From: Ted Byers (r.ted.byers_at_[hidden])
Date: 2002-06-20 16:23:56
"Victor A. Wagner, Jr." <vawjr_at_[hidden]> wrote in message
> I hope I didn't give the impression that there was a "standard" way to do
> it. The code will be definitely system dependent.
No, you didn't give that impression. But you did give the impression that
it is easy.
> It's been a while since I was on either a compiler or library team, so
> giving a hard example for any currently used architecture would be beyond
> my immediate abilities. I do recall writing them, and that they were NOT
> difficult (on the same order as "stack unwinding", but probably simpler)
> and generally took one programmer somewhat less than a week given that the
> compiler team was available for all the 'dirty little details' about stack
> frames, etc. Of course, these days, there are supporting databases to
> describe the contents of the frames (for the debuggers) which we'd want to
> use and that would likely up the programming effort required.
Alas, that doesn't help me because I do not have the backgroundr equired to
be able to help develop a compiler. I am at a loss when the debugger shows
me assembler rather than my own C++ code. So what can I do? This
"requirement" is quite orthogonal to those served by traceable exceptions,
and I see no reason why it couldn't be used both to enhance the information
provided by traceable exceptions and in contexts where exceptions are not
appropriate, but it does me no good if I don't know how to do it.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk