In this pull request I've changed a ton of method signatures to reflect the more-narrow types of Species, Move# and Form; additionally, I've narrowed other large collections that stored lists of species / permitted values, and reworked them to be more performant with the latest API spaghetti that PKHeX provides. Roamer met locations, usually in a range of [max-min]<64, can be quickly checked using a bitflag operation on a UInt64. Other collections (like "Is this from Colosseum or XD") were eliminated -- shadow state is not transferred COLO<->XD, so having a Shadow ID or matching the met location from a gift/wild encounter is a sufficient check for "originated in XD".
`Moveset` struct stores 4 moves, and exposes methods to interact with a moveset.
`IndividualValueSet` stores a 6 IV template (signed).
Performance impact:
* Less allocating on the heap: Moves - (8 bytes member ptr, 20 bytes heap->8 bytes member)
* Less allocating on the heap: IVs - (8 bytes member ptr, 28 bytes heap->8 bytes member)
* No heap pointers, no need to jump to grab data.
* Easy to inline logic for checking if moves are present (no linq usage with temporary collections).
End result is faster ctor times, less memory used, faster program.
Rewrites a good amount of legality APIs pertaining to:
* Legal moves that can be learned
* Evolution chains & cross-generation paths
* Memory validation with forgotten moves
In generation 8, there are 3 separate contexts an entity can exist in: SW/SH, BD/SP, and LA. Not every entity can cross between them, and not every entity from generation 7 can exist in generation 8 (Gogoat, etc). By creating class models representing the restrictions to cross each boundary, we are able to better track and validate data.
The old implementation of validating moves was greedy: it would iterate for all generations and evolutions, and build a full list of every move that can be learned, storing it on the heap. Now, we check one game group at a time to see if the entity can learn a move that hasn't yet been validated. End result is an algorithm that requires 0 allocation, and a smaller/quicker search space.
The old implementation of storing move parses was inefficient; for each move that was parsed, a new object is created and adjusted depending on the parse. Now, move parse results are `struct` and store the move parse contiguously in memory. End result is faster parsing and 0 memory allocation.
* `PersonalTable` objects have been improved with new API methods to check if a species+form can exist in the game.
* `IEncounterTemplate` objects have been improved to indicate the `EntityContext` they originate in (similar to `Generation`).
* Some APIs have been extended to accept `Span<T>` instead of Array/IEnumerable
* Revises legality checks to account for traveling between the three game islands (PLA/BDSP/SWSH)
* Adds conversion mechanisms between the three formats, as well as flexible conversion options to backfill missing data (thanks GameFreak/ILCA for opting for lossy conversion instead of updating the games).
* Adds API abstractions for HOME data storage format (EKH/PKH format 1, aka EH1/PH1).
* Revises some APIs for better usage:
- `PKM` now exposes a `Context` to indicate the isolation context for legality purposes.
- Some method signatures have changed to accept `Context` or `GameVersion` instead of a vague `int` for Generation.
- Evolution History is now tracked in the Legality parse for specific contexts, rather than only per generation.
* Draft checks for encounter slot mastery
* Check encounter mastery flags
* Add moves for LA static encounters that don't follow learnset
* Add moves on crossover LA static encounters
* add alpha moveset population method
Now generates and applies moves as the game does
Updates some handling of other methods to use Span
* Show better message for bad mastery init flags
* Insert descending if candidates have same level
Level 78 Yanmega:
- [01] [10] Quick Attack
- [06] [15] Gust
- [11] [20] Silver Wind
- [18] [28] Hypnosis
- [25] [35] Air Slash
- [34] [45] Ancient Power
- [43] [54] Crunch
- [43] [54] Bug Buzz
Yields:
AlphaMove
Crunch*
Bug Buzz*
Ancient Power
* Descending order due to iteration
Co-authored-by: Lusamine <30205550+Lusamine@users.noreply.github.com>
Seal ComboItem
Reorder methods for easier reading
Fix BulkGenerator living dex setting CurrentLevel -- needs to be after species,form as EXPGrowth can vary for forms.
Existing `get`/`set` logic is flawed in that it doesn't work on Big Endian operating systems, and it allocates heap objects when it doesn't need to.
`System.Buffers.Binary.BinaryPrimitives` in the `System.Memory` NuGet package provides both Little Endian and Big Endian methods to read and write data; all the `get`/`set` operations have been reworked to use this new API. This removes the need for PKHeX's manual `BigEndian` class, as all functions are already covered by the BinaryPrimitives API.
The `StringConverter` has now been rewritten to accept a Span to read from & write to, no longer requiring a temporary StringBuilder.
Other Fixes included:
- The Super Training UI for Gen6 has been reworked according to the latest block structure additions.
- Cloning a Stadium2 Save File now works correctly (opening from the Folder browser list).
- Checksum & Sanity properties removed from parent PKM class, and is now implemented via interface.
* Initial bred moveset validation logic
Unpeel the inheritance via recursion and permitted moves
* Volt tackle considerations
* Optimize out empty slot skips
* Add tests, fix off-by-one's
* Require all base moves if empty slot in moveset
* Add test to prove failure per Anubis' provided test
* Tweak enum labels for easier debugging
When two enums share the same underlying value, the ToString/name of the value may be either of the two (or the last defined one, in my debugging). Just give it a separate magic value.
* Fix recursion oopsie
Also check for scenario where no-base-moves but not enough moves to push base moves out
* Add Crystal tutor checks
* Add specialized gen2 verification method
Game loops through father's moves and pushes in one iteration, rather than checking by type.
* Add another case with returning base move
* Add push-out requirement for re-added base moves
* Minor tweaks
Condense tests, fix another off-by-one noticed when creating tests
* Disallow inherited parent levelup moves
Disallow volt tackle on Gen2/R/S
* Split MoveBreed into generation specific classes
Gen2 behaves slightly different from Gen3/4, which behaves slightly different from Gen5... and Gen6 behaves differently too.
Add some xmldoc as the api is starting to solidify
* Add method overload that returns the parse
Verify that the parse order is as expected
* Add reordering suggestion logic
Try sorting first, then go nuclear with rebuilding.
* Return base moves if complete fail
* Set base moves when generating eggs, only.
* Use breed logic to check for egg ordering legality
Don't bother helping for split-breed species
GameStrings: 10% of strings are unique; just mark everything. Stop a little early so we don't process the empty & (trade) string.
LearnsetReader: All empty entries are length==0; malformed should throw an exception (never).
EggMoves: Compute ptr inside the array fetch loop; don't use linq.
EggMoves & string shaves off 80ms of startup time according to profiling; some could be attributed to warm-up but yay more efficient.
* Handle some nullable cases
Refactor MysteryGift into a second abstract class (backed by a byte array, or fake data)
Make some classes have explicit constructors instead of { } initialization
* Handle bits more obviously without null
* Make SaveFile.BAK explicitly readonly again
* merge constructor methods to have readonly fields
* Inline some properties
* More nullable handling
* Rearrange box actions
define straightforward classes to not have any null properties
* Make extrabyte reference array immutable
* Move tooltip creation to designer
* Rearrange some logic to reduce nesting
* Cache generated fonts
* Split mystery gift album purpose
* Handle more tooltips
* Disallow null setters
* Don't capture RNG object, only type enum
* Unify learnset objects
Now have readonly properties which are never null
don't new() empty learnsets (>800 Learnset objects no longer created,
total of 2400 objects since we also new() a move & level array)
optimize g1/2 reader for early abort case
* Access rewrite
Initialize blocks in a separate object, and get via that object
removes a couple hundred "might be null" warnings since blocks are now readonly getters
some block references have been relocated, but interfaces should expose all that's needed
put HoF6 controls in a groupbox, and disable
* Readonly personal data
* IVs non nullable for mystery gift
* Explicitly initialize forced encounter moves
* Make shadow objects readonly & non-null
Put murkrow fix in binary data resource, instead of on startup
* Assign dex form fetch on constructor
Fixes legality parsing edge cases
also handle cxd parse for valid; exit before exception is thrown in FrameGenerator
* Remove unnecessary null checks
* Keep empty value until init
SetPouch sets the value to an actual one during load, but whatever
* Readonly team lock data
* Readonly locks
Put locked encounters at bottom (favor unlocked)
* Mail readonly data / offset
Rearrange some call flow and pass defaults
Add fake classes for SaveDataEditor mocking
Always party size, no need to check twice in stat editor
use a fake save file as initial data for savedata editor, and for
gamedata (wow i found a usage)
constrain eventwork editor to struct variable types (uint, int, etc),
thus preventing null assignment errors
reduce nesting (evo.RequiresLevelUp is checked twice, only check once
and handle path)
compact some methods
seal some classes
add a little xmldoc to exposed members
pull out some move data fetching to separate class
add methods to quickly fetch if the move is learnable via that method
eventual plan is this:
replace
'fetch all possible moves then look within' move validation
with:
'peek if learnable, and get info how' move validation
advantages:
- returns game the move was learned in and the level (if appropriate)
- infinitely less object creation (garbage collection)
- only looks for a given move, doesn't have to fetch everything before
checking
- faster than full fetch, can be made even faster by optimizing lookups.
subject to change :)
