(In reply to Davide Repetto from comment #0)
> I did not test the ghostscript-9.24-1.fc28.x86_64 in updates-testing, but
> that 9.24-1 release may very well be affected too.

The ghostscript-9.24-1 has some zero-day patches regarding the CUPS. Please, test it first, before we proceed here. I'm limited by time of what can I do/fix right now. Thanks!

Hi,

same here with ghostscript-9.24-1.fc28.x86_64.


Debug output:
Sep 09 17:07:11 host cupsd[24141]: [Job 281] Running command line for gs: gs -q -dNOPAUSE -dBATCH -dSAFER -dNOMEDIAATTRS -sDEVICE=ps2write -dShowAcroForm -sOUTPUTFILE=%stdout -dLanguageLevel=2 -r1200 -dCompressFonts=false -dNoT3CCITT -dNOINTERPOLATE -c \'save pop\' -f /var/spool/cups/tmp/cupsKHYBj3
Sep 09 17:07:11 host cupsd[24141]: [Job 281] Started filter gs (PID 29770)
Sep 09 17:07:11 host cupsd[24141]: [Job 281] Started filter pstops (PID 29771)
Sep 09 17:07:11 host cupsd[24141]: [Job 281] ./base/gsicc_manage.c:1244: gsicc_open_search(): Could not find ps_gray.icc
Sep 09 17:07:11 host cupsd[24141]: [Job 281] + ./base/gsicc_manage.c:255: gsicc_initialize_iccsmask(): failed to load gray smask profile
Sep 09 17:07:11 host cupsd[24141]: [Job 281] Page = 595x842; 0,0 to 595,842
Sep 09 17:07:11 host cupsd[24141]: [Job 281] slow_collate=0, slow_duplex=0, slow_order=0
Sep 09 17:07:11 host cupsd[24141]: [Job 281] This document does not conform to the Adobe Document Structuring Conventions and may not print correctly.
Sep 09 17:07:11 host cupsd[24141]: [Job 281] PAGE: 1 1
Sep 09 17:07:11 host cupsd[24141]: [Job 281] ./base/gsicc_manage.c:1244: gsicc_open_search(): Could not find default_rgb.icc
Sep 09 17:07:11 host cupsd[24141]: [Job 281] | ./base/gsicc_manage.c:2025: gsicc_set_device_profile(): cannot find device profile
Sep 09 17:07:11 host cupsd[24141]: [Job 281] GPL Ghostscript 9.24: Unrecoverable error, exit code 1
Sep 09 17:07:11 host cupsd[24141]: [Job 281] ./base/gsicc_manage.c:1244: gsicc_open_search(): Could not find default_gray.icc
Sep 09 17:07:11 host cupsd[24141]: [Job 281] | ./base/gsicc_manage.c:2025: gsicc_set_device_profile(): cannot find device profile
Sep 09 17:07:11 host cupsd[24141]: [Job 281] PID 29770 (gs) stopped with status 1!
Sep 09 17:07:11 host cupsd[24141]: [Job 281] PID 29771 (pstops) exited with no errors.
Sep 09 17:07:11 host cupsd[24141]: [Job 281] PID 29768 (/usr/lib/cups/filter/pdftops) stopped with status 1.
Sep 09 17:07:11 host cupsd[24141]: [Job 281] Hint: Try setting the LogLevel to "debug" to find out more.
Sep 09 17:07:11 host cupsd[24141]: [Job 281] GPL Ghostscript 9.24: Unrecoverable error, exit code 1
Sep 09 17:07:11 host cupsd[24141]: [Job 281] PID 29769 (/usr/lib/cups/backend/cups-pdf) exited with no errors.
Sep 09 17:07:11 host cupsd[24141]: [Job 281] End of messages
Sep 09 17:07:11 host cupsd[24141]: [Job 281] printer-state=3(idle)
Sep 09 17:07:11 host cupsd[24141]: [Job 281] printer-state-message="Filter failed"
Sep 09 17:07:11 host cupsd[24141]: [Job 281] printer-state-reasons=none


Relevant lines might be those regarding "gsicc_open_search()".
Found a similar report here: https://bugs.archlinux.org/task/59952

(In reply to David Kaspar [Dee'Kej] from comment #1)
> (In reply to Davide Repetto from comment #0)
> > I did not test the ghostscript-9.24-1.fc28.x86_64 in updates-testing, but
> > that 9.24-1 release may very well be affected too.
> 
> The ghostscript-9.24-1 has some zero-day patches regarding the CUPS. Please,
> test it first, before we proceed here.

Tested just now. :(
Unfortunately I have to confirm that ghostscript-9.24-1.fc28.x86_64 causes a “Filter failed” error in cups too.

> I'm limited by time of what can I do/fix right now. Thanks!

There is activity upstream on this one, so you may just want to wait for it, and pull when it's ready.

This should be the upstream fix:
http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=bc3df0773fcc

I'll backport it to ghostscript-9.24-1.

(In reply to David Kaspar [Dee'Kej] from comment #4)
> This should be the upstream fix:
> http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=bc3df0773fcc
> 
> I'll backport it to ghostscript-9.24-1.

Scratch that, this commit is already in ghostscript-9.24-1.

(In reply to Davide Repetto from comment #3)
> There is activity upstream on this one, so you may just want to wait for it,
> and pull when it's ready.

Could you please post the link to the upstream bug regarding this issue here? Thanks!

I think I found the bug you were mentioning:
https://bugs.ghostscript.com/show_bug.cgi?id=699721

Upstream is not sure if the 2 commits there actually fix this. So, I'll prepare a new build, and you'll have to try it (I don't have the printer to reproduce this).

(In reply to David Kaspar [Dee'Kej] from comment #6)
> I think I found the bug you were mentioning:
> https://bugs.ghostscript.com/show_bug.cgi?id=699721
> 
> Upstream is not sure if the 2 commits there actually fix this. So, I'll
> prepare a new build, and you'll have to try it (I don't have the printer to
> reproduce this).

I have observed this by printing to a PDF, i.e. using cups-pdf, so you wouldn't necessarily need a specific printer hardware to test this...

I've created a scratch-build for testing purposes here:
https://dkaspar.fedorapeople.org/share/scratch-build/fedora/

Please, download the the relevant packages (which you have currently installed), update them and test if the problem still persists for you.

If this scratch-build works for you, then I will submit the changes into Fedora. However, first I want to be sure the updated version actually fixes the issues.

(In reply to Thomas Müller from comment #7)
> I have observed this by printing to a PDF, i.e. using cups-pdf, so you
> wouldn't necessarily need a specific printer hardware to test this...

Can you provide an exact reproducer, please? It would be awesome if we could share this with upstream, so they could add this to their regression testsuite.

Created attachment 1482265 [details]
simple eps logo

It's even possible to reproduce this using epspdf on the command line:

$ epspdf Ghostscript.eps
  ./base/gsicc_manage.c:1244: gsicc_open_search(): Could not find default_rgb.icc 
| ./base/gsicc_manage.c:2025: gsicc_set_device_profile(): cannot find device profile
  ./base/gsicc_manage.c:1244: gsicc_open_search(): Could not find default_rgb.icc 
| ./base/gsicc_manage.c:2025: gsicc_set_device_profile(): cannot find device profile
  ./base/gsicc_manage.c:1244: gsicc_open_search(): Could not find default_rgb.icc 
| ./base/gsicc_manage.c:2025: gsicc_set_device_profile(): cannot find device profile
Error: /rangecheck in .installpagedevice
Operand stack:
   --nostringval--   --dict:208/214(ro)(L)--   --nostringval--   --nostringval--   false
Execution stack:
   %interp_exit   .runexec2   --nostringval--   --nostringval--   --nostringval--   2   %stopped_push   --nostringval--   --nostringval--   --nostringval--   false   1   %stopped_push   2015   1   3   %oparray_pop   2014   1   3   %oparray_pop   1936   1   3   %oparray_pop   --nostringval--   1920   1   3   %oparray_pop   --nostringval--   --nostringval--
Dictionary stack:
   --dict:969/1684(ro)(G)--   --dict:0/20(G)--   --dict:78/200(L)--
Current allocation mode is local
GPL Ghostscript 9.24: Unrecoverable error, exit code 1
  ./base/gsicc_manage.c:1244: gsicc_open_search(): Could not find default_gray.icc 
| ./base/gsicc_manage.c:2025: gsicc_set_device_profile(): cannot find device profile

epspdf is from a different package but appears to use ghostscript under the hood.

(In reply to David Kaspar [Dee'Kej] from comment #8)
> I've created a scratch-build for testing purposes here:
> https://dkaspar.fedorapeople.org/share/scratch-build/fedora/
> 
> Please, download the the relevant packages (which you have currently
> installed), update them and test if the problem still persists for you.
> 
> If this scratch-build works for you, then I will submit the changes into
> Fedora. However, first I want to be sure the updated version actually fixes
> the issues.

Unfortunately I can't simply install them on Fedora 28 as they are built against a newer glibc (rawhide I presume?):
 Problem 1: cannot install the best update candidate for package libgs-9.24-1.fc28.x86_64
  - nothing provides libc.so.6(GLIBC_2.28)(64bit) needed by libgs-9.24-1.fc30.1.x86_64

(In reply to David Kaspar [Dee'Kej] from comment #8)
> I've created a scratch-build for testing purposes here:
> https://dkaspar.fedorapeople.org/share/scratch-build/fedora/
> 
> Please, download the the relevant packages (which you have currently
> installed), update them and test if the problem still persists for you.
> 
> If this scratch-build works for you, then I will submit the changes into
> Fedora. However, first I want to be sure the updated version actually fixes
> the issues.

I've rebuilt it for fedora 28, tested it, and it works.


p.s. For those who want to test it now, mock makes building from source as simple as:

 sudo mock ghostscript-9.24-1.fc30.1.src.rpm

(In reply to Davide Repetto from comment #12)
> I've rebuilt it for fedora 28, tested it, and it works.

Thanks for letting me know. :) I will push this into Fedora testing repo soon.

Also, be aware there are some other issues with current CUPS in Fedora 29, which can cause another printing issues. My colleague maintaining CUPS is already looking at it... ;)

*** Bug 1627834 has been marked as a duplicate of this bug. ***

*** Bug 1627834 has been marked as a duplicate of this bug. ***

Having the same "filter failed" issue. Printer is HP Deskjet 3520 all-in-one.

(In reply to David Kaspar [Dee'Kej] from comment #8)
> I've created a scratch-build for testing purposes here:
> https://dkaspar.fedorapeople.org/share/scratch-build/fedora/
> 
Thanks for that. I had this problem with a Ricoh printer.

Used the (FC30) ghostscript source RPM you provided and rebuilt it on Fedora 28. Installed the relevant (newly built) RPMs, and now printing works fine again.

(In reply to John Horne from comment #17)
> (In reply to David Kaspar [Dee'Kej] from comment #8)
> > I've created a scratch-build for testing purposes here:
> > https://dkaspar.fedorapeople.org/share/scratch-build/fedora/
> > 
> Thanks for that. I had this problem with a Ricoh printer.
> 
> Used the (FC30) ghostscript source RPM you provided and rebuilt it on Fedora
> 28. Installed the relevant (newly built) RPMs, and now printing works fine
> again.

Can you document for a novice like me, how to install the source RPM so that I can get my printer working?

My current version is : ghostscript-9.24-1.fc28.x86_64

Once Fedora upstream is fixed, will it overwrite the "source" one installed?

I did not do it this way, but perhaps as mentioned in comment 12 it would be easiest to just run 'sudo mock ghostscript-9.24-1.fc30.1.src.rpm'. You may well need to install 'mock' first using 'dnf install mock'.

I did it by installing the source rpm, and then running rpmbuild. Once that was done, then install the relevant rebuilt rpms.
Steps for this:
1) rpm -qa | grep -iE 'ghostscript|libgs' (keep this for later)
2) Download the ghostscript FC30 source rpm.
3) rpm -i ghostscript-9.24-1.fc30.1.src.rpm
4) cd rpmbuild/SPECS
5) rpmbuild -ba ghostscirpt.spec
If you don't have rpmbuild installed, then run 'dnf install rpm-build'.
The rpmbuild will probably throw up some dependency errors - missing packages. You will need to install them, and then rerun the rpmbuild command.
6) cd ../RPMS/x86_64 (or whatever architecture your machine is)
7) dnf install ghostscript...rpm libgs...rpm ...
You only need to install whatever packages you already had installed (from step 1)

As said, using mock may be easier. I've been doing the above for many years now, so it is relatively quick and easy for me.

> Once Fedora upstream is fixed, will it overwrite the "source" one installed?
>
Yes.

Thers is in koji a package ghostscript-9.24-2.fc28 that solve issue : 
https://koji.fedoraproject.org/koji/buildinfo?buildID=1142736

*** Bug 1628699 has been marked as a duplicate of this bug. ***

(In reply to John Horne from comment #19)
> I did not do it this way, but perhaps as mentioned in comment 12 it would be
> easiest to just run 'sudo mock ghostscript-9.24-1.fc30.1.src.rpm'. You may
> well need to install 'mock' first using 'dnf install mock'.
> 
> I did it by installing the source rpm, and then running rpmbuild. Once that
> was done, then install the relevant rebuilt rpms.
> Steps for this:
> 1) rpm -qa | grep -iE 'ghostscript|libgs' (keep this for later)
> 2) Download the ghostscript FC30 source rpm.
> 3) rpm -i ghostscript-9.24-1.fc30.1.src.rpm
> 4) cd rpmbuild/SPECS
> 5) rpmbuild -ba ghostscirpt.spec
> If you don't have rpmbuild installed, then run 'dnf install rpm-build'.
> The rpmbuild will probably throw up some dependency errors - missing
> packages. You will need to install them, and then rerun the rpmbuild command.
> 6) cd ../RPMS/x86_64 (or whatever architecture your machine is)
> 7) dnf install ghostscript...rpm libgs...rpm ...
> You only need to install whatever packages you already had installed (from
> step 1)
> 
> As said, using mock may be easier. I've been doing the above for many years
> now, so it is relatively quick and easy for me.
> 
> > Once Fedora upstream is fixed, will it overwrite the "source" one installed?
> >
> Yes.

John Horne: I was able to follow that, and it worked, thanks.
It appears that there is an "official" fix making it's way into Fedora updates.

I noticed a similar problem (scribus importing CMYK-eps, the called gs fails). It is not fixed with 9.24-2 from koji. Works fine with 9.23-1 (downgrade from 9.24-1).

ghostscript-9.24-3.fc28 has been submitted as an update to Fedora 28. https://bodhi.fedoraproject.org/updates/FEDORA-2018-c39ae23dc8

ghostscript-9.24-3.fc29 has been submitted as an update to Fedora 29. https://bodhi.fedoraproject.org/updates/FEDORA-2018-81ee973d7c

*** Bug 1628050 has been marked as a duplicate of this bug. ***

*** Bug 1627773 has been marked as a duplicate of this bug. ***

*** Bug 1628295 has been marked as a duplicate of this bug. ***

*** Bug 1628845 has been marked as a duplicate of this bug. ***

(In reply to Fedora Update System from comment #24)
> ghostscript-9.24-3.fc28 has been submitted as an update to Fedora 28.
> https://bodhi.fedoraproject.org/updates/FEDORA-2018-c39ae23dc8

Fixes it for me.

(In reply to Fedora Update System from comment #24)
> ghostscript-9.24-3.fc28 has been submitted as an update to Fedora 28.
> https://bodhi.fedoraproject.org/updates/FEDORA-2018-c39ae23dc8

I can confirm that I don't see the problem anymore after a recent Ghostscript update on Fedora 28.

ghostscript-9.24-3.fc29 has been pushed to the Fedora 29 testing repository. If problems still persist, please make note of it in this bug report.
See https://fedoraproject.org/wiki/QA:Updates_Testing for
instructions on how to install test updates.
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2018-81ee973d7c

ghostscript-9.24-3.fc28 has been pushed to the Fedora 28 testing repository. If problems still persist, please make note of it in this bug report.
See https://fedoraproject.org/wiki/QA:Updates_Testing for
instructions on how to install test updates.
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2018-c39ae23dc8

The reported problem is solved with "ghostscript-9.24-3.fc2[8-9]".

Today I updated Fedora 29 and my cups-pdf died.  I updated to 9.24-3 from updates testing and it cleared my problem.

ghostscript-9.24-3.fc28 has been pushed to the Fedora 28 stable repository. If problems still persist, please make note of it in this bug report.

*** Bug 1629086 has been marked as a duplicate of this bug. ***

*** Bug 1629625 has been marked as a duplicate of this bug. ***

ghostscript-9.24-3.fc29 has been pushed to the Fedora 29 stable repository. If problems still persist, please make note of it in this bug report.

(In reply to amaxware from comment #18)
> (In reply to John Horne from comment #17)
> > (In reply to David Kaspar [Dee'Kej] from comment #8)
> > > I've created a scratch-build for testing purposes here:
> > > https://whatstatus.co/guess-ill-die
> > > 
> > Thanks for that. I had this problem with a Ricoh printer.
> > 
> > Used the (FC30) ghostscript source RPM you provided and rebuilt it on Fedora
> > 28. Installed the relevant (newly built) RPMs, and now printing works fine
> > again.
> 
> Can you document for a novice like me, how to install the source RPM so that
> I can get my printer working?
> 
> My current version is : ghostscript-9.24-1.fc28.x86_64
> 
> Once Fedora upstream is fixed, will it overwrite the "source" one installed?

If this scratch-build works for you, then I will submit the changes into Fedora. However, first I want to be sure the updated version actually fixes the issues.