TF2 map producing empty downloadlist file/ wont load in servers

Whoop

L1: Registered
Feb 2, 2016
17
0
I've recently been testing out a new map of mine on a VPS server my friend and I set up. Everything has been working perfectly so far and I've hosted several beta map testing events on it already, but recently something I changed in my map makes it unable to be loaded by the server when I try to switch levels. All other maps work fine and I've troubleshooted a lot with other custom maps, so it's definitely something I changed in my map.

The server appears to be not finding the file path, so I checked the tf2/tf/downloadlists to find that my maps .LST file is being written blank every time I try to load it, but it definitely doesn't do this to any other maps - even other ones I made in hammer. I can't for the life of me figure out what I did is causing this problem seeing as it works fine with every single other map.

All I did was try to make my map more symmetrical by splitting it down the middle, copying the red side, and rotating it 180 degrees to make the blue side. I made sure to hide any entities I had as to not chop up anything important except for maybe a couple trigger_hurts and a couple lights (neither of which I actually ended up cutting). I made a few tweaks, cleaned it up, compiled it and everything seemed to work fine! At least locally that is...

Here is a link to the map and log file
https://www.dropbox.com/sh/v62i7rc58eg3inq/AABpRsMLciD9fP_K0D47sGqOa?dl=0

What the hell did I do wrong? I was supposed to have a beta testing event tonight, but it looks like that's not happening. :(
 

nickybakes

You should've played Rumbleverse
aa
Jul 28, 2015
911
1,739
Did you change the iteration of you map's file name? Whenever you release a new version of your map, hangs the VMF name to gamemode_mapname_a1 for alpha. _a2 for a second alpha and so on. Not changing the name can cause people to not download the current version, and problems happen. Here is a good tutorial for iterating your map correctly:

View: https://m.youtube.com/watch?v=m7YqJ4MyG7w
 

Whoop

L1: Registered
Feb 2, 2016
17
0
Yup, I've tried with tons of different map names. Same result every time.
 

Whoop

L1: Registered
Feb 2, 2016
17
0
I've never actually renamed the file, only saved as a different name. So that's a no. Plus it works on my local client so its probably not corrupt completely?
 
Last edited:

iiboharz

eternally tired
aa
Nov 5, 2014
857
1,291
So lemme get a better understanding- the map works 100% fine locally, yes? But when you attempt to load it up on your VPS server it doesn't work? ok no i get it now

Could you provide a VMF from before you cut the map in half to mirror it? ALSO: does the server return any errors when you attempt to load the map, or what?
 
Last edited:

Whoop

L1: Registered
Feb 2, 2016
17
0

Attachments

  • arena_horsedong_b9.bsp
    7.1 MB · Views: 237
  • arena_horsedong_b9.vmf
    1.6 MB · Views: 157
Last edited:

Whoop

L1: Registered
Feb 2, 2016
17
0
I also tried copying my whole map in hammer and pasted it into a fresh .vmf with no success. Same error happens. It definitely has to be something in my map, right..?
 

Whoop

L1: Registered
Feb 2, 2016
17
0
Okay, so after hours of trouble shooting by slowly deleting things in my map I've figured out if I delete the blue side it works perfectly fine again. ಠ_ಠ
But deleting all the entities doesnt fix it, so it must be something with the copied geometry.

I'm considering just remaking the whole blue side from scratch which is what I was kinda trying to avoid in the first place, but oh well.
Thanks, hammer.
 

iiboharz

eternally tired
aa
Nov 5, 2014
857
1,291
That's very strange and I can only imagine you somehow ended up with some weird broken geometry.

I suggest watching this tutorial as it teaches you the correct methods for mirroring your map:
 

Whoop

L1: Registered
Feb 2, 2016
17
0
That's very strange and I can only imagine you somehow ended up with some weird broken geometry.

I suggest watching this tutorial as it teaches you the correct methods for mirroring your map:

That's a neat little trick, thanks you. I essentially did the same thing though so IDK where it went wrong. Still helpful in the future though.

I'm not sure if it is specifically the issue but be aware that maps require an info_observer_point or, if everybody dies (and this is arena mode so that tends to happen - maybe even in the initial waiting for players), then the server will crash.

Didn't know that could happen, we had been playing it just fine without them, but thats good to know.
I just tried my map with one set up though and it still doesn't work. :(
 

iiboharz

eternally tired
aa
Nov 5, 2014
857
1,291
Honestly this is a really peculiar issue you're running into, I can't at all think of what the issue might be because I had a look at your VMFs and the logs and I can't find anything wrong.
 

worMatty

Repacking Evangelist
aa
Jul 22, 2014
1,257
999
Hi, there.

I'm a little confused. b9 is the last known working version, but b1 is the one that doesn't load?

Your server log screenshots were a little hard to comprehend at first but eventually I gathered that the first image shows an error message when you load your map, followed very shortly by a server restart likely instigated by your host's management software. It looks like a Linux server. I searched for information about the error message...
qhull precision error: Only 4 facets remain. Can not merge another
pair. The convexity constraints may be too strong. Reduce the
magnitude of 'Cn' or increase the magnitude of 'An'. For example,
try 'C-0.001' instead of 'C-0.1' or 'A-0.999' instead of 'A-0.9'.
My understanding is that this refers to collision, possibly the collision hull of a player, which uses 'Q Physics' or 'Quake Physics' - one of the two physics simulation options in the Source engine, the other being VPhysics. QPhysics is what brush entities and players use, because it's cheap. It may be that the game performs physics calculations as it's loading, and in this case it's found something it dislikes. Some errors in your compile logs, coupled with the search result forum chatter on the web drew me towards displacements being a possible cause, so I opened up your VMFs for b1 and b9 in Hammer.

arena_horsedong_b1.vmf
6WDyvG7.png

Your displacements are using power-4 division. They are using up many more triangles than are likely necessary, especially for such simple landscape gradients. They use more memory, and they can cause crashes when a physics object hits them. It's no crime to use them but they are best avoided. The selected displacement brush to the far left is still a cube. You have made each of its faces a displacement, though only the top is ever seen. As a result, the sloping displacement's top edge is sharing the edges of two other displacement faces, which is causing the 'Displacement abutting' compile error.

arena_horsedong_b9.vmf
Y2y9fSb.png

In version b9, one side doesn't have this issue, resulting in less power-4 displacements in the map, and only one case of edge abutting.

When I ran your b1 bsp on my local dedicated Windows server, I didn't get any errors in the server console. The map was downloaded to my client, and I was able to play in it. I did get a couple of errors in my client console:
ProcessBSPDecal: Trying to project on world before host_state.worldmodel is set!!!
Warning! Static BSP decal (???), on NULL model index 0 for entity index 0.
ProcessBSPDecal: Trying to project on world before host_state.worldmodel is set!!!
Warning! Static BSP decal (???), on NULL model index 0 for entity index 0.
I'm not sure why this is happening. You have two infodecals in your map, so the quantity tallies up right. But I haven't found fault with their placement or configuration in Hammer. It may be something you can ignore but experience tells me it should be investigated.

As a side note, your map lives in a big box.
yKKSDhD.jpg

You should remove that and properly seal your map. Not that it makes a great deal of difference at this stage, but it's helpful for debugging, as it's one less potential contribution to further problems. At TF2maps, we tend to use the Beta stage for detailing, so layout and visibility optimisation would normally be settled before moving out of Alpha.

I hope we get to play test your map. I thought some of it seemed interesting, especially the chute.
 

Whoop

L1: Registered
Feb 2, 2016
17
0
Thank you all so much for your help! @worMatty helped weed it with his section on displacements. When I deleted the old dirt floors and created new and better displacements based on his recommendations everything seems to work fine now! I'm guessing the displacements got really glitched somehow when I copied them because they were actually supposed to be identical, but apparently not.

I can't wait to take all of your guys' feedback into account and get back to testing and improving my map though! It's been a several years since I touched hammer and its cool to see there's still nice community of people willing to help each other out. :)

So yea, thanks again.
I'll make sure to post the map after making some improvements.
 

iiboharz

eternally tired
aa
Nov 5, 2014
857
1,291
I'm guessing the displacements got really glitched somehow when I copied them because they were actually supposed to be identical, but apparently not.

I think it was more of a case of the vertex density being too high since you were using power 4 displacements. But still, very happy you were able to get some help :)