malloc: performance improvements and bugfixes

[Paul Eggert]

Jörn Engel wrote:
> I have not and will not.  Or at least someone with a silver tongue would
> have to spend significant time explaining the advantages of copyright
> assignment to me.

Given your later remarks, it sounds like it'd be less work for us to rewrite the 
malloc implementation ourselves, than to try to convince you to assign 
copyright. Oh well.