gaoesa

Nothing of that sort was done. Can you provide screenshots? EDIT: The crouching icons have always been on the head height.

There has not been any changes that would affect this in 0.7.1. Maybe your ping is little higher then before. It is normal and old behaviour that when two players try to occupy same space there can be small a bounce. When the ping difference is significant enough. As the low pinger will take the space first. Also, some players have been known to bind push commands with the movemet. If you can change the feeling into a video for example when colliding with a standing player that would make better description.

http://mygamingtalk.com/forums/tracker/issue-218-class-restrictions-with-values/?verfilter=26 Sorry for the inconvenience.

Added to the 0.7.2 version. Sorry it took this long.

The changelog must be missing it. These should be in the server.html cvar list though.

Documentation does have errors that we fix when we encounter them. For the plain reason that the documentation is the last thing we do after everything else and it always has the lowest priority. I don't mean to say we neglect it on purpose though. However, in this case, the documentation says correctly that it does not exist.

The g_realBody 1 does the shoulder level. Otherwise they are not that much harder.

I don't understand what you meant in your last post.

It is not a fool prof system. It can happen that he can manage to connect still. It is just better then using MAC addresses. You could consider using subnet bans if the ident doesn't stop him. Also note that you can whitelist players if they happen to have IP address from the range of the banned subnet.

MAC address is very unreliable way to ban. The mod includes ident which works better. http://mygamingtalk.com/wiki/index.php/Silent_Mod_Server_Cvar#g_identOptions

There are 2 traces. This can not be changed because otherwise some headshots would not get detected. This happens because the trace can end in the bounding box which fully contains the head box, or at least blocks it. The extra leg box is added only for one player if it hits his bounding box. Not for every player or if the hit is already in the head. The behaviour is the same regardless of the g_realbody setting. I.e. if g_realbody is used, the leg box is not added to every player. This same double trace and additional entities is necessary for all hitbox systems. Do note that we have already removed many redundant traces that are present in the etmain or in the ETPub for example. The information of the hit, once resolved, is kept as long as it is needed. The additional overhead from adding one extra leg box to one player is not mentionable enough. Furthermore, it is combined with the fact that the total count of the traces (and building bodies) is already reduced. There migh be some options to improve the performance slightly. But you shouldn't think it would reduce server CPU usage significantly, even with lot of hits to the players. It must be removed from it. Thanks.

That is not how it works. EDIT: Also, there is no such setting as g_realHead. ETPro headboxes are always used.

"kickbots" includes the following flags: k - !kick i - !list c - !cancelvote b - !ban, !unban t - !stats s - !setlevel, !levlist, !levinfo You can see all the flags from here http://mygamingtalk.com/wiki/index.php/Silent_Shrubbot#List_of_Permission_Flags

!userinfo and !useredit, check and edit his personal command flags. EDIT: Also make sure he is not using it through rcon and change it if necessary. EDIT2: I trust you are using the ALP feature to protect admin levels from GUID spoofing.

set g_floodprotect 1 set g_floodthreshold 6 set g_floodWait 1000 Works well on my server.

That would open it for abuse.

That would be possible, but not in 0.7.0.

Yes

You would need to modify the server engine so that it would be distributing different config strings to different clients.

It's not possible to force downloads to individual players. The sv_referencedPaks in the engine determine the pk3 files that the clients need to have and that is distributed to all players.

I don't have a tool to test this, but added something that can stop it. You can try stopping it with Lua script such as: function et_ClientCommand(clientNum, command) local connected = et.gentity_get(clientNum, "pers.connected") if connected == 2 then return 0 -- return without intercepting end return 1 end

How many maps do you have in the etmain folder? Do note that g_maxMapsVotedFor can still exclude it if the resulting map, after g_excludedMaps and last played restrictions, list has more then this amount of maps.

Forgot to mention that you can try issuing /rcon dboptimize command to remove bad records if there are such.

The different database files are expained in the Wiki http://mygamingtalk.com/wiki/index.php/Silent_Shrubbot#Shrubbot_Files The files with _v(xx) are results of database conversions. The database used to be from versions used before the 0.5.2 and the conversion saved the old database files with _v03 file endings. It seems that there is new record created for the player everytime he connects to the server. Do you have QMM or some Lua script that might interfere with the mod. As you can find the same player multiple times, it is not a question of database corruption, but the old record is not found when it is looked up. This effect was caused by !readadmins command in the older versions. You can make an additional check using the log that the affected players have consistently the same full (32 character) GUIDs from map to map.