Bug 32748 - [4.3 regression]: gfortran.dg/array_constructor_6.f90
Summary: [4.3 regression]: gfortran.dg/array_constructor_6.f90
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: target (show other bugs)
Version: 4.3.0
: P3 normal
Target Milestone: 4.3.0
Assignee: Janne Blomqvist
URL:
Keywords: wrong-code
Depends on:
Blocks:
 
Reported: 2007-07-12 23:48 UTC by H.J. Lu
Modified: 2007-07-16 14:54 UTC (History)
4 users (show)

See Also:
Host:
Target: ia64-unknown-linux-gnu
Build:
Known to work:
Known to fail:
Last reconfirmed: 2007-07-15 06:57:57

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description H.J. Lu 2007-07-12 23:48:09 UTC 
Revision 126278 fails:

FAIL: gfortran.dg/array_constructor_6.f90  -O3 -fomit-frame-pointer -funroll-loops  execution test
FAIL: gfortran.dg/array_constructor_6.f90  -O3 -fomit-frame-pointer -funroll-all-loops -finline-functions  execution test

and revision 126120 works.
Comment 1 H.J. Lu 2007-07-13 14:50:50 UTC 
revision 126200 works.
Comment 2 H.J. Lu 2007-07-13 18:53:06 UTC 
revision 126240 works.
Comment 3 H.J. Lu 2007-07-13 20:03:16 UTC 
revision 126260 works.
Comment 4 H.J. Lu 2007-07-13 20:50:54 UTC 
revision 126271 is bad.
Comment 5 H.J. Lu 2007-07-13 22:39:37 UTC 
revision 126264 is bad.
Comment 6 H.J. Lu 2007-07-13 22:59:28 UTC 
revision 126263 is bad. This patch

http://gcc.gnu.org/ml/gcc-patches/2007-07/msg00165.html

is the cause.
Comment 7 H.J. Lu 2007-07-14 15:12:00 UTC 
(In reply to comment #6)
> revision 126263 is bad. This patch
> 
> http://gcc.gnu.org/ml/gcc-patches/2007-07/msg00165.html
> 
> is the cause.
> 

I meant to say "revision 126263 works."
Comment 8 Janne Blomqvist 2007-07-15 06:57:57 UTC 
Assigning this bug to myself, since it seems I caused it.
Comment 9 Janne Blomqvist 2007-07-15 21:33:44 UTC 
Subject: Bug 32748

Author: jb
Date: Sun Jul 15 21:33:32 2007
New Revision: 126662

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=126662
Log:
2007-07-16  Janne Blomqvist  <jb@gcc.gnu.org>

	PR fortran/32748
	* trans-decl.c (gfc_build_builtin_function_decls): Remove
	DECL_IS_MALLOC attribute from internal_realloc, thus reverting
	part of my 2007-07-03 patch.


Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/trans-decl.c
Comment 10 Janne Blomqvist 2007-07-15 21:35:48 UTC 
With the 126662 commit, this should now be fixed, closing.
Comment 11 Richard Biener 2007-07-16 14:54:09 UTC 
The failure is gone, but the cause might just be hidden again (the IL generated
for the realloc looks safe, even with the malloc attribute set).