The bug only occurs when the Windows native L&F is turned on, and I believe it is most likely a Java bug. The code involved is extemely complex with multiple layers of Listeners interacting accross multiple classes. I am glad you will have a fix for it in the next day or two Jon.
I am working on cleaning up and standardising the ‘board-bases’ editors (Deck/Stack Positioner, Irregular Grid region Editor, Zone Shape Editor), extracting duplicated code and fixing bugs. That’s pretty much it for me apart from looking at bugs and general issues raised.
It’s odd, because it appears to happen only with Brent’s account, and it
recurs almost immediately after I change the name associated with
Brent’s account it in the database. Fixing this will require looking at
the code in Bugzilla to see under what conditions the name can be reset.
If you’d like to do that, I’d appreciate it.
I’m good with reverse-engineering, which is why I kill bugs pretty quickly. But that’s all I do.
If you ask me to “just shut up and finish up your suggested Pandemic game polish”, which I believe (perhaps mistakenly) I can technically do, I’ll be procrastinating. I promised to redo Pandemic in 2 days, and never got around to it! For the record, I think I’ll take much longer than most of you guys to redo Pandemic.
So, feel free to throw bugs at me and tell me to “just shut up and eat bugs”! Heh. Yeah, I’m willing to take it upon myself to be responsible for bugs in Vassal, cos I can do that pretty easily. Rest of you guys should help me create games I can enjoy!
[size=150]Commander, dilithium chambers are at full. Warp cores stable and ready. Engineering is good to go! Sir![/size]
So, feel free to throw bugs at me and tell me to “just shut up and eat
bugs”! Heh. Yeah, I’m willing to take it upon myself to be
responsible for bugs in Vassal, cos I can do that pretty easily. Rest of
you guys should help me create games I can enjoy!
With your ‘fix’ in place, you can no longer drag counters from GamePiece palette and drop them on a map. They just dissappear.
This existing code works perfectly with Java L&F turned on, but is only broken with Windows L&F turned on. This is why I suspect a Java bug in the Windows L&F. There appears to be a difference between the Java and Windows LaF’s in the sequence of Mouse events generated when you cancel a popup window by clicking away from it.
This existing code works perfectly with Java L&F turned on, but is only
broken with Windows L&F turned on. This is why I suspect a Java bug in
the Windows L&F. There appears to be a difference between the Java and
Windows LaF’s in the sequence of Mouse events generated when you cancel
a popup window by clicking away from it.
This might be a reason to give up on the Windows L&F. That would also
solve our font size problem.
Sorry! Submitted a new patch. Didn’t cater for the case where pieces are dragged between different maps. Actually, they don’t disappear; they just don’t get moved or dragged! Note the function “isClicked”. The problem with my patch is right there.
I don’t know if this is a Java bug. The fixes I made are all logical fixes (to do with logic). In short, it indicates that the problem isn’t at a lower (API) level, but at a higher (programmer) level.
Thanks for testing!
I didn’t test if the patch works for Java L&F. Can you check if I broke things for the Java L&F?
Now, this would most likely be an API error. Is there a tracker for this, so I can observe this difference? Currently, I’ve added codes to work around this problem. I merely used the right-click pressed to “cancel drag” (on second thought, you might wanna change this logic to “buttonPressed != BUTTON1” instead, so that anything other than left-click press will cancel drag).
Your new patch generates a NullPointerException when moving a counter from 1 map to another.
As I said, this code works perfectly With Java L&F and breaks with Windows L&F. We shouldn’t have to do anything. I think your fixes are just attempting to work around this bug.
Erm. Is it alright if you give me a little support, say in terms of testing my patches, instead of telling me I shouldn’t have to do anything? Well, if I’m breaking things, please do tell me to stop. But otherwise, is it alright for the team to allow me some indulgence in fixing things?
I don’t see that NullPointerException. Check out trunk version 7945, apply my patches, and test. Thanks!
Since we have other issues with the Windows L&F, we are considering reverting to the Java L&F, which will fix this issue also, hence my comments. Please, keep at it. I will have a look at the latest version.
No problem. Brent and I fixed it just now. Should be in v3.2. Along same lines, we probably would standardize all LookAndFeel(s) soon, so that we have less number of “variant UI environments” to handle. Enjoy!
Okay, I just looked it up too. I hadn’t realized that the “Windows Look & Feel” was just a bunch of alternate Java components rather than actually tapping into a library of Windows graphical elements. The examples given on Oracle’s site don’t look much like Windows to me… (which would seem to be the point of the entire exercise).
Brent Easton acknowledged reproducing this behavior. I don’t know if he
intends to fix it for 3.2. I’d rate the issue as low-priority, though
it can be annoying at times. (I originally reported this here:
viewtopic.php?f=5&t=4499&start=30[1] )
Do you still experience this problem with the svn8015 build?