This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Using GCC's front-end to parse SystemC
- From: Pop Sébastian <pop at gauvain dot u-strasbg dot fr>
- To: Fergus Henderson <fjh at cs dot mu dot oz dot au>, tm_gccmail at mail dot kloo dot net,gcc at gcc dot gnu dot org, Matthieu dot Moy at imag dot fr
- Date: Tue, 1 Apr 2003 22:03:23 +0200
- Subject: Re: Using GCC's front-end to parse SystemC
- References: <vpqel4wpxl8.fsf@imag.fr> <Pine.LNX.4.21.0303241224480.24829-100000@mail.kloo.net> <20030401075417.GB13035@earth.cs.mu.oz.au> <vpqbrzqhvhw.fsf@ecrins.imag.fr>
On Tue, Apr 01, 2003 at 09:23:55PM +0200, Matthieu Moy wrote:
>
> But will they do type analysis & so on ? Typically, if a global and a
> local variable have the same name, I don't want to bother about which
> one is which one, and the front-end must have done it for me. For
> example, a yacc grammar would not be sufficient.
>
> But off course, if you have a precise idea/name of software/...
>
> At the moment, the main options seem to be hacking GCC (I began some
> work on it, and I confess I'm a bit lost ...) or using the proprietary
> EDG front-end, but I would prefer to use free software.
>
You can have a look at the tree-ssa branch:
http://gcc.gnu.org/projects/tree-ssa/
It includes a simplification pass that eliminates syntactic sugar and
lowers the code to a three address form simpler to analyse or to transform.
Gimple is still a high level representation since you have all the type
information.
Gimple does not contain concurency informations, but introducing AST nodes
with this kind of semantics is in the long term todo lists.
Sebastian