Re: PHP3 and 64-bit land

On Sat, 27 Feb 1999, Rasmus Lerdorf wrote: > > > Resources use hash tables to store themselves, and they usually use the > > > numeric indices. I think it's extremely safe to use 32-bit integers as > > > unique identifiers for 64-bit pointers, by simply casting them. You'd > > > have to be extremely unlucky to bump into a problem because of this data > > > loss. Your chances of winning the lottery are much much bigger. > > > > I disagree here. The chance of having the same ptr in 32 bit is about 1 in > > 2^32 = 4,294,967,296. If you come to Germany, your chance to win the lottery > > ("6 aus 49") is about 1 to 49!/43! = 10,068,347,520. > > > > It's extremly safe, but you risk random, non repeatable crashes produced by > > playing that way. > > Would there be a big drawback to using a long here instead of an int for > unique identifiers? It doesn't change things really. The chances anybody will ever notice a problem are, as it is, pretty much non existnant. Changing it to be 'long' doesn't make it any more correct than it is now, it just gives you 64-bit safety instead of 32-bit safety (since there's no provisions sizeof(long)==sizeof(void *)). I'm happy with 32-bit safety as it is, I don't mind changing it to be a long in PHP 4.0, but I definitely don't want to start playing with that sort of stuff in the production PHP 3.x tree. Zeev -- ----------------------------------------------------- Zeev Suraski <zeev@zend.com> For a PGP public key, finger bourbon@netvision.net.il -- PHP Development Mailing List http://www.php.net/ To unsubscribe send an empty message to php-dev-unsubscribe@lists.php.net For help: php-dev-help@lists.php.net