Bug 693952 - convert(1) dumps core when processing PDF file on Oracle Solaris 11
Summary: convert(1) dumps core when processing PDF file on Oracle Solaris 11
Status: RESOLVED INVALID
Alias: None
Product: Ghostscript
Classification: Unclassified
Component: General (show other bugs)
Version: 9.00
Hardware: Other Solaris
: P4 normal
Assignee: Default assignee
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-05-07 15:15 UTC by Sunil Kumar
Modified: 2013-05-07 17:11 UTC (History)
1 user (show)

See Also:
Customer:
Word Size: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Sunil Kumar 2013-05-07 15:15:37 UTC
convert(1) dumps core when processing PDF file.

Testcase

Get any PDF file, e.g.,

http://www.oracle.com/technetwork/server-storage/solaris11/documentation/solaris11-1-whatsnew-1732377.pdf

and pass this file to convert(1).

$ ls -l solaris11-1-whatsnew-1732377.pdf
-rw-r--r--   1 root     root      349900 Jan 24 03:38
solaris11-1-whatsnew-1732377.pdf
$ convert solaris11-1-whatsnew-1732377.pdf solaris11-1-whatsnew-1732377.png
Bus Error (core dumped)
$ pstack core
core 'core' of 7810:    convert solaris11-1-whatsnew-1732377.pdf
solaris11-1-whatsnew-1732377.
 ff29df40 mutex_trylock (3c95c, 0, 0, 2944, 2800, ff) + 4c
 fde9b998 gp_monitor_enter (3c95c, fdcdb244, fe034c3c, ff3ee3c0, 1, 11db00) +
4
 fe034528 gs_heap_alloc_bytes (4be18, 6c, fe13ff7c, 0, 3c958, fe152000) + 2c
 fe039ab4 gs_lib_ctx_init (4be18, fffedf7c, fe0344fc, 12000, 408400, 4be18) +
64
 fe034c3c gs_malloc_init (0, fe15edb0, 0, fe152000, cdb0, 4be18) + 24
 fdd4b1a0 gsapi_new_instance (ffbf3874, 0, fdd4b138, fdd4b37c, fdd4b37c,
ffbf3874) + 68
 fee22ea4 ReadPDFImage (24, ffbffc00, 555e0, 4cdc8, fee28dac, fee3a000) +
10f4
 fef9d6c4 ReadImage (6050, ffbffc00, 3000, ffbfca40, ff14e050, 4cd98) + 27c
 fee77c74 ConvertImageCommand (0, 2, 2, ffbffb1c, 30ff0, 1) + 688
 00010cf8 main     (3, ffbffc84, 31668, ffbffc00, 0, 10d88) + b4
 00010c2c _start   (0, 0, 0, 0, 0, 0) + 108
$
Comment 1 Sunil Kumar 2013-05-07 15:19:55 UTC
The issue was seen only on sparc platform and was occurring due to the alignment fault.
Comment 2 Sunil Kumar 2013-05-07 15:20:51 UTC
The issue was seen only on sparc platform and was occurring due to the alignment fault.
Comment 3 Chris Liddell (chrisl) 2013-05-07 15:29:34 UTC
This looks very much like a duplicate of:

http://bugs.ghostscript.com/show_bug.cgi?id=691989

If it isn't, you need to report it in a way that doesn't rely on using "convert" or other third party package. Normally, you should report it to the package supplying the exe you actually call, and the developers of that should work out whether the problem is with their way of driving Ghostscript or a real GS problem. Then work out a simple way for us to reproduce the issue, and report that to us.
Comment 4 Sunil Kumar 2013-05-07 17:03:16 UTC
This bug is not a duplicate of 

http://bugs.ghostscript.com/show_bug.cgi?id=691989

Also the call stack is different.

convert is an ImageMagick utility.
Comment 5 Chris Liddell (chrisl) 2013-05-07 17:11:50 UTC
(In reply to comment #4)
> This bug is not a duplicate of 
> 
> http://bugs.ghostscript.com/show_bug.cgi?id=691989
> 
> Also the call stack is different.

Have you tested with the latest Ghostscript release?

> convert is an ImageMagick utility.

I know where convert comes from, we did not develop, nor contribute to ImageMagick. If you can demonstrate the problem using Ghostscript, I'll be happy to reopen this, and investigate. I do not intend to debug ImageMagick convert, too.