This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: New C (and now ObjC) parser, take 4
Joseph S. Myers wrote:
On Fri, 5 Nov 2004, Paolo Bonzini wrote:
But maybe you (or rather the people like me and Geoff that want the parser on
CVS) can ask the RM or SC about putting it on mainline as a technology
preview, optionally enabled at configure time like --enable-mapped-location.
Since it affects much more existing code than the new C++ parser, and given
how you are developing it, I guess this may make sense. It would also make it
easier for the distro makers to test it on Fedora packages, for example.
This seems reasonable to me if people desire it and with the clear
understanding that any changes to parser behavior would need to be made to
both parsers at once while both are on mainline and that the function of
having the new parser there is to facilitate getting it equivalent to and
ready to replace the old parser, not to make changes to parser actions or
optimizations such as lexing up front which aren't necessary for that
purpose until the old parser is gone. It would inevitably be the case
that some testcases would fail with the new parser because of error
recovery differences beyond those that can be addressed by wider patterns
to match diagnostics.
I'm not in favor of having two parsers active at the same time.
Instead, I think we should wait until 4.0 has branched, and then check
in the new parser for GCC 4.1. Joseph is being very thorough with his
testing -- much more thorough than I was for C++. In my opinion,
that's overkill, but that's not the point: the point is that I think
that we can be pretty sure that by the time 4.0 has branched the new C
parser will be good enough for a Stage 1 contribution GCC 4.1.
--
Mark Mitchell
CodeSourcery, LLC
(916) 791-8304
mark@codesourcery.com