(Back from a one week summercamp. )
E,
Thanks for the test of the result of overflow-underflow with respect to borders.
The effect is not random: In the example you posted the center tile is a beach-tile (TileSetIndex 13), the tile to the south-east is a plains-tile (TileSetIndex 1) so the border used is 13-01-SE.
The cutout part of the tile is 507 pixels larger than 352 pixels, so ctp2 uses the next available bytes from the tile-file-witch is the id of the next border and then as many pixels as needed from that border: in this case it is the SW and part of the NW border for 13-2 witch is beach to tundra.
The following images ilustrates the problem.
The first tile is the tile as you want it to be. The other is the tile as ctp2 displays it. I have used black for the borderID, dark red for 13-2-SW, and dark green for 12-2-NW.
E,
Thanks for the test of the result of overflow-underflow with respect to borders.
The effect is not random: In the example you posted the center tile is a beach-tile (TileSetIndex 13), the tile to the south-east is a plains-tile (TileSetIndex 1) so the border used is 13-01-SE.
The cutout part of the tile is 507 pixels larger than 352 pixels, so ctp2 uses the next available bytes from the tile-file-witch is the id of the next border and then as many pixels as needed from that border: in this case it is the SW and part of the NW border for 13-2 witch is beach to tundra.
The following images ilustrates the problem.
The first tile is the tile as you want it to be. The other is the tile as ctp2 displays it. I have used black for the borderID, dark red for 13-2-SW, and dark green for 12-2-NW.
Comment