Preconditions and the responsedescription element

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Preconditions and the responsedescription element

Werner Donné

Hi,

Section 1.6 of RFC 3253 states that in case of a multi-status
response, the error element, in which the precondition elements
are placed, appears in the responsedescription element. The
content model of the latter, however, is (#PCDATA).

Regards,

Werner.
--
Werner Donné  --  Re
Engelbeekstraat 8
B-3300 Tienen
tel: (+32) 486 425803 e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Preconditions and the responsedescription element

Julian Reschke

Werner Donné schrieb:
> Hi,
>
> Section 1.6 of RFC 3253 states that in case of a multi-status
> response, the error element, in which the precondition elements
> are placed, appears in the responsedescription element. The
> content model of the latter, however, is (#PCDATA).

Yes, and there's really no problem with that, because RFC2518 also says
that elements can be extended.

Now, when discussing RFC2518bis a few WG members claimed that extending
elements with PCDATA model with new child elements is not a good idea,
and thus RFC2518bis has changed the embedding of <error> elements. This
may be less intrusive, but it *is* an incompatible change to RFC3253,
thus, feedback appreciated (see, for instance,
<http://greenbytes.de/tech/webdav/draft-ietf-webdav-rfc2518bis-15.html#rfc.section.14.24>).

Best regards, Julian

Reply | Threaded
Open this post in threaded view
|

Re: Preconditions and the responsedescription element

Werner Donné

Adding the error element to the response element is indeed
nicer, but it would have been safer to add it after all
"old" elements in the content model, as it was done for the
location element. Current clients which don't ignore unknown
elements that intervene in an existing content model different
from "ANY", will break. It is likely that such clients simply
consume the elements they know in the known order and then stop.

Regards,

Werner.

Julian Reschke wrote:

> Werner Donné schrieb:
>> Hi,
>>
>> Section 1.6 of RFC 3253 states that in case of a multi-status
>> response, the error element, in which the precondition elements
>> are placed, appears in the responsedescription element. The
>> content model of the latter, however, is (#PCDATA).
>
> Yes, and there's really no problem with that, because RFC2518 also says
> that elements can be extended.
>
> Now, when discussing RFC2518bis a few WG members claimed that extending
> elements with PCDATA model with new child elements is not a good idea,
> and thus RFC2518bis has changed the embedding of <error> elements. This
> may be less intrusive, but it *is* an incompatible change to RFC3253,
> thus, feedback appreciated (see, for instance,
> <http://greenbytes.de/tech/webdav/draft-ietf-webdav-rfc2518bis-15.html#rfc.section.14.24>).
>
>
> Best regards, Julian
>
>

--
Werner Donné  --  Re
Engelbeekstraat 8
B-3300 Tienen
tel: (+32) 486 425803 e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Preconditions and the responsedescription element

Julian Reschke

Werner Donné schrieb:
> Adding the error element to the response element is indeed
> nicer, but it would have been safer to add it after all
> "old" elements in the content model, as it was done for the
> location element. Current clients which don't ignore unknown
> elements that intervene in an existing content model different
> from "ANY", will break. It is likely that such clients simply
> consume the elements they know in the known order and then stop.

Well, those clients are broken anyway. I don't see any point in adapting
  to them.

Are you aware of clients with that problem?

Best regards, Julian

Reply | Threaded
Open this post in threaded view
|

Re: Preconditions and the responsedescription element

Werner Donné

No, I'm not. It is just a remark about a defensive way
to extend a content model.

Regards,

Werner.

Julian Reschke wrote:

> Werner Donné schrieb:
>> Adding the error element to the response element is indeed
>> nicer, but it would have been safer to add it after all
>> "old" elements in the content model, as it was done for the
>> location element. Current clients which don't ignore unknown
>> elements that intervene in an existing content model different
>> from "ANY", will break. It is likely that such clients simply
>> consume the elements they know in the known order and then stop.
>
> Well, those clients are broken anyway. I don't see any point in adapting
>  to them.
>
> Are you aware of clients with that problem?
>
> Best regards, Julian
>
>

--
Werner Donné  --  Re
Engelbeekstraat 8
B-3300 Tienen
tel: (+32) 486 425803 e-mail: [hidden email]