[SERVER] Re: huh stack in tracebacks

• Date: Mon, 18 Mar 1996 14:50:12 PST
• From:
  ThwartedEfforts <abakun@scinc.com>
• Cc: moo-cows@parc.xerox.com
• Content-Type: text/plain; charset="us-ascii"

At 01:03 PM 3/18/96 PST, Pavel wrote:
>> list as returned by the except[ion] clause.  Any way around this short of
>> hand parsing the printed traceback strings?
>
>There isn't any other way to discover that the verb on the bottom of the stack
>is really `huh' and not whatever the first word on the command-line was.

What I ended up doing was when I got the full verb name via verb_info, if
that generated E_VERBNF and the verb name was not "", then I assume it's
"huh" and print that.  There are most likely a lot of problems with doing it
like this,  but oh well.

>after the verb is called.  It would be easy to switch over to the full set of
>verb names instead of whatever the caller used, but that would be an
>incompatible change to the callers() function (and no, I don't want to make the
>two kinds of stack lists different).

I'd suggest adding another field with the full verb name, but that is a nice
big patch to cover up for a nice big hole introduced with the implemenation
of the huh stack.  The other alternative I see is to call huh with it's
arguments being the entire command line, and not worring about the value of
verb.  This would require only one verb to be changed on most MOOs,
$root_class:huh

   $command_utils:do_huh(verb, args);
to
   $command_utisl:do_huh(@args);

Or, I can just shutup and continue to assume that E_VERBNF means "huh".

Andy.

• Prev: Re: [SERVER] question about extension.c/server.c
• Next: [SERVER] re: Production release 1.8.0 of the LambdaMOO server
• Index: MOO-cows
• Thread: MOO-cows