| Id : |
363 |
| Type : |
IMPROVEMENT |
| Reporter : |
Benjamin Jeanty-Ruard |
| Assignee : |
|
| Priority : |
MAJOR |
| Creation Date : |
07 Jan 16 02:56 |
| Last Update Date : |
07 Jan 16 02:56 |
| Resolution : |
UNRESOLVED
|
In the PipeLineManager, when a data set, a mapper or an actor are added, all the connectivity is build from scratch.
More over the connectivity computation is not optimized at all.
Indeed, if lots of data set/mapper/actor are added, each of them rebuilt the connectivity between the data set and the mapper. This is a problem, specially for EDGE.
A solution is to optimize the method addDataSet, addMaper, … This method call the method {@link VtkObjectListModel#addVtkObject(CassandraObject)}. This method calls the method fireIntervalAdded where the listeners are notified with the method {@link VtkPipeLineGraphModel#intervalAdded(javax.swing.event.ListDataEvent)}. This method updates the connectivity and must be optimized. Indeed, all the connectivity is clear and build again.
A solution is to use the input parameter of the {@link VtkPipeLineGraphModel#intervalAdded(javax.swing.event.ListDataEvent)} method. Then, you process only the connectivity about the data set. To finish, it is not necessary to clean all the data set connectivity in the current case (add a data set at the end of the data set list), but only update the good one.
Post a comment :