Performance tweak for apply_filters loop

[Quandary]

This patch uses foreach to reduce the time it takes to execute apply_filters (on my box) from an average of 120,694 ticks/standard deviation 33,038 ticks, to an average of 99,149 ticks/standard deviation 20,636 ticks. The sample size was 20 runs each, before and after.

Optimize apply_filters

Use foreach everywhere $wp_filter is iterated over

Stats on the second patch:

original		patched
total runtime (mean)	1,972,854	1,943,093
total runtime (std dev)	116,697	110,741
add_filter (mean)	14,424	14,592
add_filter (std dev)	5,458	3,224
apply_filters (mean)	120,694	103,971
apply_filters (std dev)	33,038	25,880

Sample size is still 20.

For the record, this patch is written and tested against 2.3.1, not 2.4/trunk.

Is this still applicable for 2.3.x branch? Since none of these enhancements will apply to trunk, this isn't an issue for trunk and if it isn't going to go into 2.3.x, then it should be closed.

Right?

Well, it would be unlikely to include this type of change in a maintenance release. Although the patches don't apply to trunk, do the concepts or are the issues already resolved on trunk using other code?

No. This was the ticket for 2.3.x and I made a ticket for trunk that was recently closed. If this isn't going to be applied to 2.3.x, then it just should be closed.

darkdragon, it would have been awesome if you included that other ticket # ;-)

Okay, #5338.