Been doing some further experiments on JRebel, so as a quick update, I
*think* it's doable, but requires a small enhancement to DataNucleus.

For further reading, see [1] and [2]

[1] https://issues.apache.org/jira/browse/ISIS-651
[2] http://www.datanucleus.org/servlet/jira/browse/NUCCORE-1104



On 10 January 2014 17:46, GESCONSULTOR <[email protected]> wrote:

> Many Thanks for moving forward this, Dan.
>
> It can really boost our productivity.
>
> For what I've read JRebel is the best way. It was just an alternative
> found.
>
>
> > El 10/01/2014, a las 17:54, Dan Haywood <[email protected]>
> escribió:
> >
> > On 27 December 2013 23:12, GESCONSULTOR - Óscar Bou
> > <[email protected]>wrote:
> >
> >>
> >> Hi to all.
> >>
> >> Dan, some days ago you commented the possibility to accelerate the
> >> workflow by integrating with JRebel.
> >>
> >> Just to mention, seems on the Ninja Framework they have achieved
> something
> >> similar as detailed in [1].
> >>
> >> This functionality is introduced at [2], where it references an article
> in
> >> [3].
> >>
> >> Perhaps it's a different approach to accelerate the Isis workflow.
> > Thanks for this, Oscar.
> >
> > However, in  [1], they say:
> >
> >
> > *You start Ninja’s SuperDevMode in a console. Then you edit a Java file
> in
> > your IDE and save it. Your IDE will then compile your Java file to a
> class
> > file. Ninja’s SuperDevMode recognizes that and restarts Ninja within a
> > second. You can then switch to your browser and verify that your changes
> > work at http://localhost:8080 <http://localhost:8080/> . *
> >
> >
> > So I don't think this will work, because Isis takes rather longer than a
> > second to build up its metamodel.
> >
> > I'm pretty certain that JRebel is the right way to go ... I spoke to one
> of
> > their evangelists at a conference in November, so reckon it'll work.  We
> > can invalidate Isis' metamodel ok (the
> > DeveloperUtilitiesServiceDefault#refreshLayout() method), the only real
> > unknown is how to invalidate the DN metamodel similarly.
> >
> > Let me go ask Andy Jefferson about that...
> >
> > Dan
> >
> >
> >
> >
> >>
> >> HTH,
> >>
> >> Oscar
> >>
> >>
> >>
> >>
> >> [1] http://www.ninjaframework.org/documentation/super_dev_mode.html
> >>
> >> [2] http://www.ninjaframework.org/documentation/getting_started.html
> >>
> >> [3]
> >>
> http://java.jiderhamn.se/2011/12/11/classloader-leaks-i-how-to-find-classloader-leaks-with-eclipse-memory-analyser-mat/
>

Reply via email to