This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Updating top-level autoconf to 2.59
On Fri, 9 Feb 2007, Ian Lance Taylor wrote:
> Nothing. The gain comes when you realize that I'm building a lot of
> different toolchains, both native and cross, both built with a
> cross-compiler and built with a native compiler, and I just routinely
> specify all of --build, --host, and --target. With autoconf's rules,
> I have to do an odd little dance to remove --host and/or --target in
> some cases. That dance is completely doable but it is not, I believe,
> intuitive.
I do a number of different builds too, both native and cross in various
ways, but as all the bits are buried in scripts, I do not care that much
about what has to be passed to `configure' -- all the options that may or
may not have to appear I can store in variables or conditional blocks (as
long as autoconf gets the non-canonical names right, which it did not in
the past).
> If we add options to say "build with native compiler/cross-compiler"
> and "build a native toolchain/cross-compiler" then it becomes much
> simpler to reason about what you want to happen.
OK, I see -- I have given these options example names of
"--enable-cross-to-host" and "--enable-cross-to-target", presumably with
the default of "disabled" when --host or --target were not specified,
respectively. I do not object, but please do not expect me to advocate it
in a discussion with autoconf maintainers either. ;-)
Perhaps the discussion should be carried over to the autoconf list; there
was an "autoconf-conversion" mailing list set up a while ago as well -- I
am not sure as to whether it is still out there or not.
> And note that adding such options is completely backward compatible.
> They would just override the way that autoconf makes assumptions.
Certainly.
Maciej