Re: Cel Navmesh testing for Planeshift
Sam Devlin <[email protected]>
| Newsgroups | gmane.comp.graphics.crystalspace.devel |
|---|---|
| Message-ID | <CAL0_xOqnLzgA9wW4-Mg+UE4fSmUnwkaZatFhYcG7ZTM0hHeUBQ@mail.gmail.com> |
Hi Anders, Thank you for trying out the code and for highlighting the difference in appnavmeshtest, I see what you mean and will update that today so that the two versions are consistent. I think given that you already have your code in place to work with the previous version of navmeshes and your navmeshes already constructed, there is no benefit for you to switch PS to the latest code yet. The improvements I have made this summer are beneficial to the creation of navmeshes and use within an app (i.e. you no longer need to manage the navmesh memory in an app, it is handled for you within the plugin.) These benefits are, therefore, more useful for those developing new apps. The biggest change needed, and I think we agree here, is to improve the behaviour between portals. For this I think there needs to be a significantly different approach as the current method is simply not functional. Matthieu Kraus had some detailed ideas on this when I was proposing this project, unfortunately there has not been time to explore them fully as part of this GSoC. I would be interested in hearing his thoughts on this again and perhaps designing a new solution. Kind regards, Sam. On Sun, Aug 12, 2012 at 11:43 AM, Anders Reggestad < [email protected]> wrote: > Hi, > > I was just testing the cel navmesh both in cel_ai bransh and the cel. I > really hope for a new and improved version that we can include in > Planeshift. Last time I generated the navmesh the computer used more than a > day. For the NPCs in PS the most critical error in current system is the > portal crossing issue that has been mention earlier in this list. In effect > currently the Planeshift NPCs are limited to movement within each sector. > > When running navmeshtest I notices that the DebugAgents behave different > in the two versions. In the cel_ai bransh the other end seams to disappear > after marking a new end/start point for the path. Seams to be a issue > regarding memory management of the meshes used to draw the DebugAgents. > > Since there are a number of interface changes I did not try to integrate > the new version into PlaneShift to check for performance increases yet. > Though if the portal crossing issue should be resolved it would be very > interesting to test. > > Really looking forward to see a complete version of the Navmesh system. > Keep up the good work :) > > -Anders > ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Crystal-develop mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/crystal-develop