On 21/08/2013 15:43, Mark Thomas wrote: > On 21/08/2013 14:20, Nick Williams wrote: >> >> On Aug 21, 2013, at 6:14 AM, Mark Thomas wrote: >> >>> I'm not exactly sure what the problem is. It looks like the WebSocket >>> SCI is being registered twice or is running twice but I can't reproduce >>> the problem. Anyone got any ideas? >>> >>> Currently, my plan is to add some logging to the SCI registration code >>> to see what triggers each registration of each SCI. I'd add it as an >>> info log message and then drop it to debug once this issue has been >>> resolved. >> >> If you can't duplicate it locally, I don't see any other option but the >> temporary logging. Gotta figure it out somehow. > > Bingo. I looked at the Gump logs and found inspiration. It is because > Gump adds the JARs to the class path for the test. The SCI ends up being > added twice. I'm just looking into how to stop that.
Scratch that. There is only one SCI but the @HandlesTypes finds all the annotated endpoint test classes and there are multiple endpoints that use the same path. I'm still looking into possible fixes. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org