@IanM - Did you post this in the DBDN CodeBase? (It would be useful if it isn' there)

I will as soon as I'm happy with what I've got.

I'm considering some heavy changes at the moment - writing a new function for each slightly different system I have just doesn't seem right.

Latest Changes:

A*8 and A*8R searches rolled into a single function
Added a Flood8 and a Flood8R search (in a single function)
Added a new function to enable or disable the restricted diagonal code.
Updated all search functions to allow a 'cost' based cut-off - especially important to restrict the time taken for the flood searches.

I think I'll give this a rest for now ... at least until I need to account for different cost for each cell, or someone convinces me that they really need it

The latest testbed code is in the next post.

Testbed code

Both now added to Codebase for you Rob

Excellent. The algorithm will really start to be useful when I start dealing with AI. I am using a memblock matrix so the grid is already nicely defined.

cool, i could use this for my rts

IanM, I haven't looked at the code yet but does yours stop the player from going down a dead end? My paths are usually OK but sometimes the player can get lead down a dead end corridor.

Phaelax: Help yourself

David89: You will never get a dead-end from these search routines. They either find a way to the target and populate the path, or won't.