re: can't handle 1 Gb

re: can't handle 1 Gb

[Jos van den Oever]

sorry, that was a very silly error
>From lennert@nki.nl Thu Aug 03 04:07:00 2000
From: "Lennert Ploeger" <lennert@nki.nl>
To: <gcc-bugs@gcc.gnu.org>
Cc: "Lennert Ploeger" <lennert@nki.nl>
Subject: Bug in Mingw32 C++ compiler (platform: Windows NT 4) related to calling conventions
Date: Thu, 03 Aug 2000 04:07:00 -0000
Message-id: <NBBBLIOFCEDIMJGBJJCKGEDKCFAA.lennert@nki.nl>
X-SW-Source: 2000-08/msg00067.html
Content-type: multipart/mixed; boundary="----------=_1583533112-4114-558"

This is a multi-part message in MIME format...

------------=_1583533112-4114-558
Content-length: 2453

GCC version: gcc version 2.95.2 19991024 (release)
system: Window NT 4 (SP6)
Command line: gcc -v --save-temps CallConvBug.cpp -o CallConvBug
Command output:
Reading specs from
E:\lang\mingw295\bin\..\lib\gcc-lib\i386-mingw32\2.95.2\specs

gcc version 2.95.2 19991024 (release)

E:\lang\mingw295\bin\..\lib\gcc-lib\i386-mingw32\2.95.2\cpp.exe -lang-c++ -v
 -i
prefix
E:\lang\mingw295\bin\..\lib/gcc-lib/i386-mingw32\2.95.2\ -D__GNUC__=2 -D_
_GNUG__=2 -D__GNUC_MINOR__=95 -D__cplusplus -Di386 -D_WIN32 -DWIN32 -D__WIN3
2__
-D__MINGW32__=0.2 -DWINNT -D_X86_=1 -D__STDC__=1 -D__stdcall=__attribute__((
__st
dcall__)) -D_stdcall=__attribute__((__stdcall__)) -D__cdecl=__attribute__((_
_cde
cl__)) -D__declspec(x)=__attribute__((x)) -D__i386__ -D_WIN32 -D__WIN32__ -D
__WI
N32__ -D__MINGW32__=0.2 -D__WINNT__ -D_X86_=1 -D__STDC__=1 -D__stdcall=__att
ribu
te__((__stdcall__)) -D___stdcall__=__attribute__((__stdcall__)) -D__cdecl=__
attr
ibute__((__cdecl__)) -D__declspec(x)=__attribute__((x)) -D__i386 -D__WIN32 -
D__W
INNT -D___stdcall=__attribute__((__stdcall__)) -Asystem(winnt) -Acpu(i386) -
Amac
hine(i386) -D__EXCEPTIONS -remap -Acpu(i386) -Amachine(i386) -Di386 -D__i386
 -D_
_i386__ CallConvBug.cpp CallConvBug.ii
GNU CPP version 2.95.2 19991024 (release) (80386, BSD syntax)
#include "..." search starts here:
#include <...> search starts here:

E:\lang\mingw295\bin\..\lib\gcc-lib\i386-mingw32\2.95.2\..\..\..\..\include\
g++
-3
 E:\lang\mingw295\bin\..\lib\gcc-lib\i386-mingw32\2.95.2\..\..\..\..\include

E:\lang\mingw295\bin\..\lib\gcc-lib\i386-mingw32\2.95.2\..\..\..\..\i386-min
gw3
2\include
 E:\lang\mingw295\bin\..\lib\gcc-lib\i386-mingw32\2.95.2\include
End of search list.
The following default directories have been omitted from the search path:
 /usr/local/i386-mingw32/include
End of omitted list.
 E:\lang\mingw295\bin\..\lib\gcc-lib\i386-mingw32\2.95.2\cc1plus.exe
CallConvBug
.ii -quiet -dumpbase CallConvBug.cc -version -o CallConvBug.s
GNU C++ version 2.95.2 19991024 (release) (i386-mingw32) compiled by GNU C
versi
on 2.95.2 19991024 (release).
CallConvBug.cpp: In function `int main()':
CallConvBug.cpp:20: no matching function for call to `avsquirt::avsquirt
(void (
&)())'
CallConvBug.cpp:10: candidates are: avsquirt::avsquirt(void (*)())

The problem is that the compiler mixes up the calling conventions for
function types with the same prototype. Attached is the zipped preprocessor
output, as requested.

Bye,
	Lennert.

------------=_1583533112-4114-558
Content-Type: application/zip; charset=binary; name="CallConvBug.zip"
Content-Disposition: inline; filename="CallConvBug.zip"
Content-Transfer-Encoding: base64
Content-Length: 2009
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------------=_1583533112-4114-558--
>From jb@jbacher.com Thu Aug 03 05:21:00 2000
From: J Bacher <jb@jbacher.com>
To: Alexandre Oliva <aoliva@redhat.com>
Cc: gcc-bugs@gcc.gnu.org
Subject: Re: Internal Compiler Error
Date: Thu, 03 Aug 2000 05:21:00 -0000
Message-id: <Pine.OSF.4.21.0008030715020.24942-100000@ns.shawneelink.net>
References: <orhf93w9lk.fsf@guarana.lsd.ic.unicamp.br>
X-SW-Source: 2000-08/msg00068.html
Content-length: 836

> On Aug  1, 2000, Jan Bacher <jb@jbacher.com> wrote:
> 
> > Current CC:		Dec Alpha, Tru64 5.0
> > GCC Version:		2.95.2
> 
> GCC 2.95.* doesn't support Tru64 5.0.  However, the failure mode is
> not the one you got.  Did you make any changes to the source tree?
> Which compiler did you use for stage1?  Did you try `bootstrap-lean'
> in a clean build tree?

Changes:  in ../gcc/tm.h I added an include for "alpha/osf.h"

Compiler:  stage1, Dec's CC

I tried:  make, make bootstrap, make bootstrap-lean

What I need to do is pass the cflag "-arch ev64" to the Dec compiler
during stage1 but not have it passed to the gcc compiler in successive
stages.  I have found that passing that flag results in no internal
compiler message except that it fails when the same cflags are passed to
gcc.

Apologies if my bug-complaint is premature.