Hi Eugene, Sorry for the long silence, i spend a lot of time testing my Zookeeper client embedded in a Weblogic server and i noticed a note related to ZOOKEEPER-1455 at the beginning of the clientTunneledAuthenticationInProgress() method. Actually i'm also a little bit annoyed by this method because since it is called again and again each time a response is received it floods my debug log with messages like "Could not retrieve login configuration: ... "
So here is a new patch : https://github.com/barkbay/zookeeper/compare/barkbay:branch-3.4...barkbay:weblogic-sasl-bug( you can check the final code here : https://github.com/barkbay/zookeeper/blob/weblogic-sasl-bug/src/java/main/org/apache/zookeeper/client/ZooKeeperSaslClient.java) It adds a new SaslState.NO_CONFIGURATION state : this new state is defined immediately when the ZooKeeperSaslClient is initialized and means that sasl is not authenticated. if saslState == SaslState.NO_CONFIGURATION then the clientTunneledAuthenticationInProgress() returns immediately and my logs stay clean :) There are also 2 new JUnit tests and all existing unit test have passed. Do you think that it can help ZOOKEEPER-1455 ? Thank you for you feedback. 2012/12/2 Eugene Koontz <[email protected]> > On 11/29/12 11:51 PM, Michael Morello wrote: > > Hi Eugene, > > > > Thank you for the feedback. > > I'm setting up a clean development environment (even with a clean > checkout > > some JUnit tests fail, i must understand why before i could submit a > > patch), I do more testing on our clusters and then I open a JIRA with a > > patch. > > > > 2012/11/27 Eugene Koontz <[email protected]> > > > >> Hi Michael, > >> It looks like a good idea to me and your patch looks good too. You > should > >> feel free to open a JIRA and attach the contents of your gist as a > patch. > >> -Eugene > >> > >> > Hi Michael, > Even if some unit tests fail, that's OK :) You should feel free to > submit the patch anyway. Almost all JIRAs have many iterations of a > patch before it's ready. Perhaps someone can look at your patch and help > you fix the test failure because they've seen something similar in their > own experience. > > -Eugene >
