develooper Front page | perl.perl5.porters | Postings from October 2012

Re: Eliminating the "rehash" mechanism for 5.18

Thread Previous | Thread Next
Aristotle Pagaltzis
October 30, 2012 08:51
Re: Eliminating the "rehash" mechanism for 5.18
Message ID:
* demerphq <> [2012-10-30 09:20]:
> Any "per hash" logic is going to be problematic at the hash
> function/seed level. […] There is an internal "master hash" which
> contains every key in use by perl at any given time. All other keys
> are actually pointers to the key data in the master hash. Store/update
> operations do operations in this hash as well as in the actual hash
> being used. For obvious reasons we do not want to recalculate the hash
> key twice if we can possibly avoid it, which means that every hash
> with shared keys has to produce the same hash value.

One master hash per hash algorithm?

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About