Re: $CC and intel icc/ecc compiler

[Samuel Meder]

On Wed, 2003-04-09 at 20:52, Bob Friesenhahn wrote:
> It seems to me that both the before and after code is incorrect.  The
> reason why I say is is that $CC may contain the command plus some
> arguments which are required for it to behave in some consistent way.

So it is naive to think that those things should go into CFLAGS? Off the
top of my head I can't think of a single scenario where I'd want them in
$CC rather than $CFLAGS. Do you have one?

> This means that the if $CC is 'foo -bar' then
> 
> case $CC in
>   foo)
> 
> will not match, but
> 
> case $CC in
>   foo*)
> 
> will.  Parsing out just the first word from the specification would
> solve the problem.

I still maintain that you need `basename $CC` (basename does not strip
arguments on the systems I tried). I don't really care to argue about
the CC vs. CFLAGS issue (still curious about a example though) so I've
attached a patch that adds *s.

/Sam

> You should not expect that the user won't add compiler options to the
> base compiler name since this may be required to select a compiler
> version, target architecture, or some other global option which is
> best specified via the compiler specification.
>
> Bob
> 
> On 9 Apr 2003, Samuel Meder wrote:
> 
> > On Wed, 2003-04-09 at 20:35, Bob Friesenhahn wrote:
> > > On Wed, 9 Apr 2003, Charles Wilson wrote:
> > >
> > > > Only half paying attention, but doesn't this break an earlier patch that
> > > > allowed things like "-mno-cygwin" to be included in the $CC variable?
> > > > Or am I mis-remembering?
> > >
> > > Yes, it sounds like it does.  It also would break important things
> > > like
> > >
> > >   CC="gcc -V 3.1.1"
> > >
> > > which I happen to be using at the moment.
> >
> > Please read the patch. It changes
> >
> >     linux*)
> >       case $CC in
> >       icc|ecc)
> >     _LT_AC_TAGVAR(lt_prog_compiler_wl, $1)='-Wl,'
> >     _LT_AC_TAGVAR(lt_prog_compiler_pic, $1)='-KPIC'
> >     _LT_AC_TAGVAR(lt_prog_compiler_static, $1)='-static'
> >         ;;
> >       ccc)
> >         _LT_AC_TAGVAR(lt_prog_compiler_wl, $1)='-Wl,'
> >         # All Alpha code is PIC.
> >         _LT_AC_TAGVAR(lt_prog_compiler_static, $1)='-non_shared'
> >         ;;
> >       esac
> >       ;;
> > to
> >
> >     linux*)
> >       case `basename $CC` in
> >       icc|ecc)
> >     _LT_AC_TAGVAR(lt_prog_compiler_wl, $1)='-Wl,'
> >     _LT_AC_TAGVAR(lt_prog_compiler_pic, $1)='-KPIC'
> >     _LT_AC_TAGVAR(lt_prog_compiler_static, $1)='-static'
> >         ;;
> >       ccc)
> >         _LT_AC_TAGVAR(lt_prog_compiler_wl, $1)='-Wl,'
> >         # All Alpha code is PIC.
> >         _LT_AC_TAGVAR(lt_prog_compiler_static, $1)='-non_shared'
> >         ;;
> >       esac
> >       ;;
> >
> > If you are using gcc you will never hit this case statement. Also, a
> > quick grep -r on mno-cygwin gives no hits other than ChangeLog.1 and
> > mail/cygwin32. My understanding is that the option stripping that
> > libtool does has changed a little so it may not longer need special
> > processing. It should be orthogonal in any case.
> >
> > /Sam
> >
> > >
> > > Bob
> > > ======================================
> > > Bob Friesenhahn
> > > address@hidden
> > > http://www.simplesystems.org/users/bfriesen
> > >
> >
> 
> ======================================
> Bob Friesenhahn
> address@hidden
> http://www.simplesystems.org/users/bfriesen
>

ChangeLog.patch
Description: Text Data

libtool.patch
Description: Text Data