Correctly fixes#1163 (test cases added), eg wooper->quagsire would
learn yawn at 31 before evolving; the getmoves wouldn't provide yawn
(stopped at 30) with the old code; the lvl decrement was removed in the
incorrect commit, providing a legal verdict which was obtained
incorrectly (dexlevels were inaccurate)
AreEqual checking by Reference? strip out newlines and compare with
string equals instead
add testing method for generating full dex of pkm (still WIP hence
commented out attributes)
only tested for gen7 egg->pkm (not to eventual set)
pretty sure a good chunk of the egg generation logic can be pulled out
for reuse in other IEncounterable generators (IEncounterable
extensions?)
Test parsing glaceon, input must equal output
Test generating encounters for a specific version (exists only on Moon)
Test generating encounters for the same version (add handling trainer
data to permit Tutors)
fixes error in VC generations-present fetch which allowed movesets from
generations 3-6 lul
exclude negative/invalid gameversions
Two checks are performed for Gen 3 & 4:
A: Encounterability (wurmple, ball, ability)
B: RNG Compatibility (frame)
Original code only checked 'A'
Recent code ordered by 'A', but returned anything that matched 'B'. If
'B was not satisfied, an 'A' match would get deferred, allowing for an
eventual B & !A to be returned before a A & !B being returned.
yay bool logic
we want to prefer a RNGFrameMatch as well as a SlotCompatible match
original code (no RNG checks) retrieves with SlotCompatible just fine
recent adding in framematch resulted in deferring compatible if not RNG
frame, yielding !A matches before A & !B
maintain 3 queues (A & B get yielded immediately) to track A!B, !AB, and
!A!B
not necessary to check unown case anymore as it'll fall into the A!B
group. better performant to skip that check in favor of only adding for
fail.
thanks theSlayer!
Species and Nicknamed params were never used; foreign always resulted in
true for cases when it actually mattered.
Filter to the non-fullwidth characters, check the char type (latin base
vs jp/zh/ko) in order to determine if the full/half symbols should be
squished
could be faster replacing char instead of string, so change those
add two test cases for half width & full width string
sanitization/unsanitization
extract EncounterStatic match determination; in doing so fixes "lvl" by
passing byval.
Pichu encounters (2) come before the Gift level 25 pikachu, causing lvl
to be dropped to 5. When pikachu comes around, it compares 5==25 -> bad;
pulling out the method and passing byval allows it to be 25 at the start
of every new encounterstatic checked.
dynamic level -> calc level
no sweet scent -> proc available (rock smash/fish)
change frame checking in encountergenerator to save which frame it was
found on (unused but possibly later can capture this for output)
move MarkEncounterTypeData back above the area size reduction so that
the mark-by-locationID special handling will work
prevent slottype from becoming overwritten if already defined by special
case handling
add examples provided to automated test cases
closes#1648
Refactors the Memecrypto API to improve generalized usage.
Adds unit tests that verify memecrypto works, so that future changes
don't potentially break the code.
species added requires the 'latest personal table' reference to be
USUM's; the other referencs for .SM are fine as there's no possibility
to have a species ID high enough for it to matter.
add docs, move some data fetching to more appropriate class
remove old XP memecrypto support handling, was previously removed due to
net standard/core split
refactor memecrypto to handle multiple save sizes (USUM won't be the
same size save file); placeholder -1 for USUM size
refer to stackexchange links, makes XDRNG PID/IV search so quick (2^3
instead of 2^8 or 2^16; as fast as a gen3 shiny pokemon (2^13=8192 times
as fast as last release)
throw in some auto-inlining to eliminate some overhead in repetitive
search calls
Only usable for searching Method 4 IV spreads -> seeds;
1,3's search uses the same approach as the 1,2 search
I took the 1,2 search derivation to iterate for the next step, which
allows us to not know anything about the middle rand bits.
optimize a little bit more, move the pre-loop add to the initialization
stage; moving to the precomputed section pays off after 256 calls to the
method
fixes 10ANIV pikachu having Thunderbolt twice
un-duplicates CHANNEL event data
adds seed->PIDIV generator template for M1/2/4, CXD, Channel, and BACD
improve pkm converter to update nickname of hatched eggs. isn't perfect
(farfetch'd) but works better than before
Add special ribbon detection (for wc3),
speed up seeds from IVs calc (flip top bit of seed to get the result for
other iteration)
Add a bunch of gen3 event data
Detect channel, only detect Channel PIDIV for RS origin (only really
care about method1/2/4 being used when it shouldn't)
Channel does this weird thing called not setting the met level.
Refactor set suggested met location to a method that can suppress
popups.
