Re: -1 owned a verb!

• Date: Fri, 22 Mar 1996 13:35:47 PST
• From:
  rgodard@divsun.unige.ch (Richard Godard)
• cc: "Seth I. Rich" <sir@po.cwru.edu>, moo-cows.PARC@xerox.com
• Content-Type: text/plain; charset="us-ascii"

At 15:08 3/22/96, Pavel Curtis wrote:
>Seth I. Rich writes:
>> I have no clue how it happened, but we ended up with a verb owned by #-1.
>
>I've tracked this down; it's renumber()'s fault.
>
>When renumber(OLD) is called, it first figures out what new, lower object
>number it's going to use; call it NEW.  It then runs through the entire
>database changing everything currently owned by OLD to be owned by NEW instead.
>*Also*, it changes everything currently owned by NEW to be owned by #-1.  I
>think my reasoning here had to do with not wanting OLD to silently `inherit' a
>bunch of stuff that used to be owned by the late lamented NEW (now invalid).
>
>It's clear that this is not intuitive, or even consistent (since nothing else
>can set an owner to #-1), but maybe the reasoning behind is worth saving in
>*some* form.  Ideas?  It could, for example, switch the currently owned by NEW
>things to be owned by the now-invalid OLD...
>
>        Pavel

Ok, call me paranoid but... a bunch of verbs assume that caller_perms() ==
$nothing being a server call and do a set task_perms(player)...

So having some verb owned by #-1 might introduce some security holes...

Just my 0.02 ZAp...

The Computer is your Friend. Keep your laser handy. Trust no one.

Jan-U-SSS-1

---
"640K ought to be enough for anybody."
    -- Bill Gates, 1981

• Prev: Re: <no subject>
• Next: 0-based indices for verbs in old code...
• Index: MOO-cows
• Thread: MOO-cows