openmw.org

All blank? Initial guess is that some config file is missing or corrupted.

Which version are you testing? Also what platform (e.g. windows/debian/mac) and any special language/region settings? Any installed mods?

+1 for awesome PR work everyone is doing.

Hmmm, I guess you're right. I've re-enabled step and solved it another way for now. If it does cause problems then I guess a special step for water creatures could be added later (don't they have less gravity due to buoyancy?) EDIT: allowing step seems to get them stuck in shallow waters. They are t...

EDIT4: deleted rambling Worked out the issue, water level is 0 however slaughterfish seems to spawn at higher z value. By the time they swim towards the shore they are under the water level. One of the main issue was water creatures being allowed to do things that they shouldn't (e.g. step) and that...

I honestly don't know, but I see no reason to specifically prevent it. I was kinda thinking we could have "vanilla_fly" subclass where the cliff racers seem to use pathgrids to navigate then queue up along the z axis. We could then also have "omw_fly" subclass where some new cre...

You wouldn't want this auto-dodging when moving the player. I've been comparing how vanilla works and the player indeed slides sideways unless the player is dead center towards the obstacle. Same with NPC's. My idea was to copy that that behaviour, i.e. figure out whether the direction of travel wa...

I'm not sure that makes sense. Some actors can do flying and swimming at the same time (NPCs with levitate can freely dive into water to swim and fly out of it into the air). That's fair enough. How about creatures, though? The construction set has only a few "Movement Type" defined, i.e....

edit: nvm

I don't think you want to check for collisions before moving. Collision detection isn't cheap, and it's already done for movement. The whole physics code really needs to be redone. It's a spaghetti mess, and is also using incorrect collision shapes for actors. Unfortunately when I tried to redo it,...

Yep, thanks. However I did have to tweak some values still (so it is only slightly diagonal). I'll submit a pull request shortly. EDIT: I did submit a pull request, but I got to think a bit more on this. Is it a better approach to modify MovementSolver::move() instead? Modifying that code will mean ...