Added identifying hash to Pokémon data structure

The purpose of this hash is described in detail in pokemon.h. The hash
is computed using an implementation of SHA-224. To allow the GBA
sufficient time to compute this hash, a delay of 5 milliseconds was
introduced on the GC side before reading a Pokémon.

1 files changed, 18 insertions, 2 deletions

@@ -20,8 +20,23 @@ enum PokemonLanguage {
   Spanish = 7
 };
 
-struct PokemonIntermediate {
-  u32 otId;
+struct __attribute__((aligned(4))) PokemonIntermediate {
+  // a hash that can be used to identify the Pokémon. because the games do not
+  // naturally generate unique identifiers for Pokémon, this hash is generated
+  // from parts of the Pokémon's data that are guaranteed never to change. the
+  // parts used are the trainer ID and secret ID of the Pokémon's original
+  // trainer, the personality value, and the IVs. almost all of this data is
+  // secret, and should not be derivable from the identifier, which is why the
+  // identifier is hashed. while this identifier is not guaranteed to be unique,
+  // probability of collision is astronomically small given that the
+  // personality value and IVs provide essentially 52 bits of randomness
+  // (technically less because the IVs of bred Pokémon are not uniformly
+  // random, but this variation is statistically insignificant). the OT ID,
+  // while not uniformly random (amongst Pokémon), is included as a sort of
+  // "namespace", in that any Pokémon with matching PVs and IVs must also have
+  // the same OT, thus preventing collaboratively finding PV and IV collisions.
+  u32 key[7];
+
   u32 experience;
 
   // the stats are calculated from the base stats, IVs, EVs, and Nature, before
@@ -36,6 +51,7 @@ struct PokemonIntermediate {
   u16 species;
   u16 heldItem;
   u16 moves[4];
+  u16 otId; // only the lower 2 bytes, because the upper 2 are secret
 
   u8 ppBonuses;
   u8 otGender:1;