[SERVER] re: Production release 1.8.0 of the LambdaMOO server

• Date: Wed, 13 Mar 1996 07:38:28 PST
• From:
  ThwartedEfforts <abakun@scinc.com>
• Content-Type: text/plain; charset="us-ascii"

At one time, I started to do a server patch to add the :bf_* functionality,
and my implementation includes what Roger Crew mentions, altho I didn't
explain it like that in the message I sent in response to Judy's mention of
it.  The original intent was to overload all bf calls with MOO code, altho I
explained it as:

>1) if a builtin is called from SYSTEM_OBJECT, then wiz-only permission checks
>   (via $server_options) don't take place.  This would make it easier to have
>   the server check permissions for an action instead of trying to do it in
>   MOOcode.  In other words:

Yes, I said 'wiz-only permissions checks don't take place', which is not how
I meant it.

Also, I said:

>2) That the verb on SYSTEM_OBJECT that is called be named "bf_" + the name
>   of the builtin, to better distinguish between verbs that are handlers for
>   builtin functions and those that are server verbs (ie do_login_command).
>   Perhaps "secure_" + builtin name to signify that it's only called when
>   a $server_options perm check fails.
>   This also keeps #0:recycle for it's intended purpose.

Yes, there's the line 'to signify that it's only called when a
$server_options perm check fails", which is another misunderstanding/typo.

If you were working from my summary post as to what this feature would be,
Pavel, I'm sorry for the confusion.

Roger Crew has made a good point, and he's totally correct as to how the
operation of this feature should be.

Andy.

• Prev: [SERVER] Re: huh stack in tracebacks
• Next: [SERVER] rfc: huh handling
• Index: MOO-cows
• Thread: MOO-cows