CM_GridPlane unresolvable

[Xedoh]

Hi guys, I'm having a serious problem with my map (specula), which might prevent me from completing the map. I hope i can get it fixed. I seriously don't want to just drop the map now because of some stupid error.

Whenever I load the map, I get about 30-40 of those warnings:
"WARNING: CM_GridPlane unresolvable"
I seriously have no idea what causes this problems or what they do other then being errors. I however don't want to release anything that produces a ton of warnings.

One thing that might be related to the warnings are some missing polygons (see attachment). I have tried everything that came in my mind to get rid of those (replacing the brushes, copying others of the same kind that worked, copying the whole map to a new map-file), but they were still there. One thing that worked was splitting them, but the warnings still occurred (besides that, it's an ugly method of fixing).

Anybody got ideas?

[UniqPhoeniX]

Compile log contains anything? Try copying that part to separate map and compile only that, see if either the warnings or missing polygons occur. In radiant are there any vertices at the top corner of the missing polygons? Any portals very close to those (look at .prt)? Also might want to count the warnings so you know if you have gotten rid of some of them. Mostly just guesses, this is not something that occurs often.

[Xedoh]

Thanks for the input, Uniq.

I tried to compiling only the part with the missing polygons: They are still missing, but no error-messages, the problems are most likely unrelated.
I also had the idea of compiling with NetRaiant (i normally use gtk) and it got rid of the missing polys. However, warnings still occur.

I've not seen any "wrong" nor anything in the compile -logs. The closest portal is about 8qu away, so that shouldn't be much of a problem I think. I haven't counted the number of warnings (is there any way to get a .txt of the console, so it would get easier?).

[UniqPhoeniX]

Were there vertices at top of those missing polygons?
To get a .txt: /condump
If you've been using GTK Radiant to make the map you may want to copy all in Net Radiant to a new map, and save & compile that. If that doesn't work:
To find what causes the warnings, first compile the whole map with the fastest compile option (just the BSP stage) to find the fastest way to reproduce the warnings. If all warnings still occur, compile only half the map, then if (at least some) warnings still occur, halve it again etc until you find what is causing it. It is most likely a broken brush you reused lots of times (maybe some of those thin vertical brushes on outside walls, but there are MANY possibilities in this map, and it could be several different cases).

[Xedoh]

Well, thanks again. I will try that as soon as i get some time.

Oh,I frogot to mention, there where no excessive vertices anywhere.

[gimhael]

This warning message is printed when the engine cannot compute a collision model for a patch mesh. It should happen only if there are degenerate patches in the map.