dreamstalker/rehlds

Reverse-engineered HLDS

rehlds's Issues

Amxmodx func_buyzone EngFunc_RemoveEntity problem and csdm_remove_weaponbox func. on last build

Hello,
We have problem rehlds-0.2-20150707.203020-55 bugfixed version.

I was try to using CSDM mod with reHLDS server will crash on remove func_buyzone entities.
no problem func_hostage_rescue or info_hostage_rescue just func_buyzone

CSDM mod works perfect rehlds-0.2-20150705.202431-54 bugfixed version

if ((g_MapStripFlags & MAPSTRIP_HOSTAGE) /* && (containi(mapname, "cs_") != -1) */)
{
        RemoveEntityAll("func_hostage_rescue")
        RemoveEntityAll("info_hostage_rescue")
}

if (g_MapStripFlags & MAPSTRIP_BUY)
{
        RemoveEntityAll("func_buyzone") 
}
...
stock RemoveEntityAll(name[])
{

    new ent = engfunc(EngFunc_FindEntityByString, 0, "classname", name)
    new temp
    while (ent)
    {
        temp = engfunc(EngFunc_FindEntityByString, ent, "classname", name)
        engfunc(EngFunc_RemoveEntity, ent) //crashed here
        ent = temp
    }
}

and i have got a problem csdm_remove_weaponbox function.

run_drop_wbox(id, ent, shield)
{
    new ret
    new model[32]
    ExecuteForward(g_drop_fwd, ret, id, ent, 0)
    if (ret == CSDM_DROP_REMOVE)
    {
        if (shield)
            csdm_remove_weaponbox(id, ent, 0, 1, 1)
        else
            csdm_remove_weaponbox(id, ent, 0, 1, 0)
        return 1
    } 
    else if (ret == CSDM_DROP_IGNORE) 
    {
        return 0
    }
    if (g_StayTime > 20 || g_StayTime < 0)
    {
        return 0
    }
    if (ent)
    {
        pev(ent, pev_model, model, 31)
        if (((equali(model,"models/w_usp.mdl")) || (equali(model,"models/w_glock18.mdl")))
                && (g_StripWeapons))
            csdm_remove_weaponbox(id, ent, 0, 0, 0) //boom crash when player drop usp or glock

        else if ((equali(model,"models/w_backpack.mdl")) && (g_RemoveBomb))
            csdm_remove_weaponbox(id, ent, 0, 0, 0)
        else if (shield)
            csdm_remove_weaponbox(id, ent, g_StayTime, 1, 1)
        else
            csdm_remove_weaponbox(id, ent, g_StayTime, 1, 0)
        return 1
    }
    return 0
}

static cell AMX_NATIVE_CALL csdm_remove_weaponbox(AMX *amx, cell *params)
{
    // Check index.
    unsigned int owner = params[1];
    unsigned int ent = params[2];
    const unsigned char m_rgpPlayerItems_wpnbx[] = { 34, 35, 36, 37, 38, 39 };

    if (owner < 1 || owner > (unsigned int)gpGlobals->maxClients)
    {
        MF_LogError(amx, AMX_ERR_NATIVE, "Invalid player %d", owner);
        return 0;
    } else if (!MF_IsPlayerIngame(owner)) {
        MF_LogError(amx, AMX_ERR_NATIVE, "Player %d is not in game", owner);
        return 0;
    }

    unsigned int weapon_shield = params[5];
    edict_t *searchEnt = INDEXENT(ent);

    if (FNullEnt(searchEnt))
    {
        MF_LogError(amx, AMX_ERR_NATIVE, "Invalid weaponbox %d", ent);
        return 0;
    } else if ((weapon_shield != 1) && (strcmp(STRING(searchEnt->v.classname),"weaponbox") != 0)) {
        MF_LogError(amx, AMX_ERR_NATIVE, "Not a weaponbox %d", ent);
        return 0;
    } else if ((weapon_shield == 1) && (strcmp(STRING(searchEnt->v.classname),"weapon_shield") != 0)) {
        MF_LogError(amx, AMX_ERR_NATIVE, "Not a shield %d", ent);
        return 0;
    }

    edict_t *pEdict = MF_GetPlayerEdict(owner);
    if (weapon_shield == 1)
    {
        if (searchEnt->v.owner == pEdict)
        {
            if (!params[3])
            {
                if (NotifyForRemove(owner, searchEnt, NULL))
                {
                    REMOVE_ENTITY(searchEnt);
                }
            } else {
                RemoveWeapon::SchedRemoval(params[3], pEdict, NULL, searchEnt);
            }
            return 1;
        }
        MF_LogError(amx, AMX_ERR_NATIVE, "Shield %d not found", ent);
        return 0;
    }

    edict_t *pWeapEdict;
    unsigned char i;

    for( i = 0; i < sizeof (m_rgpPlayerItems_wpnbx); i++ )
    {
        pWeapEdict = get_pdata_cbase( searchEnt, m_rgpPlayerItems_wpnbx[ i ] + EXTRAOFFSET_WEAPONS );

        if (pWeapEdict == NULL)
            continue;
        if (FNullEnt(pWeapEdict))
            continue;
        if (pWeapEdict->free)
            continue;
        if ((pWeapEdict->v.flags & FL_FAKECLIENT) || (pWeapEdict->v.flags & FL_CLIENT))
            continue;
        if (pWeapEdict->v.owner == searchEnt)
        {
//          ALERT(at_logged, "CSDM - found weapon edict in private data of weaponbox in slot %d.\n", i);
            if (!params[3] || !params[4])
            {
                if (NotifyForRemove(owner, pWeapEdict, searchEnt))
                {
                    REMOVE_ENTITY(pWeapEdict);
                    REMOVE_ENTITY(searchEnt);
                }
            }
            else 
            {
                FindWeapon *p = FindWeapon::NewFindWeapon(pEdict, STRING(pWeapEdict->v.classname), params[3]);
                g_Timer.AddTask(p, 0.1);
                return 1;

//              RemoveWeapon::SchedRemoval(params[3], pEdict, searchEnt, pWeapEdict);
            }
            return 1;
        }
    }
    ALERT(at_logged, "CSDM - weapon edict not found in private data of weaponbox!!!.\n");
    return 0;
}

Banned for move commands flooding (burst)

Hi. I tried to install rehlds on several servers on fastcup.net, but servers are banned some players with reason "Banned for move commands flooding (burst)"

Current settings:
sv_maxupdaterate 102
sv_minupdaterate 102
sv_maxrate 100000
sv_minrate 100000
sys_ticrate 1060

Rehlds build: 563

Linux binaries depend on GLIBCXX_3.4.20 which is not yet available on popular distros

Details: http://cs.rin.ru/forum/viewtopic.php?p=1064076#p1064076

Exploits and other vulnerability

1. Fix command spawn flood
2. Fix old bots connection-flood
3. integration PPS limiter 1ip+port=50 pps

Bits writing MSG_ functions may write data outside allocated buffer when sizebuf_t is overflowed

Bits writing MSG_ functions (MSG_WriteBits, MSG_WriteOneBit) are designed in a way that the byte being currently written lays outside of cursize of sizebuf on which bits writing is performed.
This could be clearly seen in MSG_StartBitWriting's code

    bfwrite.nCurOutputBit = 0;
    bfwrite.pbuf = buf;
    bfwrite.pOutByte = &buf->data[buf->cursize]; // <<<<=========

Here we set the bfwrite.pOutByte to a pointer in the buf->data buffer that lays outside buf->curside, and, possibly, may even lay outside allocated buf->data buffer (when cursize == maxsize). Notice that SZ_GetSpace() is not called here, so SIZEBUF_OVERFLOWED flag will not be set.

If we then call MSG_WriteOneBit(), data will be written into bfwrite.pOutByte pointer, without calling SZ_GetSpace(). In that case 1 byte after the allocated buf->data buffer will be corrupted.

Another piece of crappy code is hiding inside MSG_WriteBits:

    int surplusBytes = 0;
    if ((uint32)bfwrite.nCurOutputBit >= 8) {
        surplusBytes = 1; //surplusBytes could be either 0 or 1
        ....
    }
    ....
        SZ_GetSpace(bfwrite.pbuf, surplusBytes + 4); //allocate 4 or 5 bytes
        if (!(bfwrite.pbuf->flags & SIZEBUF_OVERFLOWED))
        {
            // write 4 bytes (it's ok)
            *(uint32 *)bfwrite.pOutByte = (data << bfwrite.nCurOutputBit) | *(uint32 *)bfwrite.pOutByte & ROWBITTABLE[bfwrite.nCurOutputBit];
            int leftBits = 32 - bfwrite.nCurOutputBit;
            bfwrite.nCurOutputBit = bits & 7;
            bfwrite.pOutByte += 4;
            *(uint32 *)bfwrite.pOutByte = data >> leftBits; //then write another 4 bytes! wtf?
        }

Here (in some rare cases) we allocate 4 or 5 bytes are from sizebuf and then write 8 bytes there. If sizebuf's data is the last variable allocated on stack and -fomit-frame-pointers is used, the whole return address will be rewritten and hlds will experience a (not) funny trip into nowhere when it executes the ret instruction, he-he.

Decal_Init func reads incorrect decals.wad - ignores mod

Hi
It seems that rehlds looks for decals.wad in DEFAULTGAME instead of GAME folder.
In other words it ignores MOD settings (-game modname).

How to replicate:

1. run PROCMON.EXE from Sysinternals tools to monitor which decals.wad is read.
2. filter to "Process" is "hlds.exe" and "Path" contains "decals.wad",
3. rename valve\decals.wad
4. run hlds.exe - game gearbox .. or any other mod.
5. HLDS crashes with "Couldn't find DECALS.WAD in DEFAULTGAME"
6. Procmon reveals that hlds looks for the file only in VALVE and VALVE_HD folders. It ignores mod settings.

rehlds/rehlds/engine/decals.cpp

Line 538 in 101a5f3

void Decal_Init(void)

Plz fix mod detection.

Simple anti-cheat logic

It's just a question/request.. so

Can we have a simple code implemented to break current cheats? something like to stop sending unwanted data to clients (non-visible enemies position for e.g). This feature will get you some real popular usage besides stability and improvement.

anyway, thank you guys.

Error Windows XP SP 3

Hello.
These errors appear after replacing the file swds.dll folder "bugfixed" and "pure".

An error occurred when the server starts:
"The procedure entry point K32GetModuleInformation not be located in the dynamic link library KERNEL32.dll"

Configuration OS:
Windows XP SP 3

Function Host_Kick_f is not working correctly

Oke. I'll kick myself on userid

But i am still on the server, was not kicked
And now let's write

And it's working

Please, fix "kick 'userid'" cmd, this cmd absolutely not working. "kick 'name'" cmd is working nice

It's not possible to call chain->callNext() several times in interceptors

Current implementation of hookchains assumes that interceptor function must call chain->callNext() zero or one time. Multiple calls are not allowed.

The task is to rewrite hookchains code responsible for interceptors chaining and make it possible to call chain->callNext() several times.

Autobuy doesn't buy primary weapon in TE team.

Autobuy doesn't buy primary weapon in TE team.
And rebuy doesn't buy ammo too.

Accuracy Fix

There is such a plugin Accuracy Fix.
https://forums.alliedmods.net/showthread.php?t=166648
which celebrates fire in the engine, if the sense to put it, or all the corrections is or will be?

SZ_GetSpace: overflow on

http://prntscr.com/aaygy2
проблема носит массовый характер, так что я думаю разработчикам стоит обратить на это внимание. Игроки жалуются каждый день, что тупо виснет все и кикает по таймауту.

sys_ticrate > 1000

Why rehlds not supports server fps > 1000?
When i set sys_ticrate to 1060 (for example), in stats fps still < 1000 (unstable, 850-990)

Intel Xeon E5-1620 v2, Ubuntu 14.04 with kernel 4.1.0-rt10 (realtime, 1000hz).

HLSDK NEWAPI function pfnShouldCollide doesn't work properly

This bug was discovered by hzqst https://forums.alliedmods.net/member.php?u=39607

When gamedll returns false in pfnShouldCollide (entities shouldn't collide), engine ignores collisions with all subsequent entities. This can lead to bugs in physics. The bug occurs because of a premature return from SV_ClipToLinks https://github.com/dreamstalker/rehlds/blob/master/rehlds/engine/world.cpp#L1196

        if (gNewDLLFunctions.pfnShouldCollide && !gNewDLLFunctions.pfnShouldCollide(touch, clip->passedict))
            return;

It would be correct to continue the loop instead of return.

online player count problem

Hello ,
we have little problem like this
branch : stack_align

20150628.204135-49 build some problems

Hello,
if server crashed create dir in hlds folder like this
http://i.imgur.com/YQC6Is6.png
X�j�X�j�(u5�(u5�
X�m�X�m�(�� (��
And if i use amxmodx with rehlds server crashed when player connecting

Commands with unicode don't work properly

For example:

echo Привет мир

shows empty string

But this works perfect:

echo "Привет мир"

compiling error

I got a few erros

[root@localhost
rehlds]# ./gradlew --max-workers=1 clean buildRelease
:buildSrc:compileJava UP-TO-DATE
:buildSrc:compileGroovy UP-TO-DATE
:buildSrc:processResources UP-TO-DATE
:buildSrc:classes UP-TO-DATE
:buildSrc:jar UP-TO-DATE
:buildSrc:assemble UP-TO-DATE
:buildSrc:compileTestJava UP-TO-DATE
:buildSrc:compileTestGroovy UP-TO-DATE
:buildSrc:processTestResources UP-TO-DATE
:buildSrc:testClasses UP-TO-DATE
:buildSrc:test UP-TO-DATE
:buildSrc:check UP-TO-DATE
:buildSrc:build UP-TO-DATE
:dep/bzip2:clean
:dep/cppunitlite:clean UP-TO-DATE
:flightrec/decoder:clean UP-TO-DATE
:flightrec/decoder_api:clean UP-TO-DATE
:rehlds:clean
:dep/bzip2:compileReleaseBzip2StaticLibraryBzip2Bz2_main
:dep/bzip2:compileReleaseBzip2StaticLibraryBzip2Bz2_pch
:dep/bzip2:createReleaseBzip2StaticLibrary
:dep/bzip2:releaseBzip2StaticLibrary
:rehlds:generateAppVersion
:rehlds:compileReleaseRehldsFixesRehlds_swds_engineSharedLibraryRehlds_swds_engineRehlds_pch
In file included from /home/genki/rehlds/rehlds/rehlds/public/rehlds/osconfig.h(92),
from /home/genki/rehlds/rehlds/rehlds/rehlds/precompiled.h(5),
from /home/genki/rehlds/rehlds/rehlds/rehlds/precompiled.cpp(1):
/usr/lib/gcc/x86_64-redhat-linux/4.4.7/include/smmintrin.h(32): error: #error directive: "SSE4.1 instruction set not enabled"

error "SSE4.1 instruction set not enabled"

^

compilation aborted for /home/genki/rehlds/rehlds/rehlds/rehlds/precompiled.cpp (code 2)
:rehlds:compileReleaseRehldsFixesRehlds_swds_engineSharedLibraryRehlds_swds_engineRehlds_pch FAILED
FAILURE: Build failed with an exception.

• What went wrong:
  Execution failed for task ':rehlds:compileReleaseRehldsFixesRehlds_swds_engineSharedLibraryRehlds_swds_engineRehlds_pch'.
  A build operation failed.
  C++ compiler failed while compiling precompiled.cpp.
  See the complete log at: file:///home/genki/rehlds/rehlds/rehlds/build/tmp/compileReleaseRehldsFixesRehlds_swds_engineSharedLibraryRehlds_swds_engineRehlds_pch/output.txt
• Try:
  Run with --stacktrace option to get the stack trace. Run with --info or --debug option to get more log output.
  BUILD FAILED
  Total time: 18.264 secs

---

[root@localhost rehlds]# cat /etc/release
CentOS release 6.7 (Final)
CentOS release 6.7 (Final)
CentOS release 6.7 (Final)

[root@localhost rehlds]# uname -a
Linux localhost.localdomain 2.6.32.71 #2 SMP Wed Mar 23 01:01:07 EET 2016 x86_64 x86_64 x86_64 GNU/Linux

[root@localhost rehlds]# uname -r
2.6.32.71

smmintrin.h 31-33 lines

ifndef SSE4_1

#error "SSE4.1 instruction set not enabled"

else

Cpu information

[root@localhost rehlds]# cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 44
model name : Intel(R) Xeon(R) CPU X5660 @ 2.80GHz
stepping : 2
cpu MHz : 2792.990
cache size : 12288 KB
physical id : 1
siblings : 12
core id : 0
cpu cores : 6
apicid : 32
initial apicid : 32
fpu : yes
fpu_exception : yes
cpuid level : 11
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm dca sse4_1 sse4_2 popcnt aes lahf_lm ida arat tpr_shadow vnmi flexpriority ept vpid
bogomips : 5585.98
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management:

Make ReHlds SDK compatible with original HLSDK

During reverse engineering some changes were made in SDK that now break compatibility with original hlsdk. These changes include:

• Use of osconfig.h which is rehlds-specific
• Use of types like uint32_t which are not defined in original sdk
• Use of "precompiled.h" header in public .cpp files

The task is to refactor both sdk and rehlds to make rehlds sdk compatible with original hlsdk.

Definition of done: Amxmodx is successfully compiled with rehlds sdk

Longjump not working (bugged) when using CTF powerup (item_ctflongjump)

Hi
I've been testing rehlds thoroughly on Opposing Force. There are basically two options how to get long jump - by grabbing the well known item_longjump or by taking powerup item_ctflongjump (it can be dropped)

I noticed that long jump works correctly for the default item_longjump, but is bugged for item_ctflongjump (your jump stops midair like if you hit invisible wall).

Maybe it's related to the following bool value:

rehlds/rehlds/dlls/player.h

Line 169 in 101a5f3

BOOL m_fLongJump; // does this player have the longjump module?

Plz fix it guys.

I'm running a password protected server with bugfixed version (rehlds-0.2-20160128.133207-120).
Run the following commands in opfor console to connect the server (will turn it off in a week)
password rehlds
connect 31.31.73.85

test map
longjumptest.zip

Thanks

Sometimes players start flying or fall under the ground

A screenshot for illustrating the problem:

Short description

Delta encoder, when using some entity as a baseline to encode changes of another entity, doesn't take into account that some fields could be 'switched off' by the gamedll for baseline entity when it was delta-encoded.

Steps to reproduce

(success rate is not 100%, it depends on map and position of players):

1. Start hlds on de_aztec
2. Join the server. You have to be the first player on the server (this increases chance to reproduce the bug)
3. Join CT team
4. Wait for some other players to join the server
5. When round starts go upstairs to the B bomb spot
6. Wait for other players to go upstairs
7. Start demo recording (using record command on client)
8. If you're lucky, most of the players should start flying

Technical background

Now let's see what's going on.

1. You are the first player on the server (i.e. your data is stored in svs.client[0] and your entity index is 1)
2. The second player (svs.clients[1]) have the same Z coord as you. (sv.edicts[1].v.origin[2] == sv.edicts[2].v.origin[2])
3. You starts demo recording
4. Your client doesn't send clc_delta in next cmd packet. client[0]->delta_sequence will become -1 because of that
5. Server is going to transmit entities to your client, SV_EmitPacketEntities() is called
6. Since client->delta_sequence == -1, sv_packet_nodelta is emitted
7. sv_packet_nodelta means that delta encoder should not use previous entities' states to encode current states
8. Entity of the first player (of you) is encoded using sv.baselines[1] as a baseline. However, gamedll sees that server is going transmit your own state to you, and unmarks ent->v.origin fields (because your own movement is calculated on client side)
sv.edicts[1].v.origin fields are not transmitted to the client and will be 0.0f there
9. Entity of the second player is encoded using first player's entity as a baseline (usually)
10. Since first and second player have the same Z coord, sv.edicts[2].v.origin[2] will not be marked as changed and will not be transmitted
11. When your client decodes delta for the second player, it will use origin[2] from the baseline (i.e. from edicts[1].origin[2])
12. However, server didn't transmit sv.edicts[1].v.origin[2] because gamedll unmarked it
13. sv.edicts[2].v.origin[2] will be set to 0.0f therefore
14. If ground's z coord is lower that 0 in the x,y position of the player, the player will 'start flying'
15. If ground's z coord is higher that 0 in the x,y position of the player, the player will 'fall under the ground'

Сrash after bot disconnect

OS Windows XP SP3 32 bit
ReHLDS 624 bugfixed

drwtsn32.txt

Clients are dropped with the "server failed to transmit file 'AY&SY..." message

Server sometimes kicks one or more clients with "server failed to transmit file 'AY&SY..." message in client's console.

Reason

It happens because of invalid 'reuse compressed buffer' optimization in SV_UpdateToReliableMessages():

for (i = 0; i < g_psvs.maxclients; i++) {
    client = &g_psvs.clients[i];

    if (g_psv.reliable_datagram.cursize + client->netchan.message.cursize < client->netchan.message.maxsize)    {
        SZ_Write(&client->netchan.message, g_psv.reliable_datagram.data, g_psv.reliable_datagram.cursize);
    } else {
        Netchan_CreateFragments(TRUE, &client->netchan, &g_psv.reliable_datagram);
    }
}

This code writes server's reliable message into reliable message buffer of each connected client. If there is enough space in client's reliable buffer, server reliable message is written 'as is' via SZ_Write function. Otherwise fragments are created via Netchan_CreateFragments function. The key part here is that Netchan_CreateFragments overwrites contents of g_psv.reliable_datagram with compressed version of this buffer.

Consider the following scenario:

1. g_psv.reliable_datagram doesn't fit into client N1 reliable buffer
2. g_psv.reliable_datagram is compressed and overwritten by Netchan_CreateFragments() function
3. g_psv.reliable_datagram fits into client N2 reliable buffer
4. Contents of g_psv.reliable_datagram is written into client N2's reliable buffer 'as is' by SZ_Write() function
5. Client N2 is not able to parse compressed data written by SZ_Write and will generate svc_bad error.

Info_ functions use strcpy() on overlapped buffers

Info_* functions use Q_strcpy (which is alias for strcpy) for operations on info buffer. Since copying is done on same info buffer, src and dst buffers may overlap each other, which is forbidden by C/C++ specification.

On Windows strcpy() appears to work fine with overlapped buffers while icc's optimized version of strcpy() on linux corrupts info buffer if dst and src are close to each other.

HLDS in GUI mode crashes on windows

HLDS with swds.dll from rehlds package crashes on windows if it is not started in console mode
Reason: GameServerData001 interface used in AdminServer.dll is not exposed by swds.dll

"Your .dll (cl_dlls/client.dll) differs from the server's" fix - accepting p48 clients

Hi

While I admire your work, especially the performance boost, the Re-HLDS is so good in mimicking HLDS 6153 behavior - as it keeps rejecting older clients (probably sending incorrect CRC).

I'm getting following error on client "Your .dll cl_dlls/client.dll differs from the server's". And then client disconnects.

Client and server's client.dll do match and I still get an error.
This error can be circumvented if you change few bytes in client-side hw.dll to skip the received CRC check for client.dll.

I also managed to circumvent this by replacing swds.dll with the one from build 5758 and having client.dll matching on client/server.
I tried to disassemble swds.dll and find and remove these checks but I failed.

I did couple of tests and debugging.
swds 6153 + serverside client.dll new --> error
swds 6153 + serverside client.dll old (same as clientside) --> error
swds 5758 + serverside client.dll new --> error
swds 5758 + serverside client.dll old (same as clientside) --> client can connect now, all is working
swds 5758 + serverside client.dll old + any (new/old) clientside client.dll always works.

I'm willing to be long-term tester, I'm having private server with public IP, running Opposing Force.
My goal is to make server accept all p48 clients (even those non-steam older ones, by using revemu)

All I'm asking you is to fix the infamous "your .dll(cl_dlls/client.dll) differs from the server's" error.
Could please have a look into the issue?
Let me know if you need more info, affected hl version or temporary RDP access to my server.

HLDS5758.zip

Thx!

Linux binaries are too large (>2MB)

Examine what the hell is going on with linux binaries - they are 2+ times larger than windows binaries

Animation skip during m4/ak spray

You can fix this bug?
I want to put your engine on all servers fastcup.net

Notification about unknown command in server console

В исходнике engine/cvar.cpp функции
qboolean Cvar_Command(void)
{
cvar_t *v;
const char *arg0 = Cmd_Argv(0);
const char *firstToken = Cvar_IsMultipleTokens(arg0);

if (firstToken)
{
    v = Cvar_FindVar(firstToken);
    if (v)
    {
        Con_Printf("\"%s\" is \"%s\"\n", v->name, v->string);
        return TRUE;
    }
}
else
{
    v = Cvar_FindVar(arg0);
    if (v)
    {
        if (Cmd_Argc() == 1)
        {
            Con_Printf("\"%s\" is \"%s\"\n", v->name, v->string);
        }
        else
        {
            if (v->flags & FCVAR_SPONLY && g_pcls.state >= ca_connecting && g_pcl.maxclients > 1)
            {
                Con_Printf("Can't set %s in multiplayer\n", v->name);
            }
            else
            {
                Cvar_Set(v->name, Cmd_Argv(1));
            }
        }

        return TRUE;
    }
}

return FALSE;

}

внести 2 раза код
else
{
Con_Printf("unknown command "%s"\n", arg0);
}

qboolean Cvar_Command(void)
{
cvar_t *v;
const char *arg0 = Cmd_Argv(0);
const char *firstToken = Cvar_IsMultipleTokens(arg0);

if (firstToken)
{
    v = Cvar_FindVar(firstToken);
    if (v)
    {
        Con_Printf("\"%s\" is \"%s\"\n", v->name, v->string);
        return TRUE;
    }
    else
    {
        Con_Printf("unknown command \"%s\"\n", arg0);
    }
}
else
{
    v = Cvar_FindVar(arg0);
    if (v)
    {
        if (Cmd_Argc() == 1)
        {
            Con_Printf("\"%s\" is \"%s\"\n", v->name, v->string);
        }
        else
        {
            if (v->flags & FCVAR_SPONLY && g_pcls.state >= ca_connecting && g_pcl.maxclients > 1)
            {
                Con_Printf("Can't set %s in multiplayer\n", v->name);
            }
            else
            {
                Cvar_Set(v->name, Cmd_Argv(1));
            }
        }

        return TRUE;
    }
    else
    {
        Con_Printf("unknown command \"%s\"\n", arg0);
    }
}

return FALSE;

}

Результат этому делу.
Вводим не известную команду hg
и получаем сообщение.

unknown command "hg"

Так будет удобнее :)
Хотелось бы это видеть в релизе.
С уважением индус кодерь ;)

SV_RunThink edict free bug

In sv_phys.cpp:124 edict must be cleared by FL_KILLME flag. But here must be FL_SPECTATOR flag.

Let's look into HLSDK:
https://github.com/ValveSoftware/halflife/blob/5d761709a31ce1e71488f2668321de05f791b405/dlls/util.cpp#L1558

UTIL_Remove delete entity from server by setting FL_KILLME flag, and engine cleares it. When FL_SPECTATOR is only set for spectators.

Too slow spray uploading

Client starts upload spray while he is connecting, but when client is connecting CL_GetFragmentSize always return 128. (it returns cl_dlmax only if client state is ca_active)
What about to move upload request code into SV_SendEnts_f function?

reHLDS UDP Queries /PHP/

How to get information like server name, players, max players, players list, map, os and etc.. ?
On the web there are bunch of working variants for HLDS, but now for reHLDS i cannot get properly information about the server.
The mainserverinfo is \xFF\xFF\xFF\xFFTSource Engine Query\x00 or something new ?
Can someone give me working variant for php ?

Sprays not working correctly.

Many (All?) custom sprays are not working.

Steps to reproduce:
1)Set the following example tempdecal.wad (https://www.sendspace.com/file/lfwbds) in your client cstrike directory.
2)Try to connect to ReHLDS server (make sure sv_allowuploads 1).

Client will get stuck in "Precaching resources..."

Same steps on regular HLDS result in client connecting properly and spray showing properly.

If sv_allowuploads is set to 0, then client will be able to connect (sprays won't show of course).

Also, related to this are segfault crashes after changelevel.

This line

rehlds/rehlds/engine/hashpak.cpp

Line 366 in fef6405

Q_memcpy(&newdirectory.p_rgEntries[i + 1], &olddirectory.p_rgEntries[i + 1], sizeof(hash_pack_entry_t));

Will result in segfault at random after changelevel (most likely related to what the clients' sprays are).
This gets called from

void HPAK_FlushHostQueue(void)

which gets called from

void EXT_FUNC SV_ActivateServer_internal(int runPhysics)

(Github isn't accepting my zip file)
https://www.sendspace.com/file/lfwbds

Typo in Info_IsKeyImportant

Случайно наткнулся.
"botomcolor" -> ""bottomcolor"

Metamod Plugins: DProto & VoiceTranscoder Not Working

Metamod Plugins: DProto & VoiceTranscoder Not Working ( HLDS Downloaded From SteamCMD )

DProto Not Working Ver 0.9.548 & 0.9.356

VoiceTranscoder

Do you plan to rewrite VTC for ReHLDS support?

Change maxplayers from 32 to 40 ?

Hi there, is there any chance to change this?

Add possibility to write multiline kick reasons

I want something like this:
kick #1 \nUpdate anti-cheat, you can download it here:\nhttp://github.com/
will show:

You have been disconnected from the server.
Reason: Kicked :
Update anti-cheat, you can download it here:
http://github.com/

Raise precache limit

Could you raise the precache limit? 512 is not enough for fun servers such as zombie plague and jailbreak...

Name Change

The server will change player names seemingly at random for about 5 minutes after start then it will be fine after.
Also is there some way to increase the string ban limit. Some players are getting banned for using buyscripts.

NonSteam client don't connect from favorites.

NonSteam client don't connect from favorites.
Query port changed to 0???

Here is the video of it.
https://www.youtube.com/watch?v=7VVgiZXk7mU

Bug with integer conversion in MSG_WriteBitString

If string has non-ASCII chars (code >= 0x80) then MSG_WriteBitString will write 0xFF but not char code. There is a problem with conversion from int8 to uint32 (http://en.cppreference.com/w/cpp/language/implicit_cast#Integral_conversions):

If the destination type is unsigned, the resulting value is the smallest unsigned value equal to the source value modulo 2^n where n is the number of bits used to represent the destination type.

That is, depending on whether the destination type is wider or narrower, signed integers are sign-extended or truncated and unsigned integers are zero-extended or truncated respectively.

How it works:

1. For example we have char with 0xAA (-86) code
2. It is sign-extend to 32 bits, now we have 0xFFFFFFAA
3. MSG_WriteBits checks for max value (for 8 bits it is 0xFF), but 0xFFFFFFAA is higher than 0xFF => now value is 0xFF
4. MSG_WriteBits writes 0xFF

You can test it here: http://cpp.sh/4a6i

#include <cstdio>
#include <cstdint>

using namespace std;

int main() {
    signed char a;
    scanf("%hhd", &a);
    uint32_t b = a;
    printf("%.8X\n", b);

    unsigned char c;
    scanf("%hhu", &c);
    uint32_t d = c;
    printf("%.8X\n", d);
}

ReHLDS does not handle pingboost commandline option

i don't know how should i prove it. i think it's not necessary. everyone can test it. may i'm not the only one who tried to figure out what's wrong. i compared the ReHLDS latest stable and some recent snapshots with the original HLDS. the result is that pingboost does not affects on ReHLDS. no matter what value i use and what version i try it does not set any priorities.

Voice_GetClientListening uses iReceiver in wrong way

One day I noticed that something was wrong with this function and I decided to check the correctness of its work.
I test it with this code:

for (size_t nReceiver = 1; nReceiver <= 32; nReceiver++) {
    for (size_t nSender = 1; nSender <= 32; nSender++) {
        g_engfuncs.pfnVoice_SetClientListening(nReceiver, nSender, true);

        ALERT(at_console, "IsListening? %u %u %s\n", nReceiver, nSender, g_engfuncs.pfnVoice_GetClientListening(nReceiver, nSender) ? "true" : "false");
    }
}

And the results were unexpected:

IsListening? 1 1 true
...
IsListening? 1 32 true
IsListening? 2 1 false
...
IsListening? 32 32 false

This function works properly only for receiver with index 1.

When you walk up on a ramp it get's really jerky

When i run my server same map for few hours it stars to lag when walking up ramp
I found thread with same problem, but they have no resolution
http://forums.steampowered.com/forums/showthread.php?t=142350

Prevent server/client from sending/requesting same file twice

Hi guys,

just want to ask you if you can implement a solution to avoid this issue?

I've seen this as a request and I encountered the issue myself.
ValveSoftware/halflife#1345

Thx!

Null pointer in memcpy function

https://github.com/dreamstalker/rehlds/blob/master/rehlds/testsuite/player.cpp#L530
The null pointer is passed into 'memcpy' function.
Probably, there should be memcpy(&m_CurrentTm, t, sizeof(m_CurrentTm)); ?

Gametracker not tracked.

After I moved to reHLDS, (on my top server) Gametracker isn't track it.
And after I search for it, i found this post:
https://www.gametracker.com/forums/forum.php?site=1&post=390815

which metamod version should we use ?

my question is clear.
Which metamod version should we use for the best performance ?
metamod 1.21.1-am or what ?