I’ve started using MultiTargetPath to get the shortest path from a number of different targets.
The closest endpoint was generally correct, but the path generated for it seemed odd.
The upper image is with ABPath and the lower is with MultiTargetPath run with the same single target point.
I can avoid the funnel failure in the lower picture by tweaking the recast settings (splitting the tile edge etc to reduce uneven length connections) but I’m still concerned about why it’s taking a different route in the first place.
Is this a bug, or just something unavoidable in how the multi target works?
It seems a bit silly to use MultiTargetPath to find the nearest target, then discard the rest of the result and use ABPath to generate a more correct/reliable path to get to it.
For reference, the same disparity is seen regardless of pathsForAll state and still happens if I use it as multiple starts by sending a 1-length array for the start point instead.