Integrated Path Planning issue

Don't quote me on this because I'm not typically on the giving end of info on these forums (not for a lack of want!) but I think MMF2.5 will have a path finding movement built in. If that's the case, it might be wise to wait and see how it is before going forward with something else.

Quote from ratty

Don't quote me on this because I'm not typically on the giving end of info on these forums (not for a lack of want!) but I think MMF2.5 will have a path finding movement built in. If that's the case, it might be wise to wait and see how it is before going forward with something else.

There were no such announcements at ClickConvention, and I don't know of any other public info about it other than scattered about on these forums. However, there was a new pathfinding object announced at it, though it's a different kind of pathfinding entirely.

Try reducing the size of the tile to verify the cause as I suggested in the private message. I am out of town so I can't try your example yet.

Out of curiosity, happygreenfrog, what will the new pathfinding object do?

OK, one thing to look at are the action points for both objects, are they aligned? Second quick hack until we figure out the cause is to manually skip one step once you found a path.
Note that the plan paths operation may take slightly longer time than the other operations, so if you are doing it so frequently the object might have moved by the time it is done. If I recall correctly, there was an event that was called once the planning was completed and a path was found. You can use that to skip the first step.

Alternatively, you could use a marker object and place it right at the next node, and have it update its position whenever the agent touches it. If it is bigger than the tile, it should be touched right away even if it is one tile off.

Quote from Mokhtar

Try reducing the size of the tile to verify the cause as I suggested in the private message. I am out of town so I can't try your example yet.

Out of curiosity, happygreenfrog, what will the new pathfinding object do?

The upcoming pathfinding object is called the "navigation mesh object". There is a video that shows the ClickConvention presentation on it here (which should answer your question): Please login to see this link.

Quote from happygreenfrog

The upcoming pathfinding object is called the "navigation mesh object". There is a video that shows the ClickConvention presentation on it here (which should answer your question): Please login to see this link.

Thanks. That's a neat extension, it's great to see these things coming. Although this is not a "pathfinding" object per se, this one seems focused on mesh geometric calculations which is a different (and still difficult and relevant) kind of problem. Path smoothing via splines of whatever, is a very nice feature to give realistic motion right away instead of doing it using a separate movement.

Panchos, sure. If you don't hear from me in a week, ping me again. I can't recall the events/actions now as it has been a long while, but I do recall having a chase example bundled with the extension. I am assuming you already looked at it.

Hey, Mokhtar, do you think you could perhaps port the integrated path planning to other runtimes, by any chance? That's the one thing that all of the existing pathfinding objects have lacked so far (being available for all runtimes (though there is one pathfinding object that is available for Flash, it doesn't work in Windows, which means that, if you want a Flash version and an EXE version, you need to maintain the two separately as of now)), which is why I'm wondering.

A Flash port would be a good start, if you do end up porting it to other platforms, since it shouldn't be too hard to do if you know ActionScript (and, if you don't, it shouldn't be too hard to learn it), though, if there is another port that you think would be easier, you can port it to that runtime instead (E.G. If you don't know ActionScript, and you know Objective-C incredibly well, then an iOS port would probably be easier).

OK, I'm finally back. I see what's happening now, it's crossing the node for a moment and the event to orient triggers before the path is updated. This could be fixed by setting a larger grid tolerance in the extension properties. But a more direct fix would be to simply slow down the orientation event a bit, it worked fine with a 5/100 of a second condition.

Of course, the ideal solution would be to have an invisible marker jump to the nodes until the agent collides with it, as that would allow for smother paths.

Quote from happygreenfrog

Hey, Mokhtar, do you think you could perhaps port the integrated path planning to other runtimes, by any chance?

Technically it's doable of course. I would say the XNA port would be the easiest (but probably least needed). Porting the advanced features may be difficult, but a subset of the most common may be feasible. There hasn't been much demand for a port so far though. The main problem is my severe lack of time. However, if any of the extension developers has the time and is interested in porting it, I can definitely help them out.

For the second solution: not restrict, just add an "every 0.05" condition to the orientation event.