Follow-up to the silent-drop audit. The old ABI passed every parameter through four fixed zero-page transport slots `$04-$07`, imposing a hard 4-param cap (E0506) that didn't compose with structs/arrays/u16s and fell back to "pack args into a global" workarounds whenever a function needed five things. The transport scheme also cost every non-leaf call a 4-LDA/STA spill prologue (~28 cycles, 16 bytes) to copy args out of ZP before the next nested `JSR` could clobber them. Replace it with a hybrid convention keyed on leaf-ness: - **Leaf callees** (no nested `JSR` in body, ≤4 params): unchanged. Caller stages args into `$04-$07`; body reads those slots directly for its entire lifetime. No prologue copy. Fastest path, 3-cycle ZP stores + 3-cycle ZP loads, preserves the SHA-256 leaf-primitive optimisation that motivated the original fast path. - **Non-leaf callees** (body contains a nested `JSR`, OR ≥5 params): direct-write. Caller stages each argument straight into the callee's analyzer-allocated parameter RAM slot, bypassing the transport slots entirely. No prologue copy on the callee side. Saves ~24 cycles and ~16 bytes per call vs the old transport-then-spill path, and — crucially — scales past 4 params because the per-param slots live wherever the analyzer put them rather than in a fixed ZP window. The analyzer's ceiling moves from 4 to 8. Functions with 5–8 params are silently promoted to the non-leaf convention (even if their body has no nested `JSR`), which pays the direct-write cost rather than the prologue-copy cost — still cheaper than the old ABI. Declarations with 9+ params still emit E0506. ### Implementation - `function_is_leaf` now also requires `param_count <= 4`. - `IrCodeGen::new` populates `non_leaf_param_addrs: HashMap<String, Vec<u16>>` — for every non-leaf function, the ordered list of addresses its parameters occupy. Callers use this to route each arg directly to the right slot. - `IrOp::Call` branches on presence in the map: non-leaf → direct- write, leaf (or absent — 0-arg case) → ZP transport. - `gen_function` no longer emits a prologue. Leaves didn't have one; non-leaves had a 4-LDA/STA copy that is now unnecessary because args arrive pre-written to the slot. - The previous `leaf_functions: HashSet<String>` field is removed; leaf-ness is now inferred from absence-in- `non_leaf_param_addrs` at the call site. ### Tests and regressions - `eight_param_non_leaf_function_stages_every_arg_at_its_allocated_slot` compiles an 8-param function, scans PRG for a distinct `LDA #\$NN / STA <addr>` per arg (immediates `0x11..0x88`), and asserts that STAs to the `$04-$07` range are strictly fewer than 8 — proof the old transport path is gone for this call. - `non_leaf_call_direct_writes_args_to_callee_param_slots` replaces the old `gen_function_prologue_spills_params_to_local_ram` test with a dual assertion: (a) no `LDA \$04` prologue at the callee entry, and (b) the caller-side STA lands at the analyzer-allocated param slot, not at `\$04-\$07`. - `analyze_rejects_function_with_more_than_4_params` renamed and rewritten for the new 8-param cap. - `feature_canary.ne` gains a 6-param `sum6` call (1+2+3+4+5+6 = 21) as check 8. The canary stays green (all eight checks pass), so the committed golden is unchanged. ### Blast radius - Six example ROMs change bytes (arrays_and_functions, function_chain, mmc1_banked, pong, sha256, war) because their non-leaf call sites pick up the shorter staging sequence. - Pong and war audio hashes refresh (pure layout-timing shift; no behavioural change in the 180-frame no-input window). docs/pong.gif and docs/war.gif stay byte-identical. - `examples/function_chain.ne`'s header comment updated to document the leaf vs non-leaf split it exercises. - `docs/language-guide.md` parameter-count section and E0506 entry updated to reflect the new rule. All 720 Rust tests pass; all 35 emulator goldens pass. https://claude.ai/code/session_01AoQ678uVeqpyayvWHpfDhC |
||
|---|---|---|
| .github/workflows | ||
| benches | ||
| docs | ||
| examples | ||
| fuzz | ||
| scripts | ||
| src | ||
| tests | ||
| .gitignore | ||
| Cargo.lock | ||
| Cargo.toml | ||
| CLAUDE.md | ||
| compiler-bugs.md | ||
| LICENSE | ||
| plan.md | ||
| README.md | ||
| spec.md | ||
NEScript
A statically-typed, compiled programming language for NES game development.
NEScript compiles .ne source files directly into playable iNES ROM files, with no external assembler or linker dependencies. The compiler handles everything from source text to a ROM you can run in any NES emulator.
Source: examples/platformer.ne
Source: examples/war.ne
Source: examples/pong.ne
Quick Start
# Build the compiler
cargo build --release
# Compile an example
cargo run -- build examples/hello_sprite.ne
# Run the output ROM in an emulator
# (produces examples/hello_sprite.nes)
Hello World
game "Hello" {
mapper: NROM
}
var px: u8 = 128
var py: u8 = 120
on frame {
if button.right { px += 2 }
if button.left { px -= 2 }
if button.down { py += 2 }
if button.up { py -= 2 }
draw Smiley at: (px, py)
}
start Main
Features
- Game-aware syntax -- states, sprites, palettes, backgrounds, and input are first-class constructs
- Full type system --
u8,i8,u16,bool, fixed-size arrays (u8[N]),enum,struct - Rich control flow --
if/else,while,for i in 0..N,loop,match - Functions -- with parameters, return types, real
inline funsplicing for single-return and void-body shapes, recursion detection - State machines --
statewithon enter,on exit,on frame,on scanline(N)handlers - Compile-time safety -- call depth limits, recursion detection, type checking, unused-var warnings
- IR-based optimizer -- constant folding, dead code elimination, strength reduction (incl. div/mod by power-of-two), copy propagation, peephole passes including INC/DEC fold and live-range slot recycling
- Full 16-bit arithmetic -- u16 add/sub/compare lower to carry-propagating paired operations
- Multiple mappers -- NROM, MMC1, UxROM, MMC3 (including multi-scanline IRQ dispatch per state)
- Audio subsystem -- frame-walking pulse driver with user-declared
sfx/musicblocks, builtin effects and tracks, period table, and zero-cost elision when unused - Palette & background pipeline --
paletteandbackgroundblocks, initial values loaded at reset, vblank-safeset_palette/load_backgroundruntime swaps - Asset pipeline -- PNG-to-CHR conversion, inline tile data, sfx envelopes, music note streams
- Inline assembly --
asm { ... }with{var}substitution, plusraw asm { ... }for verbatim blocks - Hardware intrinsics --
poke(addr, value)/peek(addr)for direct register access - Debug support --
--debugflag enablesdebug.log/debug.assert, runtime array bounds checks, frame-overrun and sprite-overflow counters, and thedebug.frame_overrun_count()/debug.frame_overran()/debug.sprite_overflow_count()/debug.sprite_overflow()query builtins — all stripped entirely in release builds - Sprite-per-scanline mitigations -- three layers of defense for the NES's 8-sprites-per-scanline hardware limit: compile-time
W0109static check for literal layouts, runtimecycle_spritesflicker intrinsic for dynamic scenes, and debug-mode telemetry viadebug.sprite_overflow()for playtest assertions - Compile-time diagnostics --
--dump-ir,--memory-map,--call-graphflags - Single binary -- no dependencies on ca65, Python, or any external tools
Documentation
- Language Guide -- complete reference for every language feature
- Architecture -- compiler internals and module overview
- NES Reference -- hardware quick reference for contributors
- Examples README -- how to build and run examples
Examples
| Example | Features demonstrated |
|---|---|
hello_sprite.ne |
D-pad input, sprite drawing |
bouncing_ball.ne |
Automatic movement, edge detection |
coin_cavern.ne |
Multi-state game, functions, constants, gravity |
arrays_and_functions.ne |
Arrays, functions, while loops, inline functions |
state_machine.ne |
State transitions, on enter/exit, timers |
sprites_and_palettes.ne |
Inline CHR data, scroll, type casting |
mmc1_banked.ne |
MMC1 mapper, bank declarations, multiply |
uxrom_user_banked.ne |
UxROM mapper with a bank Foo { fun ... } block — first example to put real user code in a switchable bank, called via a generated cross-bank trampoline |
uxrom_banked_to_banked.ne |
UxROM with two bank Foo { fun ... } blocks — exercises a banked→banked call (step in Logic calls clamp in Helpers) routed through the same trampoline that handles fixed→banked |
palette_and_background.ne |
Palette and background declarations, reset-time load, vblank-safe set_palette / load_background swaps |
auto_chr_background.ne |
background Stage @nametable("file.png") with automatic CHR generation — the resolver dedupes the PNG's 8×8 cells, encodes them as 2-bitplane CHR, and slots them into CHR ROM after the sprite tile range |
friendly_assets.ne |
Pleasant asset syntax — named NES colours, grouped bg0..sp3 palettes with universal:, ASCII pixel-art sprites, legend { } + map: tilemaps, palette_map: attribute grids, scalar sfx pitch:, note-name music with tempo: |
structs_enums_for.ne |
Structs, enums, for loops, struct literals |
nested_structs.ne |
Nested-struct fields (hero.pos.x) and array struct fields (hero.inv[0]) with chained literal initializers |
inline_asm_demo.ne |
Inline asm with {var} substitution, poke/peek |
audio_demo.ne |
Audio subsystem: user sfx/music blocks, builtin effects, play/start_music/stop_music |
noise_triangle_sfx.ne |
Noise and triangle channel sfx via channel: noise / channel: triangle on sfx blocks |
sfx_pitch_envelope.ne |
Per-frame pulse pitch: arrays — the audio tick walks the pitch envelope in lockstep with the volume envelope and writes $4002 on every NMI for a frequency-sweeping siren tone |
metasprite_demo.ne |
metasprite Hero { sprite: ..., dx: [...], dy: [...], frame: [...] } declarative multi-tile groups — draw Hero at: (x, y) expands to one OAM slot per tile so 16×16 sprites stop needing four hand-written draw statements |
sprite_flicker_demo.ne |
cycle_sprites — rotates the OAM DMA start offset one slot per frame so scenes with more than 8 sprites on a scanline drop a different one each frame. Turns the NES's permanent sprite-dropout hardware symptom into visible flicker, which the eye reconstructs from adjacent frames. Pairs with the compile-time W0109 warning and the debug-mode debug.sprite_overflow() / debug.sprite_overflow_count() telemetry for a three-layer defense against the 8-sprites-per-scanline limit. |
platformer.ne |
End-to-end side-scroller — custom CHR tileset, full background nametable, metasprite player with gravity/jump physics, wrap-around scrolling, stomp-or-die enemy collisions, live stomp-count HUD, pickup coins, user-declared SFX + music, and a Title → Playing → GameOver state machine with a proximity-based autopilot so the headless harness demonstrates the full gameplay loop (stomp, stomp, die, retry) inside six seconds |
war.ne |
Production-quality card game — a complete port of War split across examples/war/*.ne: title screen with a 0/1/2-player menu, animated deal, sliding face-up cards, deck-count HUD, "WAR!" tie-break with buried cards, victory screen with a fanfare, and a brisk 4/4 march on pulse 2. Pulls in nearly every NEScript subsystem (custom 88-tile sheet, felt nametable, 8-bit LFSR PRNG, queue-based decks, phase machine inside Playing, multiple sfx + music tracks). Building it surfaced seven compiler bugs, all fixed on the same branch — see git log for the details. |
feature_canary.ne |
Regression canary — a minimal program that paints a green universal backdrop at frame 180 when every memory-affecting construct round-trips correctly, and flips to red if any check fails. The committed golden is green; any silent-drop regression (state-locals, uninit struct field writes, u16 high byte, array elements, slow placement, function return values) turns it red. Built after PR #31 to close the "goldens capture whatever happens, not what should happen" failure mode that let the state-local bug survive for a year. |
sha256.ne |
Interactive SHA-256 hasher — an on-screen keyboard lets the player type up to 16 ASCII characters, and pressing ↵ runs a full FIPS 180-4 SHA-256 compression on the NES (64 rounds + 48-entry message-schedule expansion, all written in NEScript with inline-asm 32-bit primitives). The 64-character hex digest renders as sprites across eight 8-character rows at the bottom of the screen. Splits across examples/sha256/*.ne with a phased driver that runs four iterations per frame so the full hash finishes in well under a second; the jsnes golden captures SHA-256("NES") = AE9145DB5CABC41FE34B54E34AF8881F462362EA20FD8F861B26532FFBB84E0D. |
Compiler Commands
# Compile to ROM
nescript build game.ne
# Compile with custom output path
nescript build game.ne --output my_game.nes
# Type-check only (no ROM output)
nescript check game.ne
# View generated 6502 assembly
nescript build game.ne --asm-dump
# Enable debug mode
nescript build game.ne --debug
Emulator Compatibility
Output ROMs are standard iNES format and work with any NES emulator:
Project Status
NEScript implements all five planned milestones:
| Milestone | Status | Key Features |
|---|---|---|
| M1: Hello Sprite | Done | Full compiler pipeline, assembler, ROM builder |
| M2: Game Loop | Done | Functions, arrays, IR, optimizer, call graph analysis |
| M3: Asset Pipeline | Done | PNG-to-CHR, sprites, debug.log / debug.assert |
| M4: Optimization | Done | Strength reduction, ZP promotion, type casting, asm-dump |
| M5: Bank Switching | Done | MMC1/UxROM/MMC3, bank declarations, software mul/div |
694 tests across the lexer, parser, analyzer, IR, optimizer, codegen, assembler, linker, runtime, ROM, and asset modules, plus a pixel- and audio-exact emulator harness that captures a golden framebuffer + audio hash for every example. CI runs fmt, clippy, test, example compilation, and the emulator harness on every push.


