CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Tobias Soloschenko
Hi guys, Martin mentioned to me that my last integration breaks the build of Wicket-7.x. As I saw master compiled fine. After a short research we found out that the CompoundPropertyModel works slightly different. Output in wicket-7.x: Output in master: http://www.google.de/test.jpg"/> Invo

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Sven Meier
Hi Tobios, > new ExternalImage("compoundImage",compoundPropertyModel.bind("url"))); this invokes the constructor with the Serializable argument, effectively wrapping the propertyModel inside a normal model. Have fun Sven On 06.12.2015 09:31, Tobias Soloschenko wrote: Hi guys, Martin menti

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Sven Meier
BTW to really make sure this works with CompoundPropertyModel, it would be nicer if the test would be more like org.apache.wicket.model.InheritedModelTest Regards Sven On 06.12.2015 10:04, Sven Meier wrote: Hi Tobios, > new ExternalImage("compoundImage",compoundPropertyModel.bind("url")));

cors attribute

2015-12-06 Thread Sven Meier
Hi, I've noticed that the "crossorigin" attribute is popping on several places now: - Image - ExternalImage - MediaComponent The first two share the enumeration, the latter has its own inner enum. Since CORS seems to be a general topic, why not move the enum into its own file? Writing of th

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Tobias Soloschenko
Hi Sven, so the behavior of wicket-7.x is right then? Because getObect receives the actual nested model and not src as model object? Then we have an error in master. kind regards Tobias > Am 06.12.2015 um 10:04 schrieb Sven Meier : > > Hi Tobios, > > > new ExternalImage("compoundImage",comp

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Sven Meier
Hi Tobias, I'm not sure: CompoundPropertyModel#bind() has a generic type parameter, perhaps this is treated differently in Java 7 vs. 8. Sven On 06.12.2015 10:57, Tobias Soloschenko wrote: Hi Sven, so the behavior of wicket-7.x is right then? Because getObect receives the actual nested mod

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Tobias Soloschenko
But this is not good and I think a Bug of this Model. It should behave the same. kind regards Tobias > Am 06.12.2015 um 11:46 schrieb Sven Meier : > > Hi Tobias, > > I'm not sure: CompoundPropertyModel#bind() has a generic type parameter, > perhaps this is treated differently in Java 7 vs. 8.

Re: cors attribute

2015-12-06 Thread Tobias Soloschenko
Hi Sven, I would suggest to externalize it and use a separate for MediaC and Image, because it might diff in feature changes. WDYT? kind regards Tobias > Am 06.12.2015 um 10:17 schrieb Sven Meier : > > Hi, > > I've noticed that the "crossorigin" attribute is popping on several places > now:

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Sebastien
Hi Tobias, I agree with Sven about the way of testing the CPM. If you change the test like this, it works fine: test-page: ImageSrc imageSrc = new ImageSrc(); imageSrc.setUrl("http://www.google.de/test.jpg";); Form form = new Form("form", new CompoundPropertyModel(imageSr

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Sven Meier
Hi all, the Java 8 compiler is just better inferrinf the right method signature. That's nothing Wicket can do about - besides avoiding constructor overloading. I've changed the test as Sebastien has proposed. Have fun+ Sven On 06.12.2015 14:14, Sebastien wrote: Hi Tobias, I agree with Sve

Re: cors attribute

2015-12-06 Thread Sven Meier
Hi Tobias, has "crossorigin" too, and

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Tobias Soloschenko
Great! Thanks a lot for the fix! kind regards Tobias > Am 06.12.2015 um 14:21 schrieb Sven Meier : > > Hi all, > > the Java 8 compiler is just better inferrinf the right method signature. > That's nothing Wicket can do about - besides avoiding constructor overloading. > > I've changed the te

Re: cors attribute

2015-12-06 Thread Tobias Soloschenko
Sounds good to me! If it is used at so many places! kind regards Tobias > Am 06.12.2015 um 14:40 schrieb Sven Meier : > > Hi Tobias, > > has "crossorigin" too, and

Re: CompoundPropertyModel works different in wicket-7.x and master

2015-12-06 Thread Tobias Soloschenko
Hi all, I am going to add some Information about the new components the next days. :-) kind regards Tobias > Am 06.12.2015 um 14:21 schrieb Sven Meier : > > Hi all, > > the Java 8 compiler is just better inferrinf the right method signature. > That's nothing Wicket can do about - besides avo

Re: Move wicket-dashboard to wicketstuff

2015-12-06 Thread Martin Grigorov
Hi Sebastien, I've added snippets like https://github.com/wicketstuff/core/commit/5574dd29a4a2a262006c4bed29cd2bb17e10d6f1 in all -examples modules. These properties are used by their respective Maven plugins, e.g. https://github.com/wicketstuff/core/blob/5574dd29a4a2a262006c4bed29cd2bb17e10d6f1/p