PSA: Func_Brushes and Leaks.

SSX

aa
Feb 2, 2014
392
411
If you're utilizing a func_brush entity and you have a leak in your map that's shooting into the skybox, first thing I suggest you go check is any func_brushes you have in your level.

Func_brushes don't act like func_details in which when you search for the pointfile leak and it comes up with the path. No, the func_brushes act like normal brushes. So if you have a leak in the map; Hammer won't know where to put the leak and it'll just go into the skybox. (I'm sure there's a reason for this, origins maybe? Idk.)

Point is, this is something that should be looked out for if you can't find a reason for something is leaking.
 

MoonFox

L10: Glamorous Member
Mar 17, 2015
739
74
do a control+P
I think... open your problem list
 

Fantaboi

Gone and one day forgotten
aa
Mar 11, 2013
892
1,050
OK KIDS THIS IS HOW YOU FIND LEAKS
Go to Visgroup window, and uncheck everything not 'World Geometry'. Then open that and deselect Displacements. I also recommend to og over and select areaportal so they're showing. Now everything visible is relavent to sealing maps. (But if you have instances you'll have to re-add them) and should help save a ton of time.
 

henke37

aa
Sep 23, 2011
2,075
515
You need to try again, your method doesn't catch brushes with at least one face having a transparent material.
 

UKCS-Alias

Mann vs Machine... or... Mapper vs Meta?
aa
Sep 8, 2008
1,264
816
He doesnt, when you do his filter only brushes that matter for sealing are shown. If it goes through such brush you will know something is wrong with it. If it would just hide that brush you wouldnt realy find the problem and isntead build around it.

If a brush was corrupt, when trying to find the solution hiding it is not helping, you want to see the faulty brush.
 

worMatty

Repacking Evangelist
aa
Jul 22, 2014
1,257
999
Henke is saying that if you paint a world brush with a material that has the translucent property, VBSP treats it as a detail brush so it doesn't seal the map, but if you only have the World Geometry auto visgroup selected, that brush will still show up because Hammer still thinks it's a world brush. e.g. metal/metalgrate013a2.

If you disable every auto visgroup other than World Geometry, then you won't have any entities to generate a point file! The point file shows a path from the closest point entity to the leak, or the closest dynamic brush entity as shown in this screenshot of a func_brush leaking through a world brush that has been painted with a translucent material:

NsXt64J.png
 

Pocket

Half a Lambert is better than one.
aa
Nov 14, 2009
4,694
2,579
I'm pretty sure shadowslasher is talking about the func_brushes themselves being the source of a false leak. As in, the entity whose origin is outside the map and is throwing off VBSP. I had it happen recently with some windows I had grouped together around a corner; the origin was in the middle inside the void, so it looked like the pointfile arrow was glitched and just floating in the middle of nowhere. Luckily the compile log said that it was a func_brush that was leaking and I was able to figure out which one. The fix was to just turn each window into its own separate func_brush, since even just lumping the ones that were all in a row put the origin inside the brushes in between them and that counts as part of the void.
 

worMatty

Repacking Evangelist
aa
Jul 22, 2014
1,257
999
As a Deathrun mapper it's common to make very long or large trigger brush entities that are composed of multiple brushes. When tying them all together to the trigger entity, the origin can go in to the void. The point file path just shows a little red squiggle, but when you turn on Helpers and select everything in the map you can see the origin handle lining up with the red line in the 2D views. The solution for me was to grab the handle and move the origin inside the map.
 

SSX

aa
Feb 2, 2014
392
411
I'm pretty sure shadowslasher is talking about the func_brushes themselves being the source of a false leak. As in, the entity whose origin is outside the map and is throwing off VBSP. I had it happen recently with some windows I had grouped together around a corner; the origin was in the middle inside the void, so it looked like the pointfile arrow was glitched and just floating in the middle of nowhere. Luckily the compile log said that it was a func_brush that was leaking and I was able to figure out which one. The fix was to just turn each window into its own separate func_brush, since even just lumping the ones that were all in a row put the origin inside the brushes in between them and that counts as part of the void.

I don't know what happen in mine and phi's experience with this, but we were looking for a leak problem but we couldn't find anything as the compile log wasn't telling us anything. So I'm wondering it it's a certain situation in which this happens. Needless to say, it took us around an hour to finally give up trying to figure it out. I only recently had it happen again, but I figured it out this time.