Sergey Plis <[EMAIL PROTECTED]> writes Re: [bigforth] debugging floating point errors
> I have an iterative code with passes the first iteration with no problem > but then bigforth hangs. If I run it all in gdb then it tells me: [..] > Why does bigforth not catch it and just sits there eating all CPU? How > can I debug such code in order to find where the error happens on the > second run of the loop? Hypothesis: The program writes naughty bits to the system return stack (the one indexed by ESP) already on the 'successful' first iteration. These may foul up CATCH/THROW and maybe also SEH frames. What you can try is put an extra ABORT or THROW in the code that executes in the first pass, and see if that does what it should. If it doesn't, the statement is passed the point of no return. And what do you think should happen when an exception occurs in a thread or callback? (or maybe here I should say 'how' would what I want be made to happen). -marcel --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
