Project

General

Profile

Bug #5200

WorkflowRun and Reporting artifacts no longer put into run-kars

Added by Derik Barseghian almost 9 years ago. Updated over 8 years ago.

Status:
Resolved
Priority:
Normal
Category:
reporting
Target version:
Start date:
10/14/2010
Due date:
% Done:

0%

Estimated time:
Bugzilla-Id:
5200

Description

A run-kar exported from the Workflow Run Manager on trunk no longer contains the workflowRun and roml and rio artifacts. I did some testing of various revisions of trunk reporting, and it looks like the problem began with r25808: "switch to Weak References in singleton classes and add cleanup when a window is closed, which is for bug 5095".

Here's how I tested each revision:
------
  • cd kepler.modules/build-area;
  • rm -rf ~/KeplerData; ant clean-all;
  • cd ..;
  • svn update -r [rev] *;
  • cd build-area; ant run
  • create a simple sdf workflow and report, execute, export the run from the workflow run manager, and then look inside the run-kar using either the gui or by unzipping the kar from the command line.
    ------

Here are the results for different revisions:
r26078 - not working - kar only contains wf
r25841 - not working - kar only contains wf
r25840 - not working - kepler locks up during run export
r25808 - not working - kepler locks up during run export
r25807 - working - kar contains all items

History

#1 Updated by Derik Barseghian over 8 years ago

Changing the WeakHashMap back to a HashMap in ObjectManager fixes this bug -- the Reporting karEntryHandlers were unable to get the run out of OM because it had been discarded from the WeakHashMap. But instead of doing this, now that Reporting depends on WorkflowRunManager, I'm going to see about refactoring to get the run from WRM instead of OM.

#2 Updated by Derik Barseghian over 8 years ago

I wasn't able to keep the WeakHashMap, and with r26239 I've changed it back to a HashMap. In r26239 I also changed the reporting kar entry handlers to get the run from WRM instead of OM, but they still require getting the workflow from the OM, and with the WeakHashMap this has (sometimes) been prematurely discarded.
See http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5095 for more.

#3 Updated by Redmine Admin over 6 years ago

Original Bugzilla ID was 5200

Also available in: Atom PDF