Re: [Geoserver-devel] GSIP 52 - Catalog / Config DAO Refactor - Call for Vote

[Justin D. <jde...@op...>]

I guess I should have specified it more. Although historically there has
been much discussion about these issues and this approach back when
geosolutions took this on. But I am happy to clarify anything in particular
and if you can narrow down what I should document in more detail I am happy
to do so.

On Tue, Oct 19, 2010 at 9:45 AM, Andrea Aime
<and...@ge...>wrote:

> On Tue, Oct 19, 2010 at 3:54 PM, Justin Deoliveira <jde...@op...>
> wrote:
> > Hi all,
> > I am not sure if i made it clear when i posted the GSIP last week but i
> > would like to move it forward through voting. So far I have no votes on
> this
> > proposal from any PSC members.
>
> Hey there... sorry for not replying, I did not do because I did not find
> an easy to review the proposal, so I guessed I had to actually read the
> patch
> line by line to figure it out.
>
> I see the new interfaces (one is without javadoc comments btw) and I had
> a quick look at the code but the structure of the changes escapes me when
> just glancing at the code.
> (btw, why is every access to the DAO synchronized inside the catalog,
> would it be the same
>  to have the dao implementor synchronize every method if the dao is
> not thread safe?)
>

I don't think every access to the dao is synchronized. Only ones that have
to do multiple related operations. Like for instance removing a workspace or
adding a workspace also has the side affect of setting the default workspace
if it is not set. Within the old catalog implementation access to the
workspace was synchronized for this transaction.

The first priority in my mind is backwards compatibility, so i kept things
as close to the way they were before as possible. However i realize this is
an issue. But its one I think that will have to be solved incrementally.
Perhaps we need the notion of a transaction? Not too sure.

>
> I won't have time to really look into the diffs until next weekend, I
> guess having
> a bit more of description would expedite the review (things like the
> dao synch above for example).

You know, something that sits in the middle between "let's have a DAO
> approach"
> and actually going line by line through the patch.
>

Ok... i am not exactly sure what that is... perhaps going through the old
email archives and jira issue discussions for context would help?

>
> But if you can wait till the weekend I can actually sit down head
> clear and look into
> it for good as is, without any change or extra description.
>

Sure that is fine. And if all is the same I prefer that someone review the
patch in detail rather than I go through and try to document and summarize
all the implementation details. But for things that people want more
explanation for i am happy to do so.

>
> Cheers
> Andrea
>
> -----------------------------------------------------
> Ing. Andrea Aime
> Senior Software Engineer
>
> GeoSolutions S.A.S.
> Via Poggio alle Viti 1187
> 55054  Massarosa (LU)
> Italy
>
> phone: +39 0584962313
> fax:     +39 0584962313
>
> http://www.geo-solutions.it
> http://geo-solutions.blogspot.com/
> http://www.linkedin.com/in/andreaaime
> http://twitter.com/geowolf
>
> -----------------------------------------------------
>



-- 
Justin Deoliveira
OpenGeo - http://opengeo.org
Enterprise support for open source geospatial.