On 12-Jul-2002 Jamin W. Collins wrote: > On Thu, 11 Jul 2002 20:39:32 -0700 (PDT) > "Sean 'Shaleh' Perry" <[EMAIL PROTECTED]> wrote: > >> On 12-Jul-2002 Jamin W. Collins wrote: >> >> > Grrr.... I thought I'd gotten all of those. Should have use the F&R >> > function rather than eyeballing it. =( If it would help I can rerun >> > the tests. >> >> fixed patch attached (-: > > Updated bb-log attached. Window ID this time is 0x1400021. >
on my machine it asks to be 789 x 269 which becomes 791 x 288 after the frame is added. It asks this AFTER this window is visible but fast enough that you do not notice it. This is directly from a XMoveResize window call or equivalent in the swing code somewhere. I added another print directly from the values in the event we receive. While I accept that this is not occuring under other wms, I am at a great loss as to why the java gui toolkit is sending us bogus configure requests. I wonder if further testing would reveal other wms where this occurs. One thing I have noticed is that the swing toolkit sends new and different WM_NORMAL_HINTS, so without the debug messages we could not see the size the window actually loads at. Your app requests 400x150 yet it is initially mapped at 390x120 and a new WM_NORMAL_HINT and configure request later on changes the size. The problem appears to have something to do with the delay in the window loading. If I run: for i in `seq 1 10`; do /opt/java/j2re1.4.0_01/bin/java DemoApp & done it is almost always one of the last 3 that shows up an odd size. Another item of interest is the odd configure request is almost always 2x (1.5 or 2.5) the window size. Like it stacks up two requests and sends them as one almost. A reliable way to cause this would be handy. Until then your app has helped a lot. I am still waiting for a report from Chris about his problem with 1x1 windows.