Brad Allred

`IDSImporter` can also leverage this to shed a bunch of `stricmp`s.

one thing to consider are the radii used in `Map.cpp`: https://github.com/gemrb/gemrb/blob/7f26afda7e7b8be9fd800565c4de741e569a4981/gemrb/core/Map.cpp#L198-L201 and https://github.com/gemrb/gemrb/blob/7f26afda7e7b8be9fd800565c4de741e569a4981/gemrb/core/Map.cpp#L2415-L2424 I haven't dug in, but it seems suspicious that one clamps with a minimum of 1 and...

> Results for the overlap for the gemrb > >    the two actors in the middle image appear to have overlapping occupancy zones. Are they actually in...

Wow, that's great! EE's have been of great help to us. Is there a resource to see what other debugging capabilities they have?

One thing I don't understand: if pathing doesn't care about personal space, and from that debug info from EE it appears that overlap is allowed... so what does personal space...

There is probably some fairly low hanging fruit for improving the performance with pathfinding. We seem to completely recalculate paths instead of just splicing sub paths for things like bumping....

The backtrace would be useful to see what texture we are creating so large.

sure, but that should get created on the first drawing pass. Still, @mariomartinezm what resolution do you have set?

I think music already has a dedicated channel so that is covered. A first-come-first-serve pool is fine, but you still need a way to associate a source (actor, projectile, etc)...

That would match the previous behavior. Seems like that would still interrupt an in progress animation tho (if we even care). We could probably add the current cycle to `GetPlayerPortrait`.