Re: redirect_edge_and_branch questions

[Andrew MacLeod <amacleod at redhat dot com>]

On Mon, 2004-10-04 at 10:50, Jan Hubicka wrote:

> 
> Even for your trivial case it is not clear whether the edge redirection
> should decrease original destination frequency or keep it is it is.  In
> your case it should be kept, in edge forwarding it should be decreased.
> 
OK, I get it :-)

> Also you should look into the function Zdenek mentioned.
> 

I did, and it didn't see to do anything interesting. Of course, I looked
again, and discovered that *this* time around I *shouldn't* have stuck a
'tree_' in front of it when looking for the routine name :-). Doh!  
tree_make_forwarder_block isnt very interesting :-)

so yes, that looks like it might be very useful, let me move things
around a bit and combined with the cleanup_cfg() that happens after
optimization is finished, I ought to get what I want.  I think.

I will let you know if I don't :-) haha

thanks

Andrew