| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A browser has no file descriptors, so the wasm build's link port was inert.
Give it a pair of ring buffers instead: the guest's transmitted bytes pile up
in one, the host pushes received bytes into the other, and the external-clock
RX path pops from it (staying pending when empty, exactly like a quiet link).
gbw_serial_drain / gbw_serial_push hand them to JS through a fixed staging
buffer, so nothing allocates per frame.
That is enough to be a real link. SLIP-framed bytes are IP packets; unframed
bytes are console I/O - the same split tools/gbhub makes natively, which is
what lets a page type into the link port and reach the shell's stdin.
One machine per wasm instance, so the queues are plain statics: several Game
Boys on a page are several instantiations of the one compiled module, each
with its own linear memory.
The native link is untouched: the fd paths sit in the #else and, preprocessed
for a native target, gb.c comes out byte-identical to the previous commit
(verified with gcc -E). blargg cpu_instrs still passes and the golden-frame
tests are still pixel-exact.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
`make wasm` builds the core to WebAssembly (~44KB stripped) and web/ is a
front-end for it: drag-drop a ROM, pixelated canvas at any zoom, keyboard
mapped like the native build, battery saves in localStorage.
Built with zig cc, which carries its own wasm backend, linker and
wasi-libc - so no emscripten dependency, just zig on PATH. Only the
portable core goes in; render/kitty/control/main are terminal- and
POSIX-bound and stay out. src/wasm.c is the browser's whole host layer:
no terminal, no sockets, no files, and no frame pacing (the page's
requestAnimationFrame owns that). The machine is static, the RGB888 ->
RGBA conversion for putImageData happens in wasm, and nothing allocates
per frame.
The entire host interface is four WASI fd_* imports, pulled in by the
cart info banner's fprintf; the page shims them to console.log in ~20
lines. Cartridge RAM *is* the save file here, so it's exported for the
page to persist per cart title (every 5s and on unload).
Because both builds now advance on the same frame boundary, the port is
checked byte-for-byte against the native binary rather than by eyeballing
a canvas - tools/wasm_check.mjs runs N frames in node and diffs the
framebuffer against a --shot PPM:
cgb-acid2 @60, dmg-acid2 @60, cpu_instrs @600, mem_timing-2 @700,
pky.gbc @400 with an imported .sav, cgb-acid2 @120 through the full
CGB BIOS boot animation -> all pixel-identical
Also smoke-tested in headless chromium over CDP: core instantiates, ROM
loads through the file input, canvas animates (7 distinct frames over
5.6s), keyboard reaches the joypad, 60fps, and a 32KB save lands in
localStorage.
No audio: the APU is still stubbed.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Groundwork for building the core for targets that have no files, no
sockets and no terminal (a browser, next commit). Three changes, no
behaviour change for the native binary:
- cart_load() splits into cart_init_owned() (everything past the file
read) plus cart_load_mem(), which takes a ROM image already in memory.
The path version keeps owning the .sav wiring; a memory load attaches
no battery file and leaves save_path empty, which cart_save() now
treats as "nowhere to write".
- gb_load_bootrom_mem() does the same for boot ROMs, so the power-on
animation works wherever the image comes from.
- gb_run_frame() was declared in gb.h but never existed; main.c had a
private run_frame() instead. Implement it in gb.c and point main.c at
it, so every host advances the machine on one identical boundary.
The link-socket layer (--serial-sock) is the only genuinely unportable
part of the core, so under __wasm__ it becomes no-op stubs: the link is
simply always down and the call sites stay untouched. The file-backed
loaders compile out there too, which keeps the libc surface tiny.
Verified: blargg cpu_instrs (11/11), instr_timing, mem_timing,
mem_timing-2 all still Passed; --shot output unchanged.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Sixel is a 1987 printer format wearing a terminal costume: 256 colors per
image, no alpha, no compositing, one flat picture. --kitty [scale] takes
the other road on kitty/ghostty/WezTerm/Konsole:
- the framebuffer ships as raw RGB888, so no per-frame palette build and
no coarser quantization when a CGB game blows past 256 colors
- the HUD (title, render rate, frameskip, turbo, live button readout) is
a *separate* alpha-blended image on z-index 1, composited over the game
by the terminal instead of being burned into the picture; 'h' toggles
it, --no-hud starts without it
- consecutive identical images are dropped, so a pause menu or text box
stops the traffic dead
- pixels go over at 1x and the terminal scales them into a cell box, so
--kitty 6 costs the same bytes as --kitty 1 (padded to whole cells
first, so the zoom stays exact)
Two transfers: shm hands over the name of a POSIX shared-memory object
(~60 bytes/frame) when we can tell the terminal is local, b64 inlines the
bytes for ssh and tmux (which might be attached from another machine).
--kitty-shm / --kitty-b64 force either. tmux gets DCS passthrough wrapping.
--chrome and --palette work here just like in sixel mode.
Nothing about a graphics stream is human-readable, so tools/kitty_decode.py
reassembles a capture back into images - chunked payloads, tmux wrapping
and all - and can assert a frame is pixel-identical to a --shot:
tools/kitty_decode.py cap.bin -o frames --check shot.ppm
-> OK: image 1 matches shot.ppm exactly (160x144)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Built with -std=c11 + _POSIX_C_SOURCE, glibc's signal() gives System V
semantics: the disposition resets to SIG_DFL the moment the handler runs.
So a second SIGTERM - a supervisor retrying, `timeout` signalling the
process group, an impatient second Ctrl-C - killed us mid-shutdown, before
term_restore() put the terminal back and before the render backend cleaned
up after itself. Wrap sigaction() in install_sig() and route SIGINT,
SIGTERM and the serial-link SIGPIPE ignore through it.
Verified: staged tree builds clean (-Wall -Wextra, no warnings); added
lines are already clang-format clean (hook bypassed only because the
split commit leaves main.c partially staged).
|
| |
|
|
|
|
|
|
|
|
|
| |
Decimates the LCD N x N before the half-block encode so the monitor pane
takes 160/N cols x 72/N rows. Display-only: capture, recording and sixel
still read the full-res framebuffer.
Decimation is max-luma pick, not box-average: terminal fonts are 1px
bright strokes on dark background, and averaging dims them into grey
mush (the 'blurry font' regression). Keeping the brightest source pixel
of each block preserves stroke brightness and hue.
|
| |
|
|
|
|
|
|
| |
The control socket was only opened on the windowed path, so a --headless
GB couldn't be inspected. A scriptable, invisible GB is exactly the one
you most want to peek into - reading kernel WRAM over the debug socket is
how the TCP field-offset corruption got pinned down. Open it in the
headless loop too and poll it each frame; close it on exit.
|
| |
|
|
|
|
|
| |
The clean-exit opcode used to print nothing, so any guest crash that
walked into a stray $ED byte was indistinguishable from a silent
socket death (one of the kill vectors in the link-drop investigation).
Now it identifies itself on stderr with pc/af/bc/de/hl/sp.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Root-cause work for the 'idle link-socket drop' (see gbos
docs/link-drop-investigation.md): when the hub restarts or the bridge
dies, the link fd went dead forever - and before SIGPIPE was ignored,
the next console-echo write() would kill the whole emulator silently.
Now, in --serial-sock mode, EOF on read or EPIPE/ECONNRESET/ENOTCONN/
EBADF on either direction marks the link dead, closes the fd, and
retries connect() every 500ms from the serial poll path (the guest
polls constantly, so no extra plumbing). RX stays pending while down
(serial_no_eof semantics), so the guest just sees a quiet link that
comes back. Logs 'link socket lost / reconnected' either way.
Verified: SIGKILL the hub, restart it - every GB reconnects within ~1s
and keeps its lease.
|
| |
|
|
|
|
|
|
| |
Runtime chatter (link reconnect notices, input debug, cart warnings) was
landing mid-sixel-stream and corrupting the picture. Redirect fd 2 to
/dev/null for the duration of the display loop; restored on exit so the
final stats line still prints. Startup errors before the display begins
are unaffected.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The serial fds are serviced by cpu_step in every mode, so wiring them to a
connected unix socket lets networking run alongside a live display (--sixel /
terminal) - unlike --headless, which needs stdio for the serial. The emulator
connects to PATH (retrying ~1s so a launcher can listen first) and uses it
non-blocking with no EOF-exit; it overrides the headless stdio serial.
Pairs with gbos tools/netboot, which listens on PATH, brings up a TUN+NAT, and
bridges the socket <-> TUN. In windowed mode the OSK is the console and the
link port is pure network - cleanly separated. Verified: host pings the GB 3/3
over the socket link.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
- Rewrite argument handling: the ROM is now the final argument and options
precede it (parsed in argv[1..argc-2] so optional-value flags never swallow
the ROM). Add grouped --help/-h, unknown-option errors, and clear 'missing
value'/'no ROM' diagnostics.
- Add --chrome, --palette/--green, and --bios flags; set the display/capture
options before the --shot and recording paths so screenshots and recordings
honor them. Make --shot dump the composited capture frame.
- gbctl: pass the ROM last when spawning the emulator.
- gitignore the (copyrighted) bios/ directory; document all new flags.
|
| |
|
|
|
|
|
| |
The video recorder now writes render_capture_frame() instead of the raw
LCD framebuffer, so captures match what's shown: chrome frames record at
their larger dimensions (written into the .gbv header) and the LCD palette
is baked in. Button presses also show up lit in recordings.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
- --chrome: draw a DMG-style 'sl0pboy' shell around the LCD in sixel mode,
with a bitmap-font silk-screen (logo, labels), speaker grille, power LED,
and buttons that light up while held.
- LCD palette recolor (render_set_palette): remap the LCD to a 4-shade
palette by luminance for the classic monochrome look ('dmg'/'green' pea
green, 'pocket'/'gray'); applies to every render path.
- Refactor the sixel encoder into generic build_palette_buf/encode_sixel_buf
operating on any RGB888 buffer, shared by the bare LCD and the chrome canvas.
- render_capture_size/frame: produce the exact displayed frame (recolor +
chrome) as an RGB buffer, for screenshots and video recording.
|
| |
|
|
|
|
|
|
|
| |
Add optional boot ROM emulation. When a boot ROM image is loaded, gb_reset
starts execution at PC=0 with power-on register state instead of the
documented post-boot state; the ROM overlays low memory (with the CGB
0x0100-0x01FF header hole showing the cartridge through) until the game
writes 0xFF50 to hand off. gb_load_bootrom() accepts 256-byte (DMG) or
2304-byte (CGB) images.
|
| |
|
|
|
|
|
|
| |
--keys "udlrabse." plays scripted joypad taps (each char = one button
held 10 frames then released 10, so the guest sees a clean edge); pairs
with --headless --shot to drive on-screen input and capture the result.
When --keys is active, serial-RX EOF stays pending instead of sending
$04, so a shell blocked on read waits for OSK input instead of exiting.
|
| |
|
|
|
|
|
| |
Lets scripted runs feed stdin to the serial console AND capture the
resulting screen, e.g. printf 'ls\nexit\n' | gbc rom.gb --headless --shot 0 out.ppm
Previously --shot ran frames with no stdin, so anything waiting on input
(a shell) only showed its first prompt.
|
| |
|
|
|
|
|
|
| |
The unused SM83 opcode $ED now sets gb->poweroff, which breaks the headless
loop so the normal cart_free()->cart_save() path runs. Lets a guest OS shut the
emulator down cleanly (and flush its .sav) instead of relying on an external
timeout/SIGTERM, which could SIGKILL before the save. run_frame() also stops on
poweroff so exit is immediate.
|
| |
|
|
|
|
|
|
| |
When stdin is a terminal, --headless now disables canonical mode + local echo
so keystrokes reach the guest immediately and only the guest echoes (no doubled
chars). Piped/redirected stdin is untouched (scripted input still works). Keep
ISIG so Ctrl-C quits; restore termios on exit. VMIN=1 so a no-data non-blocking
read is EAGAIN, not a 0-byte read the serial layer would misread as EOF.
|
| |
|
|
|
|
| |
read()==0 completes the external-clock transfer with SB=0x04 (EOT), so a gbos
program's readc() sees EOF instead of polling forever. Enables cat/wc-style
tools and lets the shell exit on Ctrl-D / pipe end.
|
| |
|
|
|
|
| |
- external-clock transfers (SC=$80) deliver a stdin byte into SB without
echoing to stdout, and stay pending when no input is available -> programs
can poll for input cleanly. internal-clock (SC=$81) stays pure TX.
|
| |
|
|
|
|
|
|
| |
- --headless: no rendering/TUI; serial TX -> stdout, RX <- stdin (nonblocking)
- generalize serial: serial_out_fd/serial_in_fd on the GB; internal-clock
transfer now shifts in a real RX byte (from stdin) instead of always 0xFF
- paced to native speed by default, --uncapped for turbo; pipes/scripts cleanly
- README: document headless serial console
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Two socket features for building reproducible showcase clips:
- record start <path> [everyN] / record stop: append the RGB888 framebuffer of
each produced frame to a flat 'GBCV' capture file (downsample with everyN).
tools/gbgif.py turns it into a GIF (or PNG frames). Capture is decoupled from
wall-clock/turbo/frameskip, so timing is always correct.
- input play <path> [reset] / input stop: drive the joypad from a synthesized
TAS-style movie (text: '<frames> [buttons...]' per line), frame-locked so
replay is deterministic. 'reset' does an exact power-on (preserve cart
ROM+SRAM, zero all other state) so a movie replays byte-identically -- verified
by hashing two runs. tools/gbmovie.py is a Python builder for movies.
|
| |
|
|
|
| |
In --sixel mode the emulator drew a status line (title/skip/[f]ast/q:quit) under
the image; drop it so sixel output is nothing but the game graphics.
|
| |
|
|
|
|
|
|
|
| |
A client that connects and closes before reading (e.g. the HUD's liveness
probe) made the greeting write() raise SIGPIPE, terminating the whole emulator
- this was behind much of the 'it keeps dying' this session. Also: gbhud now
re-resolves the live instance from the registry on every reconnect (skipping
stale sockets by test-connecting), so the HUD follows respawns, and its box
rendering is ANSI-width-correct.
|
| |
|
|
|
|
|
|
|
| |
The APU was a black hole: reads returned 0xFF, writes were dropped. That hangs
any game that fades music out by decrementing the NR50/rAUDVOL volume register
until it reads 0 (Pokemon's StopMusic spin-wait) - the write never stuck, the
read never hit 0, infinite loop. Store the 48 sound registers and return them
with the standard hardware read-back OR-masks (NR50 mask 0x00 = exact readback).
No synthesis, just correct register behavior.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Replace the input-only FIFO's limitations with a Unix-domain socket control
channel (--sock) that is a superset of the button vocabulary plus emulator
introspection: read/write bus memory, get/set CPU context, single-step, PC
breakpoints, value-change watchpoints, run/pause. Being a socket it replies to
each command and broadcasts async 'event stop ...' lines; a stored last-stop
record (stopinfo) lets per-request clients recover a missed event.
Add gbctl: a stdlib-python CLI that spawns a tmux pane running the emulator on
a ROM, auto-resolves the socket, and drives it with terse verbs (cpu/read/write/
reg/step/break/watch/continue/pause/press/hold/screen/stop). The FIFO stays as
legacy input-only.
|
| |
|
|
|
|
|
| |
Emulation always runs at native rate (correct game speed); --frameskip N draws
only 1 of every N+1 frames so a slow terminal no longer bottlenecks the loop.
Adds --fps cap, --uncapped/--turbo, live [ ] and f keys, HUD readout, and an
exit summary of emulated-vs-drawn fps. Replaces the confusing --render-fps.
|
| |
|
|
|
|
|
| |
Adds --fifo [path] opening a named pipe that accepts button commands from any
process. Supports momentary taps (a, start, name:N), deterministic hold/release
(+name / -name), and 'release'. Merges with keyboard input; held state carries
true key-up semantics the terminal can't provide. GBC_INPUT_DEBUG traces state.
|
| |
|
|
|
|
|
| |
Only cells that changed since the previous frame are redrawn, using cursor
positioning to skip unchanged runs. Static screens write nothing after the
first frame (~108x reduction on acid2). SIGWINCH forces a full repaint.
Verified pixel-exact against the framebuffer (11520/11520 cells, 0 mismatch).
|
| | |
|
| | |
|
| |
|
|
| |
Passes blargg cpu_instrs (all 11), instr_timing, mem_timing.
|
| |
|
