[tree-ssa] bootstrap miscomparison failure on PPC

[Diego Novillo]

On Fri, 2003-07-04 at 18:16, law@redhat.com wrote:

> So how can you be testing a potential fix if you don't really understand 
> the problem?
> 
Hence the qualifier "potential".  If it works, then I'll understand the
problem.

But, I agree, this is not likely to fix the problem, unless
const_and_copies is caching non-folded constants that when replaced in
expressions result in non-gimple expressions which are tricking the
optimizers.


Diego.