Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [udig-devel] AbstractFeatureStore vs. UdigFeatureStore

You are right, I use the layer before it is added to the map. 
 
I have some temporary/memory based  layer and I have layer (sets of layer) of different wms server. The user can select one wms server out of a list. The layers are added to the map just before the map is opened. If the user changes the wms server I remove all layers from the map and add them (changing the wms layer) in the correct order. This is also done since if you close the mapeditor and opens it again, the temporary layers have been removed from the map. Perhaps its just a bug in 1.1RC8.
 
Since it works with 1.1RC8 I thought this would be ok.
 
I discoverd that it is WrapFeatureStore which intercepts the getRessource call of a layer and adds a UDIGFeatureStore. This interception is declared in plugin.xml of net.refractions.udig.project. I see that this extension exists in 1.1RC8 and in 1.1RC10 but it is only used in 1.1RC10: in RC8 the debugger tells me that UDIGFeatureStore is never used... and I get no exception.
If I remove this extension from the plugin.xml in RC10 the nullpointer exception disappears.
 
tony roth
 

Something is funny here because the mapEditor shouldn't have to be open for the layer to have a map.  Indeed as soon as the layer is added to the map it has a map.  I'll admit that there is some dependencies.  For example a layer isn't usable without a map.  There is a construct called Layer interceptors that run when a layer is created but I don't see how that could cause a problem...  

Are you sure that you are not editing a layer before adding it to a map?

Jesse


On Jul 19, 2007, at 5:19 AM, tony.roth@xxxxxx wrote:

After switching to SDK 1.1RC10 I got an odd nullpointer exception at removing features of a feature store.
I use a temporary / memory based layer which is wrapped by a class called LayerContainer. This class has some convenience methods for changing the (SLD) style and for adding / removing objects like this one:
private void removeAllContent(){
  FeatureStore fs = layer.getGeoRessource().resolve(FeatureStore.class,null);
  fs.removeFeaures(Filter.NONE);
}
Using SDK1.1RC8 the object "fs" is an AbstractFeatureStore. Using SDK1.1RC10 it is an UDIGFeatureStore which calls the method "setTransactionInternal" in every method which changes the set of features. In setTransactionInternal this line fails in my application:
Transaction transaction = ((EditManager)) layer.getMap().getEditManager()).getTransaction();
layer.getMap() returns null
getMap() invokes getMapInternal() invokes getContextModel() which returns the eContainer (type: InternalObject) which is null
I can only guess what happens: The the layer is manipulated (adding / removing features) before the map is opened in a MapEditor the first time. This is done since the user can "add objects to map" in a perspective which has no visible map / no MapEditor. The class LayerContainer is notified via eventmanagement and adds / removes features. Under 1.1RC8 this doesn't matter since another FeatureStore is used.
(The map object is created before the layer is instanced.)
what can I do?
tony roth
_______________________________________________
User-friendly Desktop Internet GIS (uDig)

_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel


Back to the top