Re: ??: Does mixin effect search result?

[Marcel Reutegger <ma...@gmx.net>]

Alexander Klimetschek wrote:
> 2008/6/16 Yonder <zy...@yahoo.com.cn>:
>> I'm weak in English, but afraid that mix is also allowed:
>>
>> mixin_opt ::= "mixin" | "mix" | "m"
> 
> Oh, sorry, you are right. Your cnd is valid. I was so used to line
> breaks in node type definitions that I thought it is required ;-)
> 
> But the actual problem is the element() function: it only allows for
> primary node types in the second argument, not for mixins. So the
> query for the mixin ocm:file2 cannot find any node. See Section
> 6.6.4.13 "element() Node Test" in the JCR spec.

looks like jackrabbit is not in line with the spec, because jackrabbit does 
consider mixin types in the element() node test.

there's something else that is strange. ocm:file2 is defined as a mixin, but is 
later used in the code as a primary type to create a node. shouldn't that be 
prevented?

regards
  marcel

Re: ??: Does mixin effect search result?

[Alexander Klimetschek <ak...@day.com>]

On Mon, Jun 16, 2008 at 10:47 AM, Marcel Reutegger
<ma...@gmx.net> wrote:
> looks like jackrabbit is not in line with the spec, because jackrabbit does
> consider mixin types in the element() node test.

Good to know! I started a wiki page for documenting these things:

http://wiki.apache.org/jackrabbit/JcrSpecImplementation

Would be cool if we could fill this page with more information, at
least when such a topic comes up on the mailing lists.

Regards,
Alex

-- 
Alexander Klimetschek
alexander.klimetschek@day.com

Re: ??: Does mixin effect search result?

[Stefan Guggisberg <st...@day.com>]

On Mon, Jun 16, 2008 at 10:47 AM, Marcel Reutegger
<ma...@gmx.net> wrote:
> Alexander Klimetschek wrote:
>>
>> 2008/6/16 Yonder <zy...@yahoo.com.cn>:
>>>
>>> I'm weak in English, but afraid that mix is also allowed:
>>>
>>> mixin_opt ::= "mixin" | "mix" | "m"
>>
>> Oh, sorry, you are right. Your cnd is valid. I was so used to line
>> breaks in node type definitions that I thought it is required ;-)
>>
>> But the actual problem is the element() function: it only allows for
>> primary node types in the second argument, not for mixins. So the
>> query for the mixin ocm:file2 cannot find any node. See Section
>> 6.6.4.13 "element() Node Test" in the JCR spec.
>
> looks like jackrabbit is not in line with the spec, because jackrabbit does
> consider mixin types in the element() node test.
>
> there's something else that is strange. ocm:file2 is defined as a mixin, but
> is later used in the code as a primary type to create a node. shouldn't that
> be prevented?

yes, it should. i just checked, currently it doesn't. i'll open a jira issue.

cheers
stefan

>
> regards
>  marcel
>