DRD Team

Sure. I see no reason for it not to be.

Latest build uses class names, and I just pushed string action specials. The translation scripts are much more self documenting now.

Cool!

I haven't given up on the stuff I was working on but it's been temporarily sidetracked for other projects.

Random question about patrol points - is it possible that a patrolling enemy would 'drift' out of a path using odd angles, or do the points correct for that? When does the enemy change direction (i.e. how do the bounding boxes work in this case) and does his position also get altered by the point?

AI is tile based, so when pathing the tile selection logic is just reduced to "use the tile in front of me unless it's blocked in which case stop moving forever". The time to reach the center of the selected tile is a constant so moving diagonally is done faster than orthogonally. When the destination tile is reached the round off error is corrected.

Does the x/y position of a patrol point inside of a tile matter? Can you have multiple points per tile?

What I'm wondering is if you can create crazy ultra-smooth curves with patrol points or if you're realisticly limited to 8-way directions and one point per center of a tile.

Smooth curves aren't currently possible. The AI code just simply can't handle it. I do at some point plan on introducing code to handle off grid (either automatically or through some kind of opt in depending on how it works out), but that's pretty far down on my list especially given the editor situation. Given the limitations of the AI for best forwards compatibility I would put the patrol points in the center of the tiles.

Off grid thing placement still works fine for decorative items and pick ups. Players will interact properly with solid objects off grid, but the AI will assume that a solid object occupies the entire tile its center exists in. I could improve the last point a bit, but it hasn't been important given the current editors.

OK, that's fine. Are 8-way direction patrol points supported or is it safer to use 4-way directions?

8 way are supported. That's part of vanilla.

What exactly are the two "modzone" entries in the Wolf 3D xlat file? I think I understand everything else.

Those handle the translation of features that were handled by floor codes, but moved elsewhere in UWMF. One is the ambush floor code which was changed to just a flag on the thing. The other is the secret exit floor code which was changed to a separate trigger action. When performing binary map translation the latter looks for triggers on the adjacent walls and changes it to the secret exit action (more specifically it disables the trigger on that side and creates a new one).

The Pacman ghosts and the Dogs are activated even when the 'no target' debug mode is on. I assume that's just a bug?

They should be active, but not seeking the player (in other words walking in a straight line until they hit a wall). If that's not the case can you send me a test map?

Here: https://dl.dropboxusercontent.com/u/91541524/demo.7z

It's a big empty square with every actor in it. If you're curious about the floor/ceiling colors, it was a visual debugging aid to to figure out which ones were Wolf 3D versus SoD actors since they are mixed in the definitions.

OK, did a little analysis. This is consistent with how vanilla would behave. This isn't something that can be fixed until the multiplayer code is stable since there's still stuff that assumes a target is always available.