On 10/03/2014 15:47, Rainer Jung wrote: > On 10.03.2014 16:08, Mark Thomas wrote: >> On 10/03/2014 15:00, Rainer Jung wrote: >>> Here's a patch: >>> >>> http://people.apache.org/~rjung/patches/TC8-SessionIdGenerator-extensible.patch >>> >>> Any concerns about applying to TC 8? Please note "Known problems". >>> >>> Some notes: >>> >>> About the patch: >>> >>> 1) New interface org.apache.catalina.SessionIdGenerator >>> - setter and getter for jvmRoute >>> - setter and getter for session id length >>> - generateSessionId() (use jvmRoute if set) and >>> generateSessionId(String route). >>> 2) Renamed org.apache.catalina.util.SessionIdGenerator to >>> org.apache.catalina.util.SessionIdGeneratorBase. >>> Not strictly needed but for consistency with other similar cases. >> >> Is it consistent? I'd expect a class with a name that ends in Base to be >> abstract rather than concrete. I'd expect the default implementation to >> be named something like DefaultSessionIdGenerator. >> >> Other than that +1 > > As always you are completely right. Sorry.
No need to apologise. As the archives will show, I make (probably more than) my fair share of mistakes around here. > So either the "Default" or the "Standard" bike shed. Pick your favourite bike shed. I have no strong preference. > We could also add a abstract base with > everything except the two generateSessionId() methods, but that would > seem a bit artificial to me. I guess it depends how pure "object orientated" you want to be. I think a base class is unnecessary but I don't feel strongly enough about it to object to a base class if you want to go that way. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
