Fix bots not being able to see ghoster behind smoke

[sunzenshen]

Description

Fixes bots not being able to see a ghost carrier from behind smoke despite the apparent glowing beacon visible by human players.

Toolchain

• Windows MSVC VS2022

[sunzenshen]

To test,

• set bots to a non-support class using bot_class 1
• use sv_neo_infinite_smoke_grenades 1 to create a smoke screen while holding the ghost
• before this patch, bots would not see you, after the patch they will see you if you are carrying the ghost.

[AdamTadeusz]

We already have ToNEOPlayer in c_neo_player.h, maybe keep the declaration here and move both definitions to neo_player_shared.cpp?

[sunzenshen]

To be honest I looked into this, and I'm not exactly sure what you mean by the relation with c_neo_player.h

I just needed the const for const CNEO_Player *, so that I can get a const pointer for my use case here.

I also wasn't sure how moving both definitions to neo_player_shared.cpp actually made things clearer, because with how I understand this, it would involve defining branches around #ifdef CLIENT_DLL, which I'm not sure is actually more readable, e.g.:

In neo_player_shared.h:

#ifdef CLIENT_DLL
class C_BaseEntity;
class C_NEO_Player;
C_NEO_Player *ToNEOPlayer(C_BaseEntity *pEntity);
#else
class CBaseEntity;
class CNEO_Player;
CNEO_Player *ToNEOPlayer(CBaseEntity *pEntity);
const CNEO_Player *ToNEOPlayer(const CBaseEntity *pEntity);
#endif

and in neo_player_shared.cpp:

#ifdef CLIENT_DLL
C_NEO_Player *ToNEOPlayer(C_BaseEntity *pEntity)
{
	if (!pEntity || !pEntity->IsPlayer())
	{
		return nullptr;
	}
#if _DEBUG
	Assert(dynamic_cast<C_NEO_Player*>(pEntity));
#endif
	return static_cast<C_NEO_Player*>(pEntity);
}
#else
CNEO_Player *ToNEOPlayer(CBaseEntity *pEntity)
{
	if (!pEntity || !pEntity->IsPlayer())
	{
		return nullptr;
	}
	return assert_cast<CNEO_Player*>(pEntity);
}

const CNEO_Player *ToNEOPlayer(const CBaseEntity *pEntity)
{
	if (!pEntity || !pEntity->IsPlayer())
	{
		return nullptr;
	}
	return assert_cast<const CNEO_Player*>(pEntity);
}
#endif

... if it's okay, maybe we can go with the original PR approach?