As described in Rex’s own site about GridMove.Direction, this expression is only valid if the movement is to a neighbouring tile. In the Slidewalk system, however, although the construction of the Slidewalk path may be straight, the target tiles may be separated from each other for efficiency sake (ie less node paths to draw in Tiled).
It is inefficient to ‘connect-the-dots’ in C2 by tagging the tiles that the path goes over, so I’m instructing GridMove to move to the Slidewalk target tiles that are not neighbours, This makes GridMove.Direction invalid.
What I did, instead, was to do my own GridMove direction by comparing GridMove’s SourceLX/Y and DestinationLX/Y expression and generate a direction using conditions.
At this point, however, the conditions assume that the nodes are positioned in a way that it follows the lines for 8 directions, as it checks how the target tile’s LX and LY are different for the source’s corresponding LX and LY.
The basic problem is that when moving diagonally, the edges are ignored because edges only exist at the sides of the square, and not at the corners.
The solution to this was to query both the start tile (slg.PreTileUID) and the target tile (slg.TileUID) and from there, determine the GridMove/Board direction index that the current path was taking.
Using the principal direction (let’s assume it was a diagonal value of 6 (as demonstrated in the above image), the two other directions flanking the main one was also queried.
In the above image, the movement from the start tile is at direction 6. Direction 2 and 3 were derived from this by a lookup that defined two flanking direction for any given direction.
Then the target tile was also queried, but this time, the principal direction was the reverse of the start tile (computing for the reverse tile was made simpler by using a lookup, although a simple modulus would have done it, too). In other words, 4 is the direction. Using the same lookup, 0 and 1 were looked up as the flanking reverse directions.
Then a check is made against existing edges: are there any edges on the 2 and 3 directions of the start tile? If so, movement cost is BLOCKING. If not, then check the target tile’s 0 and 1 direction edges if they exist. And if they do block the movement.
The method used in determining the principal direction was simply comparing the logical X and Y positions of the start and end tiles.
In the RND test project, I ran into performance issues with ZSorter due to the high number of instances I was using. I’ve easily re-implemented the C2 ‘native’ action of z sorting and the performance is much better.
However there is a gotcha.
The native version, despite what seems to be a vague reference of layering behaviour in the manual (search for Sort Z order) moves instances which might have been in other layers to layer that had been sorted last. If the sorting is done per tick, there is no way to force a particular instance to another layer and expect that to render accordingly. (For example, I created a ghost clone over my main sprite. The ghost sits on a top layer, but is an instance of that sprite. Because the ghost belongs to the same sort family in which I base the sort, the ghost also gets sorted. And because the action moves the ghost into the same layer as the other instances, it loses its layer blending purpose.
The solution looks more like a hack: in cases like the above, where instances need to be separated, the sort reference variable (the variable that the action uses as a basis for a sort), should be overridden. In the case above, the ghost instance’s sort reference variable had an extremely large value so that the sort engine will always have to place it on top.