OtherChoice

Hello everyone I was wondering which is the best way to render a color pixel at given position on screen. No 3d vector to multiply by view matrix since i need to display it only on user screen at position (x coordinate, y coordinate). Resource(memory)-wise and by execution time what would you suggest me to do: using CGraphicImageInstance like minimap npcs and monsters, loading a 1px colored external image, python turtle module, or anything else which you think its a good choice? Thanks in advice.

db/src/ClientManagerPlayer.cpp

void CClientManager::__QUERY_PLAYER_CREATE(CPeer *peer, DWORD dwHandle, TPlayerCreatePacket* packet)

2 hours ago, matyasvane said:

Hey,

I'm starting with the game server for this game and need some advice. Please help me

1.  When I want to edit (temporarily) a map,weapons etc, for example, how do I do it?
2.  Can you please recommend a client who will automatically patch? (i can buy it)
3.  How do I update a server without unnecessarily long shutdown? Can I edit files while the server is on and then restart the server?
4.  Can I ask for a good guide on how to install WorldEdit on Windows 10?
5.  Looking for a CMS for Metin2 with an Offline Store (IS) is there any?

 

Thank you,

Matthew

1. You should specific what do you mean by 'edit' and 'temporarily'. Many things can be modified including 3d elements, textures, values, mechanics. Basically anything you can and cannot see while playing. Some of those things can be changed during execution time through coding, like values, mechanics and loading different external files such as textures models and so on. The latters can only by created and modified by external programs (not during execution time) or by including a whole new system in client and/or server source (for example a Guild flag editor ingame).

To have full access over game files you should learn how to code in the following languages: c++(client and server source), python(client root), lua(server-side quests), sql(database values) and probably also php, javascript, html5, css3 for a good website. There are also a few handy libraries to learn (and maybe update) if you want to deeply understand some nasty stuff around the source: namely boost and d3dx8, windows api and std (standard c++ library) are instead mandatory.

Apart from coding you will need furthert knowledge to really get into every part of the game. This knowledge includes 3d modelling, existing plugins allow you to convert 3DS (3d studio) models in gr2 used by client, Texturing, i do prefer gimp but anything will work, Map creation, made insanely easier through WorldEditor, Effect creation, such as skill effects, projectiles, auras and so on also handled by WorldEditor.

2. Build your own

3. Usually server shutdown and restart won't take longer than a few minutes, downtime strictly depends on the operations you are performing while server is closed. You can edit the source codes and client files while server is up, compile them and then when ready replace the olders & restart server.

4.World Editor is pretty easy to setup have a D partition or mount a folder inside C as D and then create a folder named "ymir_work", extract your WorldEditor version inside that folder and you should be good (WorldEditor has a weird bug with my laptop viewport so I have to decrease Windows Resizing and Layout option from 125% to 100% to correctly display the editor).

5. Cannot help you here, I do not like CMS. If you do then make sure the one you choose is clean and flawless.

The problem was the far clip float value retrived in python application.cpp 

if (!PERF_CHECKER_RENDER_GAME)
    {
        m_kRenderTargetManager.RenderBackgrounds();
        float fAspect=m_kWndMgr.GetAspect();
        float fFarClip=m_pyBackground.GetFarClip();

        m_pyGraphic.SetPerspective(30.0f, fAspect, 100.0, fFarClip);   <---- fFarClip was smaller than dragon model far prespective 

Tried with penger camera and still nothing displays, while every other model is showed correctly at the right size (both wild dog 101 and tree boss 2307 are perfectly centered and resized).

TraceError("Actor Height: %f, Model: %d", fHeight, index); in SelectModel returns
0312 15:22:17197 :: Actor Height: 207.197571, Model: 2108 -----> for desert bandit
0312 15:22:33305 :: Actor Height: 2172.499023, Model: 2291 ----> for fire dragon

The model and its properties are all ok and still it can't render. Is there any limitation for 3d vectors size in the custom defined camera?

UPDATE: After a while i discovered that if I set normal camera CAMERA_MAX_DISTANCE to something very large (10000000.0f) then zoom all the way out the model is rendered correctly

Hello everybody, I have some problem rendering fire_dragon (2291) with CxL RenderTarget

. Checked with multiple break points and TraceError(s) and everything seems ok. The model itself is displayed correctly when spawned, and also tried with older/custom version of it. All this made me believe that something might be wrong in camera settings for such a big model, so i tweaked with fov, aspect, near, far and SetViewParams(eye, target, up) but I had no luck. Anyone knows why this specific model is not rendering?

I have some problem rendering fire_dragon (2291), I know there's should be something different about this model but I just can't figure it out. No syserr, any clue?

@Kori It will load any page that Internet Explorer 11 would (just open it, visit a web page and if you can display that embedded browser will as well). The differences between 7 and 11 are huge that I can't simply include them in a message (if you want to have a deep look into it check microsoft changelogs since version 7).

I noticed embedded ingame browser is compatibility mode IE 7 by default, looked aroud the net and it seems that there's not so much detailed information about it. The best way would be to get your hands on chromium or something like that, but its a rather time sinking activity and since internet explorer emulated version can be easily changed with a new registry key i opted toward this solution. However (apparently, I actually searched very few sites) it seems that noone has ever mentioned it thus I decided to share the know-how with you. If anyone is interested this is a very simple example (very raw i suggest to at very least modify it):

In UserInterface.cpp before bool Main(HINSTANCE hInstance, LPSTR lpCmdLine)

void SetInternetRegKey()
{
	LONG status;
	HKEY hKey;

	status = RegOpenKeyEx(HKEY_CURRENT_USER, TEXT("Software\\Microsoft\\Internet Explorer\\Main\\FeatureControl\\FEATURE_BROWSER_EMULATION"), 0, KEY_ALL_ACCESS, &hKey);
	if ((status == ERROR_SUCCESS) && (hKey != NULL))
	{
		DWORD standard = 11000;
		DWORD version;
		DWORD size = sizeof(version);
		DWORD type = REG_DWORD;
		status = RegQueryValueEx(hKey, TEXT("YOURCLIENTEXECUTABLEFILE.EXTENSION"), NULL, &type, (BYTE*)&version, &size);
		if (status != ERROR_SUCCESS)
		{
			status = RegSetValueEx(hKey, TEXT("YOURCLIENTEXECUTABLEFILE.EXTENSION"), NULL, REG_DWORD, (BYTE*)&standard, sizeof(standard));
			if (status != ERROR_SUCCESS)
			{

			}
		}
		RegCloseKey(hKey);
	}
}
then in bool Main at the beginning add SetInternetRegKey();
bool Main(HINSTANCE hInstance, LPSTR lpCmdLine)
{
	SetInternetRegKey();      <--------- here

This will add a registry value for FEATURE_BROWSER_EMULATION (YOURCLIENTEXECUTABLEFILE.EXTENSION, 11000) which will tell browser to emulate IE 11 at the launch of the executable and only once. As I said above this is very raw and won't autodetect latest avaible IE and many other lacks which you can clearly see. All documentation about IE registry key for version/svcVersion can be found here https://support.microsoft.com/en-us/help/969393/information-about-internet-explorer-versions
Do your homework (just a few hours i think) and you will have your nice and clean browser emulation to latest avaible IE version or use it as it is and have a poor compatibilty still working IE11 browser. Hope you find this useful, Have a nice day.

Show me  your ActorInstanceBattle.cpp i'll take a look too.
However those has something worng probably

void CActorInstance::__PushCircle(CActorInstance & rVictim)
{
	const TPixelPosition& c_rkPPosAtk=NEW_GetAtkPixelPositionRef();

	D3DXVECTOR3 v3SrcPos(c_rkPPosAtk.x, -c_rkPPosAtk.y, c_rkPPosAtk.z);

	const D3DXVECTOR3& c_rv3SrcPos = v3SrcPos;
	const D3DXVECTOR3& c_rv3DstPos = rVictim.GetPosition();

	D3DXVECTOR3 v3Direction;
	v3Direction.x = c_rv3DstPos.x - c_rv3SrcPos.x;
	v3Direction.y = c_rv3DstPos.y - c_rv3SrcPos.y;
	v3Direction.z = 0.0f;
	D3DXVec3Normalize(&v3Direction, &v3Direction);

	rVictim.__SetFallingDirection(v3Direction.x, v3Direction.y);
}

void CActorInstance::__PushDirect(CActorInstance & rVictim)
{
	D3DXVECTOR3 v3Direction;
	v3Direction.x = cosf(D3DXToRadian(m_fcurRotation + 270.0f));
	v3Direction.y = sinf(D3DXToRadian(m_fcurRotation + 270.0f));
	v3Direction.z = 0.0f;

	rVictim.__SetFallingDirection(v3Direction.x, v3Direction.y);
}

Check for null vectors and __SetFallingDirection function

There's something wrong in client source in either m_PhysicsObject or __Push functions, if you changed anything about physics in gamelib project you probably messed up somewhere.

Ok i checked the source and i was recalling pretty much everything but a dumb thing insted of 
 

if (fDist < fMinDist || pkVictim->IsDead())
	{
		pkVictim = pAttacker;
		fMinDist = fDist;
	}

Use this

if (fDist < fMinDist && !pAttacker->IsDead())
	{
		pkVictim = pAttacker;
		fMinDist = fDist;
	}

Tested and working. If you are willing to donate something i'll appreciate it, anyhow have a nice day!

I sent you a private message, check it out

Hello everyone, I came across a really annoying problem and I hope you guys can help me. I need to change the structure of item_proto (/locale/lang/item_proto) to commit these changes:

    <ItemDef Vnum="" Name="'" LocalizedName="" Type="" SubType="" Weight="" Size="" AntiFlags="" Flags="" WearFlags="" ImmuneFlags="" Gold="" ShopBuyPrice="" LimitType0="" LimitValue0="" LimitType1="" LimitValue1="" ApplyType0="" ApplyValue0="" ApplyType1="" ApplyValue1="" ApplyType2="" ApplyValue2="" Value0="" Value1="" Value2="" Value3="" Value4="" Value5="" Socket0="" Socket1="" Socket2="" //add here//  Socket3="" Socket4="" Socket5="" //end// RefinedVnum="" RefineSet="" AlterToMagicItemPercent="" Specular="" GainSocketPercent="" AddonType=""  />

Basically i need to add a entry for those sockets (3,4,5). Tools around the web to crypt/decrypt item_proto are very "standardized", they only work with a standard item_proto template hence packing only standard entries, and no open source code is avaible.  If you guys have any suggestion and/or solution I'll be very grateful. It seems that I was wrong i ended up solving the problem once I found this:

Uhm at first i would say you should remove it from 

HEADER_CG_CHAT
HEADER_CG_WHISPER

and try again, if this is not the problem it will need some deeper debugging

Honestly i do not know why timed_event_cancel has no declaration however just go in game/battle.h, find 
extern void    battle_end(LPCHARACTER ch); and add below 
extern bool timed_event_cancel(LPCHARACTER ch);
then check game/battle.cpp should contain this function
 

bool timed_event_cancel(LPCHARACTER ch)
{
	if (ch->m_pkTimedEvent)
	{
		event_cancel(&ch->m_pkTimedEvent);
		return true;
	}
	return false;
}

Now you should not have any compiling error.
Next you asked about packets: input_main.cpp handles packets with HEADER_CG (CG == client to game) in function int CInputMain::Analyze(LPDESC d, BYTE bHeader, const char * c_pData).
Packets with HEADER_GC (GC == game to client) are instead handled by the client itself but according to the phase the client is in (eg. phaseGame, phaseSelect...).
phaseGame packets (which are most probably what are you looking for) are recived in Userinterface/PythonNetworkStreamPhaseGame.cpp  in void CPythonNetworkStream::GamePhase().
Other Client Packet Reciever
 

And lastly there are some HEADER_GD packets which are game to database handled by db binary.

edit

If you won't succed on your own just let us know what are you exactly trying to accomplish, gl

This is not how official logout works: in game/battle.cpp you have a function bool timed_event_cancel(LPCHARACTER ch) which checks whether the argument (LPCHARACTER ch) is logging off or not, if so it will cancel the timed event and return true. This function is called in int battle_melee_attack (LPCHARACTER ch, LPCHARACTER victim) on both victim and attacker and on a true (stopped logout attempt) will send chat_packet "the battle has begun..."
If you want to stop logout on a custom defined action you can simply call timed_event_cancel(ch).

Check those things :
1) item_proto table (serverside either txt or sql) and item_proto (client side) prices matches
2) In PythonPlayerModule.cpp in PyObject * playerGetSellItemPrice(PyObject * poSelf, PyObject * poArgs) there's a line:   iPrice /= 5; and the same exact value should be set in shop_manager.cpp (server source/game) in void CShopManager::Sell(LPCHARACTER ch, BYTE bCell, BYTE bCount) like this:  dwPrice /= 5;  (If any of those two is missing or has a different value make sure to correct it).

Ok after testing the original bugfix you linked i found the problem, the m_akSimplePlayerInfo array is filled in a wrong way this is the solution:
In game/char.cpp at void CHARACTER::PointsPacket() change as following
 

void CHARACTER::PointsPacket()
{
	if (!GetDesc())
		return;

	TPacketGCPoints pack;

	pack.header	= HEADER_GC_CHARACTER_POINTS;

	pack.points[POINT_LEVEL]		= GetLevel();
	pack.points[POINT_EXP]		= GetExp();
	pack.points[POINT_NEXT_EXP]		= GetNextExp();
	pack.points[POINT_HP]		= GetHP();
	pack.points[POINT_MAX_HP]		= GetMaxHP();
	pack.points[POINT_SP]		= GetSP();
	pack.points[POINT_MAX_SP]		= GetMaxSP();
	pack.points[POINT_GOLD]		= GetGold();
	pack.points[POINT_STAMINA]		= GetStamina();
	pack.points[POINT_MAX_STAMINA]	= GetMaxStamina();
	for (int i = POINT_ST; i < POINT_IQ + 1; ++i)
		pack.points[i] = GetRealPoint(i);
	
	for (int i = POINT_IQ + 1; i < POINT_MAX_NUM; ++i)
		pack.points[i] = GetPoint(i);

	GetDesc()->Packet(&pack, sizeof(TPacketGCPoints));
}

then at void CHARACTER::Disconnect add this line

if (GetDesc())
	{
  		/*add this*/
		PointsPacket();
  		/*        */
  
		packet_point_change pack;
		pack.header = HEADER_GC_CHARACTER_POINT_CHANGE;
		pack.dwVID = m_vid;
		pack.type = POINT_PLAYTIME;
		pack.value = GetRealPoint(POINT_PLAYTIME) + (get_dword_time() - m_dwPlayStartTime) / 60000;
		pack.amount = 0;
		GetDesc()->Packet(&pack, sizeof(struct packet_point_change));
		GetDesc()->BindCharacter(NULL);
//		BindDesc(NULL);
	}

then in client source

PythonNetworkStreamPhaseGame.cpp -> bool CPythonNetworkStream::RecvPointChange()  ------->Make these changes

case POINT_PLAYTIME:
	m_akSimplePlayerInfo[m_dwSelectedCharacterIndex].dwPlayMinutes = PointChange.value;
	break;
case POINT_LEVEL:
	m_akSimplePlayerInfo[m_dwSelectedCharacterIndex].byLevel = PointChange.value;
	__RefreshStatus();
	__RefreshSkillWindow();
	break;
/*case POINT_ST:
	m_akSimplePlayerInfo[m_dwSelectedCharacterIndex].byST = PointChange.value;
	__RefreshStatus();
	__RefreshSkillWindow();
	break;
case POINT_DX:
	m_akSimplePlayerInfo[m_dwSelectedCharacterIndex].byDX = PointChange.value;
	__RefreshStatus();
	__RefreshSkillWindow();
	break;
case POINT_HT:
	m_akSimplePlayerInfo[m_dwSelectedCharacterIndex].byHT = PointChange.value;
	__RefreshStatus();
	__RefreshSkillWindow();
	break;
case POINT_IQ:
	m_akSimplePlayerInfo[m_dwSelectedCharacterIndex].byIQ = PointChange.value;
	__RefreshStatus();
	__RefreshSkillWindow();
	break;*/
case POINT_ST:
case POINT_DX:
case POINT_HT:
case POINT_IQ:
	__RefreshStatus();
	__RefreshSkillWindow();
break;

Every other thing is like original thread 

Quote

What I did with the client and the packet set in Disconnect() came from this post:

Tested and working.

Oh ok i quite didn't get it, anyhow that's the solution you are looking for
 

This is you void CHARACTER::Disconnect(...)
//ADD this
PointChange(POINT_ST, 0);
PointChange(POINT_DX, 0);
PointChange(POINT_IQ, 0);
PointChange(POINT_HT, 0);
//end

packet_point_change pack;
pack.header = HEADER_GC_CHARACTER_POINT_CHANGE;
pack.dwVID = m_vid;
pack.type = POINT_PLAYTIME;
pack.value = GetRealPoint(POINT_PLAYTIME) + (get_dword_time() - m_dwPlayStartTime) / 60000;
pack.amount = 0;
pack.point_ST = GetPoint(POINT_ST);
pack.point_DX = GetPoint(POINT_DX);
pack.point_HT = GetPoint(POINT_HT);
pack.point_IQ = GetPoint(POINT_IQ);
GetDesc()->Packet(&pack, sizeof(struct packet_point_change));

/*if (GetHorseST() > GetPoint(POINT_ST))
	PointChange(POINT_ST, GetHorseST() - GetPoint(POINT_ST));

if (GetHorseDX() > GetPoint(POINT_DX))
	PointChange(POINT_DX, GetHorseDX() - GetPoint(POINT_DX));				REMOVE ALL THIS

if (GetHorseHT() > GetPoint(POINT_HT))
	PointChange(POINT_HT, GetHorseHT() - GetPoint(POINT_HT));

if (GetHorseIQ() > GetPoint(POINT_IQ))
	PointChange(POINT_IQ, GetHorseIQ() - GetPoint(POINT_IQ));*/ 

The way you are trying to work with won't go anywhere, first, the point you are packing are in a wrong place, since the first time ComputePoints() is exectuted it will check whether your horse has higher stats than yours and then replace them if so and will however pack and send them to client. If you want to make sure you have your real stat points you need to create the packet inside CHARACTER::ComputePoints() function not in void CHARACTER::Disconnect. Next point is the packet you are using: since packet_point_change (HEADER_GC_CHARACTER_POINT_CHANGE) has already a defined behavior (it has 3 different behavior actually, in PhaseGame, PhaseSelect and PhaseLoading) i suggest you to avoid using it.
Moreover if your goal is to always show base stats and hide horse bonuses, you can skip the whole client part and do in a simpler way.
I will first explain how to get both base Stats and horse Stats.

void CPythonNetworkStream::GamePhase()
{
	...
	switch (header)
		{
			case HEADER_GC_OBSERVER_ADD:
				ret = RecvObserverAddPacket();
				break;
    		....
			case HEADER_GC_BASE_STATS:
				ret = RecvBaseStatsPacket();
					break;
  		}
}

then after
bool CPythonNetworkStream::RecvPointChange()
{
  ...
}
add
bool CPythonNetworkStream::RecvBaseStatsPacket()
{
  TPacketGCBaseStats pack;
  if (!Recv(sizeof(TPacketGCBaseStats), &pack))
	{
		Tracen("Recv Base Stats Packet Error");
		return false;
	}
  //here you can do whatever you like with those number, to make sure you everything runs smoothly you can try this
  TraceError("Point_ST = %d, Point_DX =  %d, Point_HT = %d, Point_IQ = %d \n", pack.point_ST, pack.point_DX, pack.POINT_HT, pack.point_IQ);
  // so you can run your client then check syserr.txt and look at the output
}

and in PythonNetworkStream.h after
  bool RecvPointChange();
add
  bool RecvBaseStatsPacket();

____________________________________________________________________________________________________

If you want to totally hide horse stats you can do something like this:

if (GetDesc())
	{
		struct packet_point_change pack;

		pack.header = HEADER_GC_CHARACTER_POINT_CHANGE;
		pack.dwVID = m_vid;
		pack.type = type;
		pack.value = val;

		if (bAmount)
			pack.amount = amount;
		else
			pack.amount = 0;

		if (!bBroadcast)
			GetDesc()->Packet(&pack, sizeof(struct packet_point_change));
		else
			PacketAround(&pack, sizeof(pack));
	}
}

and change it like this

if (GetDesc())
	{
		struct packet_point_change pack;

		pack.header = HEADER_GC_CHARACTER_POINT_CHANGE;
		pack.dwVID = m_vid;
		pack.type = type;
		pack.value = baseAmount ? baseAmount : val;

		if (bAmount)
			pack.amount = amount;
		else
			pack.amount = 0;

		if (!bBroadcast)
			GetDesc()->Packet(&pack, sizeof(struct packet_point_change));
		else
			PacketAround(&pack, sizeof(pack));
	}
}

Lastly  you need to edit char.cpp void CHARACTER::ComputePoints()
find
if (GetMountVnum()) 
		{
			if (GetHorseST() > GetPoint(POINT_ST))
				PointChange(POINT_ST, GetHorseST() - GetPoint(POINT_ST));

			if (GetHorseDX() > GetPoint(POINT_DX))
				PointChange(POINT_DX, GetHorseDX() - GetPoint(POINT_DX));

			if (GetHorseHT() > GetPoint(POINT_HT))
				PointChange(POINT_HT, GetHorseHT() - GetPoint(POINT_HT));

			if (GetHorseIQ() > GetPoint(POINT_IQ))
				PointChange(POINT_IQ, GetHorseIQ() - GetPoint(POINT_IQ));
		}

and change it like this

if (GetMountVnum()) 
		{
			if (GetHorseST() > GetPoint(POINT_ST))
				PointChange(POINT_ST, GetHorseST() - GetPoint(POINT_ST), GetPoint(POINT_ST));

			if (GetHorseDX() > GetPoint(POINT_DX))
				PointChange(POINT_DX, GetHorseDX() - GetPoint(POINT_DX), GetPoint(POINT_DX));

			if (GetHorseHT() > GetPoint(POINT_HT))
				PointChange(POINT_HT, GetHorseHT() - GetPoint(POINT_HT), GetPoint(POINT_HT));

			if (GetHorseIQ() > GetPoint(POINT_IQ))
				PointChange(POINT_IQ, GetHorseIQ() - GetPoint(POINT_IQ), GetPoint(POINT_IQ));
		}

Done.

In Server source game/char.cpp in ComputePoints() 

if (IsPC())
	{
		if (GetMountVnum()) 
		{
			if (GetHorseST() > GetPoint(POINT_ST))
				PointChange(POINT_ST, GetHorseST() - GetPoint(POINT_ST));

			if (GetHorseDX() > GetPoint(POINT_DX))
				PointChange(POINT_DX, GetHorseDX() - GetPoint(POINT_DX));

			if (GetHorseHT() > GetPoint(POINT_HT))
				PointChange(POINT_HT, GetHorseHT() - GetPoint(POINT_HT));

			if (GetHorseIQ() > GetPoint(POINT_IQ))
				PointChange(POINT_IQ, GetHorseIQ() - GetPoint(POINT_IQ));
		}

	}

This way horse stats will override lower stats. Those numbers are then packeted and sent to client. You could create a new packet sending base stats to client. To recive base stats from game use game/char.cpp int CHARACTER::GetPoint(type) where type is POINT_ST, POINT_DX, POINT_HT, POINT_IQ example:

game/packet.h && UserInterface/packet.h
  
  enum
{
	HEADER_CG_HANDSHAKE				= 0xdc,
  ....
    HEADER_GC_BASE_STATS			= (your_value),
  
  typedef struct base_stats
{
	BYTE	header;
	DWORD	point_ST;
	DWORD	point_DX;
	DWORD	point_HT;
	DWORD	point_IQ;
} TPacketGCBaseStats;

 then send this packet from somewhere:
 

if (IsPC())
	{
		if (GetMountVnum()) 
		{
			TPacketGCBaseStats pack;
			pack.header = HEADER_GC_BASE_STATS;
			pack.point_ST = GetPoint(POINT_ST);
			pack.point_DX = GetPoint(POINT_DX);
			pack.point_HT = GetPoint(POINT_HT);
			pack.point_IQ = GetPoint(POINT_IQ);
			ch->GetDesc()->Packet(&pack, sizeof(pack));
			if (GetHorseST() > GetPoint(POINT_ST))
				PointChange(POINT_ST, GetHorseST() - GetPoint(POINT_ST));

			if (GetHorseDX() > GetPoint(POINT_DX))
				PointChange(POINT_DX, GetHorseDX() - GetPoint(POINT_DX));

			if (GetHorseHT() > GetPoint(POINT_HT))
				PointChange(POINT_HT, GetHorseHT() - GetPoint(POINT_HT));

			if (GetHorseIQ() > GetPoint(POINT_IQ))
				PointChange(POINT_IQ, GetHorseIQ() - GetPoint(POINT_IQ));
		}

And then recive it from client and do whatever you like with those.

REMEMBER to declare both the packet header and the typedef struct inside packet.h of client binary as well.

There are a lots of input which results in riding/dismount mounts so there is a wide variety of possible reasons. The only thing i can suggest you is to check GetMountVnum() function and its occurencies inside your server game source.