As the first week of GSoC is coming to an end, I thought it best to update the blog with my progress so far. I intend to (as I did previously during GSoC 2009) post weekly on Fridays with quick updates for anyone interested in following the project.
This week has been mostly occupied by getting bttest to work again. It seems there may be a bug in trunk that is causing bttest (and graphtut, steering and walktut) to hang. I have not solved the problem but I have commited a workaround to my branch in revision 4893. If anyone else is having the problem or has suggestions to fix this please join in the conversation on the mailing list (Archived here.)
Revision 4893 also contains a workaround fix to get plgtcpnetwork to compile as DebugWithDLLs in Windows. Which may also be a general problem at this time in trunk.
The final changes in revision 4893 are what I needed to overcome compilation and linking errors in appelcmtst, plg_guiinventory and plg_guiinventory2 due to CEGUI. I think from the discussion on the mailing list these changes were needed only by me.
Since overcoming these difficulties, I have fixed a minor but significant bug in the bttest app that stopped the tree from being evaluated (see revision 4897.)
I now have a working bttest app, a better understanding of my code from 2009 and am ready to start implementing event driven behaviour trees next week.
A quick update and thank you to all who have helped me out the past couple of weeks. Thanks to the mailing list and in particular Christian's efforts I now have a fully compiled and up to date CS and CEL. I have refamiliarised myself with svn-merge.py (I typically use TortoiseSVN but wanted to stick to the tools other CS developers are using for this project), and with my code from the previous GSoC project on behaviour trees. I now feel ready and keen to start coding next week in time for the official start of GSoC.
I include below my proposed timeline and goals, with priorities labelled 1-3 with 1 being the highest:
-Update to event driven BTs (May 21st – July 1st)
(1) -- Focus on debug tools
(2) -- Allow all termination statuses
(2) -- Add load from XML option
(2) -- Create dedicated propclass (set root node, set update rate)
(1) -- Documentation
-Update to latest version R&D (July 1st – July 29th)
(1) -- Allow iPcSteer to use its methods
(2) -- Partial navmesh loading (with missing sector callback)
(2) -- Make navmesh creation parallel
(3) -- Improve function with portals
(3) -- Optimise debug mesh
(3) -- Optimise Terrain
(1) -- Documentation
-Improve life sim demo (July 29th - August 13th)
(1)-- Make behaviours work (perhaps requiring other new propclasses or
fixes to existing ones)
(1)-- Use new BTs and R&D
During this time, I will have one significant break (as mentioned in my proposal) from June 2nd-12th when I will be attending a conference in Valencia. Other than that I am excited to be back and keen to get started :)
|<< <||Current||> >>|