Infinite loop in init_alias_analysis

[Jeffrey A Law]

  In message < 199711051230.NAA01044@lslsun17.epfl.ch >you write:
  > > HJ's alpha example is actually one copy deeper, but the same
  > > basic problem:
  > 
  > I was wondering if longer cycles were likely to happen.  Now I know :-)
Oh yes :-)    The reason behind the reorganization of the alias.c
code was do deal with more longer, more complicated set chains better. 
They can get much longer than 3 sets and aren't always nice and
simple like (set (regx) (regy)) 

  > > Now that I've had a chance to look at the code, I suspect either
  > > your solution or mine will work.  Mine converges quicker, but I'm
  > > somewhat leery that for some weird example the quicker convergence
  > > might be to the wrong state :(
  > 
  > What's your solution?
Instead of wiping new_reg_base_value clean at the start of the loop,
you copy entries from reg_base_value to it.   Problem is I can't
convince myself this is the right thing to do, but I can convince
myself that your solution is the right thing to do.

jeff