What I stumbled upon seems pretty interesting and awakened my curiosity. I read in a micro optimization article that it would be smart to use the anyway present $_SERVER['REQUEST_TIME'] instead of calling time() when I need a UNIX timestamp. I would like to ask if this is firstly reliable, if I'm certain that my application will always run under HTTP, and secondly would it be smart to overwrite the default time() function to
function time(){
return $_SERVER['REQUEST_TIME'];
}
Because refactoring - ain't nobody got time fo' dat!
It makes sense, however this is not correct. Because the time you get using time() is the current accurate system time, whereas $_SERVER['REQUEST_TIME'] is a constant populated by PHP when the page was requested. If the page takes time to load and $_SERVER['REQUEST_TIME'] is called in the end, there will be a difference between the values of time() and $_SERVER['REQUEST_TIME'].
Also overriding time() wont be a good idea. Because when calling $_SERVER['REQUEST_TIME'] directly, there is no stack maintained by php, but when overriding time() (always in your code weather its called or not) will cost much more in terms of system resources from an application architectural point of view.