From: Glenn L. <pe...@ne...> - 2004-10-21 20:41:12
|
On approximately 10/21/2004 11:18 AM, came the following characters from the keyboard of Johan Lindstrom: > The mouse movement problem was related to the parameters to the mouse > movement and click event handlers. In earlier versions the first param > was kind of unused (or at least undocumented), and I guess it was > removed in this version. > > I'll just branch on version <= 0.9 to get that to work, no biggie. Hmm. I thought more parameters were added, rather than removed, as a general rule, in Laurent's reworking of the code. But I didn't look at the code for mouse movement and click handlers, so you probably have the straight scoop. Glad this one is straightforward to fix, anyway. > At 16:59 2004-10-20, Jez White wrote: > >> The global destruction issue, well, good luck tracking that one down:) >> When the Loft user closes the test/preview window, do you do anything >> special? > > > Currently I don't do anything. I guess something happened to the default > action of the Window _Terminate event. > > I can work around this, so that in itself doesn't break TGL sufficiently > to not release 1.0. But the problem is still there. What happens is this: > > The main window disappears for no reason (perhaps _Terminate returns -1 > and exits the event loop) and this is spewed: > Use of uninitialized value in subroutine entry at > C:/appl/dev/mod_perl/Perl/site > /lib/Win32/GUI.pm line 2712 during global destruction. > > That in the the Win32::GUI::DC DESTROY. The $self at that point is an > empty hash ref object. > > With a little print statements in new and DESTROY, I can see that DC > objects are created and destroyed a LOT for some reason. Maybe something > is painted. > > I don't have time to debug more right now, but I'll check it out. Sounds strange. And does this new/DESTROY happen frequently only after the main window disappears, or all through the execution of the code? And I don't know if DC's are subject to the cascading destruction of their parent window? > I suppose I should release a version of TGL that works with both > Win32::GUI pre and post 0.9 before we unleash 1.0 on the public. I'm > away tomorrow and Saturday, but maybe on Sunday. I agree this would be an appropriate path. Although I haven't used TGL, I think it is a significant application for Win32::GUI, and I wouldn't want to break it with a new release of Win32::GUI, even if it means a little delay. > /J > -------- ------ ---- --- -- -- -- - - - - - > Johan Lindström Sourcerer @ Boss Casinos johanl AT DarSerMan.com -- Glenn -- http://nevcal.com/ =========================== The best part about procrastination is that you are never bored, because you have all kinds of things that you should be doing. |