Printing (HP LJ) slow/fails

Bug #355801 reported by mogliii
28
This bug affects 4 people
Affects Status Importance Assigned to Milestone
ghostscript (Ubuntu)
Fix Released
Medium
Unassigned
gutenprint (Ubuntu)
Fix Released
Medium
Unassigned

Bug Description

Ubuntu 9.04 beta 32bit Live-CD (the same happens under fully patched ubuntu 8.10 64bit)
HP Laserjet 2200, connected via USB. The printer is functional (error does not appear under Windows)

I want to print a .jpg picture or .pdf.

When I print a document or page with picture data (.jpg, .pdf from eye of gnome, gimp, firefox ...) the print job is scheduled (symbol in notification area, ~3000 kB data) and the LED on the printer starts flashing.

Sometimes it keeps blinking for 10 Minutes, sometimes additional red LED appears and printer has to be restarted. But nothing is printed. The same .jpg File or .pdf on Windows, with same printer and computer takes less than 10 seconds to start.

Additional Notes:
1) With older Ubuntu versions (8.04, maybe 8.10 when released, don't remember) printing worked fine (as on windows)
2) No CPU or Memory load
3) This is !very! annoying. If you can't print out a scan or .pdf.

I am happy to provide debug information, but I am not sure where to look for it.

affects: ubuntu → cups (Ubuntu)
Revision history for this message
mogliii (mogliii) wrote :

Additional test

This time a different computer (IBM T40 Notebook) with fresh 9.04 beta installed. Tried to print a picture
(eg. http://www.allmoviewalls.com/pics/Harry_Potter_ATHBP_2009_02_1280x1024.jpg)
with EOG (Eye of Gnome) on HP LaserJet 4050 over Network.

After ~10 minutes the printing was aborted and a red light was blinking on the printer.

Is any person in charge reading this?

Revision history for this message
Carsten Schnober (c-schnober) wrote :

I can confirm this exact behavior on an Ubuntu 9.04 final 64 Bit installation and a Brother HL 5270 DN connected via network (AppSocket). Printing text files from OpenOffice etc. works flawlessly.

Changed in cups (Ubuntu):
status: New → Confirmed
Revision history for this message
knarf (launchpad-ubuntu-f) wrote :

Same here on Lucid as well as before on Karmic. Printing PDF to my networked Laserjet 2200 either results in a verrrrrrrry slow job or a blinking attention light and no output but mostly it just silently fails. Printing the same document using the foomatic/pxlmono driver works but gives mediocre quality compared to PS. Converting the PDF to PS and printing it via the printer's FTP server works. Printing test pages works. Printing self-test pages works.

There seems to be something amiss with the way postscript is handled in cups.

Revision history for this message
Till Kamppeter (till-kamppeter) wrote :

Got a private mail from the original poster of this bug, mogliii, that the printing-related updates of the last two weeks solved the problem. These include Ghostscript, especially with patches to not flood the CUPS error_log any more and replacing the standard floor() C function in the PDF interpreter by a more efficient inline macro. Also Gutenprint was updated to the newest upstream version, which can also be positive for performance. So I consider the bug fixed in Ghostscript and Gutenprint. Changes in CUPS and Foomatic are less probable to have delivered the fix.

affects: cups (Ubuntu) → ghostscript (Ubuntu)
Changed in ghostscript (Ubuntu):
importance: Undecided → Medium
status: Confirmed → Fix Released
Changed in gutenprint (Ubuntu):
status: New → Fix Released
importance: Undecided → Medium
Revision history for this message
Till Kamppeter (till-kamppeter) wrote :

mogliii, thank you very much for your mail. Nice to hear that the most recent updates fixed this problem which annoyed many users.

Revision history for this message
mogliii (mogliii) wrote :

Packages for 64bit not released yet?
How long does it normally take until everything gets updated?
http://packages.ubuntu.com/lucid/i386/libgutenprint2 still mentions 5.2.5-0ubuntu1, instead of 5.2.5-0ubuntu1.1

Revision history for this message
IcyJ (jeremy-michelle) wrote :

I am using Ubuntu 10.04.1 (64 bit). I cannot print PDFs or images to my networked HP Laserjet 5000 either. My computer will show pending in the printing queue and the printer display will display "processing". It will stay like this for hours or until I cancel the job. It is not the printer as five other PCs on the network (Windows based) print just fine, along with my Windows XP virtual machines. I can print plain text and OO.o documents without problems in Ubuntu. Maybe this has been addressed and I am having the same update issue as mogliii (beings I am using the 64 bit version).

Revision history for this message
mogliii (mogliii) wrote :

Please go to System -> Administration -> Synaptic Package Manager

Look for libgutenprint2 and check the exact version. My 64bit has 5.2.5-0ubuntu1, my 32bit has 5.2.5-0ubuntu1.1.

So this is the update you most probably need. But it has not been distributed yet. Thats why I asked before. But maybe between uploading (what Till Kamppeter did) and fully distributed it takes a month??? I am also waiting for the update to show up.

Revision history for this message
IcyJ (jeremy-michelle) wrote :

I have version "5.2.5-0ubuntu1.1" of libgutenprint2 installed.

Revision history for this message
Till Kamppeter (till-kamppeter) wrote :

IcyJ, if I do

apt-cache policy libgutenprint2

on a 64-bit Lucid, I get

5.2.5-0ubuntu1.1, so the update must have reached the mirrors, but the more important update is Ghostscript here. You should have version 8.71.dfsg.1-0ubuntu5.1 or newer (see also bug 539708).

Please also supply an error_log of a failing job, as described in "CUPS error_log" in https://wiki.ubuntu.com/DebuggingPrintingProblems.

Revision history for this message
IcyJ (jeremy-michelle) wrote :

Looks like I can print images and PDFs now. One of the latest updates must have addressed the issue.

Revision history for this message
Till Kamppeter (till-kamppeter) wrote :

IcyJ, thanks.

Revision history for this message
Till Kamppeter (till-kamppeter) wrote :

IcyJ, perhaps your last problem was bug 628030, which we have worked around by setting RIPCache=auto in CUPS, so the update of the CUPS package can have solved your last problem.

Revision history for this message
bruno (brunob) wrote :

Still have the problem here on 64-bit lucid (HP 2605dn) with the latest libgutenprint2 and ghostscript. If I want to print a jpeg photo (in colors), I have to produce a ps file (with eye of gnome) and print this file with evince. I see with "top" that gs is used with this workaround but pdftops is used when I try to print directly the jpeg from Eye of gnome.

N.B. I also have the problem on a karmic-32 laptop (when I try to print the photo with gimp for example).

Revision history for this message
bruno (brunob) wrote :

You can see here a fresh error_log...

Revision history for this message
bruno (brunob) wrote :

On Maverick, the problem seems to be fixed (with the HP 2605 PPD at least). However I would like a fix for lucid because it's a LTS and I have other problems with Maverick. Can this bug be reactivated or reopened?

Revision history for this message
Till Kamppeter (till-kamppeter) wrote :

bruno, the reason why evince prints PostScript without problems and not PDF, is that it passes the original PostScript files through to CUPS but regenerates PDF files, sending very ugly, inefficient PDF files to CUPS. This problem I have now reported upstream:

https://bugzilla.gnome.org/show_bug.cgi?id=636128

For the PDF prorinting in Lucid you are hitting bugs like bug 680628, bug 419143, and the interaction of these bugs with bug 668800.

Maverick contains many fixes which improve printing workflow performance here and there and they probably some up to get your printing working.

Revision history for this message
bruno (brunob) wrote :

Till: on other bugs you listed, problems seem to appear when printing a PDF file and I have this problem also. However, not only for PDF files, but also jpeg photos, like in this bug. I understand that something may be improved in the pdftops process (or pdftoraster as you mentionned in 668800?), but for me (I think) it is the pdf creation (in the printing process) that has to be fixed. is it libcairo or poppler... I really don't know. But it should affect many people???

Please tell me what I can send (and where) to help. Thanks.

Revision history for this message
Till Kamppeter (till-kamppeter) wrote :

bruno, if you print a JPEG photo out of a photo application, like shotwell, f-spot, eog, ... it will be converted to PDF and the PDF will get semnt to CUPS. For GNOME apps this process is done by Cairo. The problems appear especially with Cairo-generated PDF with embedded JPEG images, which photo applications most probably send.

Which photo application are you using?

Print output of GNOME applications is usually done by libcairo. Poppler does not create PDF, it reads PDF and converts it to other formats, or it manipulates PDF, like in the pdftopdf CUPS filter. The libcairo-generated PDF has several problems with efficiency, so that PDF interpreters (Ghostscript, Adobe Reader, Poppler) take ages to do their job or crash due to running out of memory.

Revision history for this message
bruno (brunob) wrote :

Till, I use EOG on my main system and Gimp on another one. Same problem on both systems so I suppose that Gimp also prints with Cairo...? How can I know if a photo application use Cairo or not (or you can point me some titles as a temporary workaround). I could test and confirm.

Also on post#10, you said that the ghostscript 5.1 version was the key. I tried 5.2 and 5.3 with no success (5.1 not available). Is it possible that a regression between 5.1 and 5.2 explains why I have the bug? If yes, how can I install the 5.1 version on my 64-bit lucid?

Finally, if I can confirm that my problem is purely with libcairo generated PDF, what's the best place to watch to see a fix (bug# or anything)?

Revision history for this message
Till Kamppeter (till-kamppeter) wrote :

To see whether a program uses a certain library, use the "ldd" command. Example for your case:

$ ldd /usr/bin/eog | grep -i cairo
 libcairo.so.2 => /usr/lib/libcairo.so.2 (0x00007ff3979df000)
 libpangocairo-1.0.so.0 => /usr/lib/libpangocairo-1.0.so.0 (0x00007ff3943aa000)
$

AFAIK all GNOME programs use Cairo. KDE programs do not use Cairo. So I recommend to use digikam to solve your problem. Note that KDE programs also work when you are using GNOME as your desktop. The needed libraries will get automatically installed when you install digikam via Software Center or apt-get.

In comment #10 I write about Gutenprint, not Ghostscript. There are no versions 5.1 and 5.3. The problems of the original poster are solved with Gutenprint of version 5.2.5-0ubuntu1.1 or newer and with Ghostscript of version 8.71.dfsg.1-0ubuntu5.1 or newer. Note also that Gutenprint is not recommended for your printer and you most probably do not use it. Your printer is the HP Color LaserJet 2605 and according to your previous postings you use HP's PPD to run the printer in PostScript mode.

Your discussion is mostprobably done in the wronmg bug. For the current problems of Cairo-generated PDF there is bug 680628. Please continue your discussion there and follow also the instructions there.

Revision history for this message
bruno (brunob) wrote :

What I was talking about is:
ghostscript 8.71.dfsg.1-0ubuntu5.1 -vs- ghostscript 8.71.dfsg.1-0ubuntu5.3 (or ghostscript 8.71.dfsg.1-0ubuntu5.2).

As you said, ghostscript 8.71.dfsg.1-0ubuntu5.1 solved the issue of the original poster and I can't install this (now old) version on my system. I'm probably wrong but I was asking if a regression between 8.71.dfsg.1-0ubuntu5.1 and 8.71.dfsg.1-0ubuntu5.2 is possible.

But the "libcairo theory" seems good. If I import the jpeg in Openoffice and print from here (no use of libcairo), I'm ok. I will follow and try to contribute to bug 680628 as you suggest. Thanks for all.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Bug attachments

Remote bug watches

Bug watches keep track of this bug in other bug trackers.