I had func_detail_fence, skip textured all sides except for 1, which had the AD vines.

Saw them from both sides. 

Would it be possible to have model->geo baking like q3's misc_model(static) 

I've played with that in the past, but I didn't understand QBSP as well so it didn't work very well. It's probably doable to bake in a model without converting it to brushes, but it would be illusionary only (doesn't stop bullets, no collision) so I'm not sure if it's worth the effort.

If you want collision (player / monster / bullets) qbsp needs brushes, so you can use obj2map or something to convert to .map first (turning triangles into tetrahedrons with obj2map will probably give you a bloated / inefficient bsp, though). 

Do I need to use func_detail_fence only? I've been using func_illusionary. 

no, I tested it with func_illusionary. I don't know why it's not working for you, check spelling? "_mirrorinside"
Build version should be "tyrutils-ericw-v0.15.10-12-gd86913a" 

No, not func_detail_fence only.

When I first tested I used func_illusionary and then thought, "Oh wait, he mentioned fence!" So my brain connected dots that weren't there :(

What didn't work for me was standard func_detail and func_detail_wall

hth's 

"but it would be illusionary only (doesn't stop bullets, no collision) so I'm not sure if it's worth the effort. "

Still worth it, imo. 

I think I figured out my shadow bug on my trees - it was the old "intersecting brush with func_illusionary" bug.

My trees have a little solid brush in the center as a "trunk", and making it into another func_illusionary seems to fix the issue.

Still, having a fix for that bug would be nice... 

Turns out it wasn't compiling with the ericw qbsp at all in either version. I usually just hit yes whenever JACK warns me that prt wasnt made since I know my test map isn't sealed.

Not sure what's up. The output says almost nothing to the compile dialogue. Just says the compile path and path to file which are correct and nothing else.

Do I need any sort of compile parameter to compile 220 map format? Already tried the usual admin privileges. 

Ya going back to tried and true txqbsp modified by mh and it compiled just fine.

? 

Does anyone else have trouble with using JACK, valve 220 format, and the tyrutil's qbsp? Light and vis work fine for me but not qbsp. Qbsp produces nothing and gives no output. Must be something wrong with my setup but not sure what. I'm not seeing anything obvious in the docs about compiling for 220 .MAP format. Sorry for the multiple posts but this is wierd and I really want to be able to use all the awesome features ericw has added. 

Hm.. there is no special flag for compiling valve 220, it should just work :/ 

Got it. Mirrorinside working too!! Yay! I manually draggedndropped the .map onto qbsp amd it gave me a couple errors for missing dll's.

Needed to download:
msvcp120.dll
msvcr120.dll

And put into same directory as qbsp.exe. I guess it has to do with Windows 10 missing some VisualC packages. 

Is giving me some wierd gray patches where walls aren't getting drawn...tried changing the brush work significantly with no effect. I can send you my .maps if you'd like? 

Sure, I can check it out. 

Pics included to show the problem spots. Let me know if you can't find where they are. 

This is awesome. 

Hey there,

My next map is approaching completion. I'm working on the beta build and just recently encountered a few errors when compiling that seem to have been alleviated with the use of bsp2 format.

The first time unfortunately I did not record entirely but it stated something such as: Error: "...to many vertices... try compiling with bsp2..."

Now this was most certainly not a leak. The map is sealed, there are no warnings of entities exposed to the void, no pointfile, etc. and using bsp2 fixes the problem. When I compile now with bsp2, the map compiles with vis and all is good.

However, I cannot replicate the original error message. If I remove -bsp2 the compiler reports error:
"13:C:projectstyrutils-ericwqbspsurfaces.cc:316: Q_assert<v1 == edge->v[1] & & v2 == edge->v[0] failed."

Using bsp2 gets rid of this error as well and all is good. Now, I'm not sure what this error is, but the first message was a lot more useful and allowed me to fix the problem on my own, which is great. Maybe this error message should say something similar.

Thanks 

The error reads:

"13:C:\projects\tyrutils-ericw\qbsp\surfaces.cc:316: Q_assert<v1 == edge->v[1] & & v2 == edge->v[0] failed." 

The current code to detect if bsp2 is needed isn't very good.. I think those crashes just mean that the map needs the -bsp2 flag, but qbsp didn't detect it properly. Improving that is on the todo list. 

Mostly a bug fix update:
https://github.com/ericwa/tyrutils-ericw/releases/tag/ericw-v0.15.11


- light: add "_sun" entity key to configure sunlight in an entity instead of worldspawn. More than one "_sun" entity is supported.
- light: add "_falloff" light entity key to configure light falloff in map units. Only supported on linear (delay 0) lights.
- light: add "_spotlightautofalloff".
- light: fix light cutoff on curved surfaces
- light: adjust -soft to fix regression in 0.15.10
- qbsp: add "_mirrorinside" key for mirroring the outside faces of bmodels so they are visible from inside. for func_water, or func_illusionary fences, etc.
- qbsp: fix CSG issue with overlapping off grid brushes
- qbsp: fix HOMs introduced in 0.15.10, which were caused by an attempt to fix leaks-through-solids in 0.15.10. To re-enable the buggy code that may fix leaks through solids but add HOMs, use "-contenthack"

Thanks for everyone who reported bugs and m-x-d for the contributions to light :)

Btw - I want to rename the project at some point, any suggestions? something that doesn't have my name in the title, haha. 

light-ultra-mega-extended-nopay-0.15.10

or LUMEN for short 

Just call it tyrutils-0.16 to confuse everyone and annoy tyrann (if they're still alive)

I'll have a look to see if the bugs are fixed for me yet, fingers crossed :)
Also, what's the purpose of _sun on an entity? And the purpose of having more than one? 

Lame Name ideas:
Wizann (ericw the wizard + Tyrann)
Compyre
Luxpiler
Luxspire
Map2Bsp (lame i know)
Quislitvis

I had been using an older light it seems for the black faces bug. I'll let you know if I run into anymore bugs. 

Yeah - it's time to bump the version to 0.16 soon. I don't want to use "tyrutils" since that makes it sounds like tyrann endorses the tools, I think I have a good name idea though.

I'll have a look to see if the bugs are fixed for me yet
Thanks!

Also, what's the purpose of _sun on an entity? And the purpose of having more than one?
Just convenience, really - figuring out _sun_mangle is a pain; this way lets you target an info_null to set the sunlight vector. It doesn't enable any new features that you can't do with the worldspawn keys, currently.
However, there are some things that would be easier to do with sun settings on entities, one that comes to mind is having a targetname to make them toggle-able (you could totally fade the sunlight with QC if that was implemented.)

Having more than one sun entity is for when your skybox has 2 suns on it. :D