Re: strspn() & strcspn() speed improvements

Alexander Valyalkin wrote:

On Wed, 16 Jun 2004 00:41:19 +0300, Andrey Hristov <php@hristov.com> wrote:

Privet Alex, is it possible to show some results from benchmarking?

Just compile my standalone test & view results :)

As noted before: standalone benchmarks can only tell you whether ot not it is worth the effort to implement a change within PHP *and benchmark it as a whole*. Some changes may not show a difference as the code is not on a critical path, some may only improve performance on certain platforms while hitting it on others, sometimes even in very unexpected circumstances. (i could continue almost forever ...) Besides that the good old rules of engineering applies: "If it ain't broken -> don't fix it" and "Never change a running system!" I'm sure there are lots of places within PHP that deserve code reviews and performance improvements, but especially now, in a feature freeze, it is important to act in accordance to the rules above. Any changes in code come with the risk of introducing new errors. Especially those that seem "simple" bear this risk. To keep this risk as low as possible only bug fixes are allowed in a feature freeze. A "lack of performance" is in this case only considered a bug if it *substantialy* hurts performance. PHP has become a very complex piece of software. Its developement process has also become very complex. Do you really think you can just jump in out of nowhere, throw patches at us, telling us that we are all too stupid to write efficient code and still be taken serious as the great savior? -- Hartmut Holzgraefe <hartmut@php.net>