Issues with Path

[SimonDarksideJ]

Great library and love the lack of dependencies. However, it doesn't seem to sort on best path, or could just be the viewer maybe?

As you can see, it's clearly marking a path which is not "best".

Any thoughts?

[roy-t]

Hey Simon,

Wow that looks extremely wrong. The search time is also huge. I'll try to replicate this problem asap and I'll see if I can fix it!

[roy-t]

I've found the bug and fixed it.

I store the data of the grid in a 1D array, I did not always compute the index into that array correctly (dimx * x + y instead of dimx * y + x) at one place. I've made sure this cannot happen again. I've also tweaked the distance metric. Since the squared Euclidian distance metric I used actually overestimates the cost of getting somewhere. Which also resulted in sometimes slightly sub-optimal paths. I now use the Chebyshev distance metric.

I've uploaded the changes to the repository. I'll publish a new NuGet package at the end of the day, once I'm home. Thanks so much for reporting this problem!

[SimonDarksideJ]

Thanks for catching it. Will do some more testing
At least it's best to find it early and then double down on Unit Tests :D

[roy-t]

I'll open a new issue for unit tests. Its definitely needed. But I'm having a hard time thinking of where/how to add them.

[roy-t]

The updated nuget package was release and is available at https://www.nuget.org/packages/RoyT.AStar/