>From: Mark Mitchell <mark at codesourcery.com> >I still think that hashing mangled names down to >1024-bit hashes would win Would we be able to demangle them? Or would we have to keep a database of them?