Hi,
The following short code example produces an internal
compiler error with g++ 4.4.0.
struct Void {};
template class R struct FunType {
typedef R ResultType;
};
struct WrongNumberOfSigArgs {};
template typename R struct CFunType {
template class Dummy1=Void, class Dummy2=Void struct
--- Comment #6 from sekar dot as at lntinfotech dot com 2009-07-20 06:27
---
Internal compiler error in `make_decl_rtl', at varasm.c:745
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=745
--- Comment #1 from fxcoudert at gcc dot gnu dot org 2009-07-20 06:47
---
Original bug report was for: gcc version 4.5.0 20090717 on
x86_64-unknown-linux-gnu
Confirmed with trunk rev. 148180 on i386-darwin.
--
fxcoudert at gcc dot gnu dot org changed:
What|Removed
--- Comment #10 from pault at gcc dot gnu dot org 2009-07-20 07:28 ---
(In reply to comment #9)
(In reply to comment #8)
I pointed this out to Paul already, but appearantly it is still stuck in
his
whole-file patch queue.
Last incarnation of that patch (containing this
--- Comment #2 from burnus at gcc dot gnu dot org 2009-07-20 08:32 ---
Further simplified test below. Valgrind shows first:
==29083== Conditional jump or move depends on uninitialised value(s)
==29083==at 0x4007E4: MAIN__ (aa.f90:12)
The problem is that t2's y is unallocated,
--- Comment #4 from jakub at gcc dot gnu dot org 2009-07-20 09:14 ---
The assert doesn't make any sense.
--
jakub at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #8 from janus at gcc dot gnu dot org 2009-07-20 09:24 ---
I guess everything is fixed now. Can we close this PR?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40591
--- Comment #5 from jakub at gcc dot gnu dot org 2009-07-20 09:44 ---
Patch posted: http://gcc.gnu.org/ml/gcc-patches/2009-07/msg01116.html
--
jakub at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #7 from irar at il dot ibm dot com 2009-07-20 11:18 ---
AFAIU, querying for the component type of complex type is not difficult to
implement.
I think, that loop-based vectorization is preferable here, so we should stay
with vectorization factor of 2 for doubles.
The next
--- Comment #28 from irar at il dot ibm dot com 2009-07-20 12:03 ---
I've just committed a patch that adds support of cond_expr in reductions in
nested cycles (http://gcc.gnu.org/ml/gcc-patches/2009-07/msg01124.html).
cond_expr cannot be vectorized in reduction of inner-most loop,
--- Comment #8 from ubizjak at gmail dot com 2009-07-20 12:30 ---
(In reply to comment #7)
Does __builtin_cexpi have a vector implementation? If so, does it return two
vectors?
No, only vectorized sincos is implemented (see also links at PR40766, Comment
#11).
--
--- Comment #9 from rguenther at suse dot de 2009-07-20 12:55 ---
Subject: Re: Vectorization of complex types,
vectorization of sincos missing
On Mon, 20 Jul 2009, irar at il dot ibm dot com wrote:
--- Comment #7 from irar at il dot ibm dot com 2009-07-20 11:18 ---
--- Comment #3 from pault at gcc dot gnu dot org 2009-07-20 13:30 ---
(In reply to comment #2)
Any reference, within 'fun', to 'q' clears the fault. I guess that the absence
of a reference prevents 'q' from being presented to be nullified. I'll take a
look tonight.
Paul
--
--- Comment #3 from davek at gcc dot gnu dot org 2009-07-20 14:11 ---
Created an attachment (id=18229)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18229action=view)
proposed fix
This patch fixes all the significant FAILs currently extant in the libffi
testsuite.
--- Comment #4 from davek at gcc dot gnu dot org 2009-07-20 14:31 ---
Created an attachment (id=18230)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18230action=view)
Updated patch
Now with a trivial tweak to fix a missing-prototype warning.
--
davek at gcc dot gnu dot org
--- Comment #6 from jakub at gcc dot gnu dot org 2009-07-20 14:38 ---
Subject: Bug 40792
Author: jakub
Date: Mon Jul 20 14:38:46 2009
New Revision: 149822
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=149822
Log:
PR tree-optimization/40792
* tree.c
--- Comment #7 from jakub at gcc dot gnu dot org 2009-07-20 14:39 ---
Subject: Bug 40792
Author: jakub
Date: Mon Jul 20 14:39:32 2009
New Revision: 149823
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=149823
Log:
PR tree-optimization/40792
* tree.c
--- Comment #8 from jakub at gcc dot gnu dot org 2009-07-20 14:40 ---
Fixed.
--
jakub at gcc dot gnu dot org changed:
What|Removed |Added
Status|ASSIGNED
--- Comment #5 from davek at gcc dot gnu dot org 2009-07-20 14:43 ---
Created an attachment (id=18231)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18231action=view)
final spin
Turns out that '#' makes a bad choice of comment introducer if the next word is
if and you're running
--- Comment #14 from kargl at gcc dot gnu dot org 2009-07-20 16:51 ---
Can you commit your fix
http://gcc.gnu.org/ml/gcc-patches/2009-07/msg01109.html
or revert your change that is causing the problem?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40800
--- Comment #15 from dave at hiauly1 dot hia dot nrc dot ca 2009-07-20
17:17 ---
Subject: Re: libcpp breaks bootstrap
Can you commit your fix
http://gcc.gnu.org/ml/gcc-patches/2009-07/msg01109.html
or revert your change that is causing the problem?
configure.ac needs to be
--- Comment #16 from jlquinn at optonline dot net 2009-07-20 17:42 ---
(In reply to comment #14)
Can you commit your fix
http://gcc.gnu.org/ml/gcc-patches/2009-07/msg01109.html
or revert your change that is causing the problem?
I posted the fix yesterday. I'm
--- Comment #17 from dave at hiauly1 dot hia dot nrc dot ca 2009-07-20
17:51 ---
Subject: Re: libcpp breaks bootstrap
I posted the fix yesterday. I'm write-after-approval, and I thought I needed
to wait for someone to OK the patch.
Probably could have been applied as obvious.
--- Comment #18 from sgk at troutmask dot apl dot washington dot edu
2009-07-20 17:56 ---
Subject: Re: libcpp breaks bootstrap
On Mon, Jul 20, 2009 at 05:42:50PM -, jlquinn at optonline dot net wrote:
--- Comment #16 from jlquinn at optonline dot net 2009-07-20 17:42
--- Comment #19 from jlquinn at gcc dot gnu dot org 2009-07-20 18:09
---
Subject: Bug 40800
Author: jlquinn
Date: Mon Jul 20 18:09:33 2009
New Revision: 149826
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=149826
Log:
2009-07-20 Jerry Quinn jlqu...@optonline.net
--- Comment #20 from sgk at troutmask dot apl dot washington dot edu
2009-07-20 18:28 ---
Subject: Re: libcpp breaks bootstrap
On Mon, Jul 20, 2009 at 06:09:46PM -, jlquinn at gcc dot gnu dot org wrote:
Author: jlquinn
Date: Mon Jul 20 18:09:33 2009
New Revision: 149826
--- Comment #21 from jlquinn at gcc dot gnu dot org 2009-07-20 19:00
---
http://gcc.gnu.org/ml/gcc-patches/2009-07/msg01109.html
Committed as patch is simple and fixes multiple platforms.
--
jlquinn at gcc dot gnu dot org changed:
What|Removed
--- Comment #3 from xenofears at gmail dot com 2009-07-20 19:30 ---
Subject: Re: Libstdc++ is broken for win32 host
No one is routinely testing 64-bit Windows GCC. Bootstrap fails by standard
settings, the drivers are quite broken in several ways. Further, I AM one of
the developers
--- Comment #4 from paolo dot carlini at oracle dot com 2009-07-20 19:47
---
Hey Peter, you are free to do whatever you want. Just wanted to clarify that
the GCC developers are routinely testing for many other targets, both 32-bit
and 64-bit, in general, and there the problem doesn't
The following program produces wrong results at -m64 -O3:
// a.cc
#include iostream
#define N 5
void
cvt_u4_f4(int n, unsigned int* u4, float* f4)
{
for (int j=0; jn; j++)
f4[j]=u4[j];
}
int main()
{
unsigned int u4[N];
float f4[N];
for (unsigned int j=0; jN; j++)
--- Comment #5 from xenofears at gmail dot com 2009-07-20 19:54 ---
Subject: Re: Libstdc++ is broken for win32 host
There is no need to post your lack of caring / plan of action. Or does your
company compose all of gcc's development team? If they do, that would
explain a lot.
--- Comment #6 from paolo dot carlini at oracle dot com 2009-07-20 20:02
---
No vacations, eh? I understand... ;) Anyway, since you don't trust me, about
the general points I'm making, let's CC Benjamin, I take no offense (it's the
two of us for the library, in general)
--
paolo
--- Comment #7 from paolo dot carlini at oracle dot com 2009-07-20 20:25
---
Let's also add a mingw maintainer. In fact, I think this PR should be just
recategorized as target, the generic libstdc++ code is very unlikely to be at
fault here.
--
paolo dot carlini at oracle dot com
--- Comment #8 from xenofears at gmail dot com 2009-07-20 20:51 ---
Subject: Re: Libstdc++ is broken for win32 host
Why are you going on? I told you, since you outspokenly don't care and/or
have no plans of looking at it, your post as such is unnecessary, unless you
are speaking for
--- Comment #10 from xenofears at gmail dot com 2009-07-20 21:11 ---
Subject: Re: Libstdc++ is broken for win32 host
It is target when host=linux64 target=win64 works but host=win64
target=win64 doesn't? Interesting. Again, please change the title, or if
that is too much trouble
--- Comment #1 from hjl dot tools at gmail dot com 2009-07-20 21:13 ---
Vectorizer needs to check signs when vectorizing conversions from
unsigned int to float if int float vector insn is used. For example:
cvtdq2ps u4(%rip), %xmm2
movaps%xmm2, %xmm1
Errors:
..\newmain.cpp:23: error: 'gnat' was not declared in this scope
..\newmain.cpp:24: error: 'grue' was not declared in this scope
These member variables should be in scope because they are inherited. It works
fine for normal classes but not template classes in g++. This compiles fine in
--- Comment #2 from hjl dot tools at gmail dot com 2009-07-20 21:16 ---
(In reply to comment #1)
Vectorizer needs to check signs when vectorizing conversions from
unsigned int to float if int float vector insn is used. For example:
Should it be done in vectorizer or in backend?
--- Comment #1 from pinskia at gcc dot gnu dot org 2009-07-20 21:17 ---
Please read http://gcc.gnu.org/gcc-3.4/changes.html#cplusplus. This code is
invalid as namelookup inside templates are not done the way you think they are
done.
--
pinskia at gcc dot gnu dot org changed:
--- Comment #11 from paolo dot carlini at oracle dot com 2009-07-20 21:20
---
By the way, since the first error occurs at lines 185 and 215 of type_traits,
where the library uses two new C++ front-end builtins, added the last week, are
you really sure your tree is completely up to
--- Comment #12 from xenofears at gmail dot com 2009-07-20 21:38 ---
Subject: Re: Libstdc++ is broken for win64 host
I would be happy to give it another try, sure. Thank you for asking.
--
From: paolo dot carlini at oracle dot com
--- Comment #13 from bkoz at gcc dot gnu dot org 2009-07-20 21:52 ---
Hi Peter! You say that this is broken as of rev 149763, which is a commit from
July 17, 2009. Here:
http://gcc.gnu.org/ml/gcc-cvs/2009-07/msg00645.html
But you also indicate that this has been broken for some time
--- Comment #14 from xenofears at gmail dot com 2009-07-20 22:11 ---
Subject: Re: Libstdc++ is broken for win64 host
I have (unintentionally, you have my apologies,) exaggerated, it hasn't been
a couple of weeks, time seems to have moved faster than it did. 20090620 is
the last I
--- Comment #3 from ubizjak at gmail dot com 2009-07-20 22:46 ---
PR 39943 ?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40809
--- Comment #4 from hjl dot tools at gmail dot com 2009-07-20 22:51 ---
(In reply to comment #3)
PR 39943 ?
Yes, it is. Gcc 4.5.0 revision 149104 works fine. Should the fix
backported to 4.3/4.4 branches?
--
hjl dot tools at gmail dot com changed:
What|Removed
[...@gnu-6 tmp]$ cat x.c
#define N 4
extern unsigned int u4[N] __attribute__ ((aligned(16)));
extern float f4[N] __attribute__ ((aligned(16)));
void
cvt_u4_f4()
{
int j;
for (j=0; jN; j++)
f4[j]=u4[j];
}
[...@gnu-6 tmp]$ /usr/gcc-4.5/bin/gcc -O3 x.c -S
[...@gnu-6 tmp]$ cat x.s
--- Comment #9 from ubizjak at gmail dot com 2009-07-20 23:12 ---
*** Bug 40809 has been marked as a duplicate of this bug. ***
--
ubizjak at gmail dot com changed:
What|Removed |Added
--- Comment #5 from ubizjak at gmail dot com 2009-07-20 23:12 ---
(In reply to comment #4)
(In reply to comment #3)
PR 39943 ?
Yes, it is. Gcc 4.5.0 revision 149104 works fine. Should the fix
backported to 4.3/4.4 branches?
Yes, please reopen PR 39943 as a regression for 4.3
1) Incorrect writing of formated files:
nrp = 70 000 000
open(1, file = 'tst.dat')
do i = 1, nrp
write(1, '(6i10)') (i, k=1,6)
enddo
close(1)
end
When file size reaches 2Gb, the last record is not added to file, but
overwrites the previous one. As a
--- Comment #1 from ubizjak at gmail dot com 2009-07-20 23:26 ---
What about unsigned int - double?
--
ubizjak at gmail dot com changed:
What|Removed |Added
seen while building openjdk-6 on sparc-linux-gnu with 4.4 branch. 4.3 branch
does work.
Matthias
g++-4.4 -c -fPIC -O1 -fno-strict-aliasing -fno-rtti -fno-exceptions -fcheck-new
c1_MacroAssembler_sparc.ii
In file included from ../generated/incls/_assembler_pd.hpp.incl:2,
from
--- Comment #1 from debian-gcc at lists dot debian dot org 2009-07-20
23:58 ---
Created an attachment (id=18232)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=18232action=view)
preprocessed source
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40813
--- Comment #1 from kargl at gcc dot gnu dot org 2009-07-21 00:01 ---
What's the output to 'gfortran -v'?
Since you said it's a 2 GB file size problem, I modified your
program to
nrp = huge(nrp) / 60
open(1, file = 'tst.dat')
do i = 1, nrp+1
write(1,
--- Comment #2 from apotapov at math dot ualberta dot ca 2009-07-21 00:12
---
Subject: Re: Incorrectly writes/reads formatted ASCII files greater than 2Gb
The output for gfortran -v is
gfortran -v try.f
Driving: E:\Program Files\gfortran\bin\gfortran.EXE -v
try.f -lgfortranbegin
--- Comment #10 from hjl dot tools at gmail dot com 2009-07-21 02:48
---
[...@gnu-6 tmp]$ vi v.c
[...@gnu-6 tmp]$ cat v.c
#include stdlib.h
#define N 4
void
cvt_u4_f4(int n, unsigned int* u4, float* f4)
{
int j;
for (j=0; jn; j++)
f4[j]=u4[j];
}
unsigned int u4[N] =
{
--- Comment #11 from hjl dot tools at gmail dot com 2009-07-21 02:49
---
I think the fix should be back ported to 4.3/4.4.
--
hjl dot tools at gmail dot com changed:
What|Removed |Added
--- Comment #2 from hjl dot tools at gmail dot com 2009-07-21 02:52 ---
(In reply to comment #1)
What about unsigned int - double?
We don't even support int - double.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40811
--- Comment #11 from pault at gcc dot gnu dot org 2009-07-21 04:51 ---
Subject: Bug 40726
Author: pault
Date: Tue Jul 21 04:51:30 2009
New Revision: 149846
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=149846
Log:
2009-07-21 Paul Thomas pa...@gcc.gnu.org
PR
--- Comment #3 from jvdelisle at gcc dot gnu dot org 2009-07-21 05:04
---
Works OK on i86-64-linux. I also ran the original test programs on Cygwin and
results are also OK all the way through.
I wonder if we can get another mingw attempt
Note: 4 Gb file here.
--
59 matches
Mail list logo
