:07:35 -0400] [Job 233] Read 8192 bytes of print data...
E [02/Sep/2010:11:07:35 -0400] [Job 233] Unable to write print data: Broken pipe
D [02/Sep/2010:11:07:35 -0400] [Job 233] Set job-printer-state-message to "Unable to write print data: Broken pipe", current level=ERROR

There is no equivalent set of entries in this error_log (i.e., no message says "starting renderer"). Maybe this is a PPD bug?