2015-05-27 19:11 GMT+02:00 Mark Thomas <[email protected]>: > On 27/05/2015 18:09, [email protected] wrote: > > Author: markt > > Date: Wed May 27 17:09:47 2015 > > New Revision: 1682079 > > Forgot to add that the NIO2 tests pass consistently on my laptop. We'll > see what the CI system thinks in a couple of hours. > > Ok, interesting, since this was never an issue for me.
But I still have a problem with the fill side effect in parseRequest line. fill will async read to a buffer that is in a processor that is no longer associated with the socket (parsingRequestLinePhase is not 2). I will experiment with a socket wrapper flag that avoids this non blocking IO "peeking" with NIO2. Rémy
