This is the mail archive of the
mailing list for the GCC project.
Re: [tree-ssa] RFC: Making control flow more explicit
- From: Michael Matz <matz at suse dot de>
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: Jeff Law <law at redhat dot com>,Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>,Steven Bosscher <s dot bosscher at student dot tudelft dot nl>,"gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>
- Date: Mon, 11 Aug 2003 17:52:15 +0200 (CEST)
- Subject: Re: [tree-ssa] RFC: Making control flow more explicit
On Mon, 11 Aug 2003, Diego Novillo wrote:
> However, if we are going to do this at the gimplifier then we need to
> consider whether we want GIMPLE itself be an IL with no control
> structures other than COND_EXPRs and GOTO_EXPRs.
FWIW I think GIMPLE should have no intrinsic control flow, except the
GOTO_EXPR (and maybe COND_EXPR). Everything else introduces artifacts
when dealing with almost all control flow aware optimizations (like you
repeatedly saw in the past when making the different insert-on-edges stuff
work). The control flow itself is better represented by explicit blocks
and edges, with edges coming out only at the end.