Mailing-List: contact dev-help@jackrabbit.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@jackrabbit.apache.org
Message-ID: <20682287.1162556241816.JavaMail.root@brutus>
Date: Fri, 3 Nov 2006 04:17:21 -0800 (PST)
From: "Marcel Reutegger (JIRA)" <jira@apache.org>
To: dev@jackrabbit.apache.org
Subject: [jira] Commented: (JCR-616) Completeness/Freshness of Namespace
 Registry
In-Reply-To: <19470798.1162545616576.JavaMail.root@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

    [ http://issues.apache.org/jira/browse/JCR-616?page=comments#action_12446926 ] 
            
Marcel Reutegger commented on JCR-616:
--------------------------------------

1) Allow the SPI to return the prefix as an optional part of a QName when using a namespace that the client has not yet seen

I'm not sure I understand this correctly, does that mean the QName class should provide a getPrefix() which may or may not returns a prefix?

I would rather prefer 4) and say that this method must return at least the mapping of namespace uri to prefix mapping for QNames that have been used through this instance of RepositoryService since it was created. The jcr2spi would then just refresh it's namespace registry when it encounters an unknown namespace uri.

I'd like to avoid adding methods to the RepositoryService which overlap functionality that is already there ;)

> Completeness/Freshness of Namespace Registry
> --------------------------------------------
>
>                 Key: JCR-616
>                 URL: http://issues.apache.org/jira/browse/JCR-616
>             Project: Jackrabbit
>          Issue Type: Bug
>          Components: SPI
>            Reporter: Julian Reschke
>
> We need to define the requirements on completeness and freshness of RepositoryService.getRegisteredNamespaces().
> Right now the optimistic assumption seems to be that an SPI provider is able to report all namespaces that can occur in a repository beforehand. Even if it can do that (and I know of potential targets for SPI that simply can't), this seems to be quite a waste of time if these namespace prefixes aren't actually used later on.
> Furthermore, in SPI namespace prefixes aren't really relevant, except to enable the transient layer to return "meaningful" prefixes instead of automatically generated ones.
> Therefore my propoal would be to:
> 1) Clarify that the Map returned from getRegisteredNamespaces() isn't required to be complete,
> 2) Enhance JCR2SPI to auto-generate prefixes when it encounters namespaces not in the registry.
> I expect this to also affect RepositoryService.(un)registerNamespace(...), but let's discuss the underlying issue first...

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira