Bug 11193 – Problem with GCC 3.2.1

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Bjarke Istrup Pedersen (RETIRED) 2002-11-25 05:59:05 UTC]

Is it a discribed here in details: http://forums.gentoo.org/viewtopic.php?
t=23901&highlight=glibc+python

[Comment 1 Bjarke Istrup Pedersen (RETIRED) 2002-11-26 04:45:19 UTC]

The problem seems to be when compiling Glibc 2.3.1 with GCC 3.2.1 .
When is has been compiled with GCC 3.2.1, it doesn't work, but when compiled 
with GCC 2.2.x it works.

[Comment 2 Bor 2002-12-01 15:29:58 UTC]

More specifficaly, it works when you compile glibc 2.3.1 with gcc 3.2-r4. Something changed in 
gcc from 3.2 to 3.2.1 that caused this error. As I wrote in my forum post, it was some kind of 
floating point-integer conversion error. Really bizzare, since it would alternate in about two 
minute intervals ( maybe even 128 second...). One minute it would work ok, the next it would 
bomb. 
 

[Comment 3 Matthew Kennedy (RETIRED) 2002-12-14 00:10:56 UTC]

assigned

[Comment 4 Bjarke Istrup Pedersen (RETIRED) 2002-12-18 13:19:14 UTC]

I just compiled it this way, and it works now.
Is was an error in portage, and it should have been fixed.
But, before closing this bug, will someone please test it?

[Comment 5 Matthew Kennedy (RETIRED) 2003-01-08 23:13:20 UTC]

i suppose i  can close this. i've never seen it happen in all the similar
situations i've upgraded in

[Comment 6 Bardur Arantsson 2003-01-12 02:39:20 UTC]

This has *not* been fixed (at least not for me) :(. Also see bug #12554 for details of an easy 
(although timeconsuming :)) way to reproduce the bug. 
 
As mentioned by the GCC people 
(http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=8988), 
this is a bug in the optimization code of GCC 3.2.1. 
 
There are patches at the URL given above, maybe they should be applied in the Gentoo gcc 
ebuild? 
 

[Comment 7 Aron Griffis (RETIRED) 2003-01-21 10:21:44 UTC]

This bug should not be closed.  It is blocking bug 12554 (which I don't believe
is resolved yet) and that blocks bug 13396.  There is clearly a problem here
that hasn't been fixed yet, even if it works for some people.

Has anybody looked into the patches mentioned by Bardur above?

Aron

[Comment 8 Matthew Kennedy (RETIRED) 2003-02-09 16:04:26 UTC]

if the problem described in this bug report is related to http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=8988 (as suggested in comment #6), then you may want to upgrade to gcc3.2.2. I have verified that it includes the fix.

test and let me know what happens

[Comment 9 Cong 2003-02-09 19:06:16 UTC]

Related to bug 12544,

I upgrade gcc to 3.2.2, remerge glibc and python but it did not fix bug 12544.
The versions are as follow:

# emerge gcc glibc python -Cp

>>> These are the packages that I would unmerge:

 sys-libs/glibc
    selected: 2.3.1-r2
   protected: none
     omitted: none

 dev-lang/python
    selected: 2.2.2
   protected: none
     omitted: none

 sys-devel/gcc
    selected: 3.2.2
   protected: none
     omitted: none

[Comment 10 Cong 2003-02-09 19:08:05 UTC]

I am sorry. I meant bug 12554, not bug 12544.

[Comment 11 John Davis (zhen) (RETIRED) 2003-04-04 01:21:13 UTC]

db fix

[Comment 12 John Davis (zhen) (RETIRED) 2003-04-04 01:26:06 UTC]

db fix