I don't know how everyone else is doing but I know I've virtually run out of time
To explain, I haven't got a mature enough project going that uses jME to pick up bugs and I don't know enough about game programming in general to spot bugs easily :oops:
Last weekend I had a little look but had no real clue of what areas to check and thus I didn't even find any 'leads'.
I work all week, next weekend I've got to get some Christmas shopping done. The weekend after that I'm visiting my Brother (no computer at all for a long weekend :-o) and the weekend after that is Christmas weekend! All in all I don't think I'm going to have a lot of time to go bug hunting
I realise we all have lives and are busy, I just happen to be even busier thanks to Christmas being just round the corner (as mentioned above it's eaten all my weekends!).
So I'd like to propose two potential changes to the challenge.
Extend it to include Janurary as well.
Allow dev's to submit bugs to the tracker so we can pick up points by fixing them (I know there are some there already, which is how I got this idea).
Really what I'm getting at is Decemeber probably wasn't the best time to have a challenge. Feed back!
Those are not bad ideas. December was picked because people had time off work, school, etc. but of course the converse is as you say that time also gets filled with holiday events, chores and tasks. Give us a day to discuss offline. In the meantime, any other participants have comments?
I agree December is a paradox of work/school slowing down and more free time, which is then immediately swallowed up by family and friends time (which is a good thing ).
Personally I’m extremely busy with work/school until around Dec. 15, then I have family and friends plans for some weekends, but I’m hoping to really dig into jME from Dec. 15 - 30 (although Dec. 15 - Jan 15, would probably be even better!)
I like Hal’s ideas of having dev’s perhaps submitting some issues that contestants (for a slight reduction in GP because the contestant didn’t ‘discover’ the bug) could then provide solutions to, wiki article writeups about, as well as detailed JUnit tests to allow for regression testing and greater code refactoring confidence.
I think another thing that would help everyone to find bugs is to take a look at a random class, and think up various unit tests that you could run that would test out the boundaries of the inputs to the methods, as well as any assumed behavior. I remember in a software engineering course we were given the simple task of implementing a class that represents a polynomial and the operations on a polynomial (scalar multiplication, polynomail addition/multiplication/division), we were then Tetold to write any relevant unit tests to make sure the implementation we provided is correct, during that process we found numerous obvious bugs (0 coefficients etc.). Even after that our instructor ran our code through their own unit test suite which discovered even more bugs that we didn’t even think about.
I guess the key idea is that most classes in jME are probably going to be used in a very specific manner, and unless otherwise stated (explicit documentation comments, defensive code), it should be flexible enough to handle all kinds of inputs.
So that’s going to be my plan for tackling this stuff, to try and write JUnit tests of classes that will excercise all kinds of inputs and behaviors, to see if I can catch any bugs that would normally go unnoticed in most user code (but that would probably pop up at worst time in a complex project, and be extremely difficult to track down, because that’s the whole point of doing extensive unit testing to make sure the code is robust and works in every case or at least exhaustively many cases).
Here are some good resources on automated unit testing in Java using JUnit:
I'd be happy with an extension to mid January or so. You wouldn't want it to drag on endlessly though. I can feel for you all as time is flying by and my wife seems to have accounted for the rest of my weekends until the end of the year. XD
dougnukem - that sounds like a great idea. We could have several people just with the task of writing unit tests for jME… its a big job to reverse engineer unit tests but i expect there will be a lot of benefits as far as robustness.
I could not be agree with that because we have always things to do... and because I found some bugs that I haven't posted yet :D...
...grrrr just you wait max! :P hyperkinetic rabbity things are known to be buggy creatures
i have to admit that i had no time/luck until now with bug hunting. since a few weeks all i do is update from cvs when a checkin notification arrives and quick-browse the board (brainless posting as usual - if i have the time). the bad thing is that i won't really have more time in january because i'll have a lot of exams. i don't object to extending the challenge but i doubt it will enable me to report more bugs (i hope i'm wrong about this).
).
