Re: [hlcoders] Linux dedicated game server won't load any sound scripts or soundscapes

[Ben K.]

Regarding the libraries scenefilecache and soundemittersystem:
Compile your dedicated server with -D_DLL_EXT=_srv.so instead of just 
".so". The server (game) lib loads those SOs using the file ending 
provided by the _DLL_EXT preprocessor define.


The loading of soundscapes is indeed broken. The game uses 
filesystem->LoadKeyValues(...) to parse the file but that function 
messes up the key names or whatever, I don't remember. You can get 
around that by instead using pMyKVPointer->LoadFromFile(...), as is used 
in most other places.

The offending function is used twice in game/server/soundscape_system.cpp.

It's also in game/client/c_soundscape.cpp, but I don't know if Linux 
clients suffer from the same issue, or OSX clients/servers for that 
matter as I've only compiled the Linux server thus far. Seems to work on 
Windows client/server, so I just added the workaround in an #ifdef _LINUX.


Anyway that should do it.

- Ben

Rebel Yell wrote:


Found the issue. Turns out the dedicated server depot for SDK 2013
lacks 2 libs:
scenefilecache.so and soundemittersystem.so. While games as CS:S
work fine without those, mods complain about those missing files
and server shuts down with this message:

AppFramework : Unable to load module soundemittersystem.so!
Unable to load interface VSoundEmitter002 from soundemittersystem.so

So I copied those from the client repository and that was my
mistake. You either create a symlink from their respective _srv
versions or just copy and rename them, never take the ones meant
for clients.

That solves the error related to game_sounds_manifest.txt load,
not the soundscapes manifest one though.

TL:DR linux dedicated server is broken, it won't even load and
trying to fix it may cause further issues.




File encoding was one of the things I looked at before posting here, in 
any
case I repeat that linux client is able to load sound scripts and
soundscapes when listen server is created. They are the same base sound
scripts, same game libs, etc. The SDK bin directory may differ a bit I
guess since dedicated servers require a few different libs.

I tried a gcc 4.8 compile just in case. No improvement, the stock SDK 
build
shows same issue. I don't really know what else to try.

To be honest, I'd not care anymore about this mess if it wasn't because 
the
current version of my mod on Steam is a nice engineered backdoor 
disguised
as game. It's not my fault, I merged the fixes the first day they were
available, I can't just ship a mod that gets near 2 million installs
without sounds working properly.

Thanks anyway.



>It's probably the encoding of the txt file.
>//>//>/On Tue, Oct 6, 2015 at 12:31 AM, Rebel Yell http://gmail.com>
>
wrote: />//>/>* I decided to compile the stock SDK 2013 code today and 
the
linux build />/*>* outputs the same error message when dedicate server 
is
loading: />/*>>* CSoundscapeSystem::Init: Manifest
'scripts/soundscapes_manifest.txt' with />/*>* bogus file type 
'ls/extinguish2_subrect', expecting 'file' />/*>>* Also note that these message errors 
appear no matter if />/*>* soundscapes_manifest.txt is present or not, it's like the
actual file isn't />/*>* used at all. />/*>>* My linux test machine 
uses Ubuntu 12.04.3 x86 and I
compiled the code with />/*>* gcc 4.6 />/*>>* Do you guys think there's 
an error in my setup or this
engine build isn't />/*>* functioning properly? />/*/



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlcoders

Re: [hlcoders] Sound issues

[Ben K.]

Hi Jan,

regarding your first issue I can say that Source SDK Base 2007 has the 
same issue. We had this issue in our Mod a couple of times a few weeks 
before you posted your message. I have not yet investigated it though.


Ben

On 18.08.2014 16:07, Jan Hartung wrote:


Hi list,

Some weird things are going on with the sound in our SDK 2013 MP based 
mod. First, there is a map with a looped sound in it, the entity is 
placed in the center of the map and can be heard everywhere (it's a 
small map and the volume is high enough). Now, it sometimes happens 
that when I kill a bot, this sound's origin jumps to the place where I 
currently am. It may jump to its original position when the next round 
starts or stays there for a few rounds. At times, not related to any 
kill, the sound also duplicates, so you can hear it twice while the 
two copies are out of sync.


The second thing is that we got an announcer in the game, so there are 
sounds which are played to every player. We loop through the players 
and use UTIL_EmitSoundSuit in order to do so. The announcer received 
an overhaul just recently, which also added multiple sounds for the 
same event, which are set up as rndwave in the sound script. Since 
then I notice, that whenever I stand close enough to another player, I 
can hear the announcer's line, which is played to the other player. 
However, UTIL_EmitSoundSuit is supposed to play a sound audible to the 
player itself, judging by the function's comment.


Has anyone seen these issues?

Jan

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlcoders

Re: [hlcoders] Possible 2013 Engine bug: Looping sounds stay in sound list outside of audible range

[Ben K.]

Hello Peter,

On 11.10.2014 22:24, Peter McKeown wrote:
Hello everyone, I've found what seems to be a bug with the engine 
relating to looping sounds from ambient_generics (yellow sounds on 
snd_show 2).


When testing my mod, which is based on Source SDK Base 2013 
Multiplayer, I noticed that some map sounds weren't playing, and the 
console was being filled with total_channels == MAX_CHANNELS 
messages. I investigated using snd_show 2 and it seems that every 
flavor sound ambient_generic that played a fire sound (as there was 
lots of torches in the map with a small fire sound next to it) was in 
the sound list (thus showing up on snd_show 2), regardless of whether 
or not the sound was actually audible - every ambient_generic playing 
a looping sound anywhere on the map is in the list. These sounds can't 
actually be heard audibly.


This isn't the same behavior that was in Source SDK Base 2006 ...


This then has probably been introduced with the OrangeBox engine version 
because I can see the behavior you've described in a 2007 based mod.


Ben

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlcoders

Re: [hlcoders] sdk 2006 mod, steam pipe, and debug on vs2005

[Ben K.]

You need to set the SteamAppId environment variable to your hl2.exe's 
AppId.

For example SteamAppId=218 for SDK Base 2007.

There used to be two other env vars you would need to set but apparently 
it now works without them (for me with 2007 anyway).


You can do this from within Visual Studio (debug settings for the project).

Am 29.05.2014 08:47, schrieb Nick:
Is this problem still broken? I also cannot debug on vs2005 by 
directly starting hl2.exe with the -novid -allowdebug -steam -game 
moddirectory parameters.


The only way I can debug w/ vs2005 is to manually attach. I really 
hope this problem is fixed by now.


___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlcoders

Re: [hlcoders] Hitbox calculation and sizes?

[Ben K.]

You probably haven't adjusted the bounding box. Check out your game 
rules and look for ViewVectors.


Ben

Tobias m...@tobiasbaumann.net wrote:
Already try that and it only shows the local hitboxes for me. I mean 
it displays the proper hitboxes the same way they're displayed in the 
modelviewer. However these are totally ignored for actual hit 
detection when I fire on the player. :(

Check this screenshot: http://i.imgur.com/KGJ1Kys.jpg
Thats with sv_showhitboxes 1. If I fire on the model with another 
client the bullets go right trough the model until about half the 
vertical size of it. Or in other words if I fire at the base/lower 
part I get hits and everything above is ignored. Since the model is 
pretty large/high I'm guessing that there might be some maximum height 
value for the hitbox? Or for the collision box? A standard playermodel 
reaches until about the middle of the cactus.

Best regards,
Tobias


___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlcoders

Re: [hlcoders] Font size cap?

[Ben K.]

I figured a 128x128 bitmap is all you're gonna get. I've run into this 
limit myself. Considered using a texture, maybe with $distancealpha ( 
http://developer.valvesoftware.com/wiki/$distancealpha ) yet?


Cheers,

Ben

James Deeman wrote:
Is there a cap to how large your font can go? I am finding that after 
a certain point the size of the glyphs just don't change, around 120 
'tall' value in the Scheme file. Is this right? Anyone know what the 
actual figure is?


Any replies appreciated.

--
James Deeman
Nottingham, UK



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlcoders

Re: [hlcoders] Source RCON PLIST?

[Ben K.]

A search on the interwebs for apple port 27015 reveals that 
AppleMobileDeviceService.exe or so is listening on TCP port 27015. :p


- Ben

Adam amckern McKern wrote:

How the hell is an xml output ended up in your query?


Owner Nigredo Studios http://www.nigredostudios.com

--- On *Tue, 25/1/11, Netshroud /netshr...@gmail.com/* wrote:


From: Netshroud netshr...@gmail.com
Subject: [hlcoders] Source RCON PLIST?
To: hlcoders@list.valvesoftware.com
Received: Tuesday, 25 January, 2011, 11:38 AM

I'm currently trying to build an RCON tool (in C++), but I've come
across something strange and apparently undocumented. So far I've
only implemented SERVERCOMMAND_AUTH, but this is the output I'm
getting on 127.0.0.1 (password is password, buffer is 255 bytes):

Authentication packet:
0x16 0x0 0x0 0x0 0x0 0x0 0x0 0x0 0x3 0x0 0x0 0x0 0x70 0x61
0x73 0x73 0x77 0x6F 0x72 0x64 0x0 0x0
Sent Authentication packet.
Bytes received: 16
☺

Bytes received: 254
?xml version=1.0 encoding=UTF-8?
!DOCTYPE plist PUBLIC -//Apple//DTD PLIST 1.0//EN
http://www.apple.com/DTDs/PropertyList-1.0.dtd;
plist version=1.0
dict
keyMessageType/key
stringResult/string
keyNumber/key
integer6/in

Bytes received: 24
teger
/dict
/plist


Connection closed


Since when did Source send PLIST responses? I'm unable to get a
packet capture of this as Wireshark can't capture loopback, and
Source Multiplayer Dedicated Server (running TF2) is flat out not
responding if I query on my adapter's IP, even from another
machine. Has the RCON protocol changed?




___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlcoders