Empty texture in E1m2

goldenboy · Post by **goldenboy** » Sun Jun 28, 2009 4:57 am

O_o yes, the problem are the holes and the errors.

please give me a link to the mod you use. where did you get it from?

mh · Post by mh » Sun Jun 28, 2009 7:09 pm

If you're copying and pasting the entities from one BSP to another I'm surprised that this is all that's going wrong. It's not the right way yo do it, and even if Quark claims it can do it, I still would assume that the problem here is Quark isn't doing it right.

Somewhere during whatever you're doing (or whatever Quark is doing behind the scenes) the data offset for this texture is getting lost, and whatever value it should have is being replaced with -1.

What's then happening here is that you're seeing the stock replacement texture that's used when a texture is missing from the BSP. In this case, one of two things can happen. With trad GLQuake the replacement texture is broken, so you get a hole. Any enhanced engine that fixes this will give you the texture.

QBSP does a lot more than just handle entities; it takes a raw map file, processes the geometry in it into a format that the engine can handle, adds in textures, and a whole load of other stuff. It does have a -onlyents option that can just update the entities lump, but you should be aware that a LOT of brush geometry can also live in there, and can be lost or otherwise mangled if you use it carelessly.

Another factor to be aware of is that the Quake toolset isn't exactly that robust. It works fine with the baseline ID1 stuff, but it's surprisingly easy to push it into territory where it's behaviour is "undefined" at best and downright malicious at worst. Quark is an editor that I would rate as being fairly low on the "robustness" scale (I'd call it the "Lotus Notes" of the Quake toolset); it can do it's thing, but I've seen more than enough funnies with it over the years to smell a rat everytime I read of someone having problems and it turns out they're using Quark.

goldenboy · Post by **goldenboy** » Mon Jun 29, 2009 4:22 pm

I think the problem is with the ent file and / or the QBSP he's using. I can actually vaguely remember a CTF mod like that. My money is on a broken ent file, a weird QBSP that came with the mod, or running it through Quark somehow, yeah.

It's not 100% clear what he did. More info would be desirable.

Cobalt · Post by **Cobalt** » Mon Jun 29, 2009 11:28 pm

I agree its one of those several possibilities. Instead of dissecting all that, what about just fixing the holes and errors in the map, and recompiling it into a new version of E1m2? Then would there be less of a chance of the entities effecting the map like that? Even though I now have an original ID map with the right texture appearing in every client, Quark still says there are these errors, so why QBSP or do anything with a flawed bsp? Its fixing the symptom but not the original problem.

Cobalt · Post by **Cobalt** » Mon Jul 13, 2009 3:28 am

Well, I sort of 'solved' this problem. I was running the server and the client both from the same directories as the mod. So I guess my clients were using both maps or something weird like that. I created a new directory for the mod's server, and a new directory where I use the ID1 pak files, and the special pak file separately, and the problem is gone. Still would have liked fixing that hole and the errors, but I guess its not really a problem, or is it?

Cobalt · Post by **Cobalt** » Sat Apr 16, 2011 8:02 pm

This problem was definately because I was running my local ded server from the same -game directory that I was using the client to connect from. It had a /maps seperate folder with the custom E1M2 map from which the server got its map info from, and I suspect the client was also trying to use that same file, and there was some kind of conflict that causes this problem.

Also, I noticed when I use Darkplaces as the local server, when I load the map I get these errors:

maps/e1m2.bsp: miptex #1 missing
maps/e1m2.bsp: miptex #2 missing

Would this be because of the ' hole ' in this map?

andrewj · Post by **andrewj** » Sun Apr 17, 2011 3:39 am

Cobalt wrote:maps/e1m2.bsp: miptex #1 missing
maps/e1m2.bsp: miptex #2 missing

You must have gotten "texture xxx not found" warnings when compiling the map with QBSP.

Cobalt · Post by **Cobalt** » Sun Apr 17, 2011 5:16 pm

Been a long time since I used that program to bind the entities to that map, but I dont recall that message coming up. If indeed there is a problem using Quark to place entities to maps, I have not seen anything besides this potential side effect. I pretty much use Quark all the time to change entities on maps, and dont see any other maps I edit like that giving problems. I am no mapping expert, but it does appear that there is a hole in the map at the place the slipbot texture is located, which may correspond to these positions darkplaces is pointing out, I dont know.

andrewj wrote:
Cobalt wrote:maps/e1m2.bsp: miptex #1 missing
maps/e1m2.bsp: miptex #2 missing
You must have gotten "texture xxx not found" warnings when compiling the map with QBSP.

frag.machine · Post by **frag.machine** » Sun Apr 17, 2011 7:21 pm

andrewj wrote:
Cobalt wrote:maps/e1m2.bsp: miptex #1 missing
maps/e1m2.bsp: miptex #2 missing
You must have gotten "texture xxx not found" warnings when compiling the map with QBSP.

If the map was recompiled to only change the entities lump, no.