trap_trace problems


(Nothus) #1

This is for RTCW.

Have code that checks an entitys visibility to another (the docs on beach to the player, for example).

In Q3map2 compiled maps, the player is never visible to the docs.

In normal maps, he is.

I use trap_trace with the “mask_all” flag. I know this is going to get me a lot of returns, but its needed.

I filter thru the trace results and exclude things that have a surface flag of “SURF_NONSOLID” or contents of CONTENTS_TRIGGER as blocking vis to the OBJ.

On q3map2 compiled maps, I always get an trace.entitynumber of 1022 as blocking me, and it appears like whatever it is is solid and structural, even though there is nothing there.

My question is: why? Does q3map2 handle brushes differently somehow, so that traces are affected? I dont know which version of q3map2 was used - this isnt my map.

As a side note, does anyone know how to display the bit values of trace.surfaceflags and
trace.contents? That would probably help a lot - because if I use G_Printf to show them, it only shows them as being 0 90% of the time.


(Twisted0n3) #2

It’s a wild guess, but I think this is the same thing that makes the bots think most of a map compiled with -meta is solid in Quake 3 if the bspc (bot areas) compile isn’t given -forcesidesvisible. You’ll have to do whatever bspc does when told to -forcesidesvisible I suspect. Unfortunately, I don’t think the bspc source code is available yet.