This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [Newb] grokdeclarator in gcc/cp/decl.c
- From: "Joseph S. Myers" <jsm at polyomino dot org dot uk>
- To: Zack Weinberg <zack at codesourcery dot com>
- Cc: sashan <sashang at ihug dot co dot nz>, gcc at gcc dot gnu dot org
- Date: Sun, 18 Apr 2004 20:23:37 +0000 (UTC)
- Subject: Re: [Newb] grokdeclarator in gcc/cp/decl.c
- References: <4082D077.60505@ihug.co.nz> <87fzb1qca3.fsf@egil.codesourcery.com>
On Sun, 18 Apr 2004, Zack Weinberg wrote:
> sashan <sashang@ihug.co.nz> writes:
>
> > I was tracking a bug (pr12102) and came too the grokdeclarator
> > function in gcc/cp/decl.c. The bugs not in grokdeclarator however it's
> > a 2000 line long
> > function. Given that it says here
> > http://gcc.gnu.org/projects/beginner.html that one task is to break up
> > large functions I was wondering if anyone has any objections to
> > breaking that function into smaller ones. It's something I'd like to
> > do but I want to make sure that any effort I make won't be wasted.
>
> We'd be delighted to see this happen, but please take care: that
> function has to do a very complicated job. You should get your hands
> on a copy of the C++ standard if you haven't already, so that you can
> see what it's supposed to be doing.
I would also encourage seeing if any parts that can be factored out can be
shared with the C front end. This will require a careful analysis of all
differences between the C and C++ versions of the code factored out and
whether they arise from differences between the C and C++ standards or the
front ends' internal datastructures, or whether they are arbitrary and of
no semantic effect, or whether one version is in error.
--
Joseph S. Myers
jsm@polyomino.org.uk