This is the mail archive of the
mailing list for the GCC project.
Re: [tree-ssa] COND_EXPR lowering.
On Fri, 2003-10-24 at 13:57, Zdenek Dvorak wrote:
> > > changes, but this is about where my patience ends. I really cannot
> > > continue with further infrastructure changes under these conditions.
> > > Either someone of you finds the time to review the patches reasonably
> > > fast, or I am going to just give up this effort.
> > >
> > No one has had time until recently, other things have had priority. Im
> > looking at the COND stuff now because I think its close to next on the
> > priority list (at least mine :-). I agree its dragged on, there just
> > hasn't been time. I want to get the rest of the lowering stuff in in the
> > very near future too. Whats left, SWITCH_EXPR and BIND_EXPR? I'd like
> > SWITCH now too, as it ought to resolve the ongoing problems Dan is
> > having with PRE and presplitting all the edges.
> there is SWITCH_EXPR removal. Then BIND_EXPR removal -- the first patch
> for it (http://gcc.gnu.org/ml/gcc-patches/2003-08/msg01962.html +
> perhaps one updated version, IIRC) is also laying idle for about month and half now,
> plus about two additional patches are requiered for this. Then
> replacing compound exprs with double linked lists (one simple patch).
> Then some useful cfg manipulation functions (basic block splitting, edge
> redirection, etc.). With the priority given to these fundamental
> changes, this will get in soon enough for 3.7, maybe :-((((((((
Fundamental is a relative thing. None of these lowering changes affect
SSA and our optimizations in a fundamental way. The iterators hide
everything. lowering could happen at any point and it would be
These things affect you, and thats why you have the cfg branch. We've
had other more fundamental problems relating to compile time and
correctness/runtime. We've always planned to get to lowering, you simply
got to it before anyone had time to deal with it properly. I think we're
there now as we've resolved a lot of the other issues, and the container
aspects of the IL have bubbled to the top of the pain list.