"VirtualAlloc Commit Failed"

[quakelover]

While i was making a map, when i tried to play it, it said, "VirtualAlloc Commit Failed, Attempt to Access Invalid Address"
I couldn't find anywhere on how to fix this. I'm using Windows 7, and i'm using kmqbsp3, qvis, and arghrad202.

Does anyone know how to fix this?

[knightmare]

I've encountered a similar error when loading too many player models in KMQ2 (fixed in a coming update).

How large is the BSP file? Sounds like you might have run into the hunk limit of how much memory Q2 reserves for map loading (16MB). Or it could be due to something else reserving tons of virtual memory on your system. Will the map load on another PC?

[quakelover]

The file is 6mb's, i will try to see if it will run on another system. Thanks.

[quakelover]

It doesn't seem to work on another system, any other ideas?

[knightmare]

You're loading this map under vanilla Q2 under the OpenGL renderer, right? Q2's stock loading code for models always reserves the same size (16MB) for maps. If it fails on reserving memory for one map, it'll fail for another one.

If it actually were to run out of reserved memory while loading the map, you'd get a "Hunk_Alloc overflow" fatal error.

Could you post a screenshot for the exact error?

[quakelover]

yes i am.

[knightmare]

You are adding the "-chop 240" parameter to kmqbsp3, right?

Have you tried loading the map under KMQuake2?

[quakelover]

yes i am doing -chop 240, but i have not tried kmquake2.

[knightmare]

I just remembered that I fixed some brush subdivision bugs in kmqbsp3 a couple months back.

Try this version of the tools instead:
www.markshan.com/knightmare/downloads/kmqbsp_113e_test2.zip

[quakelover]

Ok i tried kmquake 2, and it says the same thing, i could try that kmbsp out, but, just to mention when not having all the brushes to detail, qvis has an exit status of "1073741819"
I don't suppose you know what that means do you?

[knightmare]

No, unfortunately I've never seen that error.

BTW, are you also using the -oldbounds parameter with kmqbsp3? That restores the default slice/dice block size of 1024 instead of 4096, and can help with some errors with intersecting brushes.

[quakelover]

No i am not, i could try.

[quakelover]

doesn't seem to work

[quakelover]

So it's looks like it is maybe some sort of vis, and rad error since i am able to play it without the two compiling tools, and i think the reason is, is because when i compile kmbsp, it says there is multiple mixed face contents.
Entity 0, Brush 20: mixed face contents
Brush: Line 192, file D:\Quake2\baseq2\maps\lab2.map
Entity 0, Brush 149: mixed face contents
Brush: Line 1367, file D:\Quake2\baseq2\maps\lab2.map
Entity 0, Brush 150: mixed face contents
Brush: Line 1376, file D:\Quake2\baseq2\maps\lab2.map
Entity 0, Brush 151: mixed face contents

The thing is, idk how to check the brush number.

[knightmare]

Does your editor have a "find brush" function based on entity/brush number? Use that to find the offending brushes and fix their face contents. QERadiant and GTKRadiant have this function.

You could also open the map with a text editor and go to the listed lines and change the contents fields of the listed brushes, but you should back up your .map before you try that.

I'll probably have to see your map to know exactly what's wrong with it that's causing these errors. But it sounds like you have a highly complex brush structure that creates too many visportals. You'll probably have to simplify it, or else make parts of it into func_wall entities.