From the row-58 checkpoint, 30 brain minutes on the stub rotation: at
most three gym arrivals end in the fly walking straight back out inside
ten seconds, and the fly goes up the room to row 6 or above, where the
Jr. Trainer stands. Base: one arrival, back out in 309 frames, highest
row 11 -- fails. Branch: 4 arrivals, 1 back out, 30,879 frames in the
gym, highest row 2, beside the leader. Rung 11 is printed, not asserted.
The catch dump prints a small map whole with its people drawn and off
the screen, each person's ledger entries and whether a route reaches
them; outcomes are keyed by the map they finished on, the trace line
carries the seam's bytes (wCurOpponent included), and the drive ends
with the rank. How row 58's mechanism was read.
FLY_PROBE_CATCH_MAP and FLY_PROBE_CATCH_ENTRIES stop the route survey
forty frames after the fly's Nth arrival on a map (the first frames on a
new map byte still carry the old map's warps), and the dump lists every
person the macros can see with its talked, blocked and reached entries.
Row 58's pad was one door in and one door out, and what it needed read
was the room on the far side of the door.
Review round 1, N2. The gameboy-decoder-config-v1 form (channels as ordered arrays), vectors for
raw mode and the Pokemon Red macro group computed by flysim's legacy_profile_identity test from
gameboy_decoder_config_with_macros and reproduced by @flybrain/session-types from the oracle's
gameboyDecoderConfig. The example composition now carries the real macros-mode digest and
channel set.
PROF-02a and RT-01a, machine-readable half, per the operator's port decisions of 2026-09-23.
Generic (in the session schema set, so contractDigest moves):
- AgentTelemetry.stimulusRemainingMs (number|null): sugar admission reads the pulse from the
last commit.
- EpisodeRequest.kind is terminal | rollback.
- Environment.SaveSlot / Environment.RestoreSlot (capability gameboy-slots-v1) and
Agent.Rollback (capability legacy-ratchet-rollback-v1) payloads, with their scope checks;
maxSlots 4.
Legacy Game Boy (module gameboy, digested apart from contractDigest):
- registered payload schemas gameboy-readout-context-v1, gameboy-channels-v1,
gameboy-joypad-v1, gameboy-memory-inspection-v1, legacy-ratchet-rollback-v1, each SchemaRef
digest over its canonical declaration;
- the one legacy profile gameboy-legacy-fafb-v783-v1 embedding today's schema-1 fingerprint,
lif-1ms-f64-v2 and fly-kc-mbon-rstdp-v2, with its AssetRef digest;
- the composition declaration carrying the decoder and macro-channel configuration, the
executor pokered-macros-v1, gameboy-slots-v1, legacy-ratchet-rollback-v1,
legacy-transient-reset and FLYSIM01 as format of record, cross-checked against the FLYSIM01
compatibility string.
Fixtures regenerated by update_fixtures (new derived gameboy-legacy.json); valid/invalid cases
for every new type in both languages; the frame clock proven identical to the legacy f64
accumulator. flysim gains only a test (and a dev-dependency) that recomputes the pinned
fingerprint, versions, frame size, warm-up, clock and button order. The synthetic fly-session
agent reports stimulusRemainingMs null and its task names kind terminal; no runtime change.
Row 57's pocket had a way out: the road east, three tiles from the fly
and resting in the blocked window. The last resort ignores that window
but narrows to the ways toward the objective, and the only one was the
gym's door beyond the fence, so the route search refused and the road
was never tried. With the refusal withheld the pad went empty and the
fly waited out the road's window, ten brain minutes.
The narrowing is a preference and the dealer cannot search. When start's
route search cannot reach the preferred ways, it tries the rest of the
last resort, nearest reachable first, exactly as every walk chooses. The
pad is unchanged; only where the pressed macro walks.
The ROM proof now asserts the fly leaves the pocket inside a brain
minute: frame 517 (0.14 minutes), against 36,325 (10.1) on the base.
The seeding moves from scene_probe into examples/support/ledgers.rs and
trap_hunt reads it as FLY_TRAP_SEED_*, so both arms of a hunt can start
inside the state the live session was in: row 57's pad was dealt by
ledgers a restore starts empty, and a hunt from the bare checkpoint
never meets it. The report gains refusals by macro and the longest run
of one macro refused with the fly on one tile.
scene_probe's own choice seed is FLY_PROBE_RNG now, so it does not read
as one of the FLY_PROBE_SEED_* ledgers.
the_pewter_east_pad_is_never_one_dead_button_from_the_rung_ten_checkpoint,
gated on FLY_ROM and FLY_PEWTER_EAST_CHECKPOINT. Part one, from the
ratchet's rollback snapshot: GO ROUTE walks east, meets the youngster,
and the pushed ledger walls a tile he fires on, not the south entrance
the walk set out from. Part two, from the live frame with the pocket's
session ledgers rebuilt: twelve brain minutes, no button refused twice
running on one tile, and the fly leaves the pocket. On the base the same
run refuses GO ROUTE 746 holds running on one tile.
FLY_PROBE_CATCH=route drives the real PokemonPalette from a checkpoint,
one uniform choice per hold, prints every pad it deals and every refusal
with its reason, and reads the frame the pad comes down to one refusing
button on: which list emptied why, the no-window ledgers, and whether the
route search reaches each way out. FLY_PROBE_SEED_* rebuild the session
ledgers a restore starts empty; FLY_PROBE_RATCHET starts from the
ratchet's rollback snapshot; FLY_PROBE_HOLD presses raw directions first.
PokemonPalette gains inspect(), fences() and a doc(hidden) ledgers_mut()
for the survey. The loop never calls them.
Row 41 counted `YES` starts per map, which cannot tell one conversation from
another; the Pewter Gym stall was 64 `YES` and 62 `NO` in ten brain minutes at
one person. So the harness counts an answer to a box it can read as a choice,
charged to whatever the fly is facing -- `palette::facing_target`, which is
`TALK`'s own precondition and the same reading its ledger entry uses.
`the_fly_leaves_the_pewter_gym_guides_ring_from_the_rung_ten_checkpoint`: the
fly leaves map 54 on frame 1,337 against never in twenty brain minutes, spends
27 macros in the gym (one lap of the ring), answers the guide's prompt **once**
against a bound of four, re-opens the ring **never**, and deals `NEXT` on no pad
while a readable prompt is open.
The rung is printed and not asserted: neither this run nor either arm of the
trap hunt reaches rung 11 from this checkpoint, and why the fly does not walk
back through the gym door is row 12.5's ground and row 54's rather than this
row's. Named in the residuals.
Row 56 is a gym's dialog for thirty brain minutes with YES, NO and NEXT dealt
in equal thirds, so the question is which box each press is answering and
whether the seam can see it at all. `state::drawn_boxes` asks the screen
instead of a pinned rectangle: every complete `TextBoxBorder` on the frame.
`FLY_PROBE_CATCH=dialog` walks the conversation one raw pulse at a time and
prints both readings of every frame side by side, with the pad the palette
deals for it, and tallies how the candidate readings separate the frames.
`FLY_PROBE_ANSWER` picks the button a frame with a box on it is answered with,
so the two arms of a choice can be walked separately.
Three conflicts, all of them two reviews appending in the same place, all
resolved by keeping both sides: `macros.md` (row 50 keeps section 12.18, row 55
becomes 12.19), `macros-traps.md` (both trap rows, both residual pairs, both arm
sections) and `scene_probe.rs` (both survey modes, `accept` and `shop`). Every
source file auto-merged.
The two rows compose on the cartridge: from the mart checkpoint the fly leaves
the counter, walks the town, and no `MOVE n` blocks any more, so the longest
chain of one macro blocked on an unchanged frame is 1 in any scene.
From the checkpoint the stream looped in: the fly is out of the shop scene and
off the map inside twenty brain minutes, no macro reports `blocked` more than
three times in a row in the shop with the map, the tile, the wallet and the scene
unchanged, and the shop never deals an empty pad.
The chain is recorded per scene as well as overall, because the battle scene has
one of its own and it belongs to another review.
`FLY_PROBE_CATCH=shop` drives a `BUY ...` from a checkpoint the stream looped in
and prints, per frame, what the seam reads at the counter, the halves of every
purchase's precondition, the outcome of the macro three times over, and what an
A press at the counter really opens. `screen_text` decodes the tile buffer,
because the question is which figure the game drew and the bytes that would name
it are not rewritten between two of the mart's screens.
Row 50's two numbers, ROM-gated from the forest checkpoint. Before, on main:
MOVE n 940 starts and 838 blocked, 11 battles entered and 10 ended, worst 503
macros, median 48. After: 51 starts and 0 blocked, 13 entered and 13 ended,
worst 283, median 43, and the fly leaves the forest north through the gate.
The blocked share is the assertion the row is about; the median is what it
buys, and the worst battle is a tail rather than the run.
Row 50 asks which WRAM reading tells a battle menu that is accepting input
from one that is only drawn, and the honest way to answer it is to press.
FLY_PROBE_CATCH=accept drives real battles, and on every battle frame it
exports the emulator state, issues one directional pulse, reads
wCurrentMenuItem and puts the state straight back -- so the run is not
perturbed by the measurement and every frame gets a ground truth.
Beside that it asks every byte of WRAM and HRAM whether its values on
accepting frames are disjoint from its values on refusing ones, so a
reading is found rather than nominated.
The pulse releases the buttons before it presses: JoypadLowSensitivity acts
on a key's edge, so a direction the fly is already holding reads as refused
for the measurement's reason and not the cartridge's.
The operator's decision of 2026-09-22: pay the fly for keeping a wild Pokemon,
bump the adapter properly, and restart the live run from an early checkpoint
rather than from scratch.
The rule. `catch` is the catalog's ninth kind, appended so the key order
`counts` serializes in does not move. 0.30 for a species this run had never
owned, 0.10 for a repeat, three payouts per species for the lifetime of the
ledger; the `species` rule is untouched, so a first catch of a new species pays
0.80 across two kinds. The catch is read from `wCapturedMonSpecies` ($d11c),
whose comment in ram/wram.asm is "0 if no mon was captured": ItemUseBall zeroes
it before every throw and writes wEnemyMonSpecies into it only on the branch
that keeps the Pokemon, and UseBagItem's `.returnAfterCapturingMon` zeroes it
again and sets wBattleResult to 2 -- a value written on exactly two paths in
the game, that one and a link battle whose opponent ran. Both are required, so
a byte read out of a half-initialised battle cannot pay. Not wPartyCount: a
catch with a full party raises wBoxCount instead, and wPartyCount also rises
for a gift, a trade and a PC withdrawal.
"Never owned this run" is the `species` payout inside the same battle, because
nothing else can set a Pokedex bit during one. It is not read off the captured
species byte: that is the cartridge's internal index while the owned bitset is
by Pokedex number, and nothing in WRAM converts between them.
The address was resolved by tools/resolve_wram.py, not written by hand. The
tool needed NUM_TMS and NUM_HMS, which the decomp defines through its `const`
enumeration, so it now counts them from the file's own add_tm/add_hm
definitions and cross-checks NUM_TMS against the literal the same file
declares.
The feed's kinds are closed, so `catch` publishes on `wildwin` and nothing in
packages/feed or apps/stage changed. Deliberately not `pokedex`: the `species`
rule already pays for the bit the same catch sets. The stage's ticker copy is
keyed on the feed kind, so a catch row reads "wild win" -- stated in
docs/rewards-learning.md rather than left to be discovered.
v5 -> v6. STATE_VERSION stays 4: the rule adds one counter, `catchCounts`, and
changes nothing else, so a v5 state restores with it empty. That migration is
opt-in and needs all three of: the adapter segment being the only difference
between the two compatibility strings, the running adapter listing the
checkpoint's adapter in `migrates_from()`, and the deploy naming it in
FLY_ACCEPT_ADAPTERS. flysim applies the rule at restore and 05-deploy's gate
applies the same rule before it flips the symlink, writing the variable into
fly.env so the two cannot disagree.
The restart. infra/bin/fly-reset-to-milestone <N> archives both stores to a
dated directory, rewrites milestone-<N>.checkpoint with the ratchet's attempts
and recoveries at zero, installs it as the newest generation of both stores,
clears the milestone archives above N and the event log, and prints what it
did. It refuses while flysim is running and refuses a rung the run never
reached. The envelope work is in flysim::reset (`flysim
--reset-to-milestone N`); the shell script is the operator's wrapper.
Tests: catalog values and order; a synthetic WRAM trace of a catch (new,
repeat, cap, already-owned species, trainer/Safari/old-man/missed-ball
negatives, rollback replay); a v5 state restoring with the counter at zero; a
v5 checkpoint fixture accepted with the opt-in and refused without it; the
reset tool against copies of a state dir in temp directories; and a ROM-gated
catch from a rung-9 forest checkpoint, driven by the shipping THROW BALL macro.
The compatibility string differs from main's in exactly one segment, checked by
splitting both on `/`: pokered-unique8-v5 -> pokered-unique8-v6.
Forty minutes after the Boulder Badge the overworld pad had shrunk to `GO ROUTE` alone and it was
completing every 420 ms at a net of zero tiles. The trap hunt from that checkpoint on main is row
54 verbatim, one town further on: 14 distinct tiles in six brain minutes, 17 of 17 windows flagged,
`GO FRONTIER` 122 / `GO HEAL` 129 / `GO ROUTE` 126, every one done at a mean net of 0.0.
The run claims the fly leaves Pewter City for Route 3, that no chain of walks completes at a net of
zero tiles more than three times in a row, and that neither errand is offered in a town the run has
already shopped and healed in.
Six deploys in a day left the CHAT panel empty every time, because the ring
was in-memory session state and nothing on disk held it.
Every accepted line now rewrites <hot_dir>/chat-ring.json through the same
atomic sequence a checkpoint commit uses (tmp, fsync, rename, directory
fsync), and Sim::boot reads it back before the first publish. Lines older
than 24 h are dropped on load, only the newest [chat] ring of them are kept,
and an absent, unreadable, unparseable or unknown-version file is an empty
panel and a logged warning rather than a startup failure -- which is exactly
what a restart gave before this existed.
The sidecar is deliberately not part of the checkpoint: no new chunk in the
FLYSIM01 envelope, nothing added to the compatibility string, so
--print-compatibility is byte-identical and a build that refuses every
checkpoint in a directory still brings the panel back. It sits beside the hot
checkpoints because it has their lifetime, and it is written from the sim
thread on the command path, so it cannot refresh that directory's mtime --
the watchdog's liveness signal -- while the loop is wedged.
Tests: the ring round-trips through the sidecar, a corrupt or future-version
file is ignored, lines older than a day are dropped, and the integration
test's SIGKILL restart now proves the panel comes back with the lines it had,
that the durable store holds no copy, and that no chat text is in the
envelope.
Section 12.2's rule as a number on the cartridge: the longest chain of `GO ...` macros that
completed at a net of zero tiles. One is ordinary -- a walk that arrives by turning, a walk cut
short by a battle, a walk that finds its goal underfoot -- and a chain of them is the loop the
v0.4.5 after arm spent brain minutes 1.0 to 8.5 in. Three is the bound.
`FLY_PROBE_CATCH=step` holds one direction from a checkpoint and prints, per frame, the
coordinates, the grid's verdict, the tiles the cross-check disagreed on and every candidate for
"a step is in progress". It is what row 54 turns on: whether the coordinates change at the start
of a step or at its end.
The pad dump gains section 13's two errands -- which building each names, whether the ledgers have
paid it, and what the walk would aim at -- because an aim with no press settles where it stands.
From the live rung-10 checkpoint with the stub rotation: map 0x34 to map 0x36 on frame 2423, on
15 macros, against never in five and a half live hours. TALK is on the pad inside the gym -- the
fly standing in front of one of the people rung 11's place names -- and it presses it 15 times
there. BACK on a dialog or unknown frame is 0 against the 189 the hunt measured on Pewter City
alone, no frame deals a pad on an unknown with nothing drawn on it, and GO FRONTIER starts 0
times on the museum's two floors.
The unit test beside it is the other half of "the objective reaches the leader": on the gym's
own map the rung's place resolves to the person standing in it, the walk aims at the four tiles
around them, the door is not a candidate while they are there, and talking to them empties the
list. It also pins what the ROM run showed: the town's errands come first for every map in the
area, the gym included, so both have to be paid before the objective is the leader.
Two Stuck rollbacks fired on rung 10 inside half an hour, both on a fly that was walking, and
both were the ratchet working to contract: the stall window is reset by exploration -- one tile
the run has never stood on -- and a fly crossing a town it has already covered to reach the
rung's own door earns none of it. Entering a map for the first time already counts, because a
new map is ground nobody has stood on; re-entering one does not, which is what the museum was.
So the window gains a second signal, passed in by the caller and meaningless to the ratchet
itself, exactly as coverage is: the Pokemon loop answers with "nearer the objective, in map
hops, than this run has ever been", over the same map graph GO OBJECTIVE walks. It can fire at
most once per step of the road, it spends no budget, it captures nothing and it skips no
trigger. Nothing in the macro layer reads it back and no button is bound on it.
The checkpointed ratchet state is untouched: the signal is a level on one sample, not a counter.
The rung-10 review needed to tell "the whole-map grid is wrong on this map" from "this frame
was mid-step", and the refusal label cannot: both read "screen disagrees". So the decode is
split from its own cross-check, the disagreement is readable tile by tile, and the map survey
stands still for a while and counts how many of those frames decode.
On Pewter City, measured from the rung-10 checkpoint: standing still it decodes on 118 of 120
frames, and the frame the survey caught disagrees on three tiles by exactly one tile row in the
direction the fly was walking. The map is fine; the reading is refused while the fly is moving.
The first run of the new ROM test failed on YES 134, and every one past
the sixth was in Pewter City or the gym the fly had walked to -- which is
the fly playing the game, not the ring. Row 41 is about the presses spent
in the room, so the harness records starts by name while the fly is still
on the map it resumed on: 6 macros there, YES 2.
Unit: HEAL is off a full party's pad, including the rung-10 party's own
numbers, which is the assertion that the loop was never the heal's; TALK
is off a rested nurse's pad and on a hurt one's, and an ordinary
villager is unaffected; her prompt deals one answer and a plain box still
deals three; a readable prompt that is not hers deals both answers and no
NEXT; an answer whose prompt reopens is excluded and one that settled the
box is not; a completed heal and a declined prompt both retire her.
ROM-gated from the live checkpoint (FLY_CENTER_CHECKPOINT): the fly
leaves map 0x3a on a bounded number of macros, YES starts stay under
five, NEXT is on no pad while a prompt is readable, and TALK is on no pad
at a rested nurse. The run also counts dialog and prompt frames, which is
what makes the forty-six-to-one ratio visible.
Row 41 is the Pokemon Center nurse: 2,142 YES starts on one tile of map
0x3a with a party at 70/70. The survey that names it is a new probe mode
(FLY_PROBE_CATCH=nurse) that leaves the box with B and then pulses A,
printing every state the conversation passes through with the two-option
menu's own geometry beside it. It is a ring of forty-six presses and the
box is a choice on exactly one of them.
So the reading: wFontLoaded plus the border DisplayTwoOptionMenu draws,
plus the cursor it parks inside it. Both halves are needed -- the cursor
bytes survive the box closing, so all forty-six frames carry the
geometry. It claims that box and not "a choice is open" in general.
TargetKey gains Answer { at, yes } for the reopened-prompt exclusion.
A new test from the rung-10 Pewter checkpoint: the fly is on map 0x35, the
museum's upper floor, the objective is the gym where the badge is, and over
thirty-three brain minutes MENU is on no overworld pad, no overworld pad is
empty, MENU and BACK never alternate, and the fly leaves the building on frame
182. On v0.4.3 the same test fails at the first of those: MENU is on the pad of
maps 2, 52, 53, 54, 55, 57 and 58 with 147 starts.
The harness records where a macro reported blocked, which is what identified the
battle menu's swapped entries, and the forest test now asserts THROW BALL never
blocks: 15 starts and 0 blocked from the rung-9 checkpoint, against 6 of 6.
Its bound on a battle's macros goes from 400 to 700 against a re-measured worst
of 450, because a fly whose SWITCH and ITEM reach their own lists fights longer
than one that can only attack. Every battle still ends: 3 of 3.
Unit: no battle pad holds both NEXT and BACK, over every sub-state with and
without a potion and a ball, and NEXT is on none with a cursor accepting input
(the assertion the top-level menu failed before this branch). A battle frame
with a cursor accepting input is the fly's turn, against real WRAM, for all
four menus plus the two frames that correctly are not one. The bag deals
ITEM / THROW BALL / BACK, and BACK alone when the bag is empty.
ROM-gated, from the live rung-9 forest checkpoint: no pad with both buttons, no
NEXT while a menu accepts input, the longest NEXT/BACK alternation under four,
and every battle entered also left, each on a bounded number of macros. The
battle rules ask the scene the pad was dealt for rather than wIsInBattle, so
the $ff frame a lost battle passes through -- which reads Unknown, whose pad is
NEXT and BACK by contract -- is not accused of a battle rule.
The test fails on v0.4.2 from the same checkpoint with "NEXT was on the pad
while a battle menu was accepting input", NEXT dealt on battle/main, and a run
that never leaves map 51.
ROM-gated and checkpoint-gated, skipped cleanly without either. The decode
against the window predicate on every tile the window can answer for; the decode
against a survey of real presses, where every refused press has to be a wall, a
directed wall or a sprite in the way, and every step the cartridge made has to be
one the grid would have planned; GO FRONTIER aiming at ground outside the window
and walking there; and every way out of the map reachable in one plan with no
guessed tile in it.
Pallet Town: 221 walkable, 207 reachable, 90 window tiles, 120 surveyed, 58
refused presses. Viridian Forest: 719 walkable, all reachable, 90 window tiles,
120 surveyed, 74 refused presses, a 215-frame frontier walk out of the window and
ways out 27 to 149 steps away.
scene_probe prints the grid s size, its walkable count, the count reachable from
where the fly stands and the count never stood on, draws the ground with the
reading it actually used, and names the refusal when there is no grid. trap_hunt
carries the same line into every trace line and into the summary, so a stalled
walk can be read at a glance: a fly with 719 walkable tiles and 4 reachable ones
is fenced in and no re-plan will help it.
A ROM-gated run from the checkpoint the stream stalled on, gated on its
own variable because the two checkpoint tests beside it assert the map
their envelope is on and one envelope cannot be both.
What it holds: the objective is Pewter City, the road there is on the pad,
BACK is never on a battle pad with no list open, no ball is thrown at a
species the party holds, the fly gets turns, a move or a ball is chosen,
and a battle ends. Which move and which way it walks are the fly's, so
they are reported rather than asserted -- including the forest's north
gate, which this run does not reach.
A scene histogram says "battle" and a battle deals five different pads
(docs/design/macros.md 12.6, 13.1), so a loop inside one of them is
invisible in the tables the hunt already prints. Three more: frames per
sub-state, every macro channel ever bound in each one, and the macro
starts by the sub-state they started on.
