This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Configuration/Makefile cleanups
In article <20001108182036H.mitchell@codesourcery.com> you write:
>>>>>> "Zack" == Zack Weinberg <zackw@Stanford.EDU> writes:
>
> Zack> I'm not going to make a huge fuss over this, though. I can
> Zack> easily change the patch to do as you suggest. But I would
> Zack> like to hear from the *BSD people who currently hack up the
> Zack> makefile so byacc is used, first.
>I am unsympathetic to users of byacc. Bison is very portable, and is
>the standard GNU parser-generator. That's what we should use.
This is a party line.
byacc is very portable and builds just fine.
It is not encumbered with the GPL license.
I believe you will find that BSD developers won't budge one bit.
The only case in which bison is needed is the pure java parser, and
that's just fine for us, because it's a technical issue.
If you want to justify bison, you will have to do much better.
This has too many shades of political GNU correctness to stand with us.
The only trouble with byacc was *a bug* in the C++ language spec
that occurred a few months back, that happened to accidentally get through
bison (which is too lenient), but was not correct according to POSIX rules.
Trivial to fix.
The only problem we do have with byacc is that it is completely dumb to
have to jump through hoops to get gcc to use it where it can instead of
bison.
Again, putting arguments forward with little technical value, and heavy
FSF political overtones does not stand well for us.
And, yes, I do know that you don't mean it that way, but unless you give
us a good technical reason, we absolutely, definitely, won't EVER move
from byacc to bison. (Note that it's not even not to use bison, since
we need bison for java anyway. It's just because we're PROUD of our tools
and we don't see any reason to replace them with GPL clones when they're
not needed...)