jackrabbit-oak-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael Dürig <mic...@gmail.com>
Subject Re: oak-spi: NodeStore, NodeStateBuilder and CommitHook
Date Wed, 16 May 2012 11:22:59 GMT

See

> Author: mduerig
> Date: Wed May 16 11:21:08 2012
> New Revision: 1339116

for a completely different approach. Very raw still though...

Michael

On 14.5.12 17:43, Michael Dürig wrote:
>
> Hi,
>
> We had similar discussions before [1] and it keeps popping up all over
> again: The NodeStore interface and friends all operate on NodeState and
> as I said before, I don't think there is a reasonable way to implement
> this. Having NodeState all the places is too general. Have a look at my
> commits related to OAK-100. The way things are implemented there is by
> lying: NodeStore.getBuilder() and NodeStore.setRoot() only accept very
> specific instances of NodeStates as their arguments and throw exceptions
> on all other instances. I think this is wrong and we should use more
> exact types for NodeStore (and related) interfaces instead of throwing
> exceptions at runtime.
>
> Michael
>
>
> [1] http://markmail.org/message/panl3wxfekvmcfyw

Mime
View raw message