[Geesu]

Shouldn't it return CSW_C4? If you are killed by the bomb in CS it says you were killed by weapon id 0.

Test code:

public client_death( killer,victim,wpnindex,hitplace,TK )
{

    client_print( victim, print_chat, "[DEBUG] You were killed by %d with weapon %d", killer, wpnindex );
}

Result:

[DEBUG] You were killed by 1 with weapon 0

[Geesu]

Also, this is how i'm working around it:

// All we want to check is if the user was killed by the bomb
public client_death( iAttacker, iVictim, iWeapon, hitplace, TK )
{
    // Check out who the inflictor was
    new iInflictor = entity_get_edict( iVictim, EV_ENT_dmg_inflictor );

    // Check to see if the death was from the bomb
    if ( 0 < iInflictor < MAXPLAYERS && iWeapon != CSW_HEGRENADE && iInflictor )
    {
        
        if ( is_valid_ent ( iInflictor ) )
        {
            new szClassName[64];
            entity_get_string( iInflictor, EV_SZ_classname, szClassName, 63 );

            // Check the classname of our inflictor
            if ( equali( szClassName, "grenade" ) || equali( szClassName, "env_explosion" ) )
            {
                iWeapon = CSW_C4;
                iAttacker = 0;


                real_death_call( iVictim, iAttacker, iWeapon, 0 );
            }
        }
    }
}

[VEN]

I believe in the older versions of CSX it returned CSW_C4. Though the planted C4 isn't actually a "weapon_c4" but a "grenade" which isn't a weapon. [edited]

[BAILOPAN]

Indeed, when you are killed by C4, you are killed by a grenade.

However, the question is whether CSX should be abstracting past this or not.

[VEN]

Currently, IIRC, the forward return HE grenade weapon index for HE grenade kill. So why it was made to not return CSW_C4 [cut]. Seems illogical to me.

In my opinion "abstracting" wouldn't make worse since players can't kill someone with actual "weapon_hegrenade" or "weapon_c4" anyway.

[BAILOPAN]

I don't like that, and if I were to redo it (which I can't, for backwards compatibility), I would make it be consistent and return WEAPON_HEGRENADE.

That said, my thoughts: Encapsulating it to return CSW_C4 is a good idea. Saying you can't kill them with the "actual weapon" is semantics at best, implementation dependent at worst. What wpnindex should mean is "which weapon CAUSED death." By your logic, it should return the index of an M4A1 instead of WEAPON_M4A1, because it was actually a bullet that killed them ;)

As long as it doesn't break old code I don't mind making this change.

[Geesu]

yay! BAIL ftw!

[VEN]

I've made some testings and found that if you standing close to the center of the explosion it will return WID == 0 (dmg_inflictor has "env_explosion" classname). But if you are not close it will return CSW_C4 (dmg_inflictor's classname is "grenade" in this case). So the forward able to return CSW_C4 but not in the all cases.

But i personally would not use client_death forward anyway, at least because it's not called on user_kill native execution or on falldamage death.

[Geesu]

True, thats why you do the check like in my above example