Hello All
On the MELT branch, I need sometimes that cc1 be run even if there is no
input files. This is an unusual mode, but sometimes needed (This
actually is needed to have MELT lisp files translated into C; for
reasons not explained here, this is a special mode of my ./cc1 which
does that; both MELT lisp input and generated C output are specified
thru special -f flags).
So I would like that my gcc command (and hence the installed gcc by the
make install target) runs
cc1 -fbasilys=translatefile \
-fbasilys-arg=ana-base.bysl -fbasilys-secondarg=ana-base.c
when gcc is invoked as
gcc -fbasilys=translatefile \
-fbasilys-arg=ana-base.bysl -fbasilys-secondarg=ana-base.c
I could definitely afford a special option, e.g. -fforce-cc1, to be
given to gcc, eg
gcc -fforce-cc1 -fbasilys=translatefile \
-fbasilys-arg=ana-base.bysl -fbasilys-secondarg=ana-base.c
The current MELT gcc/Makefile.in just runs cc1 in that case, but I would
prefer that the advanced MELT user avoid running cc1 directly and use
gcc instead.
What is the best way to achieve that? Where should I patch my gcc/gcc.c
file? Does some equivalent of -fforce-cc1 (as a flag to gcc) which force
the cc1 to be run already exist?
BTW, I don't understand yet any details regarding spec files.
Regards
PS. The MELT branch has an incomplete doc/melt.texi file.
--
Basile STARYNKEVITCH http://starynkevitch.net/Basile/
email: basile<at>starynkevitch<dot>net mobile: +33 6 8501 2359
8, rue de la Faiencerie, 92340 Bourg La Reine, France
*** opinions {are only mines, sont seulement les miennes} ***
