[m-dev.] Re: proposed command set for the internal trace based debugger

Fergus Henderson fjh at cs.mu.OZ.AU
Fri Jun 12 13:37:41 AEST 1998


On 12-Jun-1998, Lee Naish <lee at cs.mu.OZ.AU> wrote:
...
> ... let me propose a couple of commands based on the trace
> view to see if they seem reasonable (both are to be used a an exit port):
> 
> most_recent_exit_at_next_level
> Goes to the most recent exit port at the next level (ie, the previous
> line in the full trace, like prev).

Well, `prev' is easy to implement, if you don't mind a bit of
recomputation:

	prev_event_number := current_event_number
	retry
	goto prev_event_number

> prior_exit_at_this_level
> Goes to the most recent exit port of the previous call at this level
> (like retry followed by prev;

Indeed, exactly like a retry followed by prev -- and at least for
the first implementation, that's exactly how I'd propose to implement it.

> Given that we will probably be doing recomputation anyway, the following
> could be used in addition:
> 
> initial_exit_at_next_level
> First exit *on current successful execution path* at next level.

Yep, that would be useful -- in particular, perhaps that's what the `bad'
command should do if you use it at an exit port other than the first
exit.  Not quite so easy to implement as the others, though.

> subsequent_exit_at_this_level
> Assuming we are currently at an exit on the right path, this will just
> take ust to the next exit at this level.

This is just `skip' (aka `next') followed by `creep' (aka `step'), I think.

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh>  |  of excellence is a lethal habit"
PGP: finger fjh at 128.250.37.3        |     -- the last words of T. S. Garp.



More information about the developers mailing list