Re: [apps-discuss] Feedback on draft-ietf-appsawg-json-patch-00

["Murray S. Kucherawy" <[e-mail]>]

> -----Original Message-----
> From: [e-mail] [mailto:[e-mail]] On Behalf Of Mike Acar
> Sent: Friday, March 02, 2012 2:16 PM
> To: [e-mail]
> Subject: Re: [apps-discuss] Feedback on draft-ietf-appsawg-json-patch-00
> 
> >> So to avoid depending upon undefined behavior, an implementation of
> >> Patch would have to parse the Pointer, remove the "target" value from
> >> the end, and then traverse the JSON objects to find the parent of the
> >> target.
> >
> > These specifications do very little to define how to handle errors
> > beyond terminating processing and indicating an error condition.
> 
> I don't think Pointer can even say this MAY be an error: that's one of
> the semantics an application assigns to Pointers (cf
> http://www.ietf.org/mail-archive/web/apps-
> discuss/current/msg04411.html).
> 
> For Patch, in my opinion it MUST be an error if a Pointer reference
> token can't be found in an object or array - but the consequence of
> that is that pointers have to point to parents.

+1.  This points (hah) back to the pointer document thread: That document can leave handling of failed pointer resolution explicitly undefined, but the application has to say what semantics are to be applied in such cases.  For the patch document, it MUST be an error.

-MSK