Bug 377120 - invalidaccess -7
Summary: invalidaccess -7
Status: RESOLVED DUPLICATE of bug 371887
Alias: None
Product: okular
Classification: Applications
Component: PS backend (show other bugs)
Version: unspecified
Platform: Other Linux
: NOR normal
Target Milestone: ---
Assignee: Okular developers
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-03-02 17:50 UTC by vtwin farriers
Modified: 2017-03-03 23:02 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description vtwin farriers 2017-03-02 17:50:56 UTC
CentOS 7.3.1611

okular -v
 Qt: 4.8.5
 KDE Development Platform: 4.14.8
 Okular: 0.16.5

pdftotext -v
 pdftotext version 0.26.5

I have two identically configured CentOS 7 machines. Only difference is the hardware -- e.g. one has an nvidia card, the other has an on-board intel graphics. Running Okular on postscript files on the machine with the nvidia card, I get the following errors:

Connecting to deprecated signal QDBusConnectionInterface::serviceOwnerChanged(QString,QString,QString)
kbuildsycoca4 running...
invalidaccess -7
okular(42872)/okular (Spectre) GSRendererThread::run: Generated image does not match wanted size: [0x0] vs requested [596x883]
QImage::scaled: Image is a null image
invalidaccess -7
okular(42872)/okular (Spectre) GSRendererThread::run: Generated image does not match wanted size: [0x0] vs requested [579x858]
QImage::scaled: Image is a null image
invalidaccess -7
okular(42872)/okular (Spectre) GSRendererThread::run: Generated image does not match wanted size: [0x0] vs requested [71x105]
QImage::scaled: Image is a null image
invalidaccess -7
okular(42872)/okular (Spectre) GSRendererThread::run: Generated image does not match wanted size: [0x0] vs requested [579x858]
QImage::scaled: Image is a null image
Comment 1 Wolfgang Bauer 2017-03-02 22:04:12 UTC
I would suspect this is the same as bug#371887 (caused by a security "fix" in ghostscript).

Please try the fix in comment#7 there.
Comment 2 Albert Astals Cid 2017-03-03 22:29:35 UTC
I find it very strange that you have exactly the same versions of all packages and you get different versions.

Can you confirm any related ghostscript package has exactly the same version?
Comment 3 vtwin farriers 2017-03-03 22:47:26 UTC
Ghostscript version is 7.09 on Centos 7.3.1611.

the "fix" in comment 7 works here as well, although it is line 2022 rather than line 2024
Comment 4 vtwin farriers 2017-03-03 22:51:01 UTC
(In reply to Albert Astals Cid from comment #2)
> I find it very strange that you have exactly the same versions of all
> packages and you get different versions.
> 
> Can you confirm any related ghostscript package has exactly the same version?

I can confirm I have identical configurations on multiple machines, and the "bug" only manifests itself on certain systems. Perhaps there is a library at a much lower level which differs due to the graphics card that causes the problem to manifest itself. The machines I have had difficulty with thus far are all running kmod-nvidia from the epel repository as they have NVidia cards. The few machines I have which have not had problems reported thus far have intel graphics chips on board. I have used the same .ps file for testing on both.
Comment 5 Wolfgang Bauer 2017-03-03 23:02:13 UTC
(In reply to vtwin farriers from comment #3)
> Ghostscript version is 7.09 on Centos 7.3.1611.
> 
> the "fix" in comment 7 works here as well, although it is line 2022 rather
> than line 2024

Ok, great!

Just in case it wasn't clear already from the other bug report, that is actually an upstream fix in ghostscript:
http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=99e331527d541a8f01ad5455c4eb2aabd67281a6;hp=0d4644c003067fc14ca1db9c600dce420c06e6b1

IMHO, you should ask your distribution to backport that or upgrade ghostscript to the latest version.
This will not only affect okular, but any application that uses libspectre (e.g. evince) and maybe others too.

In any case, I suppose we should rather close this as duplicate then...

*** This bug has been marked as a duplicate of bug 371887 ***