Re: Call for Adoption: SEARCH method

[Michael Douglass <[e-mail]>]

On 11/5/20 14:37, Julian Reschke wrote:
> Am 05.11.2020 um 19:05 schrieb Michael Douglass:
>> I also support not limiting the content type. I can imagine implementing
>> a search extension to existing XML based protocols.
>>
>> RFC5323 Section 3 say's
>>
>>     Clients can determine which query grammars are supported by an
>>     arbiter by invoking OPTIONS on the search arbiter.  If the resource
>>     supports SEARCH, then the DASL response header will appear in the
>>     response.  The DASL response header lists the supported grammars.
>>
>>     Servers supporting the WebDAV extensions [RFC3253 
>> <https://tools.ietf.org/html/rfc3253>] and/or [RFC3744 
>> <https://tools.ietf.org/html/rfc3744>]
>>     MUST also:
>>
>>     o  report SEARCH in the live property DAV:supported-method-set for
>>        all search arbiter resources, and
>>
>> ...
>>
>> Presumably a WebDAV compliant client knows whether or not SEARCH is
>> supported as a WebDAV service
>>
>> So:
>>
>> If it's in DAV:supported-method-set it's WebDAV SEARCH
>>
>> If it's in the "Accept-Search" Header Field it's the new SEARCH
>>
>> and you can't have both. No need to parse the content.
>> ...
>
> Yes, but...
>
> The content type in the payload is supposed to describe the query
> semantics. I would expect any new use of SEARCH to actually use a
> payload format that is more specific than text/xml or application/xml.
>
> Am I missing something here?

Sorry - I think I replied too high up in the thread.

Your later message said:

>
>> for backwards compatibility with existing WebDAV implementations, 
>> SEARCH requests that use the text/xml or application/xml content 
>> types MUST be processed per the requirements established by [RFC5323 
>> <https://tools.ietf.org/html/rfc5323>].
>
> I think this is too restrictive. If it’s not possible to relax the
> RFC5323 requirements, I would favor using REPORT instead.
> ...
>
> We can relax the requirement to apply only to */xml which has a document
> element in the "DAV:" namespace (see
> <https://greenbytes.de/tech/webdav/rfc5323.html#rfc.section.2.2.2>, or
> even restrict it to the two element names defined there).
I think my suggestion is a valid way of distinguishing between WebDAV 
search and the new search. Given the existing and proposed headers 
there's no need to parse the content to determine if it's a WebDAV search.

If it's a WebDAV search then that's handled however it's done now (or in 
the future).

If it's the new search then I don't see that we need to restrict the 
content to NOT have DAV: namespace elements.




>
> Best regards, Julian