Loop optimizer issues

[Richard Kenner]

    But (and here we go again...) not all front ends use tree-ssa, so
    while tree-ssa gets rid of it for the front ends that use it, the
    others (read: g77, Ada (Kenner! ;-)) are stuck with a crippled
    optimizer.  

I don't see that as the issue.  I am *assuming* here that everyone
will eventually use tree-ssa.  My argument is that it's not a good
idea to get into the business of having one optimization pass make too
many assumptions about what previous passes have done.

The precise set of conditions that determine whether a libcall is generated
is complex enough, let alone considering the set of such that are not 
optimizable.  Trying to do any deduction on those sets (across all machines)
is far harder than the few lines of code to support them, especially since
most of them have already been written and debugged!