[tree-ssa] edge insertion/split problem

[Diego Novillo]

On Tue, 2003-06-17 at 08:50, Andrew MacLeod wrote:

> Im not sure I'd call it brittle... Theres just the very odd case thats
> been missed, like this one. Whats there appears to work just fine, and I
> think we have all the cases now.. Its being used quite extensively.
> 
Exactly.  I meant 'brittle' in that there are several special cases to
take care of.  That makes it inherently brittle.  Not so much from your
point of view, but for the next poor soul who has to go and change
something inside :)

In any case, it's one of those changes that we can do gradually. 
There's the matter of deciding whether to represent switches as a jump
table or a decision tree.


Diego.