This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Better GCC diagnostics
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Manuel López-Ibáñez <lopezibanez at gmail dot com>
- Cc: GCC Development <gcc at gcc dot gnu dot org>, Chris Lattner <clattner at apple dot com>, Aldy Hernandez <aldyh at redhat dot com>, Tom Tromey <tromey at redhat dot com>, Gabriel Dos Reis <gdr at integrable-solutions dot net>
- Date: Fri, 15 Aug 2008 16:00:05 +0000 (UTC)
- Subject: Re: Better GCC diagnostics
- References: <6c33472e0808150709u9420a23s6391cbc669c8c562@mail.gmail.com>
On Fri, 15 Aug 2008, Manuel López-Ibáñez wrote:
> 5) Handle locations during folding or avoid aggressive folding in
> the front-ends.
I plan to delay folding for C (beyond some folding for expressions of
constants) for 4.5, probably in October. (I do not plan to do anything
for C++, and the folding will still happen before gimplification, though I
believe that in principle most of what fold-const does would be better
done on language-independent IR after gimplification.)
Note that for correct ranges you need to handle locations for 0 and (0)
and ((0)) and so on, so either need duplicate trees for constants and
parenthesised expressions or need to keep the locations outside the trees.
It's possible you can keep them in the c_expr structures or other local
variables, depending on whether you ever need locations for both an
expression and deeply nested subexpressions of it.
--
Joseph S. Myers
joseph@codesourcery.com