Re: [kitten] Some test registrations according to draft-ietf-kitten-gssapi-extensions-iana-08.txt

[Benjamin Kaduk <[e-mail]>]

On Thu, 30 Oct 2014, Shawn M Emery wrote:

> On 10/26/14 06:57 PM, Benjamin Kaduk wrote:
> > On Fri, 17 Oct 2014, Shawn M Emery wrote:
> >
> > > Could folks please review the example registry that we would like to
> > > include
> > > in the draft-ietf-kitten-gssapi-extensions-iana draft?
> > >
> > > Thanks,
> > >
> > > Shawn.
> > > --
> > > On 07/24/14 10:47 AM, Alexey Melnikov wrote:
> > > > Bindings: C
> > > > Registration type: Instance
> > > > Object Type: Context-Flag
> > > > Symbol Name: GSS_C_DELEG_FLAG
> > > > Binding of: deleg_state or deleg_req_flag
> > > > Constant Value/Range: 1
> > > > Description: On output (if set): Delegated credentials are available
> > > >               via the delegated_cred_handle
> > > >               parameter of GSS_Accept_sec_context/GSS_Init_sec_context.
> > Er, GSS_Init_sec_context does not have a delegated_cred_handle argument.
>
> Yes, this is confusing, but there are trade-offs with using abbreviated text
> and being thorough in description.  Would something like the following make
> this more clear?:
>
> Description: On output (if set): Delegated credentials are available
>              via the delegated_cred_handle parameter of GSS_Accept_sec_context
>
>              On input (if set): With the call to GSS_Init_sec_context,
> 	     delegate credentials to the acceptor

That is more clear, yes.  There are still more subtleties that are not
covered by that text (the flag can be set on a GSS_S_CONTINUE_NEEDED
return of GSS_Accept_sec_context, can be set on the initiator's returned
flags, etc.), but those are probably overkill for a registry like this.


> Note that I'm not aware of what field length constraints there are for
> registry entries.

IIRC the download links on the web are just XML files, but there is
probably some other internal representation.

-Ben