Second cleanup batch. Eight files are now gone:
asm/overlay_29_022E1618.s asm/main_0202B530.s
asm/overlay_29_022DC65C.s asm/main_0205EC38.s
asm/overlay_29_0232124C.s asm/main_02051788.s
asm/main_02042B20.s asm/main_02065BD8.s
main.lsf falls from 1175 objects to 1167.
Twenty-two functions matched across eleven files, but only seventeen land.
FixedPoint64ToInt, GetSpriteSize and RemoveItemFromKecleonShop1 did not match,
and their files therefore cannot be emptied; landing their already-matched
siblings would have converted two two-function files into two fresh
single-function fragments, so those two are held back as well. This batch
reduces the fragment count without creating any.
Three declarations follow the tree rather than the scratch: ov29_0232119C takes
struct entity * as its first parameter, ov29_023538B4 is already declared as a
u8 array, and ov29_022DC2B8 and ov29_022DC33C are declared u8 * in their
destination file. That last one is inconsistent with poly_text.h, which declares
ov29_022DC2B8 as struct unk_ov29_022DC240_t *; the file already carried the same
split for ov29_022DC240, so its local convention was followed rather than
changing pre-existing code. Worth unifying separately.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
First cleanup batch. All twenty matched at score 0, and every file they came
from is now gone:
asm/main_0204F9F4.s 2 functions
asm/main_02017C50.s 4
asm/overlay_29_02344B44.s 2
asm/overlay_11_0230CCE0.s 2
asm/overlay_11_022E9D6C.s 5
asm/main_0201BEB0.s 5
main.lsf drops from 1181 objects to 1175, the first net reduction on this
branch. Each removal also retires the matching asm/include/*.inc.
Most of these reuse types landed earlier and add nothing: the OAM group reuses
unk_0201BCCC with unk_oam_sub and unk_oam_elem, and the overlay_11 fade group
reuses unk_023890A0 and unk_0238916C. Only unk_0238A098 and unk_023531A4 are
new.
Note for later cleanup work: six of the twenty are the last function in their
file, so emptying it leaves them with no following address and the object-level
size check cannot cover them. On a cleanup batch the matching build carries
more of the verification than usual.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Ninth 5x10 batch; 48 of 50 matched at score 0 and land here.
overlay_11_022FD308 LivePerformer accessors
main_0204CF9C GLOBAL_PROGRESS_PTR bit-array accessors
main_020022C4 string, thread and card initialisation
overlay_11_022E9D6C screen-fade queries
main_0203C774 state-machine steps
New placeholder structs: unk_022FE490, unk_022FE498, unk_020AFE60, unk_020AFE70,
unk_0209C7F4, unk_0238916C, unk_0229B004, unk_0229AFC0. The overlay_11 fade
group reuses unk_02388FC0 and unk_023890A0 unchanged.
Two are NOT here. SetPositionInitialLivePerformer sits at 225, where an
address computation is scheduled earlier than the target places it; six
variants moved nothing. sub_020024E4 sits at 35, a two-register swap in a
zero-fill loop, after both declaration orderings, a block scope, an extra
temporary and a while form. Both stay as assembly.
Matching finding, now in docs/MATCHING_TIPS.md: a two-word struct copy emits
the same four instructions as two field assignments but schedules both loads
first, which is what the target has. Writing the fields separately interleaves
load and store, and hoisting them into locals merges the stores into an stm.
The OS wrappers drop their local declarations of nitro functions, which reach
every translation unit through the prefix header and disagreed on return type.
OS_ResetSystem keeps its extern; it is declared in no reachable header.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Eighth 5x10 batch; 49 of 50 matched at score 0 and land here.
overlay_11_022FAC54 LiveObject accessors
overlay_11_022E46FC coroutine table and special recruitment
main_020325AC further inventory-menu paging
overlay_17_0238A754 menu state setters
main_0201BAC8 OAM attribute copying
New placeholder structs: unk_022FD1A0, unk_02324FA0, unk_02320904,
unk_C_ROUTINES, unk_0238BE00, unk_0201BCCC with unk_oam_sub and unk_oam_elem.
struct unk_020332AC gains field_0xBC and field_0xD0, splitting two byte arrays
without changing its size; no already-landed function changes.
GetCoroutineInfo is NOT here. It reaches 30 -- two rows, a register swap between
the base pointer and the record pointer, with every instruction otherwise
identical. Eight structural variants and six declaration orderings were tried;
restructuring around a record pointer took it from 500 to 30 and nothing since
has moved it. It stays as assembly.
Two matching findings, both in docs/MATCHING_TIPS.md. A switch needs a fourth
case before MWCC emits a binary split rather than a flat compare chain, which is
now the second batch where that closed a function. And x == 4 || x == 5 folds
into a range test at every width tried, where the target keeps cmp/cmpne; the
double negation !(x != 4 && x != 5) blocks the fold and keeps the arm order.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Seventh 5x10 batch; all fifty matched at score 0.
overlay_11_022EE5E4 LiveActor accessors
overlay_11 language and string helpers
main_020325AC inventory-menu paging and scaling
overlay_29_0234B4CC alert-box wait, portrait init, message log
overlay_18 further LINK_SHOP_PTR setters
New placeholder structs: unk_022FAB44, unk_020332AC, unk_0234BAC0. struct
unk_0238C950 gains an s8 at offset 4, splitting its leading byte array without
changing its size; no already-landed function changes. struct unk_023537CC is
copied from the neighbouring sources that already define it, and struct iovec
now comes from file.h rather than a local copy.
Seven callees were already declared elsewhere in src/ and are reused verbatim
rather than redeclared: AdvanceFrame, DUNGEON_BUTTON_INPUT, MESSAGE_LOG_INFO,
GetSimpleMenuResult__0202B870, GetLanguage, LoadFileFromRom and
UnkMapRelatedFunc.
docs/MATCHING_TIPS.md gains three entries: a case label sharing default's body
still shapes the compare chain; a masked value promotes to int, so an unsigned
compare needs a cast; and the two malformed disassembly shapes, with the
reachability test that separates a merged block from an ordinary loop.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Sixth 5x10 batch; all fifty matched at score 0 and landed.
overlay_29_022DC694 HEAD team-stats forwarders, a 10-argument forwarder,
and a 22-entry table search
overlay_29_022E4714 HEAD PlayEffectAnimationEntity wrappers
overlay_29_022E37DC TAIL same family, merging into overlay_29_022E406C.c
overlay_18 mid LINK_SHOP_PTR state setters
main_020022C4 mid Nitro OS thread and mutex wrappers
New placeholder structs: unk_022DC694, unk_02353518, unk_0238C950,
unk_020027E8, unk_0229B0E0. The OS wrappers use nitro's own OSThread and
OSMutex types, which every translation unit already has through the prefix
header, rather than private tags.
Two spellings a reviewer should weigh, both recorded in the note:
_0229B0E0 is declared volatile because the target reloads it where any
ordinary spelling is common-subexpression-eliminated, and ov29_022DC78C
matches in byte-offset form rather than array indexing because the target
keeps the offset live in a separate mul.
Group selection now rejects malformed disassembly: 0.4% of asm blocks end
without a return (a .global mid-body split one function in two) and 6.5% are
several functions merged into one block (no .global marks the later entry
points). Neither is decompilable as written.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The group from asm/overlay_00_022DB1A4.s that the previous commit could not
land. It does not belong in src/: these are Nintendo Wi-Fi Connection library
functions, and src/** is built with OPTFLAGS -O4,s, which inlines a small callee
whose body is visible in the same translation unit. lib/** is built with
OPTFLAGS_SDK -O4,p, which does not. Landed into lib/src/overlay_00_022DB1A4.c
and listed in main.lsf under OVY_0, all ten are byte-identical to the original
assembly.
A decomp.me scratch cannot see this: it declares every callee extern, so no body
is visible and nothing can be inlined, and the group scored 0 either way.
lib/src was already in LIB_SRC_SUBDIR without existing on disk, so no makefile
change is needed. The object also receives the full defines, not just
GLB_DEFINES; none of these functions carries a region conditional in any case.
docs/MATCHING_TIPS.md carries the mechanism, a minimal reproducer, and a
one-object diagnostic that does not need a full build.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Fifth 5x10 batch. 48 functions matched at score 0; 38 land here.
overlay_29_022E37DC -> src/overlay_29_022E406C.c (new, 10)
overlay_29_022E41F0 -> src/overlay_29_022E4338.c (new, 10)
overlay_11_022E9B84 -> src/overlay_11_022E9A78.c (merged, 10)
main_02048764 -> src/main_02048B48.c,
src/main_02048BB4.c,
src/main_02048CB8.c (new, 8)
New placeholder structs: unk_020AFF38, unk_022AAE74. Two existing placeholders
in overlay_11_022E9A78.h are refined by this batch's evidence -- u8 field_0x4[8]
becomes s32 field_0x4 + s32 field_0x8, and u8 field_0x0[0x44] gains s16
field_0x14. Both keep their total size and no pre-existing function changed.
overlay_29_022E406C.h and overlay_29_022E4338.h gain dungeon_mode.h, which they
need for struct entity.
The fifth group -- ten DWC functions from asm/overlay_00_022DB1A4.s -- is not
here. All ten scored 0 and still changed OVY_0's size by 16 bytes, failing
main.sbin and OVY_0-OVY_9. A decomp.me scratch declares callees extern, so MWCC
cannot inline them; landing the group into a .c that defines those callees makes
the bodies visible and -O4,s inlines them. The target's calls show retail kept
these in separate translation units. docs/MATCHING_TIPS.md gains an entry with
the mechanism and a one-object diagnostic that does not need a full build.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The landing script added `#include "util.h"` to every header it generated,
across this whole branch. That was never necessary for most of them: u8/u32/s32
and the rest of the base types are typedef'd in lib/include/nitro/types.h, which
every translation unit already has through `-prefix global.mch`. 283 of the 460
headers in the tree that name u32 do not include util.h and build fine.
util.h owns bool8, bool32, fx32_8, ufx32_8 and struct fixed_point/rgba/vec2_16 --
a header naming any of those genuinely needs it, and 61 of the 110 do. The other
49 carried it for nothing and are cleaned up here.
Removal is .c-aware: util.h is kept wherever the header's paired source names a
util.h-owned symbol and does not include util.h itself, so nothing loses it
transitively. No functional change; verified byte-identical.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching build:
build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Fourth 5x10 batch.
overlay-11 shop/menu loaders 10 LoadCroagunkSwapShopOverlay .. ov11_0230CE1C
animation-table getters 10 ov10_022BFE54 .. ov10_022BFF30
move-animation flag readers 10 ov10_022BFD58 .. ov10_022BFE3C
overlay_01 menu accessors 10 ov01_0233755C .. ov01_02337624
sound veneers + playing checks 10 sub_02017B70 .. IsMePlaying
None carries a region conditional -- checked per function, not per file -- so US
alone verifies this commit. Four destination files, one .s split in total.
A bitfield read is not the same as the equivalent shift expression. Dwci_Entry-
style `(x << 28) >> 28` is folded by MWCC into `and #0xF`, while the target keeps
both shifts; declaring the member `u32 f : 4` emits them. That finding came from
the overlay_02 group, which is NOT in this commit -- see below.
GetMoveAnimation returns a 0x18-byte element of MOVE_ANIMATION_INFO. The five
flag readers in the third group index field_0x8 of that same element, so both
groups share one struct move_anim_info rather than the placeholder the flag
readers were first written against; all seven were re-scored against the unified
type.
ov10_022BFE6C needs `if (cond) { ...; return TRUE; } return FALSE;` -- the
inverted arrangement predicates the pointer load instead of branching (1100).
A sixth group of ten (Dwci_ChangeScene .. Dwci_Language__02339674, overlay_02)
matched at score 0 and is NOT landed. Extracting from asm/overlay_02.s breaks
OVY_2's checksum on its own: bisecting to a single three-instruction function
still fails, so the C is not implicated. Splitting that file also leaves four
symbols (SoAlloc, SoFree, ov02_0232FD28, ov02_0232FD40) undeclared in the tail,
which the stock extract_function.py does not carry across. Left for a separate
investigation.
Authored by Claude (Opus 5) under human direction. All fifty verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Third 5x10 batch: fifty located as five groups of ten before any matching, all
fifty matched, one build, one commit.
Dwci account bitfield accessors 10 Dwci_Acc_SetMaskBits .. Dwci_Acc_GetFlags
overlay-11 menu/overlay loaders 10 ov11_0230CB88 .. LoadRecycleShopOverlayResume
_020B0B6C accessors + copies 10 sub_0206C0D8 .. sub_0206C19C
overlay-11 veneers 10 ov11_022E9A78 .. ov11_022E9B5C
main_02065C84 wrappers 10 sub_020663C8 .. sub_020664AC
Each group is a contiguous run, costing one .s split apiece. None of the fifty
carries a region conditional -- checked per function, not per file -- so US alone
verifies this commit.
The two 8-byte copy loops need an UNSIGNED counter. With s32, `--n != 0` emits
sub followed by cmp and scores 505; with u32 it fuses into the target's single
subs. A struct assignment through a cast also scores 0 but needs a new type.
ov11_0230CB88 passes an argument to EntryOverlay13, which this tree already has
decompiled and matching as `void EntryOverlay13(void)`. Dropping the argument
scores 220 -- the target saves r0 into r4 across the InitMenu call specifically
to pass it. Retail's caller evidently had a prototype its definition did not
match; reproducing that needs the two declarations to disagree across translation
units, so this file declares `extern void EntryOverlay13(s32 a);` locally.
Dwci_Acc_GetUserId and Dwci_Acc_GetFriendKey return 64-bit values; the tell is
the pair of no-op `orr rX, rX, #0` instructions MWCC emits assembling a u64
across r0/r1. GetUserId's high half is field_0x0 masked to 0x7FF, which agrees
with Dwci_Acc_SetUserId passing that same mask to Dwci_Acc_SetMaskBits.
New placeholder structs: unk_022DB0E0, unk_020B0B70, unk_02388FC0, unk_023890A0.
Authored by Claude (Opus 5) under human direction. All fifty verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Second 5x10 batch: located as five groups of ten before any matching, then all
matched, then one build, then one commit.
inventory menu window accessors 10 ov10_022BCC7C .. ov10_022BCDBC
file/ROM transfer + bag init 10 FileRom_StopDataTransfer .. sub_0200CAE4
overlay_16 + secret bazaar 10 ov16_0238CC94 .. TalkToSecretBazaarNpcWithYesNoMenu
Gt2 socket/connection accessors 10 Gt2GetConnectionState .. Gt2SetConnectionData
forwarding veneers at +4 9 sub_0202613C .. sub_020261E4
Each group is a contiguous run, which costs one .s split apiece rather than one
per function: after the first extraction the next function is at the head of the
remainder and merges into the source just created. None of the 49 carries a
region conditional -- checked per function, not per file -- so US alone verifies.
GetDamageSource returns s16. The lsl #16 / asr #16 before its return is the
return conversion, not a cast in the body; s16 return with an s32 parameter
scores 0, while returning s32 and casting scores 510 in the shape the target
does not use.
ov29_022F05B4 needed its two arms in the other source order -- the fourth
instance of MWCC letting the first arm written fall through.
TalkToSecretBazaarNpcStandard and ov29_022F0618 differ only by `mov r3, #0`
versus `mov ip, #0`. r3 is the fourth argument register: the Standard variant
passes 0 for it, the other forwards its own. Same callee, different arity.
IsInventoryMenuActive returns s32, not bool8 -- bool8 adds an and #0xff the
target lacks.
struct unk_0202AAA8's field_0xA4[0xF4] is split to expose field_0xF8. New
placeholder structs: unk_020AF360, unk_0238CE40, unk_02026130, unk_0230E3E8 and
unk_0230E420 (the Gt2 connection and socket are distinct types -- remote IP sits
at offset 0, local IP at 4).
A fiftieth was prepared and not landed: sub_0202614C stands at 300. The target
materialises &v into r3 before the first store and moves it to r1; six spellings
(struct by value, a pointer local, storing through the pointer, an array, a
reversed store order, a hoisted destination) all compute the address after the
stores instead.
Authored by Claude (Opus 5) under human direction. All 49 verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Located as five groups of ten before any matching was attempted, then all fifty
matched, then one build, then one commit.
overlay_04/06 global accessors 10 ov04_0233CA80 .. ov06_0233CAA8
hardware register bit-fields 10 sub_02009BE8 .. sub_02009D20
save reset / play timer 10 sub_020483B8 .. sub_0204F9E0
sound veneers + window accessors 10 sub_02017C90 .. sub_02030A74
dungeon fixed-room / item 10 CloseFixedBin .. GenerateStandardItem
Every group is a contiguous run at a file head or tail, so extract_function
merges each function into an adjacent source instead of splitting a .s. None of
the fifty carries a region conditional -- checked per function, not per file --
so US alone verifies this commit.
Twelve of the fifty are interworking tail-call veneers (ldr ip, =f; bx ip). That
falls straight out of an ordinary forwarding call, and survives both argument
reordering (GenerateStandardItem) and a computed argument (sub_020483B8).
Nothing in the tree had a veneer decompiled, so the shape was proven on one
before the other forty-nine were written. Their parameter lists are guesses: a
veneer forwards registers untouched, so the emitted code is identical for any
arity and the match constrains nothing.
sub_020483D8 returns its callee's result. It sat at 40 through five spellings
with the pointer one register low; the target never reuses r0 after the call
because that value is the return value. Adding `return ok;` gives 0, and the
return is invisible in the asm because the value is already in place.
Two type corrections found by probing: the Are*Allowed family compares the
dungeon id unsigned (bcc, not blt -- 200 signed), and the register setters need a
u32 intermediate, since a u16 local forces a lsl/lsr truncation the target lacks
(615).
New placeholder structs: unk_0233F660, unk_0233EEE0, unk_020AFEF4, unk_02030A50.
src/main_02017C80.c's existing PlaySeByIdVolume(s32, s32) declaration is kept in
preference to this commit's; both affected functions were re-scored against it
and still match.
Authored by Claude (Opus 5) under human direction. All fifty verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Eleventh cleanup sweep. Five asm files and their includes removed:
asm/main_0205B6CC.s sub_0205B6CC
asm/main_0202C75C.s sub_0202C75C
asm/overlay_29_0232CDA4.s DoMoveOneShot
asm/overlay_29_0232A078.s DoMoveNightShade
asm/overlay_29_023016D8.s DisplayRunAwayIfTriggered
All five merged into an adjacent source rather than splitting one. None carries
a region conditional, so US alone verifies this commit.
sub_0205B6CC is a 64-bit unsigned >= 2 (cmp/cmpeq/movhs) on a value at offset
0xC. A u64 member sits there: MWCC aligns long long to 4, not 8. Rebuilding the
value from two u32s scores 200, and `> 1` in place of `>= 2` scores 310.
sub_0202C75C uses struct unk_0202C5E0, the collection-menu window struct its
neighbours already use; GetWindowIdSelectedMenuItemIdx immediately before it
already spells the same `GetWindowContents(window_id) + 4`. field_0x1AC[6]
covered 0x1B0/0x1B1 and is split into three members.
DoMoveOneShot and DoMoveNightShade are twins staging eleven arguments for
CalcDamageFixedWrapper, which had no declaration in the tree. The only obstacle
was the temporary holding GetMoveType's result: as enum type_id, -enum min sizes
it to a byte and emits an and #0xff the target lacks (950); as s32, 0.
DisplayRunAwayIfTriggered needs Min(999, max_hp), not Min(max_hp, 999) -- Min is
a static inline, so the argument order is not free (200). That exact expression
in that exact order already sat in the NONMATCHING block of AiMovement in
dungeon_ai_movement.c; the guess was right and had never been confirmed.
include/main_02042AF8.h's provisional `s32 sub_0202C75C(s8 a);` is retired in
favour of the real header. It was written from the one call site, which passes an
s8 field; the parameter is a window id.
A sixth was prepared and not landed: FixedPoint64ToInt (asm/main_02001CB0.s)
stands at 225. The target keeps an arithmetically redundant & 0xFFFF0000, and
whether MWCC folds it away is decided by which side of the | the masked term is
written on -- masked second folds it (735), masked first keeps it (225). Writing
it first is also what makes it evaluate first, and the target needs it preserved
but evaluated second; ten spellings found none that does both. Spelling it as a
real 64-bit shift scores 1020.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Tenth cleanup sweep. Five asm files and their includes removed:
asm/overlay_29_0232CE7C.s DoMoveReviver
asm/overlay_29_02327940.s DoMoveRapidSpin
asm/overlay_29_022EFA74.s TryActivateIqBooster
asm/main_02027AA0.s sub_02027AA0
asm/main_020568DC.s sub_020568DC
All five merged into an adjacent source rather than splitting one. Four of the
five carry a region conditional, so all three ROMs were built; sub_020568DC is
the exception and US alone would have covered it.
TryActivateIqBooster sat at 60 with r6/r7/r8 rotated one place and every
differing row an r-marked register swap. Moving the IQ local's declaration ahead
of the other two takes it to 0. This contradicts the MATCHING_TIPS entry claiming
declaration order never moves callee-saved assignment; that measurement was on
SetLeaderAction's ~75 locals, where register pressure decides the colouring. On
four locals with no pressure, first-definition order is what is left.
DoMoveRapidSpin sets its return to 1 before the random-outcome call, so a failed
random check still returns TRUE. The early-return spelling scores 570; a result
variable set inside the damage branch scores 0.
sub_02027AA0's region difference is not an offset shift: JP writes slot 3 where
the others write slot 0 and vice versa, mirrored for the second case. The rodata
a few symbols away corroborates it -- _020AFD94 and _020AFD98 are defined in
swapped order under the same #ifdef.
sub_020568DC follows GetAppointedLeaderMemberIdx, which lives in the file it
merged into and is the same function specialised to the active team. The
(&t->number_active_team_members_main)[team] idiom is already used by three
existing files.
DoMoveReviver was on the stalled list from 4d2a484c but had never been attempted;
it was set aside for needing region-conditional C, which is one #define.
sub_020568DC and TryActivateIqBooster were likewise untried rather than stuck.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Confirmed by matching builds of all three ROMs: US, EU and JP.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Ninth cleanup sweep. Five asm files and their includes removed:
asm/overlay_29_0231E990.s ShouldTryEatItem
asm/overlay_29_0232D378.s DoMoveGuardSwap
asm/overlay_29_0232A490.s DoMoveSwallow
asm/overlay_29_02329674.s DoMoveHelpingHand
asm/main_0205B738.s sub_0205B738
All five merged into an adjacent source rather than splitting one.
Three declarations the tree already carried are contradicted by these bytes and
are corrected here, touching five files this sweep did not otherwise decompile:
SubstitutePlaceholderStringTags took `u8 *string_id` in four files and `int a` in
a fifth. DoMoveGuardSwap passes a literal 1, which no pointer parameter can take
without a cast; the existing call sites all passed 0, which is why nothing caught
it. The four are now s32, agreeing with the one that was already an integer.
TryIncreaseHp took `s16 hp` in two files. DoMoveSwallow loads its argument from a
word array and passes it untouched -- s16 inserts a sign extension and scores
210, s32 scores 0. The two existing call sites load s16 data with ldrsh, so the
value is already extended and the wider declaration is byte-neutral for them.
ShouldTryEatItem's parameter is signed: the range check emits blt/ble, while
enum item_id under -enum min compares unsigned and scores 400. s16 matches the
adjacent IsItemValidVeneer and every caller's ldrsh.
DoMoveSwallow also needed its two arms in the other source order, the same rule
as sub_02009DCC last sweep: MWCC lets the first arm written fall through.
sub_0205B738's copy at +0xB0 is element [1] of struct0, not a member: 0xB0 is
sizeof(struct unkStruct_020B0A54), as sub_0205B77C (a * 176 + ...) and
sub_0205B7C8 (MemsetSimple(..., 0xB0)) already say. Reading it as a member and
appending one scored 0 on a scratch but changed the struct's size, and with it
the stride in three neighbours and the copy length in a fourth -- 29 binaries
moved. No header change is needed; the array spelling also scores 0.
Three of the five log a message whose id differs between US and JP by no constant
offset, so each is a file-scope #ifdef JAPAN define, as dungeon_ai.c and
dungeon_projectile_throw.c already do.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Three carry region conditionals, so this was confirmed by matching builds of all
three ROMs: US, EU and JP.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Eighth cleanup sweep. Five asm files and their includes removed:
asm/overlay_29_0230F840.s TeamMemberHasExclusiveItemEffectActive
asm/overlay_29_023276E8.s DoMoveTickle
asm/overlay_29_0232C538.s DoMoveSwitchPositions
asm/overlay_29_022FF8C8.s GetMobilityTypeAfterIqSkills
asm/overlay_29_0233861C.s UpdateShouldBoostHiddenStairsSpawnChance
All five merged into an adjacent source rather than splitting one.
DoMoveTickle was the stall recorded on 78b52f9d at score 15. The target reaches
ATK_STAT_IDX through SPATK_STAT_IDX's pool entry at offset 4, which is what the
two symbols' adjacency in rodata makes them. Declaring SPATK_STAT_IDX with array
type and indexing [1] emits that; taking the address of a scalar declaration and
indexing does not. DoMoveMetalSound, which shares the file, moves to
SPATK_STAT_IDX[0] for the same declaration.
TeamMemberHasExclusiveItemEffectActive replaces the provisional
`bool8 TeamMemberHasExclusiveItemEffectActive(u8);` that src/overlay_29_02338548.c
declared from its call site. The parameter is enum exclusive_item_effect_id --
u8 scores 0 in isolation but the build's -W error rejects the implicit conversion
at the inner call to ExclusiveItemEffectIsActive__0230F8AC. Both call sites move
from 0x5e/0x5f to EXCLUSIVE_EFF_MORE_KECLEON_SHOPS/EXCLUSIVE_EFF_MORE_HIDDEN_STAIRS.
DoMoveSwitchPositions raises exp_yield to 1 when below it; spelled `== 0` rather
than `< 1` it scores 405.
GetMobilityTypeAfterIqSkills is a whole function inside #ifndef JAPAN, so the
guard moves into the .c and around the prototype in the .h.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Three carry region conditionals, so this was confirmed by matching builds of all
three ROMs: US, EU and JP.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Seventh cleanup sweep. Five asm files and their includes removed:
asm/overlay_29_0232E400.s DoMoveWakeUpSlap
asm/main_0204317C.s FreeUnkStorageStruct0x18c0
asm/main_02009DCC.s sub_02009DCC
asm/main_020018A4.s sub_020018A4
asm/main_0200C4CC.s sub_0200C4CC
All five merged into an adjacent source rather than splitting one, so the sweep
removes five files and adds none.
DoMoveWakeUpSlap doubles the damage multiplier when the target is in any of the
three sleep-class states, then ends the status. EndSleepClassStatus takes five
arguments here, not four: the target loads r2=0, r3=1 and pushes a third flag on
the stack.
sub_020018A4 rounds a 64-bit value up to the next multiple of 0x100 and returns
the high 56 bits. The `and r2, r1, #0` that opens it is the high half of a 64-bit
`x & 0xFF`, not a separate test.
sub_02009DCC clears one bit of a 5-bit mask per non-zero entry in five byte
arrays at stride 5, then merges it into DISPCNT for the requested engine. The
zero-engine arm has to come first in the source or the wrong arm is predicated.
sub_0200C4CC copies 232 bytes out of _02094C18, whose size is fixed by the gap to
_02094D00.
Field types forced by the access widths in main_02042AF8.h: struct unk_02043148's
field_0x0 becomes s32 and field_0x7D5 splits into two s8s plus padding; struct
unk_020AFEE0's field_0xC becomes s32, which was the last differing row in
FreeUnkStorageStruct0x18c0.
struct unk_02094C18 is new. Its member is u32[0x3A] rather than u8[0xE8] because
the copy's alignment, and so its instruction shape, comes from the member type.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Sixth cleanup sweep. Four asm files and their includes removed:
asm/overlay_29_02328A74.s DoMoveDamagePoison40
asm/overlay_29_02329FE0.s DoMoveMudSlap
asm/overlay_29_0232B820.s DoMoveMistBall
asm/main_0200F18C.s GetEquivItemIndex
The three move effects share DoMoveBubble's shape: deal damage, and on a random
outcome apply a secondary effect, returning whether damage landed rather than
whether the effect did. The result variable doubles as an argument in two of
them -- MistBall passes it as both the stage count and the checkProtected flag,
MudSlap passes 0 as the chance.
GetEquivItemIndex walks the 50 bag slots with a cursor, testing each item's
exists flag as `(flags & 1) != 0` so the boolify emits tst/movne/moveq, and
returns the index of the first equivalent item or -1.
A fifth was prepared and not landed: GetNextIqSkill (asm/main_02058F24.s) sits at
30, the whole difference being that the target keeps its skill-list cursor in lr
where the candidate uses r5. Four spellings tried -- hoisting the value local,
moving the cursor declaration, a group temporary, and a while loop -- all 30 or
worse.
Authored by Claude (Opus 5) under human direction. All four verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Fifth cleanup sweep; five more asm files and their includes removed:
asm/overlay_29_023491C4.s IsOutlawOrChallengeRequestFloor
asm/main_0202AAE8.s sub_0202AAE8
asm/main_0202BC60.s sub_0202BC60
asm/overlay_29_023261A0.s DoMoveBubble
asm/main_0205BCC4.s sub_0205BCC4
sub_0202AAE8 and sub_0202BC60 are the same function over different window
states: check unk_0202AAA8::field_0x19C against a set, then set two flags, clear
field_0x1A4, call sub_0202830C with -5 and advance the state. Both write a byte
at 0x1A1 that the struct covered only as padding after field_0x1A0, so that gap
is now named field_0x1A1 with field_0x1A2[2] holding the remainder -- same total
size.
Three spellings that mattered:
- IsOutlawOrChallengeRequestFloor is three ORed predicates and only matches with
each operand written `!= 0`; a plain || scores 300 and early returns 1300.
- DoMoveBubble returns whether damage was dealt, not whether the speed drop
landed, so the result is funnelled through one variable which is also passed
as LowerSpeed's stage count.
- sub_0205BCC4 must not hoist the destination pointer above its NULL check; a
local shared by both branches scores 830, one declared inside the else branch
matches.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Fourth cleanup sweep; five more asm files and their includes removed:
asm/overlay_29_0232B8E4.s DoMoveCalmMind
asm/overlay_29_0232D22C.s DoMoveHammerArm
asm/overlay_29_023495E8.s IsDestinationFloorWithFleeingOutlaw
asm/overlay_29_023492D4.s IsLegendaryChallengeFloor
asm/main_02030850.s CloseAdvancedTextBox2
The two mission predicates merged into files that already declared
IsCurrentMissionType and IsCurrentMissionTypeExact with enum mission_type
parameters, so they use those declarations and the MISSION_* enumerators rather
than fresh externs with integer types. IsLegendaryChallengeFloor's floor check
is a byte subtraction written `+ 0xFF`, the same encoding DoMoveGust needed.
Both predicates only match written as a nested `if` that falls through to a
single return; the early-return spellings score 610 and 405.
CloseAdvancedTextBox2 uses the existing sub_0202836C, sub_0202812C, DeleteWindow
and MemFree declarations; only sub_020288DC needed a provisional extern.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The previous commit left src/overlay_29_022F0EDC.c carrying a provisional
`extern s32 SetActionUseMovePlayer();` because replacing it with the real header
broke three modules. The cause was not the return type, which is void, but the
parameters: the function was landed as (struct action_data*, u8, u8), and a u8
parameter makes every caller truncate.
SetLeaderAction's call site in retail passes GetTeamMemberIndex's result
unchanged and sign-extends the move index to 16 bits:
mov r1, r0
mov r2, r4, lsl #0x10
mov r2, r2, asr #0x10
bl SetActionUseMovePlayer
so the parameters are s32 and s16. The callee's own body cannot distinguish
u8, s16 or s32 -- it stores both with strb, and all three score 0 against it --
which is why the wrong types went unnoticed when it was landed.
With (struct action_data*, s32, s16) both functions match and the provisional
extern is deleted in favour of the header, leaving one declaration in the tree.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Third cleanup sweep; five more asm files and their includes removed:
asm/overlay_29_023021F0.s UpdateIqSkillsWrapper
asm/overlay_29_022EBC98.s SetActionUseMovePlayer
asm/overlay_29_02349658.s ov29_02349658
asm/main_0205EBF0.s sub_0205EBF0
asm/overlay_29_022F52B0.s ov29_022F52B0
SetActionUseMovePlayer sets ACTION_USE_MOVE_PLAYER and then two
action_parameters[].action_use_idx fields, using the enumerator and struct that
already exist. sub_0205EBF0 indexes mission_deliver_list::unk18, the struct
src/main_0205C73C.c already uses, rather than a fresh placeholder.
NOT done, deliberately: src/overlay_29_022F0EDC.c still carries the provisional
`extern s32 SetActionUseMovePlayer();`. Replacing it with an include of the real
header, which declares the function void, breaks the ROM -- SetLeaderAction's
match depends on the s32 form. The two declarations therefore still disagree.
That is a real defect in one of the two, not a formality, and it is left alone
rather than papered over; see the reasoning note.
ov29_022F52B0 is a bare veneer with no argument setup, so its arity is not
evidenced by anything and void(void) is a guess.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Second cleanup sweep. Each was the only function left in its own asm file, so
each removes an asm file and its include:
asm/overlay_29_0234B0CC.s SetMessageLogPreprocessorArgsStringToName
asm/overlay_29_023268CC.s DoMoveSmokescreen
asm/overlay_29_02325F74.s DoMoveMorningSun
asm/overlay_29_02326E20.s DoMoveGust
asm/overlay_29_022EA980.s GenerateDungeonRngSeed
Two needed a second reading:
- DoMoveGust doubles the damage multiplier when the defender's
bide_class_status.bide is 7 or 8. The target tests it as a byte with
`add #0xF9`, so the source adds 0xF9 rather than subtracting 7; `- 7` compiles
to `sub #7` and does not match.
- GenerateDungeonRngSeed runs the LCG twice, stores the second result back to
prng_state::preseed, and returns the two halves recombined. The operand order
in the final or-expression decides whether the mask or the shift is scheduled
first; taking the shifted half first matches.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Each of these was the only function left in its own asm file -- fragments left
behind when earlier work split a larger file around them. Decompiling them
removes five asm files and their includes outright:
asm/main_0204FD38.s IncrementNbSteals
asm/overlay_29_022EAC78.s ov29_022EAC78
asm/overlay_29_0234B010.s ov29_0234B010
asm/main_0202BCDC.s IsAdvancedMenuActive
asm/main_02030A2C.s SetWindowTitle
IncrementNbSteals and ov29_022EAC78 are empty in retail -- a single bx lr. Their
signatures are not evidenced by anything and are written void(void).
IsAdvancedMenuActive reads unk_0202AAA8::field_0x1A0, the same field and the same
shape as CheckParentMenuField0x1A0 next door.
SetWindowTitle writes a halfword at offset 0x20 of the same struct, which was
inside the opaque field_0x9[0x97] placeholder. That array is split into
field_0x9[0x17] / u16 field_0x20 / field_0x22[0x7E] -- same total size, and
nothing indexed the placeholder.
Authored by Claude (Opus 5) under human direction. All five verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
GetLeaderMonster, SetDungeonRngPreseed, GetDungeonRngPreseed,
SetDungeonRngPreseed23Bit, FindClosestUnoccupiedTileWithin2,
FindFarthestUnoccupiedTileWithin2, FindUnoccupiedTileWithin3,
FixedRoomActionParamToDirection, ov29_022E9FC0 and ov29_022E9FD0.
The three preseed accessors read and write prng_state::preseed at offset 8,
using the struct dg_random.h already defines. SetDungeonRngPreseed23Bit sets bit
0 and masks the top byte, which is where its name comes from.
GetLeaderMonster merged into dg_uty.h alongside GetLeader, so the provisional
`extern struct monster *GetLeaderMonster();` that src/overlay_29_022F0EDC.c
carried is removed -- that file already includes dg_uty.h, which now declares it
properly with a (void) parameter list.
The three FindNearbyUnoccupiedTile wrappers pass a displacement table and a
count (0x19 within 2, 0x31 within 3). That helper and its tables are undeclared
in the tree, so their prototypes are written from these call sites and are
provisional.
Authored by Claude (Opus 5) under human direction. All ten verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
SetMessageLogGroupStartFlag, GetMessageLogPreprocessorArgs,
InitMessageLogPreprocessorArgs, SetMessageLogPreprocessorArgsFlagVal,
SetPreprocessorArgsIdVal, SetMessageLogPreprocessorArgsNumberVal,
SetMessageLogPreprocessorArgsString, SetMessageLogPreprocessorArgsSpeakerId,
SetMessageLogPreprocessorArgsSpeakerId0x30000 and ov29_0234AFFC.
All reach a preprocessor-argument block through MESSAGE_LOG_INFO, a 12-byte
global holding a flag byte and a pointer. The block is the existing
struct preprocessor_args from main_0200C4FC.h: its member offsets -- id_vals at
+0x10, number_vals at +0x24, strings at +0x38, speaker_id at +0x4C -- land
exactly on the asm's 0xCAC, 0xCC0, 0xCD4 and 0xCE8 relative to the 0xC9C base.
Ten independent accessors agreeing with that layout is corroboration of the
existing struct rather than a new claim.
The global itself gets a placeholder, struct unk_023537CC, named for its
address; only its flag byte and pointer are known.
Authored by Claude (Opus 5) under human direction. All ten verified at score 0
(eleven candidates matched; ten landed). Confirmed by a matching build:
build/pmdsky.us/pmdsky.us.nds: OK. None carries a region conditional, so US
alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The previous commit dropped eight functions whose only difference from the
target was instruction order: retail loads the stat global before materialising
the constant destined for the stack, the candidate after. The cause is aliasing.
Declared as plain externs, MWCC must assume the stores to [sp] can affect the
global, so the load cannot move above them. Declaring the globals const frees
it, and all eight match:
DoMoveGrowl, DoMoveLowerDefense1, DoMoveFeatherDance, DoMoveFakeTears,
DoMoveMetalSound, DoMoveScreech, DoMoveCharm and DoMoveHealOrder. Also
LogMessageQuiet and LogMessageWithPopup, which pass FALSE/TRUE to LogMessage.
Every existing `extern struct StatIndex` in src/ is const-ified to match, since
the globals cannot be const in one translation unit and not another. Thirteen
files touched; the functions already landed against them still match, which is
independent confirmation of the reading.
DoMoveMetalSound lowers by 3 stages, not the 2 assumed when it was first
written; the asm says `mov r3, #3`.
Authored by Claude (Opus 5) under human direction. All ten verified at score 0.
Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a
region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Landed as one batch after verifying every function individually, rather than in
groups of ten. Five clusters:
- Sound and effect veneers that tail-call a shared helper with fixed ids:
PlaySeByIdIfShouldDisplayEntity, PlayEffectAnimationPos,
PlayEffectAnimationPixelPos, ChangeMonsterAnimation and ov29_022E67E4.
- Four DoMoveDamage__* wrappers plus DoMoveVacuumCut, all
`return DealDamage(...) != 0`.
- DoMoveExpose, DoMoveSport, DoMoveMirrorMove, DoMoveMiracleEye,
DoMoveNightmare, DoMoveSureShot, DoMoveFeint and DoMoveRockPolish, which
branch or sequence two calls rather than forwarding once.
- Minimap byte accessors over GetMinimapData, and GetTileTerrain.
- Wrappers whose argument setup evidences the callee arity:
ShouldDisplayEntityWrapper, UseSingleUseItemSelf, CalcSpeedStageWrapper,
GenerateCleanItem and three ov29_* forwarders.
Eight further candidates were prepared and are deliberately NOT included:
DoMoveHealOrder, DoMoveScreech, DoMoveCharm, DoMoveGrowl, DoMoveLowerDefense1,
DoMoveFakeTears, DoMoveFeatherDance and DoMoveMetalSound. Each passes a constant
on the stack to a six-argument stat helper, and each differs from the target only
in instruction *order*: retail loads the stat global before materialising the
constant, the candidate after. Same instructions, same count. They stay in asm.
Authored by Claude (Opus 5) under human direction. All 42 verified at score 0
individually before landing; confirmed together by a single matching build:
build/pmdsky.us/pmdsky.us.nds: OK. None carries a region conditional, so US
alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveLowerAccuracy1, DoMoveSweetScent, DoMoveNastyPlot, DoMoveSwordsDance,
DoMoveTailGlow, DoMoveBlowback, DoMoveHiddenPower, DoMoveDecoy,
DoMovePsychoShift and DebugRecruitingEnabled.
Five more stat handlers use the StatIndex helpers from move_orb_effects.h. The
rest depart from the template:
- DoMoveBlowback passes the attacker's own action.direction, read through
GetEntInfo.
- DoMoveHiddenPower forwards move and item_id to DealDamage with a 0x100
multiplier, its fifth argument on the stack.
- DoMoveDecoy passes three constants, the last on the stack.
- DoMovePsychoShift guards on attacker != defender before transferring.
DebugRecruitingEnabled returns TRUE unconditionally; dungeon_recruitment.h
already declares it, so the signature is not a guess.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate. Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
None carries a region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveTag0x1AB, DoMoveDoubleTeam, DoMoveMinimize, DoMoveAmnesia,
DoMoveBoostAttack1, DoMoveBoostDefense1, DoMoveBoostDefense2, DoMoveDefenseCurl,
DoMoveGrowth and DoMoveFamish.
Eight forward to BoostOffensiveStat / BoostDefensiveStat / BoostHitChanceStat
with ATK_STAT_IDX or SPATK_STAT_IDX and a stage count. These take
struct StatIndex by value, per move_orb_effects.h, which is why the asm loads the
global's word into r2.
src/overlay_29_0232E250.c declared those globals as s32 and BoostDefensiveStat
with an (s32, s16) tail, both contradicting move_orb_effects.h. Since the two
declarations lived in different translation units the compiler never saw the
disagreement and the ROM still matched. Replaced with an include of the canonical
header; DoMoveDefendOrder is unchanged and still matches.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate once the canonical StatIndex signature was used. Confirmed by a
matching build: build/pmdsky.us/pmdsky.us.nds: OK. None carries a region
conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DungeonRngSetPrimary, DungeonRngSetSecondary and DungeonRngUnsetSecondary write
prng_state::use_secondary and ::idx_secondary, using the struct dg_random.h
already defines and the extern src/dg_random.c already uses. No new types.
PlayLevelUpSound, PlayDungeonTipSound__022EB63C, PlayDungeonTipSound__022EB66C,
ov29_022EAC9C and ov29_022EACAC each tail-call sub_02017C50 with a fixed id
(1, 7, 7, 0, 5). PlaySeByIdIfNotSilence and PlayMeByIdIfNot998 guard on a
sentinel before forwarding -- 0x3F00 and 998 respectively, the latter matching
the name.
sub_02017C74 keeps the signature src/overlay_25_init.c already declares for it.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate. Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
None carries a region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveHealBlock, DoMoveEmbargo, DoMoveGravity, DoMoveMagnetRise, DoMoveTrickRoom,
DoMovePowerTrick, DoMoveInvisify, DoMoveMetalBurst, DoMoveGrudge and
DoMoveDamageStealItem. This exhausts the forward-and-return-TRUE family.
Three depart from the template:
- DoMoveInvisify passes the attacker as both arguments (mov r1, r0), so
Invisify targets the user rather than the defender.
- DoMoveGrudge and DoMoveDamageStealItem are tail calls rather than
call-then-return-TRUE, so they are written `return f(...)`. DoMoveDamageStealItem
forwards its own four arguments unchanged to DoMoveTakeaway, which is still asm.
DoMoveMetalBurst passes 15 to SetReflectStatus, a third distinct value for that
selector after 4 (Counter) and 10 (Rebound).
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate. Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
None carries a region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveShocker, DoMoveOneRoom, DoMoveHurl, DoMoveMobile, DoMoveSeeStairs,
DoMoveLongToss, DoMovePierce, DoMoveAquaRing, DoMoveGastroAcid and
DoMoveLuckyChant -- the same forward-and-return-TRUE template as the previous
batches. All ten helpers are undeclared in the tree and take provisional externs
written from the call site.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate. Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
None carries a region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveRebound, DoMoveStayAway, DoMovePowerEars, DoMoveSlowDown,
DoMoveSearchlight, DoMovePetrify, DoMovePounce, DoMoveTrawl, DoMoveDrought and
DoMoveHpGauge.
Two of them give a second data point for a helper argument seen earlier:
SetReflectStatus takes 10 here where DoMoveCounter passes 4, and TryWarp takes 1
where DoMoveWarp passes 0. Both third arguments are therefore selectors of some
kind; nothing here establishes what they select.
DoMoveSlowDown forwards to LowerSpeed with the same arguments as
DoMoveLowerSpeed1. The remaining helpers are undeclared in the tree and take
provisional externs written from the call site.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate. Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
None carries a region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveWrap, DoMoveMagicCoat, DoMoveProtect, DoMoveDestinyBond, DoMoveMirrorCoat,
DoMoveSnatch, DoMoveReflect, DoMoveSeeTrap, DoMoveScan and DoMoveNoMove.
Nine forward to a two-argument helper and return TRUE; DoMoveNoMove passes a
third argument of 0. None of the ten helpers is declared in the tree, so each
takes a provisional extern written from its call site.
DoMoveSeeTrap and DoMoveScan forward to RevealTrapsNearby and RevealItems, the
first helpers in this family that are not TryInflict*Status.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate. Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
None carries a region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveLowerSpeed1, DoMoveParalyze__02328230, DoMoveParalyze__0232B434,
DoMoveConfuse, DoMovePoison, DoMoveToxic, DoMoveCurse, DoMoveWarp,
DoMoveLightScreen and DoMovePerishSong -- the same shape as the previous two
batches.
Three forward to helpers already declared in move_orb_effects.h (LowerSpeed and
TryInflictParalysisStatus), which fixes their argument meanings. The rest take
provisional externs written from the call site; TryWarp's third and fourth
parameters are left unnamed because nothing establishes what they are.
DoMoveParalyze__02328230 and DoMoveParalyze__0232B434 join
DoMoveParalyze__02326E80 from an earlier batch: three separate wrappers around
TryInflictParalysisStatus with identical arguments.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate. Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
None carries a region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveEndure, DoMoveUproar, DoMoveMist, DoMoveSafeguard, DoMoveTaunt,
DoMoveConversion2, DoMoveBlock, DoMoveWish, DoMoveIngrain and DoMoveSetDamage --
the same shape as the previous batch: forward to one status helper and return
TRUE.
All ten helpers are undeclared in the tree, so each takes a provisional extern
written from its call site. DoMoveBlock forwards to TryInflictShadowHoldStatus,
the same helper DoMoveShadowHold uses, with the same argument.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate. Confirmed by a matching build: build/pmdsky.us/pmdsky.us.nds: OK.
None carries a region conditional, so US alone is sufficient.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DoMoveVitalThrow, DoMoveHealStatus, DoMoveEncore, DoMoveStringShot,
DoMoveFocusEnergy, DoMoveCounter, DoMoveParalyze__02326E80, DoMoveShadowHold,
DoMoveHaze and DoMoveBoostSpeed1 -- each a thin wrapper that forwards to one
status or stat helper and returns TRUE, following DoMoveDefendOrder's shape.
Three helpers were already declared in move_orb_effects.h, which fixes the
meaning of their constant arguments; the other seven take provisional externs
written from the call site.
Authored by Claude (Opus 5) under human direction. All ten matched on the first
candidate against a context preprocessed from the real headers. Confirmed by a
matching build: build/pmdsky.us/pmdsky.us.nds: OK. None of the ten carries a
region conditional, so US alone is sufficient here.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Two region-conditional calls the US build cannot see:
- SetDecoyAiTracker(GetLeader()) is US/EU only. Its declaration in
overlay_29_022FB538.h already sits inside #ifndef JAPAN, so calling it
unconditionally broke the JP build at compile time, not just at the checksum.
- ov29_022F2FE4() is an extra call EU makes after ov29_022E0B44() in the tail
block, from an #ifdef EUROPE in the asm that the previous commit missed.
The remaining region differences in the replaced asm need no source change. Two
are struct-layout consequences that dungeon.h already models. The other three
looked like swapped constants but are immediate encodability: in each region the
value that is a single-instruction ARM immediate is materialized inline and the
other is derived from the pooled 0xBA3 and cached in a stack slot. 0xBA0 and
0x8E0 encode; 0xBA1 and 0x8DF do not, and the JP offset of -0x2C1 swaps which of
the pair is which.
This corrects the previous commit, which stated that JP was expected not to
match and that EU was unaffected. EU was in fact broken by it, and JP needed one
guarded call rather than a rewrite.
Authored by Claude (Opus 5) under human direction. Confirmed by matching builds
of all three targets: build/pmdsky.us/pmdsky.us.nds: OK,
build/pmdsky.eu/pmdsky.eu.nds: OK, build/pmdsky.jp/pmdsky.jp.nds: OK. EU and JP
were verified matching at the parent commit as well, so both are gains rather
than a pre-existing state.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Split asm/overlay_29_022F05B4.s at 0x022F2B3C; add src/overlay_29_022F0EDC.c
and its header, and both objects to main.lsf.
Type corrections the match required:
- dungeon.field_0x1d8 / field_0x1dc: paired u16 fields -> struct position
- dungeon.field_0x614: u32 -> s32, tested >= 0
- display_data.leader_target_direction_mirror: enum direction_id -> u8; the
enum type folds the stored 0xFF to -1 and drops retail's mov r0, #0xff
- CanMonsterMoveInDirection's direction parameter: u16 -> s32; a u16 parameter
narrows the argument at the new call site, where retail passes it unchanged
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Not verified: the JP target, which is expected not to match -- the replaced asm
carries JAPAN-specific variants beyond the message-id offset. Two volatile
stand-ins remain in the new source, and this translation unit declares
DUNGEON_PTR as a scalar where other files declare it as an array; the two are
not byte-interchangeable here.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm:
IsMarowakTrainingMaze ov29_022EAEFC
FixedRoomIsSubstituteRoom ov29_022EAF20
IsSecretRoom SetUnkMusicFlag
IsDungeonEndReasonFailure ov29_0234969C
ChangeDungeonMusic ov29_023496B0
All ten reach struct dungeon through DUNGEON_PTR and every offset lands on a
named field: id at 0x748, hidden_outlaw_defeated_message at 0x2, unk_music_flag
at 0x792, dungeon_music_playing_id and field_0x2cb00 at 0x2CB06 and 0x2CB00,
gen_info.fixed_room_id at 0x40DA, and fainted_monster_dungeon_end_reason at
0x2CA66. No new types.
The constants resolve to existing enumerators that corroborate the function
names rather than being chosen to fit them: FIXED_SUBSTITUTE_ROOM is 110 and
FIXED_SECRET_ROOM is 113, which is what FixedRoomIsSubstituteRoom and
IsSecretRoom compare against; DAMAGE_SOURCE_ESCAPE is 633, the 0x279 in
IsDungeonEndReasonFailure. IsMarowakTrainingMaze tests the range
DUNGEON_NORMAL_FLY_MAZE to DUNGEON_EXPLORER_MAZE inclusive, 180 to 190.
IsDungeonEndReasonFailure reads its union member through an explicit (s16) cast.
enum damage_source_non_move declares no negative enumerator, so under -enum min
it is unsigned and compiles to ldrh with an unsigned comparison, where the target
has ldrsh and a signed one. The cast is local and reversible; adding a sentinel
to a shared enum, as was done for monster_id earlier on this branch, would not
be.
Two of the ten merged into src/run_dungeon_1.c, which already declares
DUNGEON_PTR as an array and reaches it as DUNGEON_PTR[0]. They follow that file's
convention rather than introducing a second spelling; both forms were checked on
a scratch and produce identical bytes.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm:
GetForcedLossReason GetHiddenStairsField
SetForcedLossReason SetHiddenStairsField
AreLateGameTrapsEnabledWrapper GetHiddenFloorField
GetSuccessfulExitTracker SetHiddenFloorField
SetDungeonEscapeFields GetMinimapData
All ten read or write struct dungeon through DUNGEON_PTR, and every offset lands
on a field dungeon.h already names: forced_loss_reason at 0x14,
successful_exit_tracker at 0x18, end_floor_no_death_check_flag at 0x8,
minimap_display_data at 0x1A264, and three fields of gen_info at 0x40C4 --
hidden_stairs_type, hidden_floor_type and fixed_room_id. No new types.
These are verified against the preprocessed dungeon.h context built in the
previous commit rather than a hand-written struct, which is why nine of the ten
matched on the first candidate instead of appearing to and then failing the
build. The setup cost of that context is now paid, so any overlay_29 function
reaching DUNGEON_PTR is cheap to verify; a scan finds 59 more small ones.
GetMinimapData needed the null check written as a conditional expression rather
than an if with an early return. The target predicates the whole thing --
addne twice for the non-null path, then moveq for the null one -- where the
statement form branches. That is the same predicate-ordering effect recorded on
the LFO handlers in 909484fb, in the direction that wants the expression form.
The US and JP offsets differ for the gen_info and minimap fields, which the asm
carries as #ifdef JAPAN. Nothing version-specific is needed in the C: the fields
are named the same in both, so dungeon.h resolves each build to its own offset.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm:
IsOutlawMonsterHouseFloor IsDestinationFloorWithMonster
IsGoldenChamber MissionTargetEnemyIsDefeated
IsJirachiChallengeFloor SetMissionTargetEnemyDefeated
IsDestinationFloorWithFixedRoom
IsDestinationFloorWithHiddenOutlaw
GetMissionEnemyMinionGroup
These are the functions the two previous commits could not land. The obstacle was
the scratch context, not the C, and this commit is the result of replacing it.
The context is now the real headers rather than a hand-written struct: mwccarm
-E over a file including global.pch, dungeon.h and mission.h, with
LM_LICENSE_FILE set, gives 14500 lines reproducing what the build actually sees.
Two things that took finding. global.pch is required -- without it util.h fails
at typedef s32 fx32_8, because s32 comes from nitro.h, which the build supplies
through -prefix rather than through any include in dungeon.h. And -E consumes
every macro, so TRUE, FALSE and NULL have to be restored by hand afterwards;
their values are taken from lib/include/nitro/types.h, NULL being ((void *)0) on
the C branch.
Against that context the earlier verdicts change. All eight functions landed in
the two previous commits still score 0, which is the check that the new oracle
agrees with the build. Eight of the thirteen unlanded ones score 0 unchanged and
are landed here as written. GetMissionEnemyMinionGroup does not: the real headers
let the compiler fold the +1 of enemy_species[index + 1] into the load
displacement, one instruction shorter than the target. Binding index + 1 to a
local first blocks the fold and scores 0, which is what is landed.
Four remain unmatched and are genuinely unmatched rather than mis-measured:
IsCurrentMissionType and IsCurrentMissionTypeExact at 515, where the target
branches to separate return blocks and every spelling tried so far produces
predicated moves; GetItemToRetrieve at 665 and GetItemToDeliver at 1530. Their
scores are identical under both contexts, so the context was never their problem.
The mission_type arguments use the enumerators from mission.h -- ARREST_OUTLAW,
EXPLORE_WITH_CLIENT, TAKE_ITEM_FROM_OUTLAW and CHALLENGE_REQUEST -- because the
build rejects the implicit conversion from int. The values agree with the
function names rather than being chosen to fit them.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm:
ov29_02349688 0x02349688
SetTargetMonsterNotFoundFlag 0x023496C4
GetTargetMonsterNotFoundFlag 0x023496D8
GetMissionTargetEnemy 0x02349620
The first three read and write the two dungeon flags at 0x1 and 0x3;
GetMissionTargetEnemy returns enemy_species[0]. No new types.
The previous commit left thirteen functions unlanded because they failed four
checksums with no compile error, and recorded the elimination so far. Bisecting
found a one-function reproducer, GetMissionEnemyMinionGroup, and disassembling
the object the real build emits identifies the cause.
The scratch and the real build disagree, and the scratch was wrong. For
enemy_species[index + 1] the target computes the index first --
add r0, r0, #1, then add r0, r1, r0, lsl #1, then add #0x700 and ldrsh [r0,
#0x6e]. The real build folds the constant into the displacement instead, giving
add r0, r1, r0, lsl #1, add #0x700, ldrsh [r0, #0x70], one instruction shorter
and the same address. Same compiler, same flags, same C: the difference is the
context. The scratch used a cut-down struct dungeon carrying only the offsets
these functions touch, and against that struct the compiler emits the target's
form; against the real header it folds.
That invalidates the scratch verification for the whole cluster, not just this
function. Score 0 against an invented struct is not evidence about a build that
uses the real one, even when every offset and size matches -- and they do, which
was confirmed separately by compile-time assertions against the real headers.
So the thirteen are not near-matches to iterate on; they need re-verifying
against a faithful context before any C is judged. A preprocessed context is
obtainable -- mwccarm -E on a file including dungeon.h and mission.h, with
LM_LICENSE_FILE set, yields 9200 lines that reproduce the real build's view --
but that output does not yet compile as scratch context on its own, failing at a
typedef, so wiring it up is unfinished.
The four here were each verified by a matching build rather than by the scratch:
three of them alone, GetMissionTargetEnemy alone, and all four together.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm:
ov29_02349188 0x02349188
GetMissionDestination 0x023491A4
ov29_023491B8 0x023491B8
IsDestinationFloor 0x02349208
All four reach DUNGEON_PTR->mission_destination, whose type is already fully
described in dungeon_mode.h. No new types.
This is the surviving part of a larger batch. Seventeen functions from this
cluster reached score 0 on a scratch; these four build matching in the tree and
the other thirteen do not, failing four checksums -- main, OVY_10, OVY_29 and
OVY_31 -- with no compile error. The thirteen are not in this commit.
What is known about that failure, so the next attempt does not repeat the work.
The struct layout is not the cause: compile-time assertions against the real
headers confirm mission_destination at 0x760, the two dungeon flags at 0x1 and
0x3, target_enemy_is_defeated at 0x1B, fixed_room_id at 0x16, enemy_species at
0xE, and sizeof(enum mission_type) == 1 and sizeof(enum monster_id) == 2 -- all
the offsets the scratch context assumed. A build of HEAD with no changes is
clean, so the regression is in the batch. Landing these four alone is clean.
Adding nine more, none of which use a mission_type enumerator, reproduces the
failure, which rules out the enumerator conversions that were the initial
suspicion. No pre-existing declaration of any of the seventeen names exists
elsewhere in the tree.
That main.sbin fails alongside the overlay points at something structural rather
than at a function body -- most likely the overlay size or the symbol exports in
the regenerated .inc files, since extract_function.py rewrites the .public list
when it splits an overlay .s. The byte-level diff of OVY_29 against a clean build
is the next step, using the cmp -l technique that located the rodata problem in
0f2e0ea8.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm, all reaching UNK_STORAGE_STRUCT_0X14:
sub_02043148 sub_0204329C sub_020433E0 ClearStorageSelectedItemTable
sub_0204316C sub_02043380 sub_02043400 CountSelectedStorageItems
sub_02043218 sub_02043398 sub_02043434 GetFirstSelectedStorageItemIndex
sub_0204322C sub_020433C0 sub_02043468
sub_0204323C sub_02043254
struct unk_020AFEE0 was introduced two dozen commits ago holding only the pointer
at 0x8 and filler before it. Its first eight bytes are now split into what these
functions read: halfwords at 0x0 and 0x2 and a pointer at 0x4, with a word at
0x10 added past the existing end. Nothing that already used field_0x8 moves.
The object behind that pointer is new, and its size is derived rather than
guessed. Its s16 array at 0x4 is bounded by 0x3e8 in three separate loops, which
is 1000 entries and ends at 0x7D3; the signed byte the collection-menu wrappers
read sits at 0x7D4, immediately after it. Two further fields at 0x910 and 0x18BC
come from sub_020430F4. Filling the gaps brings the struct to exactly 0x18C0,
which is the size in upstream's own name for its allocator,
InitUnkStorageStruct0x18c0. That agreement is the reason the layout is worth
trusting; the field names remain placeholders because nothing here establishes
meaning.
Fourteen of the seventeen matched on the first candidate. The other three are
loops whose index and pointer landed in swapped registers, with every
instruction otherwise correct, and all three were closed by reordering the local
declarations -- the same effect recorded in the previous two commits. All
orderings were enumerated rather than guessed: two locals for sub_02043254 and
GetFirstSelectedStorageItemIndex, three for CountSelectedStorageItems.
Eight callees remain asm and their prototypes are provisional, declared in the
shared header rather than repeated per file. None of them was declared anywhere
else in the tree. Two are typed from evidence rather than convention:
IsCollectionMenuActive and IsCollectionMenuState3 return bool8 because both
callers mask with and r0, r0, #0xff.
sub_020430F4 was attempted and is not included. It reaches the same struct and
its two extra fields are what sized it, but it stands at 495 after correcting
sub_0202C654 to the four arguments the target passes -- the fourth is a zero in
r3 that the first reading missed. Left in asm rather than landed as a near-match.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm:
SetPokemonBattled 0x0204FE58
GetNbItemAcquired 0x020503CC
SetPokemonBattled is the battled counterpart of SetPokemonJoined, already
decompiled in src/main_0204FDFC.c, and merges into that file beside it. It
differs only in which completion flag it sets and which flag array it indexes,
so the existing function is the template rather than anything derived here.
GetNbItemAcquired counts the set bits of items_acquired_flags across 0x580
entries. The pointer is hoisted to a local because the target loads it once
before the loop, and the word and bit indices use signed division and remainder,
which is what produces the asr/lsr/ror sequences the target has.
Both matched on the first candidate.
Seven adventure-log functions remain in asm. Four are large --
ComputeSpecialCounters at 144 instructions, CopyLogTo at 122, CopyLogFrom at 120
and ClearAdventureLogStruct at 74. The other three are not large but index the
region at 0x260 as an array spanning special_challenge_flags and the five
sentry_duty_game_points that follow it, mixing signed and unsigned division of
the same index within one function; expressing that without either misdescribing
the struct or writing something the compiler will not reproduce needs more care
than this batch had left.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm, all of them accessors on ADVENTURE_LOG_PTR:
SetAdventureLogStructLocation SetAdventureLogCompleted IncrementNbEvolutions
SetAdventureLogDungeonFloor GetAdventureLogCompleted GetNbEvolutions
GetAdventureLogDungeonFloor IsAdventureLogNotEmpty IncrementNbEggsHatched
IncrementNbDungeonsCleared GetNbDungeonsCleared GetNbEggsHatched
IncrementNbFriendRescues GetNbFriendRescues GetNbPokemonJoined
GetNbMovesLearned SetVictoriesOnOneFloor GetVictoriesOnOneFloor
GetNbPokemonBattled IncrementNbBigTreasureWins SetNbBigTreasureWins
GetNbBigTreasureWins SetNbRecycled GetNbRecycled
IncrementNbSkyGiftsSent SetNbSkyGiftsSent GetNbSkyGiftsSent
IncrementNbFainted GetNbFainted GetSentryDutyGamePoints
No new types. struct adventure_log already exists in include/adventure_log.h with
every field these touch named and offset-commented, and struct dungeon_floor_pair
in dungeon.h; every offset in the asm lands on a named field. That header is the
reason a batch this size was tractable: the only modelling needed was reading
which field each function touches.
Two orderings decided four of the thirty, both instances of the rule added to
MATCHING_TIPS in the previous commit. IsAdventureLogNotEmpty needed its loop
index declared before the pointer local, not after -- with the pointer first the
two land in swapped registers, at 35. The three Set functions needed the
completion-flag write placed before the clamp rather than after: written in the
obvious order the clamp is emitted ahead of the first pointer load and the two
literal-pool words come out reversed, at 250 each. Reordering the statement fixed
all three.
The counters clamp at 0x000F423F, decimal 999999, and each function sets one bit
of completion_flags[0]. The exact shape differs per function and is reproduced
rather than normalised: three increment then clamp, two clamp then increment, and
the three setters write the flag before storing the value. Those differences are
what the target does.
The clamp comparisons are signed even though the fields are u32, so each is
written with an explicit (s32) cast rather than retyping a shared struct.
ADVENTURE_LOG_PTR and _022AB69C move into adventure_log.h. src/main_0204FDFC.c
declared the pointer itself, from before any of this was decompiled; that line is
replaced by the header it already includes, and the object rebuilds unchanged.
Nine adventure-log functions remain in asm and are not in this commit:
ClearAdventureLogStruct, ComputeSpecialCounters, CopyLogTo, CopyLogFrom,
SetItemAcquired, GetNbItemAcquired, SetChallengeLetterCleared, SetPokemonBattled
and SetSentryDutyGamePoints.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Decompile from asm:
DseTrackEvent_VolumeFade 0x020723C0
DseTrackEvent_PanFade 0x02072668
Both were deferred near-matches. Each reads a signed byte as the fade target and
a little-endian pair as its duration in ticks, writes the target, and then either
snaps current to it when the duration is zero, zeroes the duration when there is
nothing to travel, or divides the distance by the duration to get the per-tick
delta. The two differ only in which struct dse_fade they address -- volume at
0x2c, pan at 0x3c -- both of which already exist in dse.h.
What closed them was declaration order, and in the opposite direction to the
obvious one. The target loads ptr_next_byte[2] first, then [0] and [1]; writing
the target expression first, so that its load comes first in the source, leaves
the loads in ascending address order at score 620. Declaring ticks first and the
target second emits them in the target's order and scores 0. So for this
scheduler the later-declared expression's loads are issued first, and the fix is
to write the declarations in the reverse of the order the asm reads them.
Seven other spellings of the same two statements were tried -- explicit locals
for each byte in the target's load order, s8/s16/s32 intermediates, an inline
cast, and a pointer-cast subscript -- and every one of them stayed at 620. Only
the declaration swap moves it.
The lib/DSE headers extract_function.py generates do not include dse.h, so both
were given it; every other header in that directory already does.
Three deferred DSE functions remain, and this commit does not close them.
DseTrackEvent_TuningFade is the same family with a bend fade plus the SetTuning
tail; hoisting container out of channel before the flag test takes it from 935 to
760, which is progress and not a match. DseTrackEvent_SetupKeyBendLfo is
unchanged in substance: its instructions have matched for some time and only
register assignment differs. All 120 orderings of its five byte locals were tried
this time, along with eleven structural variants, and the best is 55 rather than
the 65 it sat at; the earlier note suggested exhaustive permutation as the untried
move, and it has now been tried and does not close it.
No comments are added to any pmd-sky file.
Authored by Claude (Opus 5) under human direction. Confirmed by a matching
build: build/pmdsky.us/pmdsky.us.nds: OK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
