Save position of Edges

[glsp-bot]

I currently work on making the edge routing consistent. Right now edges can be moved around but when you change the position of another object they move back to the position the autolayouter assigned to them. I looked into how this is handled for nodes, where this is not happening. There (ShapeImpl) i found this code for setting the position of Shapes.

```public void setPosition(GPoint newPosition) {
        if (newPosition != position) {
            NotificationChain msgs = null;
            if (position != null)
                msgs = ((InternalEObject)position).eInverseRemove(this, EOPPOSITE_FEATURE_BASE - EnotationPackage.SHAPE__POSITION, null, msgs);
            if (newPosition != null)
                msgs = ((InternalEObject)newPosition).eInverseAdd(this, EOPPOSITE_FEATURE_BASE - EnotationPackage.SHAPE__POSITION, null, msgs);
            msgs = basicSetPosition(newPosition, msgs);
            if (msgs != null) msgs.dispatch();
        }
        else if (eNotificationRequired())
            eNotify(new ENotificationImpl(this, Notification.SET, EnotationPackage.SHAPE__POSITION, newPosition, newPosition));
    }```

This is called when a changeBounds Action is called.

The Action for moving references is rerouteConnection i believe, so i recreated the changeBounds process for it. But i do not seem to find anything on how to tailor the above code, for edges, since edges contain of multiple GPoints, which Shapes do not.

My question would be if this has been done before or if anyone could give me some insight on how this should be handled for edges.

Thanks in advance.

original thread by simongraband

[CamilleLetavernier]

Edges have a read-write list of bendpoints (GPoint[*]). So you can do something like this:

```GPoint start = GraphUtil.point(x, y);
GPoint end = GraphUtil.point(x2, y2);
List<GPoint> allPoints = new ArrayList<>();
allPoints.add(start);
... add intermediate bendpoints
allPoints.add(end);

edge.getBendpoints().clear();
edge.getBendpoints().addAll(allPoints);```

[CamilleLetavernier]

And actually, since the RerouteConnection action already contains the new list of points, it's probably sufficient to do a ...clear() ...and addAll(action.getRoutingPoints())

[tortmayr]

This problem has already bin addressed in a WIP PR that didn't got merged (yet): https://github.com/eclipsesource/graphical-lsp/pull/376/files. So you could have a look at the code there and use an adaptation of that.The ChangeRoutingPointsAction and its handlers are particular interesting

[tortmayr]

I think the ChangeBoundsTool needs to be adapted for this. In addition, to sending a ChangeBoundsAction to the server it also has to send a RerouteConnection action for each connected edge

[glsp-bot]

[simongraband]

Ok thanks for the help. The saving of the points is now working. However i still have two issues, that need some work.
First the changes seem not to apply when i reload the page. I think i need to adjust the client for it to work, since on the server side loadModel() seems to return the points.
Second is the trigggering of RerouteConnection actions from the ChangeBoundsAction, which i think also needs to be done in the client. Am I right here?

Now my question that should help me solve both of the issues is, where are those two Actions(requestModel and changeBounds) triggered on the client?

Edit: the first one seems to work now.

[tortmayr]

Yes you are right, the RerouteConnection has to be triggered on the client side. There is already a pull request on the new eclipse-glsp infra that provides exactly that: eclipse-glsp/glsp-client#25 . The relevant code is in change-bounds-tools.ts (if a change bounds action is triggered also the routingpoints for incoming and outgoing edges of the element are sent to the server)

[tortmayr]

Unfortunatley ecore-glsp is still based on the old GLSP infrastructure (https://github.com/eclipsesource/graphical-lsp) which is no longer updated. You have two options here:

[tortmayr]

a) You could wait until ecore-glsp gets migrated to the new eclipse-glsp infrastructure or

[tortmayr]

b) You bind a customized change-bounds-tool in the ecore-glsp client

[glsp-bot]

[simongraband]

@tortmayr Since i now have the new eclipse-glsp infrastructure: How am i able to get the ElementID of a ChangeRoutingPointsOperation? There seems to be no Option to get the ID. This was previosly possible by calling getConnectionElementId() on a RerouteConnectionOperation.

[tortmayr]

The ChangeRoutingPointsOperation can now handle routing point changes on multiple elements at once. The action now has a list of ElementAndRoutingPoints which is a datastructure that contains the ElementID and the updated routing points. So you have to iterate through this list to access the element id.

[tortmayr]

for reference see also: https://github.com/eclipse-glsp/glsp-server/blob/master/plugins/org.eclipse.glsp.server/src/org/eclipse/glsp/server/operationhandler/ChangeRoutingPointsHandler.java

[glsp-bot]

[simongraband]

Is the ChangeRoutingPointsOperation triggered the same way as the RerouteConnectionOperation? I updated the my method so it iterates over the list of ElementAndRoutingPoints, like the following:

```private void rerouteEdge(ChangeRoutingPointsOperationAction action, EcoreModelIndex index) {
        for(ElementAndRoutingPoints element : action.getNewRoutingPoints()){
            index.getNotation(element.getElementId(), Edge.class)
                .ifPresent(notationElement -> changeEdgePoints(notationElement, element.getNewRoutingPoints()));
        };
    }```

I changed nothing after this method, since there were no changes to how the data is stored in the notation model right? With the old operation i had it working, but with the new one i cannot get the data to be written in the .enotation file. Do you know of any changes that might affect this or is the Operation somehow triggered differently?

[tortmayr]

The operation should be triggered in the same way as before. The only difference is that it's also triggered when moving a node (to update all attached edges). Your code looks code as well, so in theory this should work just fine. If you provide me with a link to your branch I can have a quick look at your changes.

[glsp-bot]

[simongraband]

https://github.com/jonny3576/ecore-glsp/tree/flexibleEdgeRouting%2360

[tortmayr]

Looks like you did forget to bind EcoreRerouteEdgeOperationHandler in the EcoreGLSPModule bindOperationHandlers() method

[glsp-bot]

[simongraband]

Ah thanks. There must have been a problem while merging the new versions. I now got the code running and everything is working as intended. However when moving a node with a edge connected i get the following error:

```Model element not found! ID: c76dc38e-603f-4921-97ed-e0a900ffb1df org.eclipse.glsp.api.jsonrpc.GLSPServerException: Model element not found! ID: c76dc38e-603f-4921-97ed-e0a900ffb1df at org.eclipse.glsp.api.jsonrpc.GLSPServerException.getOrThrow(GLSPServerException.java:58) at org.eclipse.glsp.api.utils.LayoutUtil.applyBounds(LayoutUtil.java:50) at com.eclipsesource.glsp.ecore.handler.EcoreComputedBoundsActionHandler.execute(EcoreComputedBoundsActionHandler.java:55) at org.eclipse.glsp.server.actionhandler.AbstractActionHandler.execute(AbstractActionHandler.java:45) at org.eclipse.glsp.server.actionhandler.DIActionProcessor.dispatch(DIActionProcessor.java:44) at org.eclipse.glsp.api.action.ActionProcessor.process(ActionProcessor.java:35) at org.eclipse.glsp.api.action.ActionProcessor.process(ActionProcessor.java:46) at org.eclipse.glsp.server.jsonrpc.DefaultGLSPServer.process(DefaultGLSPServer.java:96) at jdk.internal.reflect.GeneratedMethodAccessor3.invoke(Unknown Source) at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.base/java.lang.reflect.Method.invoke(Method.java:567) at org.eclipse.lsp4j.jsonrpc.services.GenericEndpoint.lambda$null$0(GenericEndpoint.java:65) at org.eclipse.lsp4j.jsonrpc.services.GenericEndpoint.notify(GenericEndpoint.java:152) at org.eclipse.lsp4j.jsonrpc.RemoteEndpoint.handleNotification(RemoteEndpoint.java:220) at org.eclipse.lsp4j.jsonrpc.RemoteEndpoint.consume(RemoteEndpoint.java:187) at org.eclipse.lsp4j.jsonrpc.json.StreamMessageProducer.handleMessage(StreamMessageProducer.java:194) at org.eclipse.lsp4j.jsonrpc.json.StreamMessageProducer.listen(StreamMessageProducer.java:94) at org.eclipse.lsp4j.jsonrpc.json.ConcurrentMessageProcessor.run(ConcurrentMessageProcessor.java:113) at java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515) at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264) at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128) at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628) at java.base/java.lang.Thread.run(Thread.java:830) Caused by: java.util.NoSuchElementException: No value present at java.base/java.util.Optional.get(Optional.java:141) at org.eclipse.glsp.api.jsonrpc.GLSPServerException.getOrThrow(GLSPServerException.java:56) ... 22 more```

But the code is still working. I am not quite sure how the LayoutUtil works but it seems like the problem is there. @tortmayr Do you have any idea on how to alter the call of LayoutUtil to not get this error?

[tortmayr]

I did a little testing and it looks like this is a problem with the id generation in the ClassifierNodeFactory. I did a quick fix which seems to work for me and created a PR into your branch for this: jcebel/ecore-glsp#25

[glsp-bot]

[simongraband]

Thanks that fix seems to work.