Detect base algorithm (regular, regular-anti, anti, shiny) and check for sub-patterns that differ per encounter (such as OT gender).
Implement PIDIV generators within the Gift3* template, remove old branches
Add Table Weight detection (still need to test) for PCJP 5anniv
Add PCJP GCEA campaign 1-6 templates
Differentiate Mystry Mew seeds; disallow Mews that were deleted b4 cart cloning
Need to double check some gen3 events (esp hadou titans), but is 99% there
Thanks Sabresite, for your continued support
https://github.com/kwsch/PKHeX/issues/4356#issuecomment-2333031493
fix flow so it is called in format3+
Closes#4356 ty @abcboy101
add another check for EC=0 eggs from Gen4 -- game uses the PID value as the "is egg available" flag, so can't obtain a 0-pid egg from gen4!
remove python paths from gitignore, so I can call one of the test folders `Eggs`
Add Cap Pikachu const value updating for USUM, also set Magearna const for USUM (not only SM)
Remove `NotDistributed` property from WC3 as only CXD events used it, no longer WC3 type.
Delete ArrayUtil; move remaining methods into the StorageUtil class that used them for compressing/inserting bulk into boxes.
Add shiny wormhole flag
Extract byte array strings into spans, faster tests and less overhead.
Fix pk6/7 pkmeditor export, now retaining status condition instead of wiping it
Abstract b2w2 key system to a save block; better documentation on its odd mechanics
Allow gen1-3 filename language/ver detect to work if the filename is ` `->`_` (discord attachments changing spaces to underscores); also revise canadian French emerald filename pattern
others: negligible perf/using standard library functions
Handles Nidoran's shenanigans as well as more clear method names
- Add normalization for PK7->PK8 (no more 0xE... usage for the gender symbols... maybe more chars?)
- Not sure if Gen3 gamecube encoding needs sanitizing. Who is transferring Nidoran to CXD? :)
Requires some silly usage of Language passing as arguments. Future improvements can be made to revise the half/full encoding determination when setting a string. Probably has issues since we're just doing a naive check without considering nicknames w/ special chars.
Closes#4174
* Split Gen1/2 string & pokelist conversion methods
* Refactor pokelist to direct read/write, skip on save if blanked
* Add settings editor for SaveLanguage overrides
In addition to the Method 1 (and other sibling PIDIV types) correlation, an encounter can only be triggered if the calls prior land on the Method {1} seed. The RNG community has dubbed these patterns as "Method J" (D/P/Pt), "Method K" (HG/SS), and "Method H" (Gen3, coined by yours truly). The basic gist of these is that they are pre-requisites, like the Shadow locks of Colosseum/XD.
Rename/re-type a bunch of properties to get the codebase more in line with correct property names & more obvious underlying types.
Solution known for over a decade, finally reminded myself that it'd be nice to have this available. Probably will use this for displaying Method J/K/H frame info when that branch is more mature.
### Treat 'ヘ' as katakana
- For consistency with ベ/ペ/リ which are treated as katakana here and in StringConverter12Transporter.cs, map 0xCD to katakana ヘ instead of hiragana へ.
- Allow input of hiragana べ/ぺ/へ/り, mapping them to katakana ベ/ペ/ヘ/リ. Previously, they would be treated as invalid and mapped to the string terminator.
### Swap mappings for 0xE8/0xF2
- Previously, 0xE8 was mapped to U+002E `.` FULL STOP, while 0xF2 was mapped to the special character U+2024 `․` ONE DOT LEADER. Since 0xF2 is used in user input while 0xE8 is only used in `MR.MIME`, swap these mappings so that normal keyboard input maps to the user-enterable character.
- Modifies SpeciesName.cs to produce U+2024 for Mr. Mime in G1/G2.
ShowdownSet: Fix indentation, use explicit const
ItemStorage8BDSP: Rename GetAll->GetAllHeld to match others
EncounterLearn: Guard against >4 length enumerables, use explicit versions for S/V
EggMoves: Read u16[] directly rather than manually
SaveFinder: simplify expression
SAV_Database: extract func
SAV_Encounters: use RoM to match Moveset generator
With the new version of Visual Studio bringing C# 12, we can revise our logic for better readability as well as use new methods/APIs introduced in the .NET 8.0 BCL.
O(1) lookup for arbitrary species, some edge handling rules for specific game islands/forms. With HOME, there's only 3 islands of permissions. No allocation besides the singletons which aren't really necessary.
No longer need to peek within multiple hashsets, just fetch the "is possible" bit from the species listing and check if the bit is set.
Will be fun if ball shell swaps are added 🤞, get to set all bits for anything that can enter game with that feature (if ever added).
Add unit test for gen67 no-patch exclusion
More performant byte->str performance (no longer needing to hash byte and fetch bucket), unmeasured str->byte performance, but it is the same implementation as Gen3 which is not a bottleneck.
Reduces the dll size by 80KB, and RAM usage by an unmeasured (likely similar amount). Better startup speeds since multiple dictionaries do not need to be allocated and created.
Moves Gen1 trainer names (0x5D) to Transporter class, and remove Korean entry (not legally obtainable since there's only Gen2 Korean games and Gen2 Korean VC cannot trade with international).
Hard verify Gen1 trainer name for language, since 0x5D is the ROM transfer language. Nickname can still be from any of the connected games.
This refactor makes it easier to use a different charmap for byte<->str for Boxes (see #4027) and the language differences. We have a master table that "works" for all text name entry, but localizations differ for some glyphs accessible in the box naming UI.
Pokewalker: defer match if shiny
Jacq egg: recognize unhatched egg, traded
sv-pokedex-old: don't add rows for >1010 species (out of range)
static9: correctly flag scale mismatch
SAV4HGSS: expose pokegear #'s as span instead of alloc
Use the GenerateEncounters methods instead, just a single line extra for each consumer.
PokemonGenerationReturnsLegalPokemon now works for all Gen1-7 encounters->PKM (except for Korean WC6 Arceus form mismatch), still disabled from auto-test due to it taking about a minute to generate everything. Future updates can add special methods for SW/SH, PLA, BD/SP, and S/V if the species/form exists in the game.
Allow nickname->species with span
add ConstantExpected annotations
Correct Gen8b ability set calling CommonEdits instead of direct setter
Slightly better perf on gen5/gen8+ location fetch
Misc pkm ctor fixes for b2w2 trade & wc3 eggs
wurmple evo now has an enum to be more explicit
no recursion for gen5 generator
fix showdown line length check allowing 86 instead of 80
Due to how the game generates the Pokémon data, the first two (or three) RNG calls are used to set the 32-bit `EncryptionConstant` and `PID`. With 2x 32-bit and 1x 64-bit values, we can algorithmically reverse the movement & manipulation of bits to recover the initial seed. Notably, certain bits of the initial state are not captured by our first two (or three) outputs, so we must brute-force guess at the initial state, and verify the RNG's output yields the expected values.
**With the ability for real-time Xoroshiro128+ seed reversal, we can now validate RNG correlations for SW/SH raid encounters natively within the program.** For now, the legality fail error message is extremely vague and any validated seed won't be "remembered" for the Legality Parse like other RNG methods. These seeds are 64bit, while every other "remembered" `PID/IV` seed-info is 32-bit.
Co-Authored-By: SciresM <8676005+SciresM@users.noreply.github.com>
We implement simple state machine iterators to iterate through every split type encounter array, and more finely control the path we iterate through. And, by using generics, we can have the compiler generate optimized code to avoid virtual calls.
In addition to this, we shift away from the big-5 encounter types and not inherit from an abstract class. This allows for creating a PK* of a specific type and directly writing properties (no virtual calls). Plus we can now fine-tune each encounter type to call specific code, and not have to worry about future game encounter types bothering the generation routines.
