ComputationState running ...forever

May 2, 2009 at 7:12 PM
Edited May 2, 2009 at 7:15 PM
I am using the CyclePoppingRandomTreeAlgorithm and am monitoring State changes. I add a RandomTreewithRoot and sometimes it's State is Running and it just sits there forever. I can't even error handle it to see if it's broke. No error output or anything.
May 2, 2009 at 9:20 PM
Please provide a repro program for this issue.
May 3, 2009 at 3:33 AM
Edited May 3, 2009 at 3:34 AM
I wanted to ask a question first, is there a way I can set the end vertex or tree end instead of just the root beginning? I think it is erroring because the root vertex becomes separated from 0,0. And the solution always ends at 0, 0. Maybe if I allowed it to automatically select a vertex that connects to the root it would work.
May 3, 2009 at 11:41 AM
Edited May 3, 2009 at 11:42 AM
I think i will modify RandomTreeWithRoot() to handle 2 overloads, root and branch. I need to select the entrance and exit for the maze. It should always be logical plus I need to have that control for AI path-finding.
I will send up my repro soon in the issue tracker.
May 3, 2009 at 10:40 PM
Don't hesitate to turn this conversation in an issue to do this.
Jun 15, 2009 at 10:12 PM
Edited Jun 15, 2009 at 10:36 PM

Thanks so much for updating the code in the Random Walks folder. It has some nice "TryGet" features now. Good for keeping the graph unbroken and rock solid.

I merged this code with my new root plus branch code and it actually is allot more legable and manageable when looping colors in the InternalCompute method of CyclePoppingRandomTreeAlgorithm.. I was getting a vido card memory leak so maybe this will help fix it too. I will submit a patch later and see what you think of RandomTreeWithRootBranch. It's pretty cool to select entrance and exit in the graph. I am getting obsessed with this maze creation. Allot can be done with it. when just feeding it different edge formations.

Jun 16, 2009 at 4:35 AM

Probably time i make a new release.