Bug#194345: [edi@gmx.de: Re: Bug#194345: FTB the lufs package]
severity 194345 important thanks this case is not related to original report. reverting the severity of the report to the old severity. The new bug has been forwarded to http://gcc.gnu.org/PR12765 the severity of this report is not RC, as there is a workaround using g++-3.2. Eduard Bloch writes: > Forgotten copy; the file is stored on > http://people.debian.org/~blade/misc/sshfs-breaks-with-g++-3.3.tgz > > - Forwarded message from Eduard Bloch <[EMAIL PROTECTED]> - > > Date: Wed, 22 Oct 2003 11:28:47 +0200 > From: Eduard Bloch <[EMAIL PROTECTED]> > Subject: Re: Bug#194345: FTB the lufs package > To: Matthias Klose <[EMAIL PROTECTED]> > > Moin Matthias! > Matthias Klose schrieb am Wednesday, den 22. October 2003: > > > Eduard Bloch writes: > > > severity 194345 grave > > > tag 194345 + sid > > > thanks > > > > > > It became worse. The last version cannot build my lufs package because > > > of the astronomical memory usage. Even 2GB of virtual memory isn't > > > enough. > > > > it would be nice if you could tell us the file in question, the > > command line options, the architecture, if reducing the optimization > > level lets the build succeed, or if g++-3.2 succeds to finish the > > build ... > > I tried different optimisation levels, Os, O1, O2, O3, it always failed. > The memory-usage-per-time curve seems to be parabolic, it eats a normal > amount of memory in the beginning and allocates larger and larger > amounts later. The builds suceeds with the Testing version of gcc-3.3 > without any problems (*cough* the package could be built few weeks ago).
Processed: Re: Bug#194345: [edi@gmx.de: Re: Bug#194345: FTB the lufs package]
Processing commands for [EMAIL PROTECTED]: > severity 194345 important Bug#194345: g++ 3.3 needs to much memory / gets killed by OOM-killer Bug#194513: Internal error: unbalanced parenthesis in operand 1 Severity set to `important'. > thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database)
[Bug c++/12765] [3.3.2 regression] 2GB mem not enough for sucessful compile
PLEASE REPLY TO [EMAIL PROTECTED] ONLY, *NOT* [EMAIL PROTECTED] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12765 ebotcazou at gcc dot gnu dot org changed: What|Removed |Added CC|jason at redhat dot com | AssignedTo|unassigned at gcc dot gnu |jason at redhat dot com |dot org | Status|UNCONFIRMED |ASSIGNED --- You are receiving this mail because: --- You reported the bug, or are watching the reporter.
Bug#217541: fficonfig.h bogus on x86
Package: libffi2-dev
Version: 1:3.3.2-1
Severity: important
Running this little test program on my Athlon (no, not an Athlon64 ;-)):
#include
#include
int main ()
{
printf ("&ffi_type_ulong: %p &ffi_type_uint64: %p\n",
&ffi_type_ulong, &ffi_type_uint64);
}
results in:
&ffi_type_ulong: 0x8049738 &ffi_type_uint64: 0x8049738
As you can easily see, the two pointers refer to the same object, which
means fficonfig.h #defines ffi_type_ulong to ffi_type_uint64, which is
plain wrong on a 32bit platform.
Regards, Andy
-- System Information:
Debian Release: testing/unstable
Architecture: i386
Kernel: Linux ivanova 2.6.0-test7-k7 #1 Mon Oct 13 14:07:14 CEST 2003 i686
Locale: LANG=C, LC_CTYPE=de_AT.UTF-8
Versions of packages libffi2-dev depends on:
ii gcc-3.3-base 1:3.3.2-1 The GNU Compiler Collection (base
ii libc6-dev 2.3.2-8GNU C Library: Development Librari
ii libffi2 1:3.3.2-1 Foreign Function Interface library
-- no debconf information
Bug#217541: fficonfig.h bogus on x86
Andreas Rottmann <[EMAIL PROTECTED]> writes: > As you can easily see, the two pointers refer to the same object, which > means fficonfig.h #defines ffi_type_ulong to ffi_type_uint64, which is > plain wrong on a 32bit platform. > s/fficonfig.h/ffi.h/ The way ffi.h #defines the ffi_type_XX equivalents to the standard c types seems fucked up - they have them along with the #defines of the fixed-bit types (e.g. UINT32), but the two issues are orthogonal. I have attached a patch that should fix this, and furthermore adds #defines for 'long long' and 'unsigned long long'. Hope something like this can go upstream. --- /usr/include/ffi.h 2003-10-17 13:13:15.0 +0200 +++ /home/andy/include/g-wrap/ffi.h 2003-10-25 18:41:41.0 +0200 @@ -79,76 +79,99 @@ #define SINT8 signed char +/* 16 bit types */ #if SIZEOF_INT == 2 #define UINT16 unsigned int #define SINT16 int -#define ffi_type_uint ffi_type_uint16 -#define ffi_type_sint ffi_type_sint16 -#else -#if SIZEOF_SHORT == 2 +#elif SIZEOF_SHORT == 2 #define UINT16 unsigned short #define SINT16 short -#define ffi_type_ushort ffi_type_uint16 -#define ffi_type_sshort ffi_type_sint16 #endif -#endif +/* 32bit types */ #if SIZEOF_INT == 4 #define UINT32 unsigned int #define SINT32 int -#define ffi_type_uint ffi_type_uint32 -#define ffi_type_sint ffi_type_sint32 -#else -#if SIZEOF_SHORT == 4 +#elif SIZEOF_SHORT == 4 #define UINT32 unsigned short #define SINT32 short -#define ffi_type_ushort ffi_type_uint32 -#define ffi_type_sshort ffi_type_sint32 -#else -#if SIZEOF_LONG == 4 +#elif SIZEOF_LONG == 4 #define UINT32 unsigned long #define SINT32 long -#define ffi_type_ulong ffi_type_uint32 -#define ffi_type_slong ffi_type_sint32 #endif -#endif -#endif + +/* 64 bit types */ #if SIZEOF_INT == 8 #define UINT64 unsigned int #define SINT64 int -#define ffi_type_uint ffi_type_uint64 -#define ffi_type_sint ffi_type_sint64 -#else -#if SIZEOF_LONG == 8 +#elif SIZEOF_LONG == 8 #define UINT64 unsigned long #define SINT64 long -#define ffi_type_ulong ffi_type_uint64 -#define ffi_type_slong ffi_type_sint64 -#else -#if SIZEOF_LONG_LONG == 8 +#elif SIZEOF_LONG_LONG == 8 #define UINT64 unsigned long long #define SINT64 long long -#define ffi_type_ulong ffi_type_uint64 -#define ffi_type_slong ffi_type_sint64 #endif + +#if SIZEOF_INT == 2 + +#define ffi_type_uint ffi_type_uint16 +#define ffi_type_sint ffi_type_sint16 + +#elif SIZEOF_INT == 4 + +#define ffi_type_uint ffi_type_uint32 +#define ffi_type_sint ffi_type_sint32 + +#elif SIZEOF_INT == 8 + +#define ffi_type_uint ffi_type_uint64 +#define ffi_type_sint ffi_type_sint64 + #endif + +#if SIZEOF_SHORT == 2 + +#define ffi_type_ushort ffi_type_uint16 +#define ffi_type_sshort ffi_type_sint16 + +#elif SIZEOF_SHORT == 4 + +#define ffi_type_ushort ffi_type_uint32 +#define ffi_type_sshort ffi_type_sint32 + +#endif /* guess there are no 64bit shorts anywhere? */ + +#if SIZEOF_LONG == 4 + +#define ffi_type_ulong ffi_type_uint32 +#define ffi_type_slong ffi_type_sint32 + +#elif SIZEOF_LONG == 8 + +#define ffi_type_ulong ffi_type_uint64 +#define ffi_type_slong ffi_type_sint64 + #endif +/* any machines with 128bit long longs yet? */ +#define ffi_type_ulong_long ffi_type_uint64 +#define ffi_type_slong_long ffi_type_sint64 + /* System specific configurations --- */ Cheers, Andi -- Andreas Rottmann | [EMAIL PROTECTED] | [EMAIL PROTECTED] | [EMAIL PROTECTED] http://www.8ung.at/rotty | GnuPG Key: http://www.8ung.at/rotty/gpg.asc Fingerprint | DFB4 4EB4 78A4 5EEE 6219 F228 F92F CFC5 01FD 5B62 Python is executable pseudocode, Perl is executable line-noise.
Processed: retitle, patch added
Processing commands for [EMAIL PROTECTED]: > retitle 217541 ffi.h's ffi_type_[us]long bogus on x86 Bug#217541: fficonfig.h bogus on x86 Changed Bug title. > tags 217541 + patch Bug#217541: ffi.h's ffi_type_[us]long bogus on x86 There were no tags set. Tags added: patch > thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database)
[Bug target/12717] [3.4 regression] [m68k-linux] bootstrap error in stage3
PLEASE REPLY TO [EMAIL PROTECTED] ONLY, *NOT* [EMAIL PROTECTED] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12717 pinskia at gcc dot gnu dot org changed: What|Removed |Added Keywords||build, ice-on-valid-code --- You are receiving this mail because: --- You reported the bug, or are watching the reporter.
