Re: [Pnet-developers] The stack trace bug revisited

[Gopal V]

If memory serves me right, Rhys Weatherley wrote:

verify_except.c:109

                /* If execution falls off the end of the matching code,
                   then throw the exception to the calling method */
                ILCoderThrow(coder, 0);

                /* Mark the end of the handler */
                ILCoderTryHandlerEnd(coder);

                /* Advance to the next region within the code */
                offset = end;
        }

        /* If execution gets here, then there were no applicable catch blocks,
           so we always throw the exception to the calling method */
        ILCoderTryHandlerStart(coder, 0, IL_MAX_UINT32);
        ILCoderThrow(coder, 0);

So why is this outputting a set_stack_trace ?. ILCoderThrow does indeed
generate a set_stack_trace ..

IMHO, we should have a new coder function ILCoderSetStackTrace() to have 
a bit more fine-tuning in our hands, Or at least one more arg to the 
ILCoderThrow ..

I tried to add one , but ended up having an undebuggable stack smash 
right now ... The positional struct thing is getting on my nerves :-)

> Reproducible at last!

It's REALLY hard to debug without a stacktrace and line number info ...

Gopal
-- 
The difference between insanity and genius is measured by success