[tree-ssa] Computed gotos

[Andrew Haley]

Robert Dewar writes:
 > > Is this kind of thing ever likely to work reliably and efficiently?
 > > I could probably rearrange things so we end up with a switch statement 
 > > containing normal gotos. This would eliminate the need for indirect jumps 
 > > altogether. Maybe this is the way to go?
 > 
 > Absolutely, computed gotos are an abomination in most cases

I use indirect jumps in a rather appalling way in the Java bytecode
compiler.  I guess I'll plead that support should be retained until I
can think of a better way to do it.

Andrew.