gcc version 4.2.0 20060304 (experimental)
[EMAIL PROTECTED]:/tmp% cat test.c
void *memcpy(void *dest, const void *src, __SIZE_TYPE__ n);
struct timeval {
long tv_sec;
};
struct outdata {
char seq;
struct timeval tv __attribute__((packed));
};
void send_probe(struct outdata
--- Comment #8 from ebotcazou at gcc dot gnu dot org 2006-03-05 09:39
---
Subject: Bug 25885
Author: ebotcazou
Date: Sun Mar 5 09:39:40 2006
New Revision: 111732
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=111732
Log:
Backport from mainline:
2006-02-13 Â Ed
--- Comment #9 from ebotcazou at gcc dot gnu dot org 2006-03-05 09:42
---
Fixed on the 4.1 branch.
--
ebotcazou at gcc dot gnu dot org changed:
What|Removed |Added
With -O2 or higher, the following little C loop does not terminate:
short i;
for (i=-1; i; i--) ;
The generated assembler code is:
.L2:
jmp .L2
which means that the exit condition has been completely eliminated.
The same loop works as expected (terminates after 0x
--- Comment #1 from falk at debian dot org 2006-03-05 10:06 ---
Signed integer overflow is undefined in C. Use -fwrapv to force twos-complement
semantics.
--
falk at debian dot org changed:
What|Removed |Added
--- Comment #7 from fxcoudert at gcc dot gnu dot org 2006-03-05 10:08
---
(In reply to comment #6)
Alright, I'll try that... rebuilding GMP with base/low level CFLAGS and see
how it goes...
You may want to configure GMP with --host=none --target=none --build=none to
disable
--- Comment #1 from laurent at guerby dot net 2006-03-05 11:07 ---
Fixe on x86 as of 4.2.0 20060304 (experimental)
--
laurent at guerby dot net changed:
What|Removed |Added
--- Comment #6 from laurent at guerby dot net 2006-03-05 11:10 ---
Still fails as of 4.2.0 20060304
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18817
--- Comment #10 from th dot r dot klein at web dot de 2006-03-05 13:38
---
Ping to remember
Fails on Version 3.2.3; 3.4.5; 4.0.2; 4.1.0; 4.2.0 (snapshot).
Earlier versions with additonal other buggs. So useless to test.
Bugfix still exists and working but ignorded at GCC side;
Why
--- Comment #6 from olaf dot dietsche at gfa-net dot de 2006-03-05 14:21
---
Sorry, if my example wasn't clear enough :-(. I'll try to clarify.
M::value_type is std::mapint, double::value_type, which is std::mapint,
double::pairint, double in my first example.
g++ doesn't recognize
--- Comment #1 from pinskia at gcc dot gnu dot org 2006-03-05 14:22 ---
Hmm:
outdata-tv
isn't that semi wrong as the pointer just lost the alignment information?
A workaround is to do:
memcpy((void*)outdata-tv, tp, sizeof outdata-tv);
--
--- Comment #2 from falk at debian dot org 2006-03-05 14:40 ---
(In reply to comment #1)
Hmm:
outdata-tv
isn't that semi wrong as the pointer just lost the alignment information?
I'm not sure what you mean.
A workaround is to do:
memcpy((void*)outdata-tv, tp, sizeof
--- Comment #7 from belyshev at depni dot sinp dot msu dot ru 2006-03-05
15:27 ---
seen also on amd64 with r111705
--
belyshev at depni dot sinp dot msu dot ru changed:
What|Removed |Added
--- Comment #7 from pinskia at gcc dot gnu dot org 2006-03-05 15:43 ---
Comeau C++ does the same thing as GCC does (at least in strict mode).
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26512
--- Comment #10 from pinskia at gcc dot gnu dot org 2006-03-05 15:51
---
(In reply to comment #7)
Bootstrap comparison failure!
./attribs.o differs
./c-aux-info.o differs
./c-common.o differs
Yes this is because HPUX's as puts in timestamps into the .s file
Please read:
--- Comment #1 from pinskia at gcc dot gnu dot org 2006-03-05 15:54 ---
One thing is that if libmudflap generates references to
std::ios_base::_S_local_word_size and they are not in libstdc++, then libstdc++
is broken as this would be invalid C++.
--
--- Comment #8 from olaf dot dietsche at gfa-net dot de 2006-03-05 16:22
---
Which doesn't make this inconsistent behaviour any better. I don't know Comeau
C++, is it some sort of reference implementation?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26512
--- Comment #9 from pinskia at gcc dot gnu dot org 2006-03-05 16:26 ---
(In reply to comment #8)
Which doesn't make this inconsistent behaviour any better. I don't know Comeau
C++, is it some sort of reference implementation?
It is just another implementation to compare against.
--
this line of code causes an ICE if compiled with -C option:
/**/#define foo
--
Summary: [4.2 regression] ICE in c_lex_with_flags, at c-lex.c:472
with -C
Product: gcc
Version: 4.2.0
Status: UNCONFIRMED
Keywords:
--- Comment #1 from pinskia at gcc dot gnu dot org 2006-03-05 16:42 ---
Confirmed, not a regression.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #12 from pinskia at gcc dot gnu dot org 2006-03-05 16:43
---
(In reply to comment #6)
The previous was because i'm an idiot.
That ICE is now reported as PR 26567.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13726
--- Comment #2 from pinskia at gcc dot gnu dot org 2006-03-05 16:44 ---
This is minor because you have to directly invoke cc1 and you cannot get this
ICE when using the driver.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #11 from h dot m dot brand at xs4all dot nl 2006-03-05 16:46
---
To Steven: I'm not pushing, but IMHO, I think it could be said that it is a bug
in gcc if gcc does not generate valid HP PA-RISC assembly code *as understood
by the native compiler*.
To Andrew: For the last
--- Comment #2 from pinskia at gcc dot gnu dot org 2006-03-05 17:28 ---
Here is a new testcase so that we don't thread the jump as we currently do:
int g(int i, int j)
{
int t = 0;
int i1;
if (i == j)
t = 3;
for(i1=0;i11;i1++) h();
if (t != 5)
return 0;
else
--- Comment #7 from reichelt at gcc dot gnu dot org 2006-03-05 17:43
---
Mine.
--
reichelt at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #4 from pinskia at gcc dot gnu dot org 2006-03-05 17:43 ---
In 4.2.0 and above we get:
baz:
.LFB2:
mulsd %xmm1, %xmm0
mulsd %xmm0, %xmm0
mulsd %xmm0, %xmm0
ret
Which is what I recommend but we don't get that on the tree level:
return
--- Comment #6 from jvdelisle at gcc dot gnu dot org 2006-03-05 17:54
---
Subject: Bug 26554
Author: jvdelisle
Date: Sun Mar 5 17:54:07 2006
New Revision: 111738
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=111738
Log:
2006-03-05 Jerry DeLisle [EMAIL PROTECTED]
PR
--- Comment #2 from hjl at lucon dot org 2006-03-05 18:00 ---
std::ios_base::_S_local_word_size is in libstdc++.so. But it is marked as
local via linker version script. It is true for x86, x86-64 and ia64.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26532
--- Comment #7 from jvdelisle at gcc dot gnu dot org 2006-03-05 18:00
---
Subject: Bug 26554
Author: jvdelisle
Date: Sun Mar 5 18:00:22 2006
New Revision: 111739
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=111739
Log:
2006-03-05 Jerry DeLisle [EMAIL PROTECTED]
PR
--- Comment #6 from kargl at gcc dot gnu dot org 2006-03-05 18:01 ---
NAG's compiler jumps to 999.
kargl[203] f95 -o z -O l.f90
kargl[205] ./z
read error occurred
This is most likely processor dependent behavior.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26509
On Mar 5, 2006, at 1:00 PM, hjl at lucon dot org wrote:
--- Comment #2 from hjl at lucon dot org 2006-03-05 18:00 ---
std::ios_base::_S_local_word_size is in libstdc++.so. But it is marked
as
local via linker version script. It is true for x86, x86-64 and ia64.
There is a bug in
--- Comment #3 from pinskia at gcc dot gnu dot org 2006-03-05 18:03 ---
Subject: Re: [4.1]: libmudflap failures on ia64
On Mar 5, 2006, at 1:00 PM, hjl at lucon dot org wrote:
--- Comment #2 from hjl at lucon dot org 2006-03-05 18:00 ---
--- Comment #4 from pinskia at gcc dot gnu dot org 2006-03-05 18:09 ---
Reduced testcase:
#include iostream
struct a : std::ios_base
{
static const int*f;
};
const int *a::f = std::ios_base::_S_local_word_size;
int main(void)
{
return *a::f;
}
--
pinskia at gcc dot gnu dot org
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Ever Confirmed|0 |1
Last
--- Comment #6 from tkoenig at gcc dot gnu dot org 2006-03-05 18:35 ---
Fixed on 4.1 as well.
Closing.
--
tkoenig at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #6 from tkoenig at gcc dot gnu dot org 2006-03-05 18:36 ---
Fixed on 4.1 as well.
Closing.
--
tkoenig at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #5 from pcarlini at suse dot de 2006-03-05 19:07 ---
I would like to know why the original problem doesn't exist in 4.2. Because, in
general, we don't want to export _S_local_word_size, only an implementation
detail. In other terms, strictly speaking, the reduced testcase is
--
pcarlini at suse dot de changed:
What|Removed |Added
Status|NEW |WAITING
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26532
--- Comment #4 from eedelman at gcc dot gnu dot org 2006-03-05 19:24
---
Subject: Bug 16136
Author: eedelman
Date: Sun Mar 5 19:24:48 2006
New Revision: 111741
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=111741
Log:
fortran/
2005-03-05 Erik Edelmann [EMAIL PROTECTED]
--- Comment #5 from tkoenig at gcc dot gnu dot org 2006-03-05 20:03 ---
Subject: Bug 23092
Author: tkoenig
Date: Sun Mar 5 20:03:08 2006
New Revision: 111742
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=111742
Log:
2006-03-05 Thomas Koenig [EMAIL PROTECTED]
PR
--- Comment #6 from tkoenig at gcc dot gnu dot org 2006-03-05 20:03 ---
Fixed on 4.1.
Closing.
--
tkoenig at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #6 from pcarlini at suse dot de 2006-03-05 20:14 ---
Created an attachment (id=10973)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=10973action=view)
Removal of _S_local_word_size
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26532
--- Comment #7 from pcarlini at suse dot de 2006-03-05 20:18 ---
In yet other terms: I could easily apply the attached, without affecting the
functionality of the library and binary compatibility. No, we don't want to
export that symbol and the real reason of the link failure on those
--- Comment #15 from pinskia at gcc dot gnu dot org 2006-03-05 21:27
---
Oh, the one thing we would miss with unrolling early is some vectorization.
That could be fixed with a basic block vectorizer.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18754
--- Comment #7 from pinskia at gcc dot gnu dot org 2006-03-05 21:30 ---
(In reply to comment #2)
This is not a bug. You need IPA to figure out that y is never changed. For
example:
int x, y = x, z;
int bar() { y = z; }
This actually does:
*y = z; and not y = z;
--
--- Comment #9 from ebotcazou at gcc dot gnu dot org 2006-03-05 21:35
---
Still there on 4.2
Right, let's get rid of it.
--
ebotcazou at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #2 from ebotcazou at gcc dot gnu dot org 2006-03-05 21:39
---
VRP is out of the loop, they are all present at -O0.
--
ebotcazou at gcc dot gnu dot org changed:
What|Removed |Added
This is reported:
recursor_cache.hh:58: warning: comparison between signed and unsigned integer
expressions
syncres.cc: In function â(static destructors for syncres.cc)â:
syncres.cc:649: note: file syncres.gcda not found, execution counts estimated
*** glibc detected *** corrupted double-linked
--- Comment #3 from ebotcazou at gcc dot gnu dot org 2006-03-05 23:30
---
c34004a, c46033a and cxg2024 have been introduced/exposed by
2006-02-24 Roger Sayle [EMAIL PROTECTED]
* fold-const.c (fold_comparison): New subroutine of fold_binary
containing transformations
--- Comment #1 from nmiell at comcast dot net 2006-03-05 23:31 ---
On FC4 AMD64, compiling:
#include algorithm
void rotate(int x[3])
{
std::rotate(x[0], x[1], x[2]);
}
int main(int argc, char* argv[])
{
int x[3] = { 0, 1, 2 };
rotate(x);
return 0;
}
--- Comment #2 from pinskia at gcc dot gnu dot org 2006-03-05 23:39 ---
(In reply to comment #1)
On FC4 AMD64, compiling:
This is a different issue for sure as the ICE is in
rtl_value_profile_transformations which does not exist in 4.1.0.
--
--- Comment #3 from ahu at ds9a dot nl 2006-03-05 23:42 ---
Thanks Nicholas,
I can run the command you suggest without errors (I saved the file as
exploit.cc). However, if I compile first with -fprofile-generate, then run it
once, and then add -fprofile-use, it reports:
$ g++ -O2
--- Comment #4 from pinskia at gcc dot gnu dot org 2006-03-05 23:53 ---
Ok, compiling comment #1 with -O2 -fprofile-generate -fprofile-use, I do get
the failures.
Here is the valgrid report:
==5355== Invalid free() / delete / delete[]
==5355==at 0x11B1DA33: free
--- Comment #5 from pinskia at gcc dot gnu dot org 2006-03-05 23:56 ---
It does not ICE on the mainline but valgrind is still unhappy:
==5383== Invalid free() / delete / delete[]
==5383==at 0x11B1DA33: free (vg_replace_malloc.c:235)
==5383==by 0x80680D:
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
CC||hubicka at gcc dot gnu dot
|
--- Comment #6 from pinskia at gcc dot gnu dot org 2006-03-06 00:04 ---
Confirmed.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #7 from ahu at ds9a dot nl 2006-03-06 00:10 ---
The crasher bugs go away from release 4.1.0 when never running with
-fprofile-use and -fprofile-generate simultaneously.
There is another 'checksum mismatch' issue with PowerDNS:
base64.cc: In function
--- Comment #8 from reichelt at gcc dot gnu dot org 2006-03-06 02:01
---
Subject: Bug 15759
Author: reichelt
Date: Mon Mar 6 02:01:29 2006
New Revision: 111754
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=111754
Log:
PR c++/15759
* tree.c (bot_manip): Don't
--- Comment #9 from reichelt at gcc dot gnu dot org 2006-03-06 02:07
---
Fixed on mainline.
--
reichelt at gcc dot gnu dot org changed:
What|Removed |Added
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
Target Milestone|--- |3.4.6
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15155
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
Target Milestone|--- |3.4.6
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18592
--- Comment #5 from pinskia at gcc dot gnu dot org 2006-03-06 02:19 ---
Since this was not a bug in GCC and Steve could not reproduce this, closing as
works for me.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #5 from pinskia at gcc dot gnu dot org 2006-03-06 02:20 ---
Fixed in 4.0.0.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #2 from pinskia at gcc dot gnu dot org 2006-03-06 02:20 ---
Fixed in 4.0.0.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
Target Milestone|--- |3.4.6
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25133
--- Comment #6 from pinskia at gcc dot gnu dot org 2006-03-06 02:21 ---
Fixed so closing.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #1 from pinskia at gcc dot gnu dot org 2006-03-06 02:26 ---
Confirmed, this did not ICE in 2.95.3.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #2 from pinskia at gcc dot gnu dot org 2006-03-06 02:29 ---
This works for me and many other people.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
CC||pinskia at gcc dot gnu dot
|
--- Comment #1 from pinskia at gcc dot gnu dot org 2006-03-06 02:32 ---
Fixed in 4.0.0.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
CC||pinskia at gcc dot gnu dot
|
The invalid testcase
===
struct A {};
unsigned A a;
===
triggers the following bogus error message since GCC 3.4.0:
bug.cc:2: error: 'A' does not name a type
The error message before wasn't very helpful
bug.cc:2: error: syntax error
The following invalid code snippet is accepted since at least GCC 2.95.3:
==
templateint void foo()
{
struct A;
struct B : A {};
}
template void foo0();
==
As A is incomplete, the definition of the local class B should trigger
The following invalid code snippet
==
void foo()
{
struct A { static int i; };
}
==
triggers a duplicate error message since GCC 3.1:
bug.cc: In function 'void foo()':
bug.cc:3: error: local class 'struct foo()::A' shall not
--- Comment #8 from diskman at kc dot rr dot com 2006-03-06 03:56 ---
Well I went ahead and reinstalled GMP-4.1.4:
CC=gcc-3.2.3 CXX=g++-3.2.3 CPP=cpp-3.2.3 \
CFLAGS=-O2 -mieee -mtune=ev56 -w -pipe \
CXXFLAGS=-O2 -mieee -mtune=ev56 -w -pipe \
./configure \
--build=alpha-alpha-linux
On Solaris 9 (x86), the error occurs in the libjava.
I fixed line 11597 of gcc-4.1.0/libjava/classpath/configure
from
QT_INCLUDE_DIR=$($PKG_CONFIG --variable=includedir QtGui)
to
QT_INCLUDE_DIR=$(`$PKG_CONFIG --variable=includedir QtGui`)
Is this a right fix
--- Comment #1 from pinskia at gcc dot gnu dot org 2006-03-06 05:00 ---
Did you set CONFIG_SHELL to be /bin/ksh as directed on the installation
directions?
http://gcc.gnu.org/install/specific.html#x-x-solaris2
The Solaris 2 /bin/sh will often fail to configure libstdc++-v3, boehm-gc or
On Solaris 9 (x86), the error occurs in the libjava.
I fixed line 11597 of gcc-4.1.0/libjava/classpath/configure
from
QT_INCLUDE_DIR=$($PKG_CONFIG --variable=includedir QtGui)
to
QT_INCLUDE_DIR=$(`$PKG_CONFIG --variable=includedir QtGui`)
Is this a right fix
--- Comment #1 from pinskia at gcc dot gnu dot org 2006-03-06 05:06 ---
*** This bug has been marked as a duplicate of 26574 ***
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #2 from pinskia at gcc dot gnu dot org 2006-03-06 05:06 ---
*** Bug 26575 has been marked as a duplicate of this bug. ***
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26574
On Solaris 9 (x86), the error occurs in the libjava.
I fixed line 11597 of gcc-4.1.0/libjava/classpath/configure
from
QT_INCLUDE_DIR=$($PKG_CONFIG --variable=includedir QtGui)
to
QT_INCLUDE_DIR=$(`$PKG_CONFIG --variable=includedir QtGui`)
Is this a right fix
--- Comment #1 from shanwill44 at yahoo dot com 2006-03-06 05:16 ---
Reported twice. Sorry.
--
shanwill44 at yahoo dot com changed:
What|Removed |Added
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
Status|UNCONFIRMED |WAITING
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26574
--- Comment #6 from jvdelisle at gcc dot gnu dot org 2006-03-06 06:23
---
I have found the fix for this. It is simple. Do we want to allow the missing
and then give an error on -pedantic?
Intel allows the missing and continues at the next non blank character.
How strict shall we
Submitter-Id: net
Originator:Clifford Heath
Organization: ManageSoft
Confidential: no
Synopsis: g++ ICE in cp_expr_size expanding template with inline
assembly
Severity: critical
Priority: medium
Category: g++
Class: ice-on-legal-code
Release: 4.0.3
--- Comment #9 from diskman at kc dot rr dot com 2006-03-06 06:51 ---
Did some hunting around on the net, it seems the 'hanging' problem is happen to
others as well. Here's a Solaris user with the same issue:
http://gcc.gnu.org/ml/fortran/2005-05/msg00456.html
--
--- Comment #10 from fxcoudert at gcc dot gnu dot org 2006-03-06 07:35
---
Could you try to compile a test program under gdb to see where it freezes?
$ cat a.f90
integer :: a = 45
real(8) :: x = 1.78d2
print *, hello, x / a
end
$ gdb -args
--- Comment #3 from page dot christian at gmail dot com 2006-03-06 07:49
---
It was because of bad kernel version. Works fine now...
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24237
88 matches
Mail list logo
