1
0
Fork 0
mirror of https://github.com/imjasonh/nescript synced 2026-07-08 17:06:04 +00:00
nescript/.github/workflows/ci.yml
Claude 12b54a715e
ci: emulator job must still rebuild ROMs from source
The previous commit wired the emulator job to read the committed
examples/*.nes directly, with the rationale that the `examples`
job's reproducibility diff would catch "stale committed ROMs"
at PR time. That reasoning is wrong and defeats the point of
the emulator harness.

Failure mode: a compiler change lands with correctly-rebuilt
ROMs (passing the examples job) but introduces a rendering
regression that flips a golden. The emulator job would catch
this today because it runs the harness against freshly-compiled
ROMs. With the previous commit's shortcut, the harness would
boot the committed ROMs — which the PR author just rebuilt to
match — so the mismatch wouldn't show up until a second commit
touched anything else. That's exactly the kind of silent-failure
hole the golden harness exists to plug.

Fix: emulator job rebuilds the compiler (toolchain + cache +
`cargo build --release`) and compiles every .ne into the
workspace before running the harness, same as the pre-ROMs-
committed era. The committed ROMs keep their review/demo role
(clone-and-play, diff visibility in PRs) but the test job
always validates the working compiler, not a frozen snapshot.

CLAUDE.md updated to match: the harness runs against whatever
sits in examples/*.nes, so iterating locally still means
rebuilding the ROM(s) you care about first.

https://claude.ai/code/session_01BcCcHi6FUmTh8jC7UgkA3A
2026-04-13 15:21:38 +00:00

162 lines
5.4 KiB
YAML

name: CI
on:
push:
branches: [main]
pull_request:
branches: [main]
env:
CARGO_TERM_COLOR: always
RUSTFLAGS: "-D warnings"
jobs:
check:
name: Check
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: dtolnay/rust-toolchain@stable
- uses: Swatinem/rust-cache@v2
- run: cargo check --all-targets
fmt:
name: Format
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: dtolnay/rust-toolchain@stable
with:
components: rustfmt
- run: cargo fmt --all -- --check
clippy:
name: Clippy
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: dtolnay/rust-toolchain@stable
with:
components: clippy
- uses: Swatinem/rust-cache@v2
- run: cargo clippy --all-targets -- -D warnings
test:
name: Test
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: dtolnay/rust-toolchain@stable
- uses: Swatinem/rust-cache@v2
- run: cargo test --all-targets
examples:
name: Build Examples
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: dtolnay/rust-toolchain@stable
- uses: Swatinem/rust-cache@v2
- run: cargo build --release
- name: Verify committed ROMs match their .ne sources
# Every example has a committed .nes file under examples/. The
# compiler is deterministic (verified: same input → byte-identical
# output), so rebuilding each example into a tmp directory and
# diffing against the committed ROM catches "someone edited a .ne
# file but forgot to regenerate the ROM" at PR time.
run: |
mkdir -p /tmp/reprotest
fail=0
for f in examples/*.ne; do
name=$(basename "$f" .ne)
./target/release/nescript build "$f" --output "/tmp/reprotest/$name.nes"
if ! cmp -s "examples/$name.nes" "/tmp/reprotest/$name.nes"; then
echo "::error file=examples/$name.ne::committed examples/$name.nes is stale; rerun \`cargo run --release -- build examples/$name.ne\` and commit the new ROM"
fail=1
fi
done
if [ $fail -ne 0 ]; then
exit 1
fi
- name: Verify ROMs are valid iNES
run: |
for f in examples/*.nes; do
echo "==> Checking $f"
# iNES magic: first 4 bytes must be "NES\x1a"
header=$(od -A n -t x1 -N 4 "$f" | tr -d ' ')
if [ "$header" != "4e45531a" ]; then
echo "FAIL: $f is not a valid iNES ROM (header: $header)"
exit 1
fi
size=$(stat -c%s "$f")
echo " OK: $size bytes"
done
emulator:
name: Emulator Smoke Test
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: dtolnay/rust-toolchain@stable
- uses: Swatinem/rust-cache@v2
- uses: actions/setup-node@v4
with:
node-version: '22'
- name: Install Chrome runtime deps for puppeteer
# puppeteer ships its own Chrome for Testing binary via the
# package's postinstall hook, but the runner image is missing
# a handful of shared libs that Chrome needs to launch.
run: |
sudo apt-get update
sudo apt-get install -y --no-install-recommends \
libnss3 \
libatk-bridge2.0-0 \
libatk1.0-0 \
libatspi2.0-0 \
libcups2t64 \
libdbus-1-3 \
libdrm2 \
libgbm1 \
libpango-1.0-0 \
libxcomposite1 \
libxdamage1 \
libxfixes3 \
libxkbcommon0 \
libxrandr2 \
libasound2t64
- name: Build compiler
run: cargo build --release
- name: Compile all .ne examples
# Overwrites the committed examples/*.nes in the CI workspace
# (the workspace is ephemeral; the committed ROMs aren't
# touched). This is load-bearing: the goldens must be diffed
# against what *this* compiler produces, not against the ROMs
# an earlier commit happened to ship, otherwise a broken
# compiler change can still pass the emulator job as long as
# the contributor forgot to rebuild the ROMs. The `examples`
# job catches that second case via its reproducibility diff.
run: |
for f in examples/*.ne; do
echo "==> Compiling $f"
./target/release/nescript build "$f"
done
- name: Install emulator harness dependencies
working-directory: tests/emulator
run: npm ci
- name: Diff each ROM's framebuffer against its golden
working-directory: tests/emulator
run: node run_examples.mjs
- name: Upload actual + diff PNGs on failure
if: failure()
uses: actions/upload-artifact@v4
with:
name: emulator-diff
# `actual/` holds the run's output and red-highlighted diff
# PNGs for any ROM that didn't match its golden. `report.json`
# has the per-example pass/fail/diff counts. `goldens/` is
# included too so reviewers can compare the three side by side
# without cloning the repo.
path: |
tests/emulator/actual/
tests/emulator/goldens/
tests/emulator/report.json