> >Yes, but I think we had better plan some time and effort for > peer review. > >I'd also like any related test classes to be reviewed to see > if they should > >be extended to test the new changes. > > Excellent points but the files that require review need not > be in the sandbox, > do they?
No, they do not. I guess I am looking for the fine line. As you have said before, the biggest threat to log4j stability is us making changes that have unforeseen ramifications. Or being stuck with designs we would have changed if we had thought about it a bit longer. Given the number of changes that have just been applied, some time for peer review is in order. I'm not putting any blame any where. I have been bad about this, and Paul and Scott have been pretty much begging for feedback and discussion. Given that I have asked for feedback in the past for stuff I have committed, I am breaking my own rules and guidelines. I'm going on a vacation end of this week (for 2 weeks), so I don't know how much time I can devote between now and then. But, yes, I am going to be reviewing the changes, with an eye towards the changes in the "core" classes. (I've had this feeling that the Socket/SocketNode related classes might need some refactoring, but I need to think about this some more). Obviously, test cases in these areas help. I also want to use the jDiff tool to track api changes. I plan to attempt a jDiff report between v1.2.8 and the current cvs in the near future. -Mark --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
