Bug 692464 - Ghostscript 9.04 fails to build on ~x86 without SSE2 support
Summary: Ghostscript 9.04 fails to build on ~x86 without SSE2 support
Status: RESOLVED INVALID
Alias: None
Product: Ghostscript
Classification: Unclassified
Component: Build Process (show other bugs)
Version: master
Hardware: PC Linux
: P4 normal
Assignee: Chris Liddell (chrisl)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-08-25 14:52 UTC by Timo Gurr
Modified: 2011-08-26 10:23 UTC (History)
0 users

See Also:
Customer:
Word Size: ---


Attachments
build.log (809.93 KB, text/x-log)
2011-08-25 14:52 UTC, Timo Gurr
Details
ghostscript-9.04-sse2-fix.patch (356 bytes, patch)
2011-08-25 14:54 UTC, Timo Gurr
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Timo Gurr 2011-08-25 14:52:12 UTC
Created attachment 7831 [details]
build.log

System uname: Linux-2.6.38.7-yoruichi-i686-AMD_Athlon-tm-_XP_3200+-with-gentoo-2.0.2

build failure:

In file included from ./base/gxht_thresh.c:45:0:                                
/usr/lib/gcc/i686-pc-linux-gnu/4.5.3/include/emmintrin.h:32:3: error: #error
"SSE2 instruction set not enabled"

Offending Ghostscript git commit:
http://git.ghostscript.com/?p=ghostpdl.git;a=commit;h=4132ef5d36ff63b2b84f9864b7240609ac84ac38

Downstream Gentoo bug with more information and complete build.log:
https://bugs.gentoo.org/show_bug.cgi?id=379597
Comment 1 Timo Gurr 2011-08-25 14:54:48 UTC
Created attachment 7832 [details]
ghostscript-9.04-sse2-fix.patch

Proposed patch by Andrew Savchenko, source: https://bugs.gentoo.org/show_bug.cgi?id=379597#c4
Comment 2 Chris Liddell (chrisl) 2011-08-25 15:13:51 UTC
It works correctly for me, and the patch makes no sense - if the compile test fails, then the compile and link test should also fail (which it does for me).

I'm using autoconf 2.67, but I'm not aware of anything that should cause problems with the stated required version (2.52).
Comment 3 Timo Gurr 2011-08-26 10:10:13 UTC
(In reply to comment #2)
> It works correctly for me, and the patch makes no sense - if the compile test
> fails, then the compile and link test should also fail (which it does for me).

Thanks for investigating, in the meantime we found out that it has to be some CFLAGS/LDFLAGS combination causing the error. If it's some valid combination causing the problem the reporter will report and reopen this bug.
Comment 4 Chris Liddell (chrisl) 2011-08-26 10:23:15 UTC
Thanks for updating this bug, much appreciated.

Indeed, if you do established a sane set of flags that reproduces the problem, I'll be happy to investigate.