User panel stuff on forum
  32 posts on 2 pages  First page12Last page
Server Talk
2006-04-28, 19:20
Member
151 posts

Registered:
Jan 2006
There was new version of MVDSV - 0.19.20-develop. In the nearest weeks there will be release version MVDSV 0.20. For this purpose it is necessary to test this and to inform on bugs on sourceforge.net. (Report bugs here: https://sourceforge.net/tracker/?group_id=118599&atid=681546)

In the given release there are collected versions MVDSV and QWDTOOLS for following platforms:
FreeBSD4 i386 (select, kqueue), FreeBSD5 i386 (select, kqueue), Linux GLIBC 2.1 i386, Linux GLIBC 2.3 i386, SunOS 5.10 i386, Win32 i386.

D/L: http://quakeworld.ru/files/vvd/mvdsv/
Full Changelog: http://quakeworld.ru/files/vvd/mvdsv/changes.txt
MD5: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b.md5

Changelog since 0.19 (last official release):
MVDSV v0.19.20-develop build 2320 (2006.04.28)
- replaced all:
atoi -> Q_atoi
atof -> Q_atof
- bugfix from 0.179b: SunOS on SPARC-64 didn't determine big endian (Thanks to Krok)
- for developers: move all common tools functions from common.{c|h} and qwdtools/tools.{c|h} to bothtools.c and bothdefs.h
[server]
- new cvar sv_forcespec_onfull:
0 – old connect algorithm, i.e you got full connect message and continue trying connect to server as player
1 – force player enter server as spectator if all players's slots are busy and if there are empty slots for spectators
default: 0
(Thanks to qqshka)
- bugfix from 0.1727b: possibility to crash server from client via upload function (Thanks to jhodge)
- added kqueue instead select for FreeBSD target: 'make -DKQUEUE' (EXPERIMENTAL)
- fixed number of sended packets while connect with sv_getrealip > 0 to 10
- improved spectator support in qvm mods:
"seems spectator is sort of hack in QW, I let qvm mods serve spectator like we do for normal player" (by qqshka)
- bugfix ktpro: forbid ';' as commands separator if mod is ktpro, because it don't quote arguments from admin users; example: cmd kick "X;quit" => kick X;quit => server will exit (EXPERIMENTAL)
- returned back sys_select_timeout:
MUST be set to ~ (sv_mintic / 1.3) * 1 000 000 = 10 000
(else can occur packets lost if sv_minping > 0)
if set too low then occur higher CPU usage
Example:
1. sys_select_timeout 10000 + FreeBSD 4 = 4-5% of CPU usage on empty server + sv_minping 39 + low packets lost
2. sys_select_timeout 1000000 + FreeBSD 4 = 0% of CPU usage on empty server + sv_minping 39 + ~20 packets lost
Administrator can tune this cvar self.
May be to make auto change sys_select_timeout on connect 1st user and disconnect last or/and set sv_minping...
can set value only between 10 (10 microseconds) and 1 000 000 (1 second)
default: 10000 (10 milliseconds)
- returned back and made cvar registered CVAR_ROM and seted value to 1; can't delete it because ktpro didn't allow to go at some placeses of, for example, start map
- bugfix: spectators can't hear say_team in qvm, qc-based mods not affected by this fix (by qqshka)
- ZQuake's pmove code (by disconnect)
- some ZQuake's Protocol Extensions: PM_TYPE, PM_TYPE_NEW, SERVERTIME, PF_ONGROUND (by disconnect)
- ZQuake-like minping --> no more sys_select_timeout (by disconnect)
- new cvar sv_kickuserinfospamtime – time in seconds for checking userinfo changes spam
0 – disable kick of the userinfo spaming player
default: 3
- new cvar sv_kickuserinfospamcount – number of permitted userinfo changes,
server will kick player if count exceed in sv_kickuserinfospamtime seconds
0 – disable kick of the userinfo spaming player
default: 30
- bugfix from 0.160b: possibility to crash server from client (Thanks to anm-zip)
- bugfix ktpro: very important security fixes (Thanks to anm-zip)
- bugfix from 0.174b: in ktpro command 'klist' show incorrect info in section "Not full connected clients" (Thanks to Renzo)
- new cvar sv_old_status_for_ktpro:
0 – always show new status
1 – show old status in server console if ktpro was detected (need for correct 'klist''s output)
default: 1
Note: 'rcon status' show new status always
- bugfix: didn't compile on AMD64/EM64T – disabled assembler optimization for this platform (Thanks to Renzo and kingiz)
- no more ".gz" strings in sources – used sv_demoRegexp
- bugfix from 0.19.10-develop: didn't close files not from paks (by disconnect)
- removed cvar registered (by disconnect)
- new code in determination of demo existing for easyrecord command
- bugfix from 0.1728: posibility to crash server via lastscores (Thanks to 100Nosao)
- fixed insanly dangerous bug with master_rcon_password :E. Now it can be set only from server configs. (by disconnect)
- replaced bliP's fix for 'connecting...challenge' with Tonik's. In some cases i can't connect for 60 secs to mvdsv servers, never have this problem with zq/fq/ezq servers; recent qw262 server versions use Tonik's fix too. (by disconnect)
- added "scroll" console output on Win32
- bugfix from 0.19.10-develop: incorrect demo numbering while game proceed if used extracts by substrings in command 'cmd dlist', example: 'cmd dlist aerowalk' (Thanks to gLAd)
- give back from EXPERIMENTAL changes in 0.19.10-develop: 'serverinfo key *version changed from "2.40" to "mvdsv QWE_VERSION" and *qwe_version was removed (EXPERIMENTAL)', because ktpro can't determine version of the server and said " command is only for qwe > 2.40" (commands: autotrack, tracklist and etc.) (Thanks to Renzo)
[qwdtools]
- bugfix from 0.19.10-develop: crashed

MVDSV v0.19.10-develop build 2227 (2006.01.25)
- hardcore cleanup. Removed lot of client-onlt code (by Angel)
- work on SunOS SPARC and ix86 targets now
- added MSVC 7.1 solution/project file
[server]
- serverinfo key *version changed from "2.40" to "mvdsv QWE_VERSION" and *qwe_version was removed
(EXPERIMENTAL) (QWE_VERSION = "0.19.10-develop" (by disconnect)
- bugfix: disabled download from gamedir. So only subdirs of gamedir are allowed.
- new variable: allow_download_other (0 by default) - allows/disallows downloading of stuff
- improved timer accuracy for win32. use -nohwtimer to use old-style timing
(by disconnect from ZQuake)
- bugfix to bliP's sv_speedcheck: round to nearest value instead of rounding down
This bug cause to wrong frametime --> wrong /fpslist (ktpro/ktx commands) output.
- To run server you only need a map (.bsp file), qwprogs and mvdsv executable.
You even dont need pak0.pak, etc (by disconnect from ZQuake)
- A thousand bugfixes i found at ZQuake (like merge pmtrace_t with trace_t, etc),
but I dont think I have to write it here, since only developers
can understand it } (by disconnect)
- sv_nailhack is 1 by default. Dunno why it was set 0 at begin of 'mvd-era'. (by disconnect)
- new cvar sv_maxuserid - limits userid in 1...99 range
default: 99
(by disconnect)
- server-side flash for TeamFortress (oldgrens mode) (ported by Angel from QW262)
- added possibility to set small speed for objects on map (by qqshka)
- added possibility to set non integer volume for sounds from mods (by qqshka)
- added possibility to record multiple demos at same time (from FTE by disconnect)
- FTE's QTV (FTV) support (from FTE by disconnect) :
- mvd_streamport - cvar, to control a port for MVD data streaming
- mvd_maxstreams - cvar, to control the maximum number of proxies to allow to connect
directly to the server.
- build: removed define REGEX - can't build only with external regex lib
- new cmd dlistr, cmd dlistregex, cmd demolistr, cmd demolistregex -
same as demolist, but use parameters as regex
- new commands:
lastscores - same as cmd lastscores, but in server console
dlist - same as demolist
dlistr, demolistr, dlistregex, demolistregex - same as cmd dlistr
- new connectionless commands:
dlist, demolist - same as cmd dlist (packet IPORT dlist)
dlistr, demolistr, dlistregex, demolistregex - same as cmd dlistr
- added OpenBSD targets (by Claymore and VVD)
- fix for gratuitous frag updates when v.frags is fractional (from ZQuake by disconnect)
(Thanks to Spike)
- bugfix from 0.19: fullserverinfo was not sent to client if there was lw or ls userinfo keys
- 'cmd dlist' and 'cmd demolist' can show about 4000 demos in list now, but if serverinfo key
status is not "Standby" or demo recording is in progress then showing only last 100 demos;
was made some additional changes:
- MAX_BACK_BUFFERS increased from 16 to 128 (EXPERIMENTAL)
- MAX_REDIRECTMESSAGES increased from 4 to 128 (EXPERIMENTAL)
- bugfix from 0.160b: pause with sv_minping > 0 cause all players disconnected
(Thanks to Siberex)
[client]
- removed client sources. MVDSV now consist of server and qwdtools only (by Angel)
kill me now and burn my soul
2006-04-28, 19:25
Member
151 posts

Registered:
Jan 2006
Direct Links:
FreeBSD4/poll: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_FreeBSD-4_i386.tar.bz2
FreeBSD4/kqueue: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_FreeBSD-4_i386_kqueue.tar.bz2

FreeBSD5/poll: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_FreeBSD-5_i386.tar.bz2
FreeBSD5/kqueue: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_FreeBSD-5_i386_kqueue.tar.bz2

Linux GLIBC 2.1: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_Linux_GLIBC_2.1_i386.tar.bz2
Linux GLIBC 2.3: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_Linux_GLIBC_2.3_i386.tar.bz2

SunOS 5.10: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_SunOS-5.10_i386.tar.bz2

Win32: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_Win32_i386.rar

^^^ All binaries are for x86 hardware, in case of Sparc/PPC hardware - fetch sources and build it (./configure && make) ^^^

Sources: http://quakeworld.ru/files/vvd/mvdsv/mvdsv_0.19.20b_sources.tar.bz2

P.S. For FreeBSD server admins: kqueue much better that poll - so please use it. If you FreeBSD server admin and dont know what is kqueue - quickly google it
kill me now and burn my soul
2006-04-28, 19:59
Member
113 posts

Registered:
Apr 2006
There was no mention of the spectator 2 bug but it looks like it's been taken care of now =)
biomass
2006-04-28, 20:25
Member
151 posts

Registered:
Jan 2006
spectator 2 bug was fixed
kill me now and burn my soul
2006-04-28, 22:42
Moderator
1329 posts

Registered:
Apr 2006
disconnect wrote:
spectator 2 bug was fixed

Nice. So when do we see it again?

Well updated it for qw.cs.fi too, let's see how many bugs are still left.
Servers: Troopers
2006-04-28, 22:55
Administrator
2058 posts

Registered:
Jan 2006
f.nu:27500 runs the newest mvdsv (and ktx mod)
2006-04-29, 02:47
Administrator
2058 posts

Registered:
Jan 2006
- new cvar sv_forcespec_onfull:
0 - old connect algorithm, i.e you got full connect message and continue trying connect
to server as player
1 - force player enter server as spectator if all players's slots are busy and
if there are empty slots for spectators
default: 0
(Thanks to qqshka)

-- server log --

[...]
MVDSV 0.19.20-develop (l), build 2320
Build date: Apr 28 2006, 21:41:16
[...]
Unknown command "sv_forcespec_onfull"
[...]
2006-04-29, 02:47
News Writer
493 posts

Registered:
Jan 2006
what a small release. (hehe gj guys)
2006-04-29, 12:28
Moderator
1329 posts

Registered:
Apr 2006
Empezar wrote:
- new cvar sv_forcespec_onfull:
MVDSV 0.19.20-develop (l), build 2320
Build date: Apr 28 2006, 21:41:16
[...]
Unknown command "sv_forcespec_onfull"
[...]

renzo@IceFyre ~/temp/mvdsv_0.19.20b/source $ cat * |grep -E sv_for
cat: pcre: Is a directory
cat: qwdtools: Is a directory
extern cvar_t sv_forcenick;
if (sv_forcenick.value && cl->login)
cvar_t sv_forcenick = {"sv_forcenick", "0"}; //0 - don't force; 1 - as login;
Cvar_Register (&sv_forcenick);
extern cvar_t sv_forcenick, sv_login;
if (sv_forcenick.value && sv_login.value && host_client->login)
renzo@IceFyre ~/temp/mvdsv_0.19.20b/source $

It isn't there at all?
Servers: Troopers
2006-04-29, 12:50
Member
1011 posts

Registered:
Feb 2006
it doesn't appear to be in latest cvs either?
2006-04-29, 19:49
Member
805 posts

Registered:
Mar 2006
I'm receiving this:

======== QuakeWorld Initialized ========
execing server.cfg
Unknown command "sv_aim"
Unknown command "sv_forcespec_onfull"
https://tinyurl.com/qwbrasil - QuakeFiles
2006-04-29, 20:26
Member
151 posts

Registered:
Jan 2006
It seems VVD made bianres without sv_forcespec_onfull =:-)
sv_aim is gone - remove it from your server configs
kill me now and burn my soul
2006-04-29, 22:46
Member
25 posts

Registered:
Jan 2006
Bad Luck Troopers #1-3 :: KTeams Allround
bltpowah.mine.nu:27500-27502

Servers updated!
This update rocks and I've been longing for it since 19.10.
This is soo great.. solves all known problems on bltpowah!! :>

- THANK YOU _o_ -o- _o_
My Rocket Will Soon Be In Your Pocket!!! (c) xazqe - 1999
2006-04-30, 18:18
Member
805 posts

Registered:
Mar 2006
Is possible to have different flood protection rules for say and say team?
https://tinyurl.com/qwbrasil - QuakeFiles
2006-04-30, 22:41
Member
1754 posts

Registered:
Jan 2006
this version rocks(don't know any difference)
2006-05-01, 03:44
Member
79 posts

Registered:
Jan 2006
Peppe wrote:
this version rocks(don't know any difference)

for starters, some major exploits to get admin access to the server has been fixed
i just joined the blt server to try some of the old exploits to gain admin access and all of the ones i've tried seemed to be fixed. also, the server sends out warning messages when someone try to use them, which is great.

i guess you can see this version as a bug fix version - not any big new features for the players, but small improvments which really makes it safer to run a qw server and also makes the overall player experience better in the long run.

this kind of work (developing/coding) should be more appreciated by the community
zaiko of BaconMen
2006-05-01, 07:09
Member
1026 posts

Registered:
Feb 2006
quaketv!
god damn hippies >_<
2006-05-01, 10:45
Member
1754 posts

Registered:
Jan 2006
ok, nice
2006-05-02, 21:54
Moderator
1329 posts

Registered:
Apr 2006
Would it be possible to consider the following things?

1) check for vip_addip when banning. If the VIP's ip has "vip level 63" (rcon_admin) then no ban is allowed.

This allows rcon admins not to be banned and able to join server when some idiot (that has gotten admin access somehow) makes a temporary ban. Should make server admin's life much easier if they didn't have to go through server console to remove such bans.


2) I was thinking about disallowing real admins to kick someone with not_kick rights, but I started to think about it again. The point was that if someone has been given not_kick rights, no "admin" should be able to kick him (not_kick affects only elected admins at the moment) even if you could do that kick from the console/rcon rights. This would also make "vip level 63" (rcon admin) more protected/powerful than a mere admin.
Servers: Troopers
2006-05-04, 02:10
Member
123 posts

Registered:
Mar 2006
Idea - minping for individual players. This would be helpful in NA vs EU matches, etc where one player or a few ping much higher. Say you are playing 2v2, you have 2x 39ms vs 39ms and 117ms. If you minping you can raise it to 117ms, and everyone gets same high ping. Or you could raise one person to 117ms, or both to 78ms, etc so the averages equal but not everyone has to lag. You can also qizmo reroute, but individual minping would be much faster, and easier.
2006-05-04, 07:46
Administrator
2058 posts

Registered:
Jan 2006
why do we even have minping when people can just reroute through qizmo
2006-05-04, 07:51
Moderator
1329 posts

Registered:
Apr 2006
Empezar wrote:
why do we even have minping when people can just reroute through qizmo

Indeed. The whole damned command should be disabled by server admins since it only fucks up the whole server making aim feel like shit, movement being jerky and shots fired feeling laggy. (read: real/rerouted 52ms is actually better than sv_minping 40).
Servers: Troopers
2006-05-04, 10:45
Member
151 posts

Registered:
Jan 2006
Few words about minping:
Minping delays only "incoming" packets (packets send by player to a server. They are incoming for server. Packets send by server to a player are outcoming for server. Here i'll talk only about server packets). But in real word we have half delay on message receive and half dellay on send.

When is going when we have sv_minping 40:
incoming packets delayed up to 40 ms. If you have _REAL_ 41 ms, then your packets delayed on receive by ~20~ms and ~20~ms on send. If you have 39 ms, ten your packets delayed on receive by ~40ms~ and ~20ms~ on send.

With sv_minping 40. You have ping about 40+(half of your ping to this server), but "incoming" delay have much power on "button reaction'. When you jump or fire with sv_minping 40, it acts like you have 80ms ping. Thats why sv_minping sucks.

Solution: Split sv_minping delays. MVDSV _MUST_ delay 0.5 * sv_minping.value on receive and 0.5 * sv_minping.value on send. I dont know so well this part of MVDSV sources, so i cant do it :E.
kill me now and burn my soul
2006-05-05, 04:51
Member
113 posts

Registered:
Apr 2006
disconnect wrote:
It seems VVD made bianres without sv_forcespec_onfull =:-)

Actually it seems that the functionality is there by default but since the var was left out of the code you can't disable it ?
biomass
2006-05-05, 09:30
Moderator
1329 posts

Registered:
Apr 2006
biomass wrote:
disconnect wrote:
It seems VVD made bianres without sv_forcespec_onfull =:-)

Actually it seems that the functionality is there by default but since the var was left out of the code you can't disable it ?

Yeah I noticed it actually being working too. I was "browsing" qw.cs.fi servers for free port with spectator 0 and I got connected as a spectator few times since there were some games going on.
Servers: Troopers
2006-05-05, 09:32
Member
693 posts

Registered:
Jan 2006
I think Wargamez has been updated now, at least Autotrack works anyway. Is there an rcon command that tells you the server version?
2006-05-05, 11:44
Member
1011 posts

Registered:
Feb 2006
you can just look at 'serverinfo' i think
2006-05-05, 13:35
Administrator
2058 posts

Registered:
Jan 2006
yeah its updated
2006-05-05, 20:55
Member
151 posts

Registered:
Jan 2006
If you want to know OS at server, connect to it as player. It'll print you:
mvdsv 0.19.20 (f)
f = FreeBSD
o = OpenBSD
w = Windows
l = Linux
kill me now and burn my soul
2006-05-06, 13:19
Member
810 posts

Registered:
Jan 1970
s = SunOS
  32 posts on 2 pages  First page12Last page