This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Why gimplification regimplifies things?
- From: Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>
- To: Richard Henderson <rth at redhat dot com>, gcc at gcc dot gnu dot org,s dot pop at laposte dot net, dnovillo at redhat dot com, jason at redhat dot com
- Date: Fri, 30 Jan 2004 01:26:01 +0100
- Subject: Re: Why gimplification regimplifies things?
- References: <20040129222238.GA11501@atrey.karlin.mff.cuni.cz> <20040130001609.GB11315@redhat.com>
Hello,
> > why is the gimplification done the way it is done, i.e. so that
> > once gimplified things are sometimes regimplified?
>
> I've avoided this as much as possible. What specific case
> are you thinking of?
I am speaking just generally, without any specific case in mind
(I wanted to try the possibility to represent gimple statements
by different type than 'tree', just to check how difficult it
would be to change the current representation, and the way
gimplification is done turned out to be one of the problems I
encountered).
> > It does not seem somehow principially difficult to produce
> > directly gimple statements.
>
> >From what? The front end? Yes, that is difficult.
no, in the gimplification.
Zdenek