vinsci@ra.abo.fi (Leonard Norrgard) (07/13/89)
Today we encountered a slight problem with CloseWindow(). When we close one window (window A), the result is a change in another windows Flags (window B, a backdrop window), namely the REPORTMOUSE flag, which is cleared. Window B becomes active when window A is closed. This is what happens in the code: A gadget in the window is double-clicked. This eventually leads to the gadgets window being closed. The gadget has the flags GADGIMMEDIATE, RELVERIFY and *FOLLOWMOUSE* set, and is of type boolean (hit select). After the double-click, CloseWindow(A) is called. When window B:s Flags is examined it's REPORTMOUSE flag has been cleared. If we insert a Delay(100); before the CloseWindow(A) call, everything works OK. A Delay(10); is not long enough. For now, we'll just set the flag after the call, but we'd like to know why this happens, and if there's some other way around it. Clues? -- Leonard Norrgard, vinsci@abo.fi, vinsci@finabo.bitnet, +358-21-654474, EET.
usenet@cps3xx.UUCP (Usenet file owner) (07/14/89)
In article <VINSCI.89Jul13032359@ra.abo.fi> vinsci@abo.fi writes: > > If we insert a Delay(100); before the CloseWindow(A) call, everything works >OK. A Delay(10); is not long enough. This reminds me of a bug I found, but have yet gotton around to reporting :-( (shame on me). It may be that this is the same bug. If you make type and run resident, have NEWCON: mounted and do the following: in RAM: have a script and a big file. The script is just a bunch of run type >newcon:0/0/640/100/BUG! ram:bigfile and execute it. You should probly have 30-40 "types" to assure that the bug will show. This will usually crash the machine even when there is plenty of memory (i think, but I only have 512K ). I think that some of the windows started closeing when it crashed every time I tried it. I found this out when I was testing the residentability of a program I was writing. REAL NAME: Joe Porkka jap@frith.cl.msu.edu