Re: NO! autoconf cross-tools detection is really useful!

[Balint Joo]

These are compilers specially  modified/adapted to be cross compiled
for the Cray XT3. There also similar ones for BlueGene/L although they
sort of adhere to the host triplet (well -quadruplet anyway - 
bgl-blrts-gnu-llnux-gcc )
They are not necessarily standard or necessarily GNU blessed GCCs.
Think also of people who use weird and wonderful embedded systems,
who may or may not adhere to the host triplet conventions.

Yes, I could make life simple by making syminks in my own 'bin'-dir.
However, that would be no use to anyone who wants to use autoconf
and automake on other systems where my directory is not available.

But as I said, this is getting away from the default host triplet thing, 
since
I explicitly specify CC, CXX etc. It's just I'd like a surefire way to 
trip cross
compile mode. Currently this  is done very effectively by
using a different --host and --build value. (Even if 
--build=none-none-none is what
I use)

Incidentally, the  whole point of our switch to learn autoconf and 
autmake (which
is not always easy especially the blasted aclocal version checking
macro) was to make the code compile easily anywhere. It is a credit
to autoconf and automake that when a new machine comes along (Cray,
BG/L, QCDOC or other) we can often compile it appropriately within
the first afternoon. Our user experience varies tho - some people love it,
--  some won't touch are code because of autoconf and automake.
Some would have loved to contribute but nearly ran a mile when we 
suggested they use
autoconf/autmake and --enable to make a feature optional, or pushed
that burden back to us.

Since more and more we are cross compiling (even on machines which have
the same front and back end architecturally, but the back end is behind 
a queueing
system so we can't check for libraries there etc) we never really test 
'system parameters' but give choices
with --enable and --with.  Autoconf and automake give us a flexibility
that justifies the sometimes very real pain to use them. It would be a 
shame if we needed
to look for another build system since there are not really any other very
good, free and freely available, open source de-facto standards.

   Ciao,
      B




Ralf Corsepius wrote:
> On Wed, 2007-04-18 at 12:02 -0400, Balint Joo wrote:
>   
> > Hi,
> >     This is a slightly different issue, but here is a feature to please 
> > keep,
> > or give a new way of doing if it will be changed.
> >
> >     I often have situations where I have to give some dummy host triplet
> > and then explicitly specify the compiler eg for Cray XT3 builds:
> >
> > CC="qk-gcc" CXX="qk-g++" --host=x86_64-none-none --build=x86_64-suse-linux
> >
> > Here I want cross compilation and the compilers have nothing to do with the
> > host triplet etc.
> >     
> Why? Where do these gcc/g++'s come from.
>
> The GNU toolchain by defaults installs cross tools with the host prefix,
> so the person having built these "qk-"prefixed toolchains must have done
> something nasty or hasn't understood how prefixing the tools works.
>
> Anyway, if you want to make your like simpler you an simply install
> symlinks using the correct prefix.
>
> Ralf
>
>
>
>
> _______________________________________________
> Autoconf mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/autoconf
>