Fishy build system: make_exports.pl called on darwin?

[Marcin Dalecki]

On 2005-10-03, at 00:42, Marcin Dalecki wrote:

> I'm not quite sure if this is correct since guessing the GCC  
> version interpretation here I would expect this to came out as:
>
> ppc-apple-darwin7.9.2
>
> aka: Tiger darwin 10.4.2

Trying to find out where the actual name mangling occured I have  
found the following inside config.sub:

# Here we handle the default operating systems that come with various  
machines.
# The value should be what the vendor currently ships out the door  
with their
# machine or put another way, the most popular os provided with the  
machine.
...

case $basic_machine in
...
     *-apple)
         os=-macos
         ;;
...
esac.

This seems to be inadequate to the current reality at least. Similar  
references found there like for example:
             -aux*)
                 vendor=apple
                 ;;
             -hms*)
                 vendor=hitachi
                 ;;
             -mpw* | -macos*)
                 vendor=apple
                 ;;

are at best tribute to history. However grepping a bit showed that  
it's config.guess, which is responsible for
the host name mangling. It contains the following tidbit: $ 
{UNAME_PROCESSOR}-apple-darwin${UNAME_RELEASE}.
This suggests that one has to look at the UNAME_RELEASE. Well the  
initial setting is derived from uname -r.
This shows on my system:

  uname -r
8.2.0

This would let us expect:

powerpc-aaple-darwin8.2.0

Thus still there is something strange going on later. The following  
doesn't mangle it either:

dzc14:~/BUILD/gcc-4.1-20051001 dalecki$ ./config.guess
powerpc-apple-darwin8.2.0
dzc14:~/BUILD/gcc-4.1-20051001 dalecki$ ./config.sub powerpc-apple- 
darwin8.2.0
powerpc-apple-darwin8.2.0

So where the hell does it happen? Thoes are the symptoms which  
justify my strong personal aversion against autoconf/automake...