It looks like it's doing it when it's calculation collisions. Any ideas on what that is?
Adding prop_static for fizzler adds a lot to compile time!
Well 28 Seconds isn't much of a big deal, but for that area of compilation it seems funny.
Incidentally, I did try it as a prop_dynamic. The compile is fast as usual, but at runtime they don't appear! The fizzler is just in the air by itself.
What could it be?
Incidentally, I've moved my map to another PC and now during compile it says the following error when I create a prop_static for portal_cleanser_1.mdl:
Error loading studio model ""!
So WTF is going on! Any help? 
The error went away apparently randomly after a couple restarts.
So they only mystery left is why adding them as prop_static triples my compile time? 
@ the fizzler destroying the model
As for why the prop_static is taking longer to compile, i have no clue, that really doesn't make any sense. Which program (vbsp, vvis, or vrad) is giving you the additional compile time? That could help figure out why it's doing it.
The extra compile time is happening during vbsp. It gets to these lines really quickly:
Building Physics collision data...
done (1) (106651 bytes)
So that's fine. But then it takes it's long time to think, which, when it's done, outputs this:
Placing detail props : 0...1...2...3...4...5...6...7...8...9...10
Compacting texture/material tables...
Reduced 478 texinfos to 191
Reduced 65 texdatas to 52 (2795 bytes to 2374)
Writing e:\bin\steam\steamapps\[email protected]\sourcesdk_content\portal\mapsrc\d03o.bsp
26 seconds elapsed
Does that mean anything?
Load up your map in glview and see what the area looks like there. I wonder if something weird is happening and vbsp is cutting it up strangely because of the prop, or maybe the clipping hull is getting mucked up.
Maybe the problems are related?
Also, it happens no matter where it is on the map, so it's not a visleaf problem. It also happens on an empty map with just one room.
