Hi Thanks for the example. It helps to clarify the problem. The definition of the flow has a problem:
@Produces @FlowDefinition public Flow defineFlow(@FlowBuilderParameter FlowBuilder flowBuilder) { String flowId = "flux"; flowBuilder.id("", flowId); flowBuilder.viewNode(flowId, "/" + flowId + "/" + flowId + ".xhtml"). markAsStartNode(); /* flowBuilder.returnNode("returnFromFlux"). fromOutcome("#{flowScopedBean.return}");*/ return flowBuilder.getFlow(); } In this part: flowBuilder.viewNode(flowId, "/" + flowId + "/" + flowId + ".xhtml"). markAsStartNode(); you are using the flow id as a viewNodeId. That confuses the navigation algorithm because the spec says that an outcome can be a flowId, and in that sense a flowId is a global identifier that cannot be reused. The fix is just set another id: flowBuilder.viewNode("start", "/" + flowId + "/" + flowId + ".xhtml"). markAsStartNode(); The example only requires some small changes in FlowScopedBean and that's it. regards, Leonardo Uribe 2014-02-13 6:03 GMT-05:00 l.pe...@senat.fr <l.pe...@senat.fr>: > On 13/02/2014 01:44, Leonardo Uribe wrote: > [...] > > >> This is an edge case. Theorically it should work. Could you please provide >> the >> stack trace, so I can investigate it further?. > > > Please find attached a mini maven project reproducing the «bug» (or maybe a > PBKAC there :-) ). > > If you just build and debug it, you will be in the situation I just > described. > > If you disable the Flux description bean (I just comment the @Produces and > @FlowDefinition annotations to do that) and rename > > src/main/webapp/flux/flux-flow-empty.xml > or > src/main/webapp/flux/flux-flow-simple.xml > > to > src/main/webapp/flux/flux-flow.xml > > you will be able to test the other situations I described. > > Thanks again ! > > > Ludovic > > > | > | AVANT D'IMPRIMER, PENSEZ A L'ENVIRONNEMENT. > |