It was recently discovered by LRXC that Pokémon Emerald has a bug where if the player has a Pokémon with Pressure and the opposing non-ghost Pokémon uses Curse, two PP are deducted instead of just one. Since a non-ghost Curse only affects the user, I believe that it should not interact with Pressure. This is the case when the player's non-ghost Pokémon uses Curse against the opponent's Pressure Pokémon - there, only one PP is deducted. This asymmetric interaction indicates that this is a bug. I believe that the reason is that when the player uses Curse, there is a check in battle_controller_player.c (line 486) that makes the user the target of the move when it not a ghost type, but there is no corresponding check in player_controller_opponent.c, so presumably when the AI uses Curse the target always remains (one of) the player's Pokémon, since in battle_moves.h, the target is defined as "MOVE_TARGET_SELECTED".
To make sure that non-ghost Curse never interacts with Pressure as was most likely intended, I added a bugfix in battle_script_commands.c, Cmd_ppreduce(), starting line 1224, where ppToDeduct doesn't increase if the move is Curse and the user is not a ghost type.
I have tested this myself in Emerald where my Torkoal's Curse against a wild Absol with Pressure on Route 120 only deducted 1 PP, but without the bugfix, the wild Regice's Curse against my Absol with Pressure deducted 2 PP (I caught it with a Master Ball right after it used the move to check.) After my bugfix, Regice's Curse now only deducted 1 PP.
Jaizu
commented
10 months ago
Description
It was recently discovered by LRXC that Pokémon Emerald has a bug where if the player has a Pokémon with Pressure and the opposing non-ghost Pokémon uses Curse, two PP are deducted instead of just one. Since a non-ghost Curse only affects the user, I believe that it should not interact with Pressure. This is the case when the player's non-ghost Pokémon uses Curse against the opponent's Pressure Pokémon - there, only one PP is deducted. This asymmetric interaction indicates that this is a bug. I believe that the reason is that when the player uses Curse, there is a check in battle_controller_player.c (line 486) that makes the user the target of the move when it not a ghost type, but there is no corresponding check in player_controller_opponent.c, so presumably when the AI uses Curse the target always remains (one of) the player's Pokémon, since in battle_moves.h, the target is defined as "MOVE_TARGET_SELECTED". To make sure that non-ghost Curse never interacts with Pressure as was most likely intended, I added a bugfix in battle_script_commands.c, Cmd_ppreduce(), starting line 1224, where ppToDeduct doesn't increase if the move is Curse and the user is not a ghost type. I have tested this myself in Emerald where my Torkoal's Curse against a wild Absol with Pressure on Route 120 only deducted 1 PP, but without the bugfix, the wild Regice's Curse against my Absol with Pressure deducted 2 PP (I caught it with a Master Ball right after it used the move to check.) After my bugfix, Regice's Curse now only deducted 1 PP.
Discord contact info
rogerwrightshoe