transfer level exp and current exp not being different enough, which
limits the amount of natures possible. Could handle the 100&<=2 cases
the same way, but keep the branches for faster cases.
fix bad test cases
Not version specific; I'm assuming the 00020003 PID is a result of
hacking the save file to change OT gender. Legitimate saves haven't
yielded the Hardy PID. Changing gender probably results in a bad
pointer, fetching the wrong value.
Closes#2290
Remove usage of linq for sorting, use array.sort for in-place sorting.
Wasteful creating temp sorted buffers then enumerating them to dump into
the list.
existing logic had been overwriting the inner error messages with the
generic fail message
Remove backwards conversion precondition check; the ConvertPKM method
has no branches that allow conversion so it'll naturally fail -- end
result is the same with less upfront logic.
Fresh day, fresh tests of the new code :)
Fix assert conditions
Add extra considerations
Detected clones should be skipped for PID/EC/TID checks -- should
probably detect this a different way rather than checking a computed
string (bool array of flags for the clone check results)
FileInfo inherits from FileSystemInfo, which initializes a bunch of
strings that we don't need
this isn't a great optimization, just a fun exercise finding a quicker
approach to get these values
EvolutionSet was just an object that hides an array, with no logic
was planning to use the EvolutionSet to do the evolution checks, but
EvolutionTree was required to connect the evolution chains together
Removes empty trashbyte array allocation (less objects)
Change int[] to byte[] (less filesize/mem) (-256*6)
Change int[] to ushort[], precompute reverse table in saved space
removes dictionary lookup for array index fetch (faster, less memory, no
temp obj allocations!)
could make the ushort[] arrays into byte[] by changing them to be value
shifts? Not worth saving filesize for cpu.
reduces loading time (don't have to allocate conversion arrays when
launching a gen7 game), and separates things to easier to manage
locations
reworks gen3 string encode/decode, no longer does 3->4->5 and 5->4->3;
instead goes straight to the end result without an intermediary format.
String sanitization should probably be broken up rather than reused, oh
well.
if conditional checks >=3, and all subsequent checks require Valid; add
this to the continue condition, and remove unnecessary comparisons
update some comments
PK2 doesn't like getting passed null for the byte[] parameter (rightly
so); don't depend on the src code order having the first constructor
being the one with the least arguments
fix it in pk1/2 so that the first constructor has the least args, anyway
remove Identifier param. 99% of the time, identifier is not provided,
resulting in a useless call
end result:
provide 'create new' and 'create from' constructors
For xd, getfinaldata returned a standard 344KB save, but then slapped it together with the inner savedata (0x28000 bytes) rather than using the GCI header. Don't use 'Data'
replace linq for Encrypt/Decrypt pkmdata fetch with faster length check
& optional resize (Span pls!?)
update pk6 comment for why affection is not cleared
use 'u' to denote uint values, so that less IL ops are required
~(21->17) in the case of IVs
no functional change, just a fun exercise looking at generated IL and
finding ways to use less ops
no functional change, just rewrite some logic for better flow
fix subform popup count restriction (OwnedForms requires forms to be
added/removed manually; it doesn't automatically update)
sanitize box names when dumping boxes with separate folders (and bad box
names)
hide setters for LegalInfo outside assembly (shouldn't overwrite the
stored values)
update GenNumber to Gen* if appropriate; GenNumber checks for first Gen*
to match, is a little more explicit and quicker than calling twice when
checking a range
kinda slow, if a save has 900 mons, it has to check 900*900 which is
O(n^2), takes a few seconds
as opposed to just doing a hashset checking contains, would be O(n)
gotta do some fun stuff to make it faster
relocate program language enum
test that test case is proper
move getcompatible & gettemplate to core
remove catch rate changing for pk1 (catchrate editor now used instead)
replace Friendship control click to match shortcut description (used to
do nothing)
Remove some duplicate translation strings
fix gameselect string formatting
apparently systemsounds changed in a recent win10 update, Question is
now silent?? Exclamation is now the same as Asterisk (same as Beep?) Use
hand instead of Exclamation.
only use asterisk for notifications
Set default values to prevent any properties from being null
extract ClearNickname logic for hot path & reusable method
extract item conversion and move to appropriate logic class
SetPINGA{S}(pk, criteria) -> pid,iv, nature, gender, ability (stats
future?)
memes aside, this is where criteria data is forced
Split N encounterstatic (has fixed PID and special flag) from
EncounterStaticPID
remove double string replace (now less allocations); only add up values
if they're valid numbers.
builtin int.Parse throws with "1 2", but users can type stupid stuff
like that in the program (so we parse as 12).
be explicit that the string is empty rather than possibly missing
disallow encrypted export for BK4 (they're not encrypted), removes type
check
simplify replaceall in showdownset (don't call ReplaceAll 4x, just get
valid chars and rebuild)
simplify get ribbon sprite name (precompute ToLower and appended values
disable GeneratePKMs from throwing an exception (it's a testing flag,
not to leak out if someone wants a debug dll I guess)
expose FixedGenderFromBiGender
expose ribbon interfaces
Reuse some code in netutil for fetching stream (provide useragent for
all uses)
make sav4ranch use dynamicly read offsets to handle both variants
(quicker) #2248
still isn't close enough, but removes the usage of decimals for cp calc.
redd.it/ahac9i
CP calc doesn't flag anything I have that is legal, but the calc is
still off by 2 bits at most... whyyyy? Even the same code run in cpp
results in the same inaccuracy? Maybe this is some arm64 difference?
sav1: assign personal table based on (detected/provided) version
catchrate editor: allow to use savefile personal
remove gb era prompt whenever gb/blank loaded; have it as a setting (can
be overridden for VC now). keep allowgbera as autodetect
personaltable: track format for debugging purposes/info
update translations
showdownset: contains char instead of string
gameinfo: static readonly array instead of redefining new on update
legality: order of operation / value reuse / simplification / comments
pkx: compare char instead of single character string
header/footer: move assignment into method as an out instead of ref
Provide gen1/2 gb only version origin fetch for personal/moves ty
shadowmario (i assume this fixes things)
make listbox use ingame font for special characters
2f6e30bb3c (r31924945)
ty @pmarkive
Update handler after friendship update (CurrentHandler needs to update
afterwards)
Fix argument order (gender & nonpresent region data)
#2235 side effect
don't continue; instead seek forward
#2235 part 1, need surgery for part 2 (starter/locked prevents writing
data to the slot, but the slot can still be swapped)
24 bytes per object -> 16 bytes
2,624 objects are made for legality checking; reduces from 73KB to 41KB
removing unnecessary clone recreation -> count from 2624 -> 414 (41KB to
6.6KB)
yay 10x reduction; not huge in the big picture but a fun exercise
reducing allocations, increasing clarity by removing some magic numbers
probably can rewrite some of the evo loading/checking for even less, but
good for now.
All IVs are flagged in LGPE even if they're perfect :(
Might have to revisit this for future games and if LGPE transfer out and
maintain these bad flags (this is bad for inspecting parents before
breeding).
Closes#2229 , thanks @SteelPhase (report) and @sora10pls (verified)!
kinda have to restrict these to 'valid' choices; should be enough of a
selection->result to be adequate for changing version
restart the program after changing the setting; won't take effect until
rebooted.
rearrange pkm version groups for clear break between
twins/siblings/family
clear pokelist for sav7b on non-exportable saves (prevents b1s1 from
showing markings)
rename horohoro (pgo) to non romaji lol, hide daycare/party tabs on
nonexportable save
remove some unnecessary operations
rewrite getallcontrolsoftype to remove ToList() and make generic
remove linq for characteristic calc (get max IV) and others
remove some reliance on new[] for iv/ev
pk5 replace bitmagic with property get->set copy for cleanliness
remove alloc on gc key advancement; advance keys in place
make g6 sizes an explicit comparison instead of new[].Contains
replace Any/Count with Array.Find calls (removes linq dependency from
savetype detection!)
footer is different for Korean saves; checked JP/EN and they use the
2006 date. Probably the same for everyone-but-KO.
Simplify check further, only check the 0x4.... sav as it's the first
region saved to. No point wasting time checking the other!
a previous commit alluded to general/storage being initialized in
different slots
e8644280a1
, gonna ignore what was stated as I can't find any proof with the few
tests I did today (all init on second, and found no backups from that
time)
the ereader mons are nature/gender locked too, so unroll a little prior
to rechecking the overall team (with ereader mon included). Have to do
it this way as a lock can pop if the ereader data matches a prior spread
before the prior teammate can be generated.
footer identifiers work just fine for any legitimately created save
file; I might have had the checksum check code for pokesav blank
savefiles
dropping support for bad-blanks in favor of a non-negligible speedup is
favorable :)
how did this get forgotten so easily lol
added legality check; eggs can't participate in pokestar studio.
not editable from GUI, use ExtraBytes editor.
Re-add 0x86 (array was copied from pk4?) and reorder/comment
Other games may vary the flawless IV count so best to have consistent
declarations
(the vc transfer encounter has 5 or 3 depending if it's a mew/celebi)
settled on the class name; looks like it'll be roughly readable as a
mystery gift so inherit some properties. maybe can get r/w in the
mystery gift editor :)
maximum is always 3; sometimes we know the max size so specify.
bcl has default size = 0, which then resizes to 4 on first addition
wow such optimization
expand string array check to cast and check against length; batch editor
Nature>=25 causes problems in hover tooltips. Previous commit fixed for
pb7 stat calcs for invalid natures, pretty sure the actual game is
exploitable for stat calcs and invalid natures (bad access), but isn't
exploitable as it's only used to retrieve an enum to switch against.
Didn't really feel like pulling out the shared structure get/set
operations, only the derived properties & overall fixer methods. Should
be easier to maintain if any handling updates.
Clean up pb7 unused stuff; removes ribbon logic since none are used.
c# 8.0 interface defaults can maybe pull out more logic (ex making
FixRelearn and relearn moves into an interface); multiple inheritance
can be useful as features are added/dropped between generations.
Make extrabytes a pkm property (don't mutate array pls)
reconfigure startup loading to only initialize after initial load of sav
& pkm (using blanks if not provided)
there's a value in trainer pokes that gets set to an unused pkm byte
this byte is fetched in the stat/cp calc, but is never used (param
unused)
plans for future? it's right next to friendship in trpoke7b, maybe there
might be another stat amp?
https://www.serebii.net/magikarpjump/trainerrank.shtml
stronger trainer => stronger monz /speculation
remove lockedslots, store & save all teamslots for sav7
differentiate locked slots from overwrite protected slots by emitting an
enum containing info about the slot
locked = can't be replaced, period.
starter/battleteam = can't be pasted over by bulk-pastes
can eventually show more specific sprite layers to denote party
indexes/team#/starter
2d array -> 1d
skip copy by cloning instead
eliminate %24 (number is 0-31) by duplicating 0-7 as 24-31
dropping a modulo, >1 bounds check, and a half loop unroll should be
less cpu cycles :)
Closes#2160 , use Batch Editor to modify slots; it's not worth GUI
presence as it's only saved for party format mons and is undocumented
for anything besides affliction flags
Closes#2167
Based off sub_71001D16A0 with shortcut modifications:
* code marks new entries for evolution processing later using the low
bit (hence the << 1); in save files I can't see any bits set, but I
can't see anywhere in the code that DOESNT set this bit. Thus the
assumption of post-processing.
As noted in gameplay, a new capture doesn't set Min and Max; it only
sets it if it is above the usual sizing values.
Megas apparently don't set until seen in battle, so don't worry about
setting the other forme based indexes.
Starter Pikachu & Eevee have different growth rates than their base
forms (seriously WHY?)
remove old api surface in PKX as a breaking change as adding the
parameter is necessary.
Derived from WC7 with minor alterations (dynamic OT/Nickname which is
stored in the Full data). Since the 0x108 is not stored in the save
file, just keep the full data around.
Not sure if there'll be ones for other languages, hopefully not.
dupe checker:
var langs = new string[] {"en", "ja", "ko", "es", "it", "de", "fr",
"zh"};
foreach (var l in langs)
{
var names = new GameStrings(l).itemlist;
var dupe = names
.Select((z, i) => new {Index = i, Value = z})
.GroupBy(z => z.Value)
.Where(z => z.Key != "???")
.Where(z => z.Key != "？？？")
.Where(z => z.Key != "(?)")
.Where(z => z.Count() >= 2)
.ToArray();
if (dupe.Length == 0)
continue;
l.Dump();
dupe.Dump();
}
now iterates for all boxes, and restores to the correct offset at the
end
update direct offset pointers to box specific (so that it can work on
gen3 rsefrlg saves if ever used on them).
there's some "AHTB" files which contain the internal enum names; dump
out a few that seem useful
note they come from different tables, will need to figure out spacing &
if their order is absolute or uses that weird hash in front???
LGPE uses a list of pkm for boxes, and has pointers indicating where
each party member is at
need some logic for handling the sorting. Had suspected this a while
back so this pre-work came in handy (untested tho)
make eventflags offset ptr virtual, I might end up doing things
differently for event flags
One field added to indicate the GO species or whatever; the 891/892
meltan values match here. Might not be only for GO species translation,
might be for fetching models idk
* Rewrite tests with XUnit and .Net Core
* Add better "because" message
* Skipping test that was not ready & convert the fact to a theory
* Tweak casing
* Convert select date tests to theories
* Make the GetStringList load lock safer
Update comments / xmldoc
Add a savefile storage compressor (ie array[] with empty interstitials
-> list); return true if the compression moved anything (repopulate
views), and the count of occupied slots in the list.
Add saveblock base class; I haven't really liked how SAV6/7 do all the
logic; I'll still expose properties that will then point to a saveblock.
Cuts down clutter.
Add template Dex manipulator, with gen6/7 implementations
Speculate sequential gp/ge gameversion IDs
Unused after assignment, but if I rework the validation check to return the teamlock result instead of a bool I can spit out the Species-PID generated list as extra info. Helps identify the as-generated conditions if there are multiple teamlocks possible for an encounter (multiple shadow mons, seen vs unseen).
Now accounts for the shiny locks & associated restrictions, abusing the
recursion by setting & unsetting the required CPU Trainer's shiny value.
Repoint all test methods to use the new obj methods
A lot of internal information is nice to have saved; create a copy of
the static class's logic and pepper in some state variables (Required
CPU Shiny Value) to account for the weird edge cases.
Code is essentially the same as the static methods, except for more
fleshed out logic in the AllLock interrupt checking and the eventual
Trainer checks.
Now gets count=2 tests working
yield return new SeedFrame { FrameID = ctr + (l.Seen ? 5 : 7), PID = pid
};
used to be ctr + 6, needs to alternate even-odd
all 2 lock cases pass, but didn't make lock3's pass. Something else is
still amiss :)
break out some classes for the solidified lock objects
Check max case for level first, then iterate upwards (eliminates inner
if)
change wc* nature to sbyte (update comparisons, 0xFF too magic-y)
use Rand.Next(x) instead of (0, x)
More languages than Italian have a rule-breaking name. Rework checks a
bit
Flag eggs more accurately
Update force hatch to move Link Trade met location to Egg Location
fix Ho-oh -> Ho-Oh text strings (correctness is key!)
Thanks @ijuintekka ! Closes#2148
Will check later for commented out tests that should pass.
May end up turning on the first-shadow-lock check feature prior to
release.
Currently toggled off, can be optionally toggled on (I'm not sure that
it's perfect? Random save files have stuff flagged since they don't
match the lock; will have to investigate later)
commented out test calls = not working
will have to debug in more detail later; the first lock is always
working at least.
data sourced from
https://github.com/ijuintekka/Eligor/blob/master/Eligor/Spread.cs , not
sure if it's perfect.
can't cleanly condense deviating team appearances as double-shadow leads
can't be selectively encountered. didn't feel like modding lockfinder's
recursive algo to be smarter.
rearrange getmove with respect to tutors; verifycurrentmoves was getting
tutor moves in levelup
extract not-found hint fetch
Expose gen5 block objects (so I can play with saves in linqpad)
Move logical checks using encounter data out of Core.cs
might be possible to excise legality checking objects from
Legal/PKHeX.Core so that a 'slimmer' core can be made
location ID was originally added by:
afde4514e2 (diff-83a32d69355a64376bc77bd07e338a33R446)
I
have another save file that uses met location 110 (like this one) so I
assume it was originally a copypaste error from Sudowoodo.
Change Trade strings to not allocate separate empty arrays.
can't just re-call base.SetEventFlag with an adjusted flag ID since the
exception checks max.
just override the geteventflag method with a silly adjustment -- we
can't be sure if Block2 is immediately after Block2, so just adjust the
flag r/w to a different offset&flag when appropriate.
fixes r/s badgeflag get/set, oops, meant to modify FRLG's value in the
last commit.
was block2+0x64
8*((0xF80-0xEE0)+0x64) = 0x820
this highlights a critical flaw in event flag reading; it expects the
region to be contiguous in the savedata. it's not. that's super bad, and
event flag r/w for flags >1280 won't work until I work something out
#2128
it's still an imprecise thing, as ingame gifts aren't counted by either
release count isn't tracked, so the number should be roughly >= total
box+party count.
#2126
Receiving a lower leveled species from GSC will evolve it away from the
required species, ruling it unavailable.
Receiving a Machoke will evolve it, so can't reuse for Haunter trade
results in:
- precomputed GetTable() reference
- no casting when iterating for matches
- no db null checks, as they are initally Array.Empty<T>();
reduces noise a little & slightly faster
Remove unused sav properties
Add translatable strings for autodetect SAV error suggestion
add speculated gg dex formtable
- no eevee forms
- increments 1 for pikachu since it already exists in table
- stops at mewtwo (no other formed-species available)
im currently setting up skeleton stuff for lgpe since we're <70 days
from release; by doing some pre-work like prior games I can get a good
idea of what needs to be adjusted prior to crunchtime!
can now check individual strings / identifiers in a LegalityAnalysis
object by:
var la = new LegalityAnalysis(pkm);
var parse = la.Info.Parse;
if (parse.Any(z => ReferenceEquals(z.Comment, LegalityCheckStrings.X))
...
should be faster than doing string.Contains
Closes#2109
Program boots -> load extra locations rather than constantly fetching.
These files won't change location or presence (not an external drive).
just store the locations instead of constantly fetching, we already call
detect once when loading the form anyway.
extract final pkm manipulation logic from pkm editor
add xmldoc to boxmanipulator, add utility class for enum->manip fetching
fix xmldoc referencing removed enum member
relocate boxmanip overview logic to utility class
move SAVPaths file pointer to Main for reuse #2109 , will add to
autodetect in a later commit
open/save dialog with extra extensions: ignore ones already present
V### names weren't enjoyable to work with; use similar verbose style as
the program message strings.
updating the translation files with the remapped variable names shortly
remap list: https://pastebin.com/jybkVDAK
click ball sprite/label to cycle through legal balls
hold alt to use old behavior (set to pokeball)
hold shift to cycle through legal balls randomly (default behavior goes
poke->color->poke)
want to randomize setting balls? use the ApplyBallLegalRandom method at
the end :)
can't really get too fancy as only one color is stored in the personal
data, pkm aren't one single color.
0, [9,19] are set (didn't see 15), but there's also a switch case which
references all values 0-19.
these values are used by a 80 sbyte array (4*20) in the Resort.cro
remove usages of "goto case"
use shared class for pk1/2 setnotnicknamed
fix extendedeurope values (copypaste from extendedAmericas)
move ball out of verifiers, move nature/movetype with ball
move encountertype datasource providing to core
fix rerolling EC not updating characteristic
remove some repeat logic calls
relocate geolocation name fetch to separate class, add tests to ensure
functionality, add languageID->country/region fetch method
could use concurrentdictionary but after reading some perf drawbacks vs
manual locks, just do it manually. regex checking is the real slow part,
whatever
add some comments to describe
hides pid & ec in pkmeditor, gamesync & securevalue in saveditor
easier to share screenshots/stream without leaking identifying
information
fix auto-centering of pkmeditor subwindows when a savefile is detected
on program load(tab_otmisc isn't initialized and reads an incorrect
width)
egg sprites can be a little difficult to see what's underneath (glance
value)
make the new egg sprite method the new default; adds setting to revert
to old behavior
adds translation text for the mega rayq control from #2098
Closes#2095
split out unsketchable moves (struggle/chatter) to separate array
remove trailing spaces (auto fixed by my visual studio plugin), add
space before comment //
use trainerstat editor control in gen6 editor
fill in some details from the disassembly via setrecord usages
still slightly fuzzy on some:
fureai (based on usum idb name, similar logic)
soaring sky counts
looks like there's some other usages of the fields which were outside
the stat range, throwing an exception when loading to NumericUpDown,
added bypass logic
Thanks Holla!
calling Japanese ? X : Y on every single access is too excessive, just
compute one layout for each type whenever the first sav1 requires it.
could probably reduce the amount of properties by reusing others + shift
value; offsets are different due to string lengths being different
Logic is essentially identical; implement a base class and have the
generation specific structures implment the differences.
Reduce the verbosity a little
