> Do we have any tests or evidence to show that the new code is faster than > the old code? > > My 5 minute check shows the old code is considerably faster. I'm an idiot. I wrote a test before applying the patch. I ran the test. I looked at the results. But I failed big time to actually *read* the results. Yes, my test says clearly that the old code is faster, but I somehow mixed up the result labels in my head and thought it's the other way round. Thanks for testing it again. I rolled back the patch. Andi -- splitbrain.org -- DokuWiki mailing list - more info at http://wiki.splitbrain.org/wiki:mailinglist