Debugging unexpected failure (was http server)

> Typically, the

way to debug is

   ?- tspy(doit/1).

...  Now step through it



Whenever I do this I invariably think that the computer could be doing this for me.
I dream at times of an additional tool. In the debugger this would be an additional
command (maybe ? key) that says 'step over this (as s), but if the predicate fails, reset the debugger to the point at which the stack was deepest.
(If that's not unique, you probably want the last one encountered).

Since frequently that's ultimately all the info I need from stepping through, is
'why is this thing failing?', and usually that's at the point where the stack is deepest.
-------------- next part --------------
HTML attachment scrubbed and removed