Tom Lane wrote:
> Kevin Brown <kevin@sysexperts.com> writes:
> >> I think it's best to leave well enough alone. The tarball ships with
> >> working bison output files anyway, so all of this really only matters
> >> to people trying to build 7.2.* from a CVS pull.
>
> > Okay, fair enough, but if we intend to continue to maintain 7.2.*,
> > shouldn't we at least fix the .y files?
>
> What for? If you've bought the assumption that we aren't going to port
> 7.2's grammar forward indefinitely, why shouldn't we say it stops with
> bison 1.35 rather than 1.75? 1.75 and later are not widely deployed
> anyway, as yet.
Okay, fair enough. I'll take your word for this (no idea how to
determine how widely bison 1.75 and later are deployed).
> But the more compelling point is *it does not matter*
> to our customers, and only barely to us. No one but a very few
> developers will ever again build the .y output files for 7.2.* (even
> assuming that there are more 7.2.* releases, which I doubt). Doesn't
> seem worth expending any effort on, to me. We have other, more
> productive things to do ...
Agreed. I wasn't aware, previously, that our source distributions
shipped with .c files generated from the .y files.
That said, if the parser's gram.y file is anything to go by, the fixes
to make it work with bison 1.75 are trivial, so if you change your
mind I'll be happy to do the work.
--
Kevin Brown kevin@sysexperts.com
