Operator confirmed scripts/build_switch.sh --fetch --fused --version
0.0.0-test produced gen1recomp-0.0.0-test-switch.nro (9147a64) and
copied it to sdmc:/switch/gen1recomp/ per switch-install.md.
Co-authored-by: Cursor <cursoragent@cursor.com>
Yellow's IsSurfingPikachuInParty swaps the player's overworld sheet to
a Pikachu-on-a-surfboard when the party mon that knows SURF is a
Pikachu. The recomp was missing both halves of this: the sheet was
never extracted, and the engine had no seam for the swap.
Extraction: SurfingPikachuSprite (gfx/sprites/surfing_pikachu.2bpp)
loads outside SpriteSheetPointerTable via
LoadSurfingPlayerSpriteGraphics2, the same bypass RedBikeSprite uses.
Added the symbol to the Yellow manifest and a parallel extract in
RomExtractor / build_rom_data / extract/sprites, minting
SPRITE_SURFING_PIKACHU. PaletteFX colors it (player OBP palette, same
as the bike).
Engine: new field.playerSprites.surfPikachu (default
SPRITE_SURFING_PIKACHU, guarded so before extraction the ride keeps
the Seel). Player.new caches surfPikachuSprite; pose() picks it when
surfing and the SURF-mon is a Pikachu. New
OverworldState:syncSurfingPikachu derives the flag from partyKnows at
every surf-state toggle (mount, dismount, fly, teleport, blackout,
forced-surf tile, boot-restore). Runtime-only, re-derived at load so a
party change between save and load is honored.
Lane B: RFC at docs/rfcs/0001-surfing-pikachu-sprite.md.
Backward-compatible — existing mods see no change (surf still defaults
to SPRITE_SEEL; surfPikachu only resolves on a Yellow import after
regeneration). Parity tests in tests/parity_surfing_pikachu_sprite.lua
(12/12) and tests/mod_world_tests.lua (19229/19229 with the new
boot-seed checks). tests/parity_cinnabar_east_surf.lua (24/24)
unchanged.
Regeneration required: re-run make_yellow_manifest.py against a
pret/pokeyellow checkout, then re-import the Yellow ROM.
Clarify that Mac+OpenMTP is contributor tooling (not the final product
contract), record Dusklight method references, and surface how/what we
tested so reviewers and the community can help finish the Switch port.
Co-authored-by: Cursor <cursoragent@cursor.com>
Ignore hidden ._*.gb in the ROM inbox like zips, keep mixed-rescan success notices while appending sibling errors, and cover FileData mount / PK rejection in unit tests.
Co-authored-by: Cursor <cursoragent@cursor.com>
Mac OpenMTP leaves ._*.zip sidecars that fail PhysFS mount and hide a good install; prefer FileData mount on Horizon and keep success notices when a sibling fails.
Co-authored-by: Cursor <cursoragent@cursor.com>
T24 closed: fused boot/import, Red+Blue Play after Blue mount fix,
and NRO-only replace keeps saves on OLED.
Co-authored-by: Cursor <cursoragent@cursor.com>
NX Play for Blue failed because mountVersion relied on absolute
PHYSFS_mount first. Prefer love.filesystem.mount of blue|yellow, overlay
generated trees by version prefix, and align CacheFs.prefix in bootGame.
Co-authored-by: Cursor <cursoragent@cursor.com>
PhysFS does not merge archive data/ with save-dir data/generated, so
fused NX Play crashed after import. Prepend-mount generated trees,
fall back to CacheFs.read in Data:load, keep multiline lua-error logs,
and skip Boot.run when network is unvalidated. T24 stays open.
Co-authored-by: Cursor <cursoragent@cursor.com>
Expose a generic seam so a mod can drive a second screen without the
engine owning any dual-screen layout policy:
- render.compose hook in Renderer:endFrame hands a mod the finished
world + UI canvases, their SGB zones, the frame metrics,
Renderer:blitCanvas (lifted from the internal blit closure) and the
SecondScreen bridge. Return true to take over the window; no wrap (or
calling next) runs the normal single-window composite byte-for-byte.
- SecondScreen.lua + the Android Presentation bridge (love_android_
secondary_* in common/android.cpp, GameActivity secondary display)
as the optional physical-second-display transport.
No battle-render changes: a mod lays out the two screens (including any
battle split) itself. Ships with a unit test, no-mod parity via
gate_hooks, and docs/modding.md (D14).
OLED pass at 2699c9a: physical A confirms, B cancels; game.love
a208b21e… recorded. T19 suspend×10 still open.
Co-authored-by: Cursor <cursoragent@cursor.com>
SDL labels south as a and east as b; on Switch remap so physical A
confirms and physical B cancels in launcher and NamingScreen.
Co-authored-by: Cursor <cursoragent@cursor.com>
Introduce an OPTIONS -> PERFORMANCE setting that scales the port's
optional presentation extras down for weaker hardware, so older/lower-end
devices can run the game smoothly.
The tier governs the three heaviest non-faithful extras -- the 3D TILT,
the GBC FX post-process shader, and survey ZOOM (which renders connected
neighbor maps) -- plus a hard FPS ceiling. It never touches game logic,
which is fixed-step off dt, so every tier plays identically.
- src/core/Performance.lua: tiers (auto/high/balanced/low), a conservative
device auto-detect (ARM handhelds -> low, phones -> balanced, normal
desktops -> high), per-tier caps, and the option-row cycle. Zero
requires, like GameVersion.
- Game:applyOptions clamps the *live* presentation state against the tier
without rewriting stored options, so a lower tier hides the player's
TILT/GBC FX/ZOOM/FPS choices and a higher tier restores them exactly.
- Zoom.offsetRange floors the range at FIT when survey is disallowed, so
the option row, hotkey, and mouse wheel all stop at close-up on LOW.
- New save.options.performance default "auto"; OPTIONS row heads the
display group and re-applies live.
- Tests: tests/engine/performance_tiers.lua (ROM-free); mod_ui_tests row
golden updated for the spliced row.
AUTO resolves to HIGH on a normal desktop and on every options.lua that
predates the option, so the common case is unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Q6bFAiQyZ5jDmewsbB4LG9
Follow-up to #452, from dogfooding the merged build on a real iPhone.
A mod fused into game.love sits in the read-only app bundle: the mod
manager's Delete can't remove it and it reappears every launch, and an
imported newer version can never fully replace it. iOS now packs no
mods, matching every other platform - mods install as .zips at runtime
with a real install/upgrade/delete lifecycle.
With nothing fusing it, mods/pokewalker moves out of this tree to its
canonical home, https://github.com/mresnick67/Gen1ReComp-Pokewalker
(already at 0.2.0 there with move learning and a paged credit dialog;
in-tree 0.1.0 was stale). docs/ios-install.md points at it. The
event:save.loaded coverage-debt entry returns to gate_meta_coverage:
its coverage came from the mod's test suite and leaves with it.
Local reproduction of the ci workflow on this branch: all tiers, mod
lint, fingerprint + meta-coverage gates green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Makes the iOS build a first-class citizen: ROM/mod/save import through
the system document picker (the README's missing "UIDocumentPicker
handoff"), Files-app drop-in support, and an opt-in Apple Health
step-sync seam consumed by a new gallery mod (Pokewalker).
Native layer (mobile/ios/native/, wired by mobile/ios/patch_love_src.py
on every build, so the fetched love-src tree stays pristine + re-patchable):
- GRPickerBridge.swift: love.system.pickFile("rom"|"mod"|"sav") and
love.system.createFile on iOS with the same contract as love-android's
SAF picker (picked_rom.gb / picked_mod.zip / picked_save.sav /
export_done.flag in the save dir). Reached from wrap_System.cpp via the
ObjC runtime, so liblove needs no Swift interop.
- GRBootstrap.m: sweeps .gb/.gbc/.zip/.sav dropped in Documents (Files
app / Finder sharing) into the save dir on every activation;
UIFileSharingEnabled + LSSupportsOpeningDocumentsInPlace in the plist
overlay. Drop a ROM, open the app, it imports with zero taps.
- GRHealthBridge.swift: love.system.syncHealthSteps() -> read-only
HealthKit step query anchored to the last sync, delivered as
steps_pending.json (merge-not-overwrite). HealthKit entitlement +
usage description included.
Lua:
- RomImporter: iOS rides the Android mobile flows; a 0.5s poll consumes
picker deliveries (iOS pickers are in-process modals, so the Android
refocus rescan never fires); failed pick copies surface as an
on-screen notice via pick_error.txt.
- main.lua: on iOS, stop forwarding touchpressed to the Importer - LOVE
already synthesizes a mousepressed for the primary touch, and the
resulting same-frame double-present made the document picker
auto-dismiss with zero documents (silent import failure).
- mods/pokewalker: opt-in Pokewalker mod (manifest v2, MECHANIC,
permissions declared, mod.card, CHANGELOG, headless test suite 9/9,
modkit validate --base imported + lint clean). Fused into iOS
game.love only; loads dormant anywhere without the bridge.
Build (scripts/build_ios.sh):
- Fix Xcode 26: the global PRODUCT_NAME override also renamed liblove.a
and broke the app link; the app bundle is renamed after the build
instead.
- Fix nondeterministic pack failures: grep -q + pipefail races SIGPIPE
on the game.love content checks.
- Simulator builds sign ad-hoc so entitlements embed (HealthKit works in
the simulator).
- Device builds: signing team auto-detected from the keychain,
CODE_SIGN_STYLE=Automatic + -allowProvisioningUpdates for CLI-only
provisioning, per-team derived bundle ID (explicit App IDs are
globally unique, so third parties can't sign the project default),
gitignored mobile/ios/bundle_id.local pin, and --install to push to a
connected iPhone.
- docs/ios-install.md: a zero-knowledge walkthrough from bare Mac to
playing on an iPhone.
Backward compatibility: no behavior change on desktop or Android. The
new love.system functions exist only under LOVE_IOS; RomImporter's
mobile flag simply includes iOS alongside Android; the main.lua change
is iOS-gated; the Pokewalker mod is packed only by the iOS build script
and its option defaults off.
Verified on an iPhone 17 Pro simulator and an iPhone 16 Pro device:
scripted ROM import to title screen, Files-drop zero-tap import,
picker-driven mod install and save import/export, HealthKit permission
sheet + step credit (4000 steps -> +200 EXP at the default rate through
the engine growth curve).
On a window too short for the stacked single-column layout -- a phone, or a
narrow desktop window -- the ROM / SAVE FILES / Play / SAVE SLOT stack ran
past the bottom of the window while the footer stayed pinned there and drew
over it. Nothing clipped the panel and nothing scrolled it, so the overflow
was unreachable.
Everything under the tab bar (panel, updater banner, footer) is now one
scrolling column, used only when it is taller than the room below the tab
bar. The strip, logo and tab bar stay pinned, so navigation is always on
screen, and the footer is laid out downward from footerTop right after the
content instead of upward from the window bottom.
- RomImporter.pageScrollFor is the whole decision, pure and covered by
tests/engine/launcher_page_scroll.lua. A window that grows back drags the
offset down with it, so the page never stays parked past its own end.
- The panels return their natural height as they draw, so the measurement is
the previous frame's: the same one-frame settle the slot and mod lists
already rely on.
- One scroll axis at a time. While the page scrolls, the panels draw paged:
the slot and mod lists take their natural height, keep no inner scroll
region and report a max of 0, so wheel, right stick and drag all move the
page. Two-column layouts do not overflow, paged stays false, and every one
of these behaves exactly as before.
- inside() and _ptIn() reject a rect that scrolled out of the viewport, so a
control that slid under the tab bar cannot be clicked through it. Tab chips
are pinned and exempt. pageScroll resets on a tab change.
Android had no scroll gesture at all: the launcher is handed no move events
(main.lua forwards neither touchmoved nor mousemoved while it is up) and its
mouse emulation was never trusted, which is what "no reliable pointer
polling" referred to. That was survivable while every scroll region was an
inner list, and useless once the page itself scrolls, since a phone is
exactly where it overflows. love.touch is pollable, so _pointerHold reads the
first active touch there and hands _updateSlotDrag the same (held, y) pair
the mouse gives on desktop. Slot rows and mod toggles consequently arm on
press and commit on release on Android too, matching desktop, so a swipe that
starts on a card scrolls instead of selecting it. All of it is gated on
touchPollable: without love.touch every Android path is exactly what it was.
conf.lua also grows minwidth/minheight (480x360) for the desktop window,
under which the cards stop being readable. Mobile is fullscreen and ignores
it.
Verified on the Android emulator (1080x2400) against a build of the parent
commit: before, the footer painted over the SAVE SLOT card with no way to
reach it; after, the page pans by touch and the footer is reachable and
intact.
Andrew Quenehen
Andi Miller
bryanthaboi
Bart in 't Veld
erereck
Claude
Myles Resnick
Marcus Pereira