Did find one tiny bug in the page-caching logic - if two people request the same uncached page in the same second, Minx will try to create two entries in the cache and the second one will fail. This gets logged but doesn't affect the output at all. The reason I'm going to bother fixing it is that it's bad practice to have any known errors; if there's a SQL update error in the log file it should be an abnormal condition.

Actually, what I'll do is just replace the SQL-based page caching with memcached. I didn't do that before because it's another server process I have to worry about, but it was something I was planning - memcached is designed for exactly this use.

Posted by: Pixy Misa at 10:01 AM | No Comments | Add Comment | Trackbacks (Suck)
Post contains 152 words, total size 1 kb.