hello i've tested the patch and it worked fine. how do you check that access.log is updated? (tail -f .../logs/access.log?)
there is a possibility that writes to access.log are deferred i will find out what is going on there... few words about performance: there is a really small overhead (one if ) in clientWriteComplete() first time this function executes logging code (and calls accessLogLog()) but the second and subsequent just skip logging code so this should not consume much time ;) regards michal matusiak
