| Summary: | add fork observer race condition | ||
|---|---|---|---|
| Product: | frysk | Reporter: | Sami Wagiaalla <swagiaal> |
| Component: | general | Assignee: | Unassigned <frysk-bugzilla> |
| Status: | NEW --- | ||
| Severity: | normal | CC: | pmuldoon, swagiaal |
| Priority: | P1 | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Host: | Target: | ||
| Build: | Last reconfirmed: | ||
| Bug Depends on: | |||
| Bug Blocks: | 1516, 1586, 1654 | ||
public Action updateForked(Task task, Proc child) {
//...
child.observableTaskAdded.addObserver(new Observer(){
public void update(Observable arg0, Object task) {
Task myTask = (Task) task;
myTask.requestAddForkedObserver(customObserver);
}
});
//...
This does not actually work... it just does not throw an exception.
I tested this by trying to trace the execution tree of firefox;
only the very first fork is caught.
(title changed, it isn't an unhandled state) is this still broken? known problems with the observer code have been fixed frysk.proc.TestTaskForkedObserver.testTaskForkedObserver seems to test for this and passes. |
In the update function of a fork when one tries to add an observer to the child process the state machine does not handel the transition. This however works: public Action updateForked(Task task, Proc child) { //... child.observableTaskAdded.addObserver(new Observer(){ public void update(Observable arg0, Object task) { Task myTask = (Task) task; myTask.requestAddForkedObserver(customObserver); } }); //...