Re: NEW PREFERENCE - depth-noroot (request for expert review)

[Julian Reschke <[e-mail]>]

On 2017-01-12 12:44, Amos Jeffries wrote:
> On 13/01/2017 12:16 a.m., Julian Reschke wrote:
>> Hi there,
>>
>> draft-murchison-webdav-prefer-13 defines a new "prefer" directive (RFC
>> 7240) in
>> <https://tools.ietf.org/html/draft-murchison-webdav-prefer-13#section-4>.
>>
>> Please review and provide feedback.
>>
>> Best regards, Julian
>>
>
> After reading the preference section I am left wondering:
>
> 1) what does Appendix B.4 have to do with it?
>  there is no use of the preference in the examples.

My copy says Appendix B.1...

> 2) what happens when Depth:0 is given in the request with depth-noroot?

It says:

>    This preference is only intended to be used with WebDAV methods whose
>    definitions explicitly provide support for the Depth [RFC4918] header
>    field.  Furthermore, this preference only applies when the Depth
>    header field has a value of "1" or "infinity" (either implicitly or
>    explicitly).

So it would be ignored.

> 3) what happens when the Depth header is omitted in the request with
> depth-noroot?

Defaults apply, as stated above...

> 4) what methods this preference is actually valid for?
>  The depth-noroot text references Depth header existence *or implicit*,
> but the Depth header leaves it open and explicitly says any method can
> re-define Depth. That does not bode well for implementations getting
> interoperability correct. Particularly for the impliciat-Depth methods.

I don't see a problem here. The preference is defines for methods that 
use the Depth: request header field (including future ones). Clients can 
always check whether the preference was applied.

Best regards, Julian

(But thanks for checking!)