21.5-b28 strange behaviour: buffer modification flag set, even though I can't remember having changed anything, patch for configure on aix

[Thomas Mittelstaedt]

Stephen J. Turnbull schrieb:
> Thomas Mittelstaedt writes:
>
>  > I also had to change configure to use -pthread instead -mthreads:
>
> I guess this is a version dependency of AIX.  Do you have any idea
> what the right thing to do is?
>
>   
Just built xemacs without -pthread (or -mthreads) and it runs. The gcc 
4.2.2 man page says about -mthreads that it's needed
for multi-threaded exception handling on mingw.

We use -pthread to link with the correct, multi-threaded version of the 
libstdc++ library, which on aix 5.2
are in separate directories:
tmstaedt at buildaix3$ find /opt/gcc-4.2.2/lib -name libstdc++*
/opt/gcc-4.2.2/lib/gcc/powerpc-ibm-aix5.2.0.0/4.2.2/pthread/ppc64/libstdc++.a
/opt/gcc-4.2.2/lib/gcc/powerpc-ibm-aix5.2.0.0/4.2.2/pthread/ppc64/libstdc++.la
/opt/gcc-4.2.2/lib/gcc/powerpc-ibm-aix5.2.0.0/4.2.2/pthread/libstdc++.a
/opt/gcc-4.2.2/lib/gcc/powerpc-ibm-aix5.2.0.0/4.2.2/pthread/libstdc++.la
/opt/gcc-4.2.2/lib/gcc/powerpc-ibm-aix5.2.0.0/4.2.2/ppc64/libstdc++.a
/opt/gcc-4.2.2/lib/gcc/powerpc-ibm-aix5.2.0.0/4.2.2/ppc64/libstdc++.la
/opt/gcc-4.2.2/lib/gcc/powerpc-ibm-aix5.2.0.0/4.2.2/libstdc++.a
/opt/gcc-4.2.2/lib/gcc/powerpc-ibm-aix5.2.0.0/4.2.2/libstdc++.la

tmstaedt at buildaix3$ gcc -v
Using built-in specs.
Target: powerpc-ibm-aix5.2.0.0
Configured with: ../gcc-4.2.2/configure 
--enable-version-specific-runtime-libs --enable-static --enable-shared 
--enable-threads --with-as=/usr/bin/as --without-gnu-ld 
--with-ld=/usr/bin/ld --prefix=/opt/gcc-4.2.2 --disable-nls --with-pic 
--disable-symvers --enable-symvers=no --enable-languages=c,c++,objc
Thread model: aix
gcc version 4.2.2




> 1.  If POSIX threads were never reported as buggy, maybe a feature
>     check for the pthreads library?
>
>   
> 2.  If there was an AIX version where pthreads became the "official"
>     way to go, do you know how to check for version on AIX, and what 
>     the version to check for would be?
>
> 3.  If "mthreads" is truly ancient we could deprecate it, but I'd
>     rather not.
>
> Regards,
> Steve
>