jackrabbit-oak-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jukka Zitting <jukka.zitt...@gmail.com>
Subject Re: [NEWS] Oak newsletter (2nd edition)
Date Mon, 17 Sep 2012 13:11:34 GMT

On Fri, Sep 14, 2012 at 4:47 PM, Michael Dürig <mduerig@apache.org> wrote:
> A bit late here is the 2nd edition of the almost weekly newsletter on
> what is going on in Apache Jackrabbit Oak.

Nice work, keep it coming!

> - Jukka started to document the internal node state model used in Oak
> core. This is a great place to start for newcomers:
> http://svn.apache.org/viewvc/jackrabbit/oak/trunk/doc/nodestate.md?view=markup

I'm using Markdown, which makes the documents render nicely on github:
Later on if (or when) we migrate the Jackrabbit and Oak web sites to
the Apache CMS we'll be able to use the Markdown sources there as

There are still plenty of TODOs and a number of areas that probably
should have their own separate articles. Feel free to fill in or just
add more TODOs on areas that you think should be documented.

> - A MongoDB based Microkernel implementation has been contributed by
> Philipp Marx. This is currently under review but we expect it to make
> its way into the code base.
> See https://issues.apache.org/jira/browse/OAK-293

The contribution has just hit svn in revision 1386591. For now you'll
need the -Pmongomk profile to enable the MongoMK as a part of the
build, but I'm hoping to integrate the code more tightly with the rest
of the build over the next few days.

> - Basics of node type validation are in place now. There are still
> some loose ends however.
> See
> https://issues.apache.org/jira/browse/OAK-66?focusedCommentId=13454846&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13454846


The better validation causes some trouble with the fake
same-name-sibling code I'm using in some places. I'll see if I can fix
such code or if we should at least temporarily relax validation of
such cases.

> - There is a new validator for checking uniqueness of the jcr:uuid
> property. Some of this work is duplicated by the indexing and query
> mechanism. We will probably have to unify these further down the line.
> See:
> http://svn.apache.org/viewvc/jackrabbit/oak/trunk/oak-core/src/main/java/org/apache/jackrabbit/oak/plugins/unique/

Agreed. The main reason why I went for a separate index for now was
that the uniqueness check needs to be able to not just update the
index but also reject commits that violate the uniqueness constraint.

Also on this topic, currently the uniqueness check only uses a single
global index. A more flexible solution would be to allow uniqueness
constraints to be specified on a per-subtree basis, possibly with an
oak:unique mixin type associated with related configuration under an
./oak:unique child node. For example, the root node of a workspace
would define the usual uniqueness constraint on jcr:uuid properties,
while an application could use the same mechanism for
application-specific uniqueness constraints on selected subtrees.


Jukka Zitting

View raw message