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.
Big thanks to @SciresM @sora10pls @Lusamine @architdate @ReignOfComputer for testing and contributing code / test cases. Can't add co-authors from the PR menu :(
Builds will fail because azure pipelines not yet updated with net6.
* Add more memory legality checks
Includes the following new memory checks:
- 4: met in link-trade, allows all possible genlocs except dangerous place
- 7: successful fishing, allows fishable species in gen 6 and 8
- 9: paying attention to another mon, allows only available species for gen 8
- 29: encountering legendary Pokémon, only seen on Zacian, Zamazenta, Calyrex so far
- 32: riding a bike, only in genlocs where biking is possible
- 75: taken to Nursery and placed with a mon, allows only available species for gen 8
Unstubs lotto check and applies it to gen 6.
Co-Authored-By: Skadiv <62726360+Skadiv@users.noreply.github.com>
Co-Authored-By: Matt <17801814+sora10pls@users.noreply.github.com>
* Loto-ID only has one t
* Crown Shrine has another genloc outside
* Consistent Loto name for array
Co-authored-by: Skadiv <62726360+Skadiv@users.noreply.github.com>
Co-authored-by: Matt <17801814+sora10pls@users.noreply.github.com>
The middle table in poke_memory.prmb contains the analogue of Gen6's memory table data. All existing memory data is the same, with 20 memories added.
Feeling bitflags are the same as before, but as we've noticed, feeling 0 is not obtainable. All the feelings are value upshifted by 1. This is why the game shows a blank line for Gen6 Feeling-0, as that game was [0,23] not [1,24] for the span of feeling strings.
Memories with 1% chance are actually 0% due to how they if-abort in the game code. Nice bug, GameFreak! @Lusamine had previously committed the unobtainable memories, derived from the community's empirical data.
Apparently they don't use the bit-permission table in Bank to get a random feeling, and instead just do rand(0,10).
Our logic to set a random feeling for bank transfers is still fine, because we set [0,10) within the bit table.
Can definitely be refined as these memories can restrict to capture/hatch/gift encounters. The multi-value arrays can also be restricted for non-hatches too (maybe first element if WasEgg?)
Co-Authored-By: Lusamine <30205550+Lusamine@users.noreply.github.com>
Since we have more metadata with move learn sourcing, we can check if it was traded to gen2 to get new moves / deleted.
Adjust call sites appropriately
might have some issues, to be ironed out maybe
Check per-evolution min level rather than basing it on the original encounter data.
Correctly parses stuff like a captured level 23 Nidorina that immediately evolves into Nidoqueen, then learns Body Slam.
Passing zero is better than passing 1, as 0 requires no value pushing.
tidy up method signatures in MoveLevelUp so that species is followed by form, and the lookups are textually aligned. Clarify "maxLevel" instead of max/lvl, same as min.
Regex objects aren't cheap, and the RegexCache has DefaultMaxCacheSize = 15.
We're checking 4,000 regexes for each unique string, so just keep the created regexes around instead of cycling new through the cache.
+4MB passive consumption, but each IsFiltered call no longer generates >4MB of discarded objects.
My unit tests run >25% faster now... nice?!
