remove_unnecessary_notes breaks libjava build

Wed Mar 7 03:54:00 GMT 2001

 > But the point is that BLOCK_DEAD is being set *precisely* in the
 > case where the note is being deleted. So how can somebody be using it?
 Ah, but the insn notes for the *sub*-blocks are not deleted, 
remove_unnecessary_notes claims to not delete a block if there is a 
sub-block and the code indeed appears to be avoiding that. Can you show
me the case where this happens?
 Does this look right?
Not completely because you have to do something similar in integrate.c.
But I don't think BLOCK_DEAD should be set in this case, so this is confusing.