Re: toplevel makefile question

[Stan Shebs <shebs at apple dot com>]

Nathanael Nerode wrote:

> I'd like to eliminate the differences between these two makefile targets in
> the top level:
>
> .PHONY: install-gcc
> install-gcc:
> 	@if [ -f ./gcc/Makefile ] ; then \
> 	  r=`${PWD}`; export r; \
> 	  s=`cd $(srcdir); ${PWD}`; export s; \
> 	  $(SET_LIB_PATH) \
> 	  (cd gcc; $(MAKE) $(GCC_FLAGS_TO_PASS) install); \
> 	else \
> 	  true; \
> 	fi
>
> .PHONY: install-gcc-cross
> install-gcc-cross:
> 	@if [ -f ./gcc/Makefile ] ; then \
> 	  r=`${PWD}`; export r; \
> 	  s=`cd $(srcdir); ${PWD}`; export s; \
> 	  $(SET_LIB_PATH) \
> 	  (cd gcc; $(MAKE) $(GCC_FLAGS_TO_PASS) LANGUAGES="c c++" install); \
> 	else \
> 	  true; \
> 	fi
>
> Note that the only difference is the extra restriction of LANGUAGES for
> a cross install.  This seems frankly bogus to me.  What is the purpose of
> this?  If it is truly necessary, would it suffice to include a
> @maybe_restrict_languages@ bit which configure set to restrict languages
> if host != target ?
I remember that at one time ObjC didn't work for crosses, and that might 
have
been the original reason for this bit.  Don't think it's still an issue 
though,
and these days I think we'd want to fix the compiler rather than try to work
around it in top level configury.

Stan