max_mapfile_planes

kat · 2003-10-31 03:27:13 UTC

I’m not 100% sure this is the problem I’m having but I’m trying to create the AAS files for a RtCW SP map and the BSPC compiler keeps bottoming out when it’s reading & listing the func_static/explosives contained in the map.

I’ve tried BSPC’ing maps compiled with q3map2 that came with GTK 1.2.13 and q3m2 v2.2.32, 2.3.38, 2.5.4, all BSPCs result in the same ‘error’. I did do a compile with the q3map version of said above GTK and it threw up that error - max_mapfile_planes. (compiled without using the -meta BSP switch)

I did a search on this but didn’t find anythng conclusive, what did crop up seemed to have something to do with the size of brushes in the map (topic in question actually related to clipModel) which is odd becasue non of the brushes I’m currently using are even half the size I used in the last SP series…

I’m confused… :???:


Opened log bspc.log
BSPC version 2.1c by Mr Elusive
GtkRadiant 1.2.1-nightly Jan 13 2002
threads = 1
using cfg file gtk1213\aascfg_sm.c
using cfg file e:\progra~1\return~1\gtk1213\aascfg~1.c
bsp2aas: E:\Program Files\Return to Castle Wolfenstein\Main\maps\generator.bsp to E:\Program Files\Return to Castle Wolfenstein\Main\maps\generator.aas
-- Q3_LoadMapFromBSP --
Loading map from E:\Program Files\Return to Castle Wolfenstein\Main\maps\generator.bsp...
creating planar surface planes...
searching visible brush sides...
     0 brush sides
 74739 brush sides
55827 brush sides textured out of 74739
Ignoring script_mover brush..
Ignoring script_mover brush..
Ignoring script_mover brush..
Ignoring script_mover brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_explosive brush..
Ignoring func_invisible_user brush..
Ignoring func_invisible_user brush..

it just stops here and blonks me back at the command prompt…

Also tried BSPC from GTK 1.2.9… same thing

Also gone through the map and removed func_explosives that wen’t necessary or converting some conplex ones to func_statics… is there a limit to the number of these you can use?

ydnar · 2003-10-31 08:00:10 UTC

Convert the func_statics to func_groups and try again.

Alternatively, convert the most complex ones to ASE models, and attach them to func_explosives via targetting to eliminate the planes.

y

PS
Why are you not using -meta?

kat · 2003-10-31 14:37:46 UTC

thnx ydnar… was in the process of doing just that after a bit more research turned up some more info about the number of textured faces.

the func_statics are needed becasue they are removed at runtime as part of the gameplay. I’ll let you know how the model conversion goes…

PS -meta seems to sometimes make the aas files go a bit screwy (don’t know why), I don’t nomally use it untill I’ve got solid AI action, then it’ll be used… :drink:

ydnar · 2003-10-31 15:25:16 UTC

If -meta is used in the BSP phase, then you must pass -forcesidesvisible to BSPC.

y

kat · 2003-10-31 16:05:38 UTC

aye I know… this is above and beyond -forcesidesvisible.

I’m not 100% sure on this but I think using that causes part of the problem in terms of how it relates to BotClipping and ‘gaps’ in maps between those and structural walls.

I’m probably not understaning it correctly but from what I’ve read only clip and structural brushes (plus a few other things) are used for BSPC calculation?? So if you clip a wall of detail without going all the way thru it to the structural behind you leave ‘gaps’ and it’s these bits that seem to cause the problem. I’ve only noticed it as a ‘general rule of thumb’ when correcting AAS issues in the past so pulling the clip right back to the S wall usually corrects any problems. It’s probably more an incorrect clip usage problem than a -FSV one…!

ydnar · 2003-10-31 18:44:18 UTC

You definitely don’t want gaps, both for BSPC reasons and lighting reasons. Fill them up with caulk (structural or detail).

y

kat · 2003-11-03 22:01:53 UTC

ok, this is still giving me some major problems…

so far I have

converted all the fiddly details to ase models (mainly a walkway with a handrail which is used quite a bit),
converted a couple of big curved patchmesh objects to ASE.
MonsterClipped everything but the kitchen sink, inc dead space in the middle of some of the rooms,
taken out all the func_statics and func_explosive but the compiler bottomed out after listing the func_invisible_user triggers (so put them back in)
taken a large(ish) section out of the map (didn’t like it anyway…! phooee)
‘flattened’ the caulk hull as much as possible

BSPC compile was listing around 70k textured surfaces and using about 35k or so, after all this it’s still lists about 60k and uses about 30k… what I don’t understand is that regardless as to whether -meta is used during compile of the BSP, BSPC lists the same amount of faces regardless of this… I always thought -meta physically removed those dead planes and faces from the map files, this error suggests otherwise??

oh also I loaded up the portal file to look at how they were being distributed when I noticed a shit load of them outside the map… is this normal?? I know that sometimes happens when a leaked map is compiled but I’m not getting any errors to indicate this is the case… mucho confusido

:???:

Detoeni · 2003-11-04 22:32:15 UTC

not sure if this will help with the problem but, you should be using “common/clip” to hide your complex brushwork. Bots see “clip” as a solid brush.
“common/monsterclip” just defines a space where the bot can see into but cant go into.

kat · 2003-11-04 22:49:30 UTC

yah that’s what I’m doing… the maps structure is pretty ‘open’ so AI can see everything… clipping is relative to that.

kat · 2003-11-06 19:47:35 UTC

well FFS this took a lot of work to fix and the ‘thing’ causing the error wasn’t what I expected, usually a bad entity.

To track down ‘where’ it was happening I BSP’d sections of the map until I got an approx area. More compiles were done by taking out various ‘types’ of things and doing new versions of the effected area, no detail, no entites, no models, no lights, no clip, etc… I got a clean BSPC doing a no clip compile so it was one of the clip brushes causing the error. Got rid of that and so far - touch wood - getting clean BSP and BSPC… what a t*t to figure out…! and it wasn’t a dodgy brush so far as could be told… bobtoolz didn’t pick up anything and there arn’t any funny 3pt clipped brushes in use…

kat · 2003-11-07 02:41:08 UTC

spoke too soon… that wasn’t the problem it was this…

No scripting had been done (yet) so a couple of the doors had a complex entity setup that allowed them to be triggered by some buttons on both sides of the doors… this had a lot of same named entites (func_statics) which were set to hide/shown on command, they were connected to a couple of target_delays/relays and once I’d disconnecting them BSPC ran all the way thru, it wasn’t a ‘bad’ entity but an overly complex setup causing the error…

I had to version save and compile 30 or so differnt map files, deleting more specific sections of the map to find this out mind you and it had nothing to do with the initial ‘correction’ mentioned above… what an pain in the butt.