Hello again, thanks for your replies.
2007/3/30, Joe Huber <[EMAIL PROTECTED]>: > At 4:56 PM +0200 3/30/07, Bastian Bense wrote: > >Just a thought, but is it possible that the RB compiler causes > >problems on really long methods? The suspicious one here is around > >1240 lines, so I wonder if that might be the cause (if I comment out > >some of the code, it seems to work fine). > > RB has a couple of problems with very long methods. If your method > uses complex logical statements, lots of local varables or is simply > too big, you may get a compile time message saying that the method > uses too much stack space. > > There's another issue on Windows and Mac Intel where runtime > exception handlers don't get called if the method is too big. Does > your big method cause any exceptions that you try to handle locally? > That would be a problem in many recent versions up to and including > 2007r1. I don't get any exceptions nor compile-time errors. The application simply quits without further notice or warning. I tried to catch this crash in WinDbg and I get memory access exceptions (app is trying to access memory not belonging to itself). And sometimes I even get INVALID INSTRUCTION exceptions (at OS level). Here are a few other things I noticed: - Sometimes the crash happens further down the stack (in a method called by the really big method). - I sometimes can't re-construct the crash if I step through the code with the debugger (different stack?) - Sometimes breakpoints work, sometimes they simply get ignored or the app crashes before reaching them. - The crash only occours on Windows. It doesn't matter if I compile the code on Mac Intel or Windows itself using the IDE on Windows. - I NEVER get any REALbasic exceptions or warnings. - The application works just fine on Mac OS X - no crashes, nothing. Does this sound familiar, or... am I on the right path? -- Thanks, Bastian _______________________________________________ Unsubscribe or switch delivery mode: <http://www.realsoftware.com/support/listmanager/> Search the archives: <http://support.realsoftware.com/listarchives/lists.html>
