This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [ast-optimizer-branch] [patch]: SSA for trees
- To: Diego Novillo <dnovillo at redhat dot com>, law at redhat dot com
- Subject: Re: [ast-optimizer-branch] [patch]: SSA for trees
- From: Daniel Berlin <dan at cgsoftware dot com>
- Date: Mon, 23 Jul 2001 16:16:18 -0400
- cc: Nathan Sidwell <nathan at codesourcery dot com>, gcc-patches at gcc dot gnu dot org
- References: <20010723160017.A1959@tornado.cygnus.com>
--On Monday, July 23, 2001 4:00 PM -0400 Diego Novillo
<dnovillo@redhat.com> wrote:
> On Mon, 23 Jul 2001, Jeff Law wrote:
>
>> In message <3B5C728D.2949ED0C@codesourcery.com>you write:
>> > * optimize_tree is not a good name for an ssa specific optimization
>> One could argue that trees should always be in SSA form. I'm not sure
>> how practical that is, but maybe that's what Diego had in mind.
>>
> Actually, this is a single-entry point that should probably
> disappear. Its only purpose is to call the CFG builder, the
> varref finder and the SSA builder in sequence.
>
> I've been thinking that it might be better to do all these steps
> explictly in c-decl.c. There really is no need for the
> additional indirection. It's obviously confusing already.
Except that then you'd have to do the same in cp/optimize.c, and make sure
the steps stay in sync.
And for every other front end that operates on trees, too.
>
> Diego.