Bug 11798 - ICE in Open Inventor GeneralizedCylinder.c++
Summary: ICE in Open Inventor GeneralizedCylinder.c++
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: c++ (show other bugs)
Version: 2.95.3
: P2 normal
Target Milestone: 3.0.x
Assignee: Not yet assigned to anyone
URL:
Keywords: ice-on-valid-code
Depends on:
Blocks:
 
Reported: 2003-08-05 00:05 UTC by Scott McPeak
Modified: 2005-07-23 22:49 UTC (History)
1 user (show)

See Also:
Host: i386-slackware-linux
Target: i386-slackware-linux
Build: i386-slackware-linux
Known to work:
Known to fail:
Last reconfirmed:


Attachments
This is the preprocessed input that triggers an ICE (72.49 KB, text/plain)
2003-08-05 00:07 UTC, Scott McPeak
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Scott McPeak 2003-08-05 00:05:36 UTC
$ g++ -o /dev/null -c -fPIC -O GeneralizedCylinder.ii
GeneralizedCylinder.c++: In method `void
GeneralizedCylinder::initSpineKeyPointRotations()':
GeneralizedCylinder.c++:478: Internal compiler error.
GeneralizedCylinder.c++:478: Please submit a full bug report.
GeneralizedCylinder.c++:478: See <URL:http://www.gnu.org/software/gcc/bugs.html>

$ g++ -v
Reading specs from /usr/lib/gcc-lib/i386-slackware-linux/2.95.3/specs
gcc version 2.95.3 20010315 (release)

Note that the optimization settings, "-fPIC -O", are necessary to trigger
the ICE.  I'll attach the preprocessed input file as the next step (assuming
this works like Mozilla's bugzilla...)
Comment 1 Scott McPeak 2003-08-05 00:07:25 UTC
Created attachment 4563 [details]
This is the preprocessed input that triggers an ICE
Comment 2 Andrew Pinski 2003-08-05 00:11:35 UTC
I could reproduce this on 2.95.3 but it has been fixed since at least 3.0.4 so closing as fixed.
3.3.1 is about to come out, I think it is time for an upgrade.
Comment 3 Scott McPeak 2003-08-05 00:27:13 UTC
I have several versions of gcc-3.x installed, and confirm that the ICE does
not occur with them.  However, 2.95.3 is *four times* faster than 3.3 at
compiling, and generates code that is only a little slower.  Even when
optimization is disabled, 2.95.3 is still at least twice as fast.  Therefore,
it's my preferred development compiler.  I only use 3.x to test that my
releases compile on it.

If I fix this myself (not that I'm confident I can), will there be an avenue
for getting this into a 2.95.x release?  Or has the decision already been
made that 2.95.{4,5} will never exist?
Comment 4 Andrew Pinski 2003-08-05 00:29:52 UTC
There are no plans to update anything before 3.3 any more, Sorry.
Comment 5 Nathanael C. Nerode 2003-08-05 00:41:36 UTC
We are, of course, trying to make the 3.x series faster.  Significant speedups
should be in 3.4 when it comes out, so please try again then.  :-)