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]>
149 lines
7 KiB
YAML
149 lines
7 KiB
YAML
name: WAI operation catalogue
|
|
|
|
# WAI's toolchain has one catalogue of operations, `noun.verb` (wai/wai-rs/src/catalogue.rs),
|
|
# and every door names them by it. This workflow fails on drift:
|
|
#
|
|
# - the export, wai/catalogue/ops.json, is the table (`wai ops export --check`);
|
|
# - the table keeps its naming rules (the library's catalogue tests);
|
|
# - the command line reaches operations through the catalogue alone, and every operation
|
|
# keeps the agent rules: its help, the exit codes, the JSON envelope, --dry-run on every
|
|
# operation that is not read-only, --yes (tests/cli_ops.rs, and the binary's own tests);
|
|
# - the other doors name operations as the catalogue does: MCP tool names, the services'
|
|
# OpenAPI operationIds, the C header, and the Swift and Kotlin sinks
|
|
# (scripts/check-wai-ops-doors.py, with its selftest).
|
|
#
|
|
# Push filters match the push's delta, so every path a door or a test reads is listed.
|
|
|
|
concurrency:
|
|
group: ${{ github.workflow }}-${{ github.ref }}
|
|
cancel-in-progress: true
|
|
|
|
on:
|
|
push:
|
|
branches: [main]
|
|
paths:
|
|
- 'wai/wai-rs/**'
|
|
- 'wai/catalogue/**'
|
|
- 'wai/examples/**'
|
|
- 'wai/http-conformance/**'
|
|
- 'wai/score-conformance/**'
|
|
- 'wai/live-conformance/fleet/**'
|
|
# tests/cli_ops.rs builds the revenue share's --dry-run sample from this corpus.
|
|
- 'wai/live-conformance/share/**'
|
|
- 'wai/live-conformance/channels/**'
|
|
- 'wai/REGISTRY.md'
|
|
- 'wai/LICENSE'
|
|
- 'wai/wai-quantum/**'
|
|
- 'wai/wai-ops/**'
|
|
- 'wai/wai-app/**'
|
|
- 'wai/wai-sink-swift/**'
|
|
- 'wai/wai-sink-android/**'
|
|
- 'joule-code/joule-code-rs/crates/*/openapi.json'
|
|
- 'scripts/check-wai-ops-doors.py'
|
|
- 'scripts/check-wai-ops-doors-selftest.sh'
|
|
- 'rust-toolchain.toml'
|
|
- '.github/workflows/wai-catalogue.yml'
|
|
pull_request:
|
|
paths:
|
|
- 'wai/wai-rs/**'
|
|
- 'wai/catalogue/**'
|
|
- 'wai/examples/**'
|
|
- 'wai/http-conformance/**'
|
|
- 'wai/score-conformance/**'
|
|
- 'wai/live-conformance/fleet/**'
|
|
# tests/cli_ops.rs builds the revenue share's --dry-run sample from this corpus.
|
|
- 'wai/live-conformance/share/**'
|
|
- 'wai/live-conformance/channels/**'
|
|
- 'wai/REGISTRY.md'
|
|
- 'wai/LICENSE'
|
|
- 'wai/wai-quantum/**'
|
|
- 'wai/wai-ops/**'
|
|
- 'wai/wai-app/**'
|
|
- 'wai/wai-sink-swift/**'
|
|
- 'wai/wai-sink-android/**'
|
|
- 'joule-code/joule-code-rs/crates/*/openapi.json'
|
|
- 'scripts/check-wai-ops-doors.py'
|
|
- 'scripts/check-wai-ops-doors-selftest.sh'
|
|
- 'rust-toolchain.toml'
|
|
- '.github/workflows/wai-catalogue.yml'
|
|
|
|
jobs:
|
|
# The path filters gate the trigger; this job gates the work from the pull request's own
|
|
# diff (scripts/pr-touches-paths.py, reading the list above), as wai-pack.yml explains.
|
|
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-catalogue.yml \
|
|
"${{ github.event.pull_request.base.sha }}" \
|
|
"${{ github.event.pull_request.head.sha }}")
|
|
echo "this change touches the operation catalogue: $r"
|
|
echo "run=$r" >> "$GITHUB_OUTPUT"
|
|
|
|
catalogue:
|
|
name: the catalogue and every door · linux
|
|
needs: relevant
|
|
if: needs.relevant.outputs.run == 'true'
|
|
runs-on: linux
|
|
timeout-minutes: 60
|
|
steps:
|
|
- uses: actions/checkout@v5
|
|
- name: Install Rust 1.98.0
|
|
shell: bash
|
|
run: |
|
|
export PATH="$HOME/.cargo/bin:$PATH"
|
|
command -v rustup >/dev/null 2>&1 || \
|
|
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --profile minimal --default-toolchain 1.98.0
|
|
rustup toolchain install 1.98.0 --profile minimal
|
|
rustup default 1.98.0
|
|
SUDO=""; [ "$(id -u)" -ne 0 ] && SUDO="sudo"
|
|
command -v python3 >/dev/null || { $SUDO apt-get update -qq; $SUDO apt-get install -y -qq python3; }
|
|
echo "$HOME/.cargo/bin" >> "$GITHUB_PATH"
|
|
# The export other tools and this workflow's door check read is the table, byte for
|
|
# byte; the message says how to regenerate it.
|
|
- name: The catalogue's export is up to date
|
|
working-directory: wai/wai-rs
|
|
run: cargo run -q --locked --no-default-features --bin wai -- ops export --check ../catalogue/ops.json
|
|
- name: Every door names its operations as the catalogue does
|
|
shell: bash
|
|
run: |
|
|
set -euo pipefail
|
|
bash scripts/check-wai-ops-doors-selftest.sh
|
|
python3 -I scripts/check-wai-ops-doors.py
|
|
# The library's catalogue tests: the naming rules, each broken in a copy of the table
|
|
# and refused. The binary's own tests: every implementation names an operation of the
|
|
# catalogue and every operation with a command line has one; help is the catalogue's.
|
|
# tests/cli_ops.rs: wai ops, help, the exit codes, the envelope, --dry-run and --yes (and
|
|
# --confirm, its deprecated spelling),
|
|
# first in the archive's build (quickstart), then with every command-line feature that
|
|
# needs no system library, so --dry-run is held on generate, score and stage too, with
|
|
# Create's (app), so it is held on wai app export, with the revenue share's
|
|
# (revenue_share), so it is held on wai contract create|accept and wai statement create, and
|
|
# with the opportunity report's (opportunity_report), so it is held on wai opportunity
|
|
# create and wai statement check.
|
|
- name: The catalogue's rules, and the command line through it
|
|
shell: bash
|
|
working-directory: wai/wai-rs
|
|
run: |
|
|
set -euo pipefail
|
|
ran() { grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$1" || { echo "::error::no test ran in $1"; exit 1; }; }
|
|
cargo test --locked --no-default-features --lib catalogue 2>&1 | tee "$RUNNER_TEMP/lib.log"
|
|
ran "$RUNNER_TEMP/lib.log"
|
|
cargo test --locked --no-default-features --features quickstart --bin wai --test cli_ops 2>&1 | tee "$RUNNER_TEMP/cli.log"
|
|
ran "$RUNNER_TEMP/cli.log"
|
|
cargo test --locked --no-default-features --features quickstart,generate,score,staged,video_int,audio_int,splat_int,exact_upscale,exact_tonemap --test cli_ops 2>&1 | tee "$RUNNER_TEMP/cli-full.log"
|
|
ran "$RUNNER_TEMP/cli-full.log"
|
|
cargo test --locked --no-default-features --features quickstart,app --test cli_ops 2>&1 | tee "$RUNNER_TEMP/cli-app.log"
|
|
ran "$RUNNER_TEMP/cli-app.log"
|
|
cargo test --locked --no-default-features --features quickstart,revenue_share --test cli_ops 2>&1 | tee "$RUNNER_TEMP/cli-share.log"
|
|
ran "$RUNNER_TEMP/cli-share.log"
|
|
cargo test --locked --no-default-features --features quickstart,opportunity_report --test cli_ops 2>&1 | tee "$RUNNER_TEMP/cli-opportunity.log"
|
|
ran "$RUNNER_TEMP/cli-opportunity.log"
|