[Request] 1.13.4 GetPVPRankProgress()

[flawblure]

I'm very interested in finding out more about this function and how it gets its values and if I can determine a more accurate rounding error.

What I believed is that there is a .01% error which ended up as 5 RP, and this could round up to 6, and since you calculate RP gains with two values you could get a maximum error of 12. However occasionally someone gets far less than the expected value. This week someone that was supposed to get 12997 actually got 12956. Even with the maximum error thatd only be 12968 so I'm wondering whats going on here.

My guess is that since GetPVPRankProgress() returns a value based on the image of your progress % on the character screen there are certain %s that have a higher rounding error than others.

What I'm most interested in is the code being run by GetPVPRankProgress and just what it's returning so I figured this was the right place. I wish I had more to post about this.

[ejt]

signed __int64 __fastcall Script_GetPVPRankProgress(__int64 L)
{
  __int64 active_player; // rax

  sub_170C10(L);
  if ( active_player )
    lua_pushnumber(
      L,
      (float)((float)*(unsigned __int8 *)(*(_QWORD *)(active_player + 0xB190) + 0x3A15i64) * 0.0039215689));
  else
    lua_pushnumber(L, 0.0);
  return 1i64;
}

Obviously active_player + 0xB190 is a pointer to the active player's descriptors.

psudo-code would be something like

int Script_GetPVPRankProgress(lua_State* L)
{
  CGActivePlayer* player = GetActivePlayer(L);
  if (player)
    lua_pushnumber(L, player->Descriptors.PVPRankProgress * 0.0039215689);
  else
    lua_pushnumber(L, 0.0);
  return 1;
}

[Jadd]

Just FYI, 0.0039215689 is the constant value to convert one byte integers (0x00 to 0xFF) to a float (0.0 to 1.0). It's the same as doing X/255.0f (although less accurate...)