Re: Tree-SSA self checking infrastructure

[Jan Hubicka <hubicka@ucw.cz>]

> On Wed, Nov 19, 2003 at 12:13:06PM -0700, law@redhat.com wrote:
> >  >This brings me into questions about my tail call updating code.
> >  >Perhaps I need to re-do SSA form on call cobbered variables after
> >  >removing the call?
> > Are you changing the CFG?  Are you changing the dominator tree?  Those
> > are the key questions.
> 
> Yes, he is, but only by adding a new block on the ENTRY->BB0 edge,
> and then edges into that block from the tail recursion sites.
> 
> I may be mistaken, but I thought the new phi nodes he was adding 
> there would be correct.  Perhaps this is a good test for a verify_ssa
> pass?
Actually good motivation to write it.
However I didn't find anythign wrong about it, just double checking
everything.

The problem appears to be must_alias related.  If I do tailcall/tail
recursion before it, everything works, after it it breaks, but of course
disabling must_alias may alter decisions of later passes.
I've collected several fixes to must_alias/tree-dfa and I re-testing now
so perhaps the bug will just go away....

Honza
> 
> 
> r~