SSA backend

[Stephen Weeks]

> How would a second pass of remove at the end of the optimize cycle cause
> problems?

Nothing in particular -- it just creates a different program than running one
pass.

> Let me know if you have any good techniques for debugging SSA functions on
> the size of mllex's main_0.  dot dies pretty quickly on the .dot file, and
> I get lost in the SSA file pretty quickly.

I don't have any offhand.  It might be nice to have a graph function that takes
a node and a depth and returns all nodes reached in a dfs up to the depth.

> Well, the sooner we can assume that SSA optimizations will never see Set*
> statements, the better.  I'd suggest getting at least a trivial one going
> soon.

The problem is that we need to keep them around until the exceptions paper, so
we're going to have to handle them.