*Many* cachebuster requests

I’ll probably going to have a workaround by doing $wpdb instead of get_option for the AO version, this should bypass the alloptions stale cache issue, hope to have this on github within an hour approx.

so there is no version-issue @rvola?? I’m baffled … :-/

the bulk of people who have the problem (which seems to be a minority given the amount of complaints + I have upgraded approx. 10 sites myself and the problem did not occur once) saw a huge amount of “cachebuster”-requests in their webserver logs, so this is definitively related to the cache being emptied automatically continuously.

the only part of AO that calls autoptimizeCache::clearall automatically is autoptimizeUpdateCode.php and the only reason that gets included & executed is if the version in the code (2.3.0) is different from the one in the database (well, it checks the major version actually 2.2 !== 2.3), which is why it occurred only upon the upgrade from 2.2 (which introduced “speedupper” which includes the cache preload) and 2.3. the filter you mention simply stops speedupper, which stops the cache being preloaded, stopping the high number of requests.

as (almost) everyone reporting this problem indeed is running redis or memcached to do object-caching, I’m 99% sure the root cause indeed is a stale value for autoptimize_version in the object cache somehow.