This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Configuration/Makefile cleanups
- To: Marc Espie <espie at quatramaran dot ens dot fr>
- Subject: Re: Configuration/Makefile cleanups
- From: Gerald Pfeifer <pfeifer at dbai dot tuwien dot ac dot at>
- Date: Fri, 24 Nov 2000 15:57:35 +0100 (CET)
- cc: Mark Mitchell <mark at codesourcery dot com>, <zackw at stanford dot edu>, <gcc-patches at gcc dot gnu dot org>
On Fri, 24 Nov 2000, Marc Espie wrote:
> [byacc vs bison]
> It is not encumbered with the GPL license.
Please, let's keep politics from this list. My primary goal (at least)
is helping to develop a free compiler that is as good as possible for as
broad a user base as possible, and I don't think anyone here will
disagree.
> In fact, you heard about it a few months back, when a change in C++ turned
> out to be cutting corners and playing too chummy with bison (missed ;
> in a rule... bison supplies it, byacc doesn't), so this is a good thing.
I agree. And it's good that we/you fixed this bug.
Whereever there are standards, we should try to adhere to them (even if
bison, or some other GNU tool for that matter is more error-tolerant),
unless there is a strong reason not to do so (as is the case with Java
parser you mentioned). Again, I don't think anyone here will disagree.
> More generally, there is something fundamentally bogus with the way
> those autoconf rules work: they do insist on asking for bison first,
> and getting tested for bison, which means that at autoconf level,
> there is no way to distinguish between `I want a parser generator' and
> `I want a pure parser generator'. If you want to use byacc, you have
> to check manually that what you've got is working with byacc (no pure
> parser). There ought to be two separate macros, and they ought to be
> used sensibly.
You seem to make a valid point. The following questions are non-rhethoric,
I'm really interested!
The two cases
o only bison is present
o only yacc (byacc) is present and Java should not be built
are trivial, but what should be the behavior, that is, which "YACC"
should be used for which parser or how should configure behave, in
your opinion, if
o both yacc (byacc) and bison are present and Java should not be built?
o both yacc (byacc) and bison are present and Java should be built?
o only yacc (byacc) is present and Java should be built?
And, how does the current behavior differ?
Gerald
--
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/