Martin Grigorov <mgrigo...@apache.org> wrote: >Hi Mark, > >Thanks for the fix! > >On Fri, Aug 2, 2013 at 10:28 PM, <ma...@apache.org> wrote: > >> Author: markt >> Date: Fri Aug 2 20:28:12 2013 >> New Revision: 1509847 >> >> URL: http://svn.apache.org/r1509847 >> Log: >> Use the same Map implementation for user properties to prevent >> implementation conflicts >> >> Modified: >> >tomcat/trunk/java/javax/websocket/DefaultClientEndpointConfig.java >> >> >tomcat/trunk/java/javax/websocket/server/DefaultServerEndpointConfig.java >> > >Is it OK to change the JSR API jars ?
For the sake of clarity, that code is not a copy of the implementation from the EG. It is a clean room implementation of the same API. Therefore we are free to change it as long as the API is in unchanged and functionality is consistent with the spec. >The change is binary compatible per se, but I wonder whether such >changes >in classes would not lead to different behavior in different web >containers. The Java WebSocket spec just says Map. It does not fix discuss nulls or thread safety. That is perhaps something that should be clarified in the spec. Please raise a Jira issue against the spec for this. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
