6517401f01
Every path-gated workflow begins with one small job -- `does this change touch
it`, or `which workspaces this change touches` for the standards build -- that
all of its real jobs declare in `needs`. All 55 of those gate jobs ran on
`linux-arm64`.
That is the one label the fleet cannot spare. Sampled while 267 runs were
queued, every waiting run had the same shape:
[waiting] linux-arm64 does this change touch it
[blocked] linux member · dataflow device, simulated
[blocked] macos member · dataflow device absent
237 runs, each holding a gate in the longest queue, and every real job blocked
behind it. Not only `linux` jobs: `macos`, `linux-hw`, `linux-arm64-hw` and
`macos-m3-ultra` jobs were blocked too. The measured effect was `linux-arm64`
263 deep against `linux` **zero**, with jetson-hub's 20 cores taking no job for
a full hour -- starved not for want of work but because the work could not be
unblocked. A gate waiting on the resource it is rationing is head-of-line
blocking, and it makes every other runner in the fleet idle on its behalf.
The gates move to `linux`. Nothing else changes: the diff is 55 `runs-on` lines
and no others, and every touched file still parses.
Why `linux` and not elsewhere. The gate is not free -- it compares
`BASE...HEAD`, which needs the merge base, so `fetch-depth: 0` is a full clone
rather than a ten-second script. `macos` is served by hosts on a Mac Studio
sitting at load 185; `linux` is a separate 20-core box that was measurably
empty. This moves a serialisation point to the least contended label; it does
not remove it, and the gate is still a single point of failure for its
workflow. The structural fix is an early-exit first step inside each job, which
costs one checkout per leg and is a larger change.
Follow-up, deliberately not bundled: jetson-hub runs at capacity 2 with 15 GB of
RAM, so 55 clones still serialise two at a time. Raising capacity helps gates
and hurts the heavy emulated-architecture jobs that share the label, so it wants
measuring on its own.
Co-Authored-By: Claude Opus 5 <[email protected]>
204 lines
10 KiB
YAML
204 lines
10 KiB
YAML
name: wai-quantum editors
|
|
|
|
# The engine in editors:
|
|
# - the language service (wai/wai-quantum/src/engine/editor): its unit tests,
|
|
# each module by name, on Linux and macOS;
|
|
# - the language server, `wai-quantum lsp`, driven over standard input and
|
|
# output by a scripted session from initialize to exit (tests/lsp.rs), on
|
|
# Linux and macOS;
|
|
# - the debugger (wai/wai-quantum/src/engine/debug.rs, with the visual edits
|
|
# and the statistical tests beside it), each module by name, and the debug
|
|
# adapter, `wai-quantum dap`, driven by scripted sessions from initialize
|
|
# to disconnect (tests/dap.rs), on Linux and macOS;
|
|
# - `wai-quantum fmt` on every program in the repository: formatting the
|
|
# formatted program changes nothing, and the engine draws the same circuit;
|
|
# - the tree-sitter grammar (wai/tools/tree-sitter-openqasm): regenerated
|
|
# with the pinned command line and identical to the committed parser, its
|
|
# corpus and highlighting tests, and every program in the repository's
|
|
# conformance material and examples parsed with no error;
|
|
# - the editor extension (wai/tools/wai-quantum-editor): built, packaged as a
|
|
# .vsix with the pinned packaging tool, and checked (its manifest against
|
|
# the server, its highlighting, its message when the command line is
|
|
# missing, the other editors' configurations, and the package's files);
|
|
# - the extension for editors with WebAssembly extensions, built for
|
|
# wasm32-wasip2.
|
|
|
|
concurrency:
|
|
group: ${{ github.workflow }}-${{ github.ref }}
|
|
cancel-in-progress: true
|
|
|
|
on:
|
|
push:
|
|
branches: [main]
|
|
paths:
|
|
- 'wai/wai-quantum/**'
|
|
- 'wai/wai-quantum-cli/**'
|
|
- 'wai/tools/tree-sitter-openqasm/**'
|
|
- 'wai/tools/wai-quantum-editor/**'
|
|
- 'wai/quantum-studio/src/examples.js'
|
|
- 'wai/quantum-ops-conformance/**'
|
|
- 'rust-toolchain.toml'
|
|
- '.github/workflows/wai-quantum-editors.yml'
|
|
pull_request:
|
|
paths:
|
|
- 'wai/wai-quantum/**'
|
|
- 'wai/wai-quantum-cli/**'
|
|
- 'wai/tools/tree-sitter-openqasm/**'
|
|
- 'wai/tools/wai-quantum-editor/**'
|
|
- 'wai/quantum-studio/src/examples.js'
|
|
- 'wai/quantum-ops-conformance/**'
|
|
- 'rust-toolchain.toml'
|
|
- '.github/workflows/wai-quantum-editors.yml'
|
|
|
|
jobs:
|
|
relevant:
|
|
name: does this change touch it
|
|
runs-on: linux
|
|
outputs:
|
|
run: ${{ steps.gate.outputs.run }}
|
|
steps:
|
|
- uses: actions/checkout@v5
|
|
with:
|
|
fetch-depth: 0
|
|
- id: gate
|
|
run: |
|
|
set -euo pipefail
|
|
r=$(python3 scripts/pr-touches-paths.py .github/workflows/wai-quantum-editors.yml \
|
|
"${{ github.event.pull_request.base.sha }}" \
|
|
"${{ github.event.pull_request.head.sha }}")
|
|
echo "this change touches the editors' support: $r"
|
|
echo "run=$r" >> "$GITHUB_OUTPUT"
|
|
|
|
editors-linux:
|
|
name: editors · linux
|
|
needs: relevant
|
|
if: needs.relevant.outputs.run == 'true'
|
|
runs-on: linux
|
|
timeout-minutes: 60
|
|
# The grammar's tree-sitter command (from npm, a prebuilt binary) needs glibc
|
|
# 2.39, which the runner's default image (act-22.04, glibc 2.35) lacks; this
|
|
# image has it, as wai-neural-e2e.yml's job does for its libraries.
|
|
container:
|
|
image: catthehacker/ubuntu:act-24.04
|
|
steps:
|
|
- uses: actions/checkout@v5
|
|
- name: Install Rust 1.98.0 with the wasm32-wasip2 target
|
|
shell: bash
|
|
run: |
|
|
export PATH="$HOME/.cargo/bin:$PATH"
|
|
command -v rustup >/dev/null 2>&1 || \
|
|
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --profile minimal --default-toolchain 1.98.0
|
|
rustup toolchain install 1.98.0 --profile minimal
|
|
rustup default 1.98.0
|
|
rustup target add --toolchain 1.98.0 wasm32-wasip2
|
|
echo "$HOME/.cargo/bin" >> "$GITHUB_PATH"
|
|
- uses: actions/setup-node@v4
|
|
with:
|
|
node-version: '22'
|
|
# The filters are module paths, and a filter that matches nothing passes, so
|
|
# each module must show a test that passed.
|
|
- name: The language service's tests
|
|
shell: bash
|
|
working-directory: wai/wai-quantum
|
|
run: |
|
|
set -euo pipefail
|
|
cargo test --locked --features full --lib engine::editor:: 2>&1 | tee "$RUNNER_TEMP/service.log"
|
|
for m in lex index gates format tests; do
|
|
grep -qE "^test engine::editor::$m::.* \.\.\. ok$" "$RUNNER_TEMP/service.log" || { echo "::error::no test of engine::editor::$m ran"; exit 1; }
|
|
done
|
|
grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$RUNNER_TEMP/service.log"
|
|
- name: The language server over standard input and output, and the verbs' parsing
|
|
shell: bash
|
|
working-directory: wai/wai-quantum-cli
|
|
run: |
|
|
set -euo pipefail
|
|
cargo test --locked --release --test lsp 2>&1 | tee "$RUNNER_TEMP/lsp.log"
|
|
grep -qE '^test result: ok\. 2 passed; 0 failed' "$RUNNER_TEMP/lsp.log" || { echo "::error::the protocol session did not pass"; exit 1; }
|
|
cargo test --locked --release --bin wai-quantum -- lsp:: the_command_line_parses_as_documented 2>&1 | tee "$RUNNER_TEMP/unit.log"
|
|
grep -qE '^test result: ok\. 4 passed; 0 failed' "$RUNNER_TEMP/unit.log" || { echo "::error::the server's unit tests did not all run"; exit 1; }
|
|
cargo build --locked --release
|
|
- name: The debugger's tests, and the debug adapter over standard input and output
|
|
shell: bash
|
|
run: |
|
|
set -euo pipefail
|
|
cd wai/wai-quantum
|
|
cargo test --locked --features full --lib -- engine::debug:: engine::visual:: engine::stats:: 2>&1 | tee "$RUNNER_TEMP/debug.log"
|
|
for m in debug visual stats; do
|
|
grep -qE "^test engine::$m::.* \.\.\. ok$" "$RUNNER_TEMP/debug.log" || { echo "::error::no test of engine::$m ran"; exit 1; }
|
|
done
|
|
grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$RUNNER_TEMP/debug.log"
|
|
cd ../wai-quantum-cli
|
|
cargo test --locked --release --test dap 2>&1 | tee "$RUNNER_TEMP/dap.log"
|
|
grep -qE '^test result: ok\. 2 passed; 0 failed' "$RUNNER_TEMP/dap.log" || { echo "::error::the debug adapter's sessions did not pass"; exit 1; }
|
|
- name: The grammar, regenerated as committed; its corpus, its highlighting, and every program in the repository
|
|
shell: bash
|
|
working-directory: wai/tools/tree-sitter-openqasm
|
|
run: |
|
|
set -euo pipefail
|
|
npm ci --no-audit --no-fund
|
|
npx tree-sitter generate --abi 14
|
|
git diff --exit-code -- src || { echo "::error::src/ is not what grammar.js generates: regenerate and commit it"; exit 1; }
|
|
npx tree-sitter test
|
|
node script/check-programs.mjs --out "$RUNNER_TEMP/programs"
|
|
- name: wai-quantum fmt on every program in the repository
|
|
shell: bash
|
|
run: |
|
|
set -euo pipefail
|
|
w="$PWD/wai/wai-quantum-cli/target/release/wai-quantum"
|
|
n=0
|
|
for f in "$RUNNER_TEMP"/programs/*.qasm; do
|
|
"$w" fmt "$f" > "$RUNNER_TEMP/once.qasm"
|
|
"$w" fmt "$RUNNER_TEMP/once.qasm" > "$RUNNER_TEMP/twice.qasm"
|
|
cmp -s "$RUNNER_TEMP/once.qasm" "$RUNNER_TEMP/twice.qasm" || { echo "::error::formatting $f again changed it"; diff "$RUNNER_TEMP/once.qasm" "$RUNNER_TEMP/twice.qasm" || true; exit 1; }
|
|
# The engine draws the formatted program as it draws the program, or refuses
|
|
# both (a refusal names a line, which formatting may move).
|
|
args=(); grep -qE '^\s*input\s+u?int' "$f" && args=(--input n=5)
|
|
a_out=$("$w" "${args[@]}" draw < "$f" 2>/dev/null) && a=0 || a=$?
|
|
b_out=$("$w" "${args[@]}" draw < "$RUNNER_TEMP/once.qasm" 2>/dev/null) && b=0 || b=$?
|
|
[ "$a" = "$b" ] || { echo "::error::the engine reads $f (status $a) and its formatted form (status $b) differently"; exit 1; }
|
|
[ "$a" != 0 ] || [ "$a_out" = "$b_out" ] || { echo "::error::the engine draws the formatted $f differently"; diff <(echo "$a_out") <(echo "$b_out") || true; exit 1; }
|
|
n=$((n + 1))
|
|
done
|
|
[ "$n" -ge 20 ] || { echo "::error::only $n programs were formatted"; exit 1; }
|
|
echo "fmt: $n programs, each stable, each drawn the same or refused alike"
|
|
- name: The editor extension, packaged and checked
|
|
shell: bash
|
|
working-directory: wai/tools/wai-quantum-editor
|
|
run: |
|
|
set -euo pipefail
|
|
npm ci --no-audit --no-fund
|
|
npm run package < /dev/null
|
|
WAI_QUANTUM_BIN="$GITHUB_WORKSPACE/wai/wai-quantum-cli/target/release/wai-quantum" node test/check.mjs --vsix wai-quantum.vsix
|
|
- name: The extension for editors with WebAssembly extensions, for wasm32-wasip2
|
|
shell: bash
|
|
working-directory: wai/tools/wai-quantum-editor/editors/extension
|
|
run: cargo build --locked --release --target wasm32-wasip2
|
|
|
|
editors-macos:
|
|
name: editors · macos
|
|
needs: relevant
|
|
if: needs.relevant.outputs.run == 'true'
|
|
runs-on: macos
|
|
timeout-minutes: 60
|
|
steps:
|
|
- uses: actions/checkout@v5
|
|
# launchd's PATH omits cargo on the macOS host runner.
|
|
- name: Put cargo on PATH
|
|
shell: bash
|
|
run: echo "$HOME/.cargo/bin" >> "$GITHUB_PATH"
|
|
- name: The language service's tests
|
|
shell: bash
|
|
working-directory: wai/wai-quantum
|
|
run: |
|
|
set -euo pipefail
|
|
nice -n 10 cargo test --locked --features full --lib engine::editor:: 2>&1 | tee "$RUNNER_TEMP/service.log"
|
|
grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$RUNNER_TEMP/service.log"
|
|
- name: The language server over standard input and output
|
|
shell: bash
|
|
working-directory: wai/wai-quantum-cli
|
|
run: |
|
|
set -euo pipefail
|
|
nice -n 10 cargo test --locked --release --test lsp 2>&1 | tee "$RUNNER_TEMP/lsp.log"
|
|
grep -qE '^test result: ok\. 2 passed; 0 failed' "$RUNNER_TEMP/lsp.log" || { echo "::error::the protocol session did not pass"; exit 1; }
|
|
nice -n 10 cargo test --locked --release --test dap 2>&1 | tee "$RUNNER_TEMP/dap.log"
|
|
grep -qE '^test result: ok\. 2 passed; 0 failed' "$RUNNER_TEMP/dap.log" || { echo "::error::the debug adapter's sessions did not pass"; exit 1; }
|