This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Patch ping
Andrew MacLeod <amacleod@redhat.com> writes:
> On Wed, 2003-12-03 at 14:03, Zdenek Dvorak wrote:
>> Hello,
>>
>> > > The only real
>> > >way to do this is to have another TREE_CODE, and I dont like that idea.
>> > I'm not sure how much I like/dislike this. The only way this wins is if
>> > the old COND_EXPR node goes away completely, which it probably would.
>> > Then it becomes a question of whether or not the memory savings of our
>> > new node is enough to overcome the increased memory fragmentation and
>> > loss in locality (assuming we had any locality to begin with! :(
>>
>> there is one more possibility that seems somewhat tempting to me, but
>> that would work well only if we were expanding to rtl without throwing
>> away the CFG. We might then just forget on COND_EXPR_THEN and
>> COND_EXPR_ELSE (i.e. set them to NULL), since the information is
>> represented in cfg anyway.
>
> Right now Im not interested in keeing the CFG right through to rtl
> expansion, but then you know that :-)
Representing control flow only via CFG, during RTL expansion and
beyond - all the way until final, in fact - would get rid of a number
of issues with RTL. (*cough* tablejumps)
zw