Cool bust of luck with the debugging session :-) I went and checked and the developers guide[1] had this to say: > Before you Start a word on Thread Safety > Because this is a dynamic application we ask you not to hack away at these > data structures as the system is running (to do so often freezes the screen). > Instead we ask you to assemble acommand that will be issued in the user > interface thread. > > uDig is highly threaded, as benefits a client application, ensuring that > "work" happens in the correct thread is a large benefit provided to you by > the GIS Application. > >
So let me know if we can make this warning appear in a more obvious location for the next developer :-) Jody [1] http://udig.refractions.net/confluence/display/DEV/02+Project+Map+and+Layer On 22/09/2010, at 6:00 PM, Stanislas Nanchen wrote: > Hi Jody, hi list, > > Thanks for the answer. I indeed did not use a command to make this change. > However the error still appears; debugging session ahead :) > > stan > > On 2010 Sep 21, at 22:33, Jody Garnett wrote: > >> I think you have probably just found a bug; since I don't myself have the >> problem I will need to ask you to spend some time in the debugger to see >> what is going on and create a patch. >> >> I am going to guess that the renderer is still trying to draw as the map is >> changed; we probably need to stop the renderer and then change the map? >> >> I am going to just confirm that you are using a command to make this change >> (a command is like a small runnable that you pass to the map; it will run >> the command to update the internal state in a threadsafe manner; this is how >> we avoid changing the state of the map while the renderer or other code is >> making use of it). >> >> Specifically calling any internal methods should be done only from inside a >> command. >> >> Jody >> >> On 21/09/2010, at 11:59 PM, Stanislas Nanchen wrote: >> >>> Hello everyone, >>> >>> At my company, we have embedded uDig for our map displaying purposes and I >>> have the following problem with the MapViewer >>> when setting a map for the second time via the method setMap. >>> >>> In setMap, if an old map is present, its renderManager is kind of >>> 'disposed' with the following two lines of code: >>> >>> (in MapViewer) >>> 170 renderManager.setMapInternal( null ); >>> 171 renderManager.setMapDisplay(null); >>> >>> In certain circumstances, unfortunately unclear, the old renderManagner >>> tries to refresh an image. This fails with an NPE in the following >>> line as its MapDisplay is null. >>> >>> (in CompositeRendererImpl.java) >>> 377 Dimension displaySize = >>> getContext().getMapDisplay().getDisplaySize(); >>> >>> the stack trace is the following: >>> >>> java.lang.NullPointerException >>> at >>> net.refractions.udig.project.internal.render.impl.CompositeRendererImpl.refreshImage(CompositeRendererImpl.java:377) >>> at >>> net.refractions.udig.project.internal.render.impl.CompositeRendererImpl.refreshImage(CompositeRendererImpl.java:296) >>> at >>> net.refractions.udig.project.internal.render.impl.RenderExecutorComposite$CompositeRendererJob.incrementalUpdate(RenderExecutorComposite.java:119) >>> at >>> net.refractions.udig.project.internal.render.impl.RenderExecutorComposite$CompositeRendererJob.startRendering(RenderExecutorComposite.java:166) >>> at >>> net.refractions.udig.project.internal.render.impl.RenderJob.run(RenderJob.java:213) >>> at org.eclipse.core.internal.jobs.Worker.run(Worker.java:55) >>> >>> The problem is quite puzzling as this NPE is not raised systematically. >>> Should I take any special precaution when setting a map to a MapViewer for >>> the second time in order to avoid this error? >>> >>> Thank you for your help. >>> Stanislas Nanchen. >>> _______________________________________________ >>> User-friendly Desktop Internet GIS (uDig) >>> http://udig.refractions.net >>> http://lists.refractions.net/mailman/listinfo/udig-devel >> >> _______________________________________________ >> User-friendly Desktop Internet GIS (uDig) >> http://udig.refractions.net >> http://lists.refractions.net/mailman/listinfo/udig-devel > > _______________________________________________ > User-friendly Desktop Internet GIS (uDig) > http://udig.refractions.net > http://lists.refractions.net/mailman/listinfo/udig-devel
_______________________________________________ User-friendly Desktop Internet GIS (uDig) http://udig.refractions.net http://lists.refractions.net/mailman/listinfo/udig-devel
