Bug 655403 - Second page of printout garbled on HP DeskJet f2280
Summary: Second page of printout garbled on HP DeskJet f2280
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Fedora
Classification: Fedora
Component: ghostscript
Version: 13
Hardware: Unspecified
OS: Unspecified
low
high
Target Milestone: ---
Assignee: Tim Waugh
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks: 639593
TreeView+ depends on / blocked
 
Reported: 2010-11-20 20:13 UTC by James
Modified: 2010-11-26 15:56 UTC (History)
3 users (show)

Fixed In Version: ghostscript-8.71-21.fc13
Clone Of:
Environment:
Last Closed: 2010-11-25 17:32:09 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
Typical second-page output (up to the point where I cancelled the job). (48.90 KB, image/jpeg)
2010-11-22 18:58 UTC, James
no flags Details

Description James 2010-11-20 20:13:11 UTC
Description of problem:
When printing out a multi-page black-and-white PDF document from evince, the first page prints fine, but the second page shows up as a series of horizontal bars, with what should be the text streaking down the page (eventually resulting in a solid block of black).

Version-Release number of selected component (if applicable):
hpijs-3.10.9-5.fc14.x86_64

How reproducible:
Always.

Steps to Reproduce:
Print a multi-page PDF file.

Comment 1 Tim Waugh 2010-11-22 10:53:48 UTC
If you print the PDF from the command line ("lp -d queuename file.pdf"), does that have the same effect?

Does it also happen when printing a different multi-page black-and-white PDF document (e.g. a simple one made using OpenOffice.org), or only that particular document?

Are you able to attach this document?

Have previous versions of Fedora been able to print correctly to this printer?

Comment 2 James 2010-11-22 18:57:07 UTC
(In reply to comment #1)
> If you print the PDF from the command line ("lp -d queuename file.pdf"), does
> that have the same effect?

Yes. Scan attached for illustration.

> Does it also happen when printing a different multi-page black-and-white PDF
> document (e.g. a simple one made using OpenOffice.org), or only that particular
> document?

Happens with every document I've tried.
 
> Are you able to attach this document?

No, but for an example, just pull one off the arXiv at http://arxiv.org/list/hep-th/new

> Have previous versions of Fedora been able to print correctly to this printer?

Fedora 13 worked fine.

Comment 3 James 2010-11-22 18:58:12 UTC
Created attachment 462108 [details]
Typical second-page output (up to the point where I cancelled the job).

Comment 4 James 2010-11-22 19:44:00 UTC
Additional note: printing out the document each page at a time works.

Comment 5 James 2010-11-22 20:40:04 UTC
I've just tried downgrading to hplip-3.10.6. Didn't work. So presumably there's trouble elsewhere in the chain...

Comment 6 Tim Waugh 2010-11-23 09:28:50 UTC
Next suspect would be ghostscript.

Comment 7 James 2010-11-23 15:38:21 UTC
(In reply to comment #6)
> Next suspect would be ghostscript.

Right. I'll try a downgraded ghostscript this evening. I tried hpijs with a LaserJet 4000 today (on the config broken with mk f2280), and it worked OK for that model of printer.

Comment 8 James 2010-11-23 19:40:00 UTC
Suspicions confirmed... I have tried a few different ghostscripts in Koji, all with hpijs-3.10.9-5.fc14.x86_64. The results are:

  ghostscript-8.71-16.fc14.1.x86_64  WORKS
  ghostscript-8.71-17.fc13.x86_64    WORKS
  ghostscript-8.71-17.fc14.x86_64    WORKS
  ghostscript-8.71-19.fc13.x86_64    BROKEN
  ghostscript-8.71-20.fc14.x86_64    BROKEN

So there seems to be a bad interaction between hpijs-3.10.9-5.fc14.x86_64 and recent builds of ghostcript which affects at least the DeskJet f2200 series (and presumably other "lightweight" printers), but not the LaserJet 4000 series.

Comment 9 Tim Waugh 2010-11-24 09:29:36 UTC
OK, that points to this change:

* Thu Oct 14 2010 Tim Waugh <twaugh> 8.71-19
- Apply some fixes from upstream to avoid gdevcups
  segfaults (bug #639593).

Thanks for narrowing it down.

Comment 10 Tim Waugh 2010-11-25 17:32:09 UTC
I've reverted that change, as I couldn't find a smaller subset from upstream that would avoid the problem while fixing bug #639593.

Closing as CURRENTRELEASE because the faulty package never made it out through updates.

Comment 11 Charlie Bennett 2010-11-26 15:11:27 UTC
So this and the segfaults are fixed in F14?

(I'd been on the segfault bug but see this problem also on PhotoSmart 1315)

Comment 12 Tim Waugh 2010-11-26 15:56:06 UTC
No, the segfault is not fixed. (The fix for that caused this garbled printing as a side-effect.)

Until the correct fix is found I've reverted it.


Note You need to log in before you can comment on or make changes to this bug.