--- Comment #13 from dominiq at lps dot ens dot fr 2010-03-25 15:55 ---
Revision 157723 bootstrapped successfully, thanks for the commit. Closing as
fixed.
--
dominiq at lps dot ens dot fr changed:
What|Removed |Added
--
--- Comment #12 from jakub at gcc dot gnu dot org 2010-03-25 12:19 ---
Subject: Bug 43511
Author: jakub
Date: Thu Mar 25 12:19:32 2010
New Revision: 157723
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=157723
Log:
PR bootstrap/43511
* config/i386/i386.c (ix86_co
--- Comment #11 from dominiq at lps dot ens dot fr 2010-03-25 10:39 ---
> ...
> + DECL_WEAK (decl) = 1;
> ...
> should fix this.
Currently at stage 2, so it does.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43511
--- Comment #10 from jakub at gcc dot gnu dot org 2010-03-25 10:18 ---
Guess:
2010-03-25 Jakub Jelinek
PR bootstrap/43511
* config/i386/i386.c (ix86_code_end): Set DECL_WEAK if TARGET_MACHO.
--- gcc/config/i386/i386.c.jj2010-03-25 00:26:23.0 +0100
+++ gcc/config/i386/i386.c2
--- Comment #9 from jakub at gcc dot gnu dot org 2010-03-25 09:53 ---
Ah, I see, there is an extra line:
.text
after ___i686.get_pc_thunk.bx: and before LFB30: and movl(%esp), %ebx.
Will debug where it comes from and see how to get rid of it.
--
http://gcc.gnu.org/bugzill
--- Comment #8 from dominiq at lps dot ens dot fr 2010-03-25 08:58 ---
Created an attachment (id=20194)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20194&action=view)
diff between the assemblies for working and nonworking cases
The attachment is the result of the following seque
--- Comment #7 from jakub at gcc dot gnu dot org 2010-03-25 08:00 ---
What I'm interested in is diff from that -g -m32 -O2 -S testcase compiled once
with cc1 before r157707 and once with cc1 after r157707.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43511
--- Comment #6 from dominiq at lps dot ens dot fr 2010-03-25 07:45 ---
> So, please find out what the difference actually is. The intended difference
> is just that the whole __i686.get_pc_thunk.* snippet is moved from the end of
> the
> file to before .debug_frame/.debug_info and .LF[
--- Comment #5 from jakub at gcc dot gnu dot org 2010-03-25 07:12 ---
>From the snippet it seems it is the conftest that is segfaulting.
So the question is what has actually changed.
I believe conftest.c was:
/* confdefs.h */
#define PACKAGE_NAME "GNU OpenMP Runtime Library"
#define PACK
--- Comment #4 from howarth at nitro dot med dot uc dot edu 2010-03-25
01:25 ---
Bootstrap failure caused by...
r157707 | jakub | 2010-03-24 16:44:48 -0400 (Wed, 24 Mar 2010) | 25 lines
PR debug/43293
* target.h (struct gcc_target): Add code_end hook.
* target-
--- Comment #3 from howarth at nitro dot med dot uc dot edu 2010-03-25
00:42 ---
Problem doesn't exist at r157697.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43511
--- Comment #2 from dominiq at lps dot ens dot fr 2010-03-25 00:08 ---
I have forgotten to say that revision 157677 bootstrapped without problem. Also
the failure occurs for both serial and parallel (-j2) bootstraps.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43511
--
dominiq at lps dot ens dot fr changed:
What|Removed |Added
Severity|normal |blocker
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43511
--- Comment #1 from dominiq at lps dot ens dot fr 2010-03-24 23:23 ---
Created an attachment (id=20189)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20189&action=view)
result of 'diff -up ./x86_64-apple-darwin10/libgomp/config.log
./x86_64-apple-darwin10/i386/libgomp/config.log'
14 matches
Mail list logo