Calling module_implements loads up files. We don't want to do this. Thus this needs to be cached.
Shaves ~10ms off of the request for an advagg cache hit. Following patch has been committed
I've gotten this down to core: 7ms, advagg: 9ms using microtime(TRUE) timers. Using xdebug it shows core: 32ms, advagg: 42ms. ~30% slower when we are talking about 2ms isn't a big issue in my opinion. If I wanted to get more out of it I could cache advagg_merge_plans or _advagg_aggregate_css/js but that would give me a lower cache hit ratio for that cache due to inlined output. Nonetheless something to think about.
core: 7ms, advagg: 9ms
core: 32ms, advagg: 42ms
By adding caching at the highest level I got this to be faster than core.
I'll build a setting to turn caching on and off.
The following patch has been committed.
Automatically closed -- issue fixed for 2 weeks with no activity.
Drupal is a registered trademark of Dries Buytaert.