Wave 351/509: Kernel rig: every area and the ladder as a parallel gate on pull requests that touch the kernel (and 3 more)

4 commits, squashed; the originals stay reachable from tag archive/merge-main-develop.

- 6c5537df6 jyviko: Kernel ladder: a gate run fails on any case whose status or counts differ from the golden
- ee2a1ea06 jyviko: CI: maintainers' pull requests run the tests, and the reference tests run on pull requests
- b4fcd5910 jyviko: Kernel rig: every area and the ladder as a parallel gate on pull requests that touch the kernel
- 18d2dddac jyviko: Release gate: every check in full and in parallel before a branch replaces main

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ
This commit is contained in:
jyviko
2026-10-29 09:48:29 -04:00
parent e89b8528cd
commit 524b25b8ba
16 changed files with 401 additions and 24 deletions
+29 -12
View File
@@ -12,8 +12,9 @@
# checks (no archive to upload and download). The GPU path tracer's tests run one at a time
# (.config/nextest.toml), so they have a job of their own in four shards (GPU path tests). A few
# heavy tests (see .config/nextest.toml) are skipped on pull requests and run on pushes to main and
# nightly. The reference tests and the doc tests run in the merge queue, on main and nightly, not
# on pull requests (the fast lane, in Select checks); a break there is fixed forward on main. The
# nightly. The reference tests run on a pull request that selects solidcraft-cli (they replay
# recipes through it). The doc tests run in the merge queue, on main and nightly, not on pull
# requests (the fast lane, in Select checks); a break there is fixed forward on main. The
# Windows jobs (D3D12 shaders and app startup) run nightly, and in the merge queue only for a group
# that changes what they exercise. Only main saves the build caches; pull requests restore them.
# sccache is on a Blacksmith sticky disk that only main commits.
@@ -21,8 +22,9 @@
# commit, or when nothing a build reads changed since the last push to main whose run passed.
# Otherwise it runs everything.
# Heavy checks wait for review: a pull request without the `approved-for-ci` label runs only
# Select checks, the format and repository checks and a quick compile. Adding the label (after
# review) runs everything above, and every later push to the PR does too, except a push that only
# Select checks, the format and repository checks and a quick compile, unless its author is an
# owner, member or collaborator of the repository (they could push the code to a branch here
# anyway). Adding the label (after review) runs everything above, and every later push to the PR does too, except a push that only
# merges main into a change that already passed (scripts/ci/main_merge.py): it carries that
# result and runs the repository checks. Merge groups, the nightly run and Run workflow are not
# pull requests and always run the full suite.
@@ -31,6 +33,8 @@
# `ci` job still runs and reports the commit's real result (see the `ci` job): the merge queue
# reads `ci` from the newest run of a commit, so a run without it would block a PR that passed.
# Cross-compilation is opt-in through Run workflow; release builds remain in release.yml.
# The release gate (release-gate.yml) calls this workflow (workflow_call) as a Run workflow with
# cross-compilation on.
# Every Linux job runs on Blacksmith's Ubuntu 24.04 runners (blacksmith-<N>vcpu-ubuntu-2404), sized
# per job: 32 vCPU for the tests, 16 for Clippy, the GPU path tests, the archive build (nightly) and the merge queue's, 8 for compiles, doc and reference tests, 4 for the light checks, 2 for Select checks and ci.
# The Windows shader job runs on Blacksmith's Windows Server 2025 runners (blacksmith-8vcpu-windows-2025);
@@ -64,6 +68,14 @@ on:
description: "Also check WebAssembly and RISC-V (before a release)"
type: boolean
default: false
# The release gate (release-gate.yml): the event stays the caller's (workflow_dispatch), so a
# call runs exactly what Run workflow runs.
workflow_call:
inputs:
cross_compile:
description: "Also check WebAssembly and RISC-V (before a release)"
type: boolean
default: false
permissions:
contents: read
@@ -137,6 +149,9 @@ jobs:
HEAD_REF: ${{ github.event.pull_request.head.ref }}
REPO: ${{ github.repository }}
APPROVED: ${{ contains(github.event.pull_request.labels.*.name, 'approved-for-ci') }}
# An owner, member or collaborator could push the change to a branch here: their pull
# requests need no label.
TRUSTED: ${{ contains(fromJSON('["OWNER", "MEMBER", "COLLABORATOR"]'), github.event.pull_request.author_association) }}
GH_TOKEN: ${{ github.token }}
run: |
native=false
@@ -230,16 +245,18 @@ jobs:
if [ "$EVENT_NAME" = pull_request ] && [ "$full" != true ]; then
case " $packages " in *" solidcraft-cli "*) ;; *) oracle=false ;; esac
fi
# The pull request fast lane: Windows, the reference tests and the doc tests run on main
# (and the nightly), not on every pull request push; a break there is fixed forward on
# main. Over three days of pull requests the doc tests were the only failing job 3 times
# in about 7,900 runner-minutes, and the reference tests never were.
# The pull request fast lane: Windows and the doc tests run on main (and the nightly),
# not on every pull request push; a break there is fixed forward on main. Over three
# days of pull requests the doc tests were the only failing job 3 times in about 7,900
# runner-minutes. The reference tests run when solidcraft-cli is selected (above): a
# kernel change that broke them went unseen until a full run on develop.
doctests=$native
if [ "$EVENT_NAME" = pull_request ]; then windows=false; oracle=false; doctests=false; fi
# Until a maintainer adds approved-for-ci, a pull request gets a quick compile instead of
# Clippy, the tests and the doc tests.
if [ "$EVENT_NAME" = pull_request ]; then windows=false; doctests=false; fi
# Until a maintainer adds approved-for-ci, a pull request from outside the repository's
# owners, members and collaborators gets a quick compile instead of Clippy, the tests,
# the reference tests and the doc tests.
quick=false
if [ "$EVENT_NAME" = pull_request ] && [ "$APPROVED" != true ]; then
if [ "$EVENT_NAME" = pull_request ] && [ "$APPROVED" != true ] && [ "$TRUSTED" != true ]; then
quick=$native
native=false
oracle=false
+175
View File
@@ -0,0 +1,175 @@
# The kernel rig and the complexity ladder as a regression gate. A pull request that touches the
# kernel surface (the paths below) builds solidcraft-cli once (release, SMLib prebuilt), hands it
# to a matrix of shards as an artifact, and each shard runs one area of tests/kernel-cases/ (or
# the ladder) in parallel, so the wall time is the build plus the slowest shard. See
# tests/kernel-cases/README.md.
#
# - An area fails on a regression (a case that fails and its tests/kernel-cases/<area>/KNOWN_FAILING
# does not list) or on a listed case that passes (it is fixed: remove its line in the same pull
# request, so the list only shrinks).
# - The ladder fails on any case whose status or counts differ from
# tests/kernel-cases/ladder/golden.json; a right change is re-recorded (`--record "why"`).
#
# GitHub filters a workflow by whole files, so any change to crates/doc/src/eval*.rs or
# crates/engine/src/cmd/features*.rs selects it. Pushes to main and develop that touch the same
# paths run it too and commit the compiler cache (a Blacksmith sticky disk) that every other run
# mounts read-only. A merge group (once the queue is on) always runs it: it is the last check
# before the change lands. The release gate (release-gate.yml) calls it (workflow_call).
name: Kernel rig
on:
pull_request:
types: [opened, synchronize, reopened, ready_for_review]
paths:
- "crates/kernel/**"
- "crates/geom/**"
- "crates/smlib/**"
- "crates/step/**"
- "crates/doc/src/eval*.rs"
- "crates/engine/src/cmd/features*.rs"
- "tests/kernel-cases/**"
# The rig itself: the grading, the runner and this workflow.
- "apps/solidcraft-cli/src/kernel_cases.rs"
- "xtask/src/kernel_cases.rs"
- "xtask/src/kernel_ladder.rs"
- "scripts/fetch-smlib.sh"
- ".github/workflows/kernel-rig.yml"
push:
branches: [main, develop]
# The same paths as pull_request.
paths:
- "crates/kernel/**"
- "crates/geom/**"
- "crates/smlib/**"
- "crates/step/**"
- "crates/doc/src/eval*.rs"
- "crates/engine/src/cmd/features*.rs"
- "tests/kernel-cases/**"
# The rig itself: the grading, the runner and this workflow.
- "apps/solidcraft-cli/src/kernel_cases.rs"
- "xtask/src/kernel_cases.rs"
- "xtask/src/kernel_ladder.rs"
- "scripts/fetch-smlib.sh"
- ".github/workflows/kernel-rig.yml"
merge_group:
workflow_call:
workflow_dispatch:
permissions:
contents: read
concurrency:
# A new push to a pull request replaces its old run.
group: kernel-rig-${{ github.event_name }}-${{ github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
env:
CARGO_TERM_COLOR: always
jobs:
build:
name: Build the CLI
# Draft pull requests run no CI, as in ci.yml.
if: github.event_name != 'pull_request' || !github.event.pull_request.draft
runs-on: blacksmith-8vcpu-ubuntu-2404
timeout-minutes: 60
env:
CARGO_INCREMENTAL: "0"
RUSTC_WRAPPER: sccache
SCCACHE_DIR: /home/runner/.cache/sccache
SCCACHE_CACHE_SIZE: 10G
steps:
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
- name: Size parallel compiles to the job's memory
run: scripts/ci/build-jobs.sh
- uses: dtolnay/rust-toolchain@stable
# Only pushes to main and develop commit the cache; pull requests, merge groups and forks
# mount a clone and never write back.
- name: Mount the compiler cache
uses: useblacksmith/stickydisk@3f6be1451e3cad893b778bc93495c80452a5da26 # v1.7.2
continue-on-error: true
with:
key: ${{ github.repository }}-sccache-kernel-rig
path: /home/runner/.cache/sccache
commit: ${{ github.event_name == 'push' && (github.ref == 'refs/heads/main' || github.ref == 'refs/heads/develop') && 'on-change' || 'false' }}
- uses: Swatinem/rust-cache@6323deb102c322ba6fcbdcafc7e3dddab59af2b6 # v2.9.2
with:
key: kernel-rig
save-if: ${{ github.event_name == 'push' && (github.ref == 'refs/heads/main' || github.ref == 'refs/heads/develop') }}
- name: Fetch SMLib (crates/smlib/SMLIB_REV)
shell: bash
run: scripts/fetch-smlib.sh
- name: Set up compiler cache
uses: mozilla-actions/sccache-action@fc920bf0ec8de6ee65d409111f7ec508035751ba # v0.0.11
with:
version: v0.16.0
- name: Build the CLI (release, SMLib) and the rig's runner
run: |
cargo build --release --locked -p solidcraft-cli --features smlib
cargo build --locked -p xtask
# A tar keeps the executable bits, which an artifact drops. xtask finds the repository from
# the path it was built in (CARGO_MANIFEST_DIR), which is the same checkout path in every
# job of this workflow.
- name: Pack the binaries
run: tar -cf "$RUNNER_TEMP/kernel-rig-bin.tar" -C target release/solidcraft-cli debug/xtask
- uses: actions/upload-artifact@cf430e030ddbb5b0abf93d22962f4752f3646cd9 # v7.0.2
with:
name: kernel-rig-bin
path: ${{ runner.temp }}/kernel-rig-bin.tar
retention-days: 1
- name: Compiler cache statistics
if: always()
run: sccache --show-stats
shard:
name: Kernel rig (${{ matrix.shard }})
needs: build
runs-on: blacksmith-4vcpu-ubuntu-2404
timeout-minutes: 30
strategy:
fail-fast: false
matrix:
# Every area of tests/kernel-cases/ and the ladder. On the build box (aarch64, 4 cases at
# a time) the slowest areas took about 1 minute (booleans, 212 s of cases with one of
# 84 s; threads, 161 s) and the ladder, one case at a time, 92 s: no area needs splitting.
shard: [booleans, continuity, crashes, fillets, import, ladder, sheet-metal, shell, sweeps, tessellation, threads]
steps:
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
- uses: actions/download-artifact@9000827ccba6bdab643e8b6fd33ac0654aef8333 # v8.0.2
with:
name: kernel-rig-bin
path: ${{ runner.temp }}
- name: Unpack the binaries
run: mkdir -p target && tar -xf "$RUNNER_TEMP/kernel-rig-bin.tar" -C target
- name: Run the cases against the known-failing list (the ladder against its golden)
env:
SHARD: ${{ matrix.shard }}
run: |
if [ "$SHARD" = ladder ]; then
set -- --ladder --gate
else
set -- --area "$SHARD" --known-failing "tests/kernel-cases/$SHARD/KNOWN_FAILING"
fi
target/debug/xtask kernel-cases "$@" --cli target/release/solidcraft-cli --out "$RUNNER_TEMP/kernel-rig"
- name: Keep the results
if: always()
uses: actions/upload-artifact@cf430e030ddbb5b0abf93d22962f4752f3646cd9 # v7.0.2
with:
name: kernel-rig-results-${{ matrix.shard }}
path: ${{ runner.temp }}/kernel-rig/results.json
if-no-files-found: ignore
# One check for the whole gate, for branch protection and the merge queue.
kernel-rig:
name: kernel-rig
needs: [build, shard]
if: always() && (github.event_name != 'pull_request' || !github.event.pull_request.draft)
runs-on: blacksmith-2vcpu-ubuntu-2404
steps:
- name: Every shard passed
env:
BUILD: ${{ needs.build.result }}
SHARDS: ${{ needs.shard.result }}
run: |
test "$BUILD" = success
test "$SHARDS" = success
+70
View File
@@ -0,0 +1,70 @@
# The release gate: everything, in full and in parallel, on one branch's head, before that branch
# replaces main as a release (develop, rebased onto main). Run it by hand:
# gh workflow run release-gate.yml --ref develop
# It calls the workflows that own each check instead of copying their jobs:
# - ci.yml as Run workflow with cross-compilation: every native suite in full (not scoped), the
# reference tests, the doc tests, the GPU path tests, the action matrix (all 16 shards) and its
# determinism check, the property tests, the repository checks (web-offline among them),
# WebAssembly and RISC-V, Windows D3D12 FXC and Windows app startup;
# - kernel-rig.yml: every rig area against its KNOWN_FAILING and the ladder against its golden;
# - release.yml as a dry run (draft false): the Linux, Flatpak, Snap, FreeBSD, macOS, Windows and
# web packages with their smoke tests and package checks.
# The gate passes only when all three pass. Before main is replaced, tag the old main
# (main-before-develop-<date>) and push the tag.
name: Release gate
on:
workflow_dispatch:
permissions:
contents: read
concurrency:
group: release-gate-${{ github.ref }}
cancel-in-progress: false
jobs:
ci:
name: CI
uses: ./.github/workflows/ci.yml
permissions:
contents: read
# Select checks reads this commit's other runs.
actions: read
with:
cross_compile: true
kernel-rig:
name: Kernel rig
uses: ./.github/workflows/kernel-rig.yml
permissions:
contents: read
packages:
name: Packages
uses: ./.github/workflows/release.yml
# release.yml's last job asks for contents: write (it creates releases); a dry run only keeps
# the files as an artifact.
permissions:
contents: write
actions: read
secrets: inherit
with:
draft: false
release-gate:
name: release-gate
needs: [ci, kernel-rig, packages]
if: always()
runs-on: blacksmith-2vcpu-ubuntu-2404
steps:
- name: Every part passed
env:
CI: ${{ needs.ci.result }}
KERNEL_RIG: ${{ needs.kernel-rig.result }}
PACKAGES: ${{ needs.packages.result }}
run: |
echo "ci=$CI kernel-rig=$KERNEL_RIG packages=$PACKAGES"
test "$CI" = success
test "$KERNEL_RIG" = success
test "$PACKAGES" = success
+13
View File
@@ -20,6 +20,7 @@
# releases/latest/download/<name> links in the README keep working across releases.
#
# A dry run: dispatch with draft=false (the default). Only the release branch may create drafts.
# The release gate (release-gate.yml) calls it as such a dry run.
# A dry run from any branch also test-signs the Windows files when SIGNPATH_API_TOKEN exists:
# gh workflow run release.yml --ref <branch> -f draft=false [-f signpath_signing_policy=test-signing]
name: Release
@@ -40,6 +41,18 @@ on:
description: "SignPath signing policy for the Windows files (empty: the repository variable, else test-signing)"
type: string
default: ""
# The release gate (release-gate.yml): a dry run (draft false) of every package and its checks.
# The event stays the caller's (workflow_dispatch), so a call runs as a dispatched dry run does.
workflow_call:
inputs:
draft:
description: "Create/update a draft (release branch only; off: build artifacts only)"
type: boolean
default: false
signpath_signing_policy:
description: "SignPath signing policy for the Windows files (empty: the repository variable, else test-signing)"
type: string
default: ""
permissions:
contents: read
@@ -0,0 +1,4 @@
# The booleans cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
@@ -0,0 +1,4 @@
# The continuity cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
+4
View File
@@ -0,0 +1,4 @@
# The crashes cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
+4
View File
@@ -0,0 +1,4 @@
# The import cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
@@ -0,0 +1,4 @@
# The sheet-metal cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
+4
View File
@@ -0,0 +1,4 @@
# The shell cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
+4
View File
@@ -0,0 +1,4 @@
# The sweeps cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
@@ -0,0 +1,4 @@
# The tessellation cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
+4
View File
@@ -0,0 +1,4 @@
# The threads cases that fail today, one a line: `<case> <why it fails today>`. Sorted. The kernel
# rig gate (.github/workflows/kernel-rig.yml) fails on a case that fails and is not listed here,
# and on a listed case that passes: a fix removes its line in the same pull request.
# Baseline: not seeded yet; seeded from a develop run after the SMLib pin bump.
+1 -1
View File
@@ -276,7 +276,7 @@ fn write_report(dir: &Path, list: &[PathBuf], results: &[Value]) -> Result<Strin
);
}
let mut md = String::from(
"# Kernel regression rig\n\nGenerated by `cargo xtask kernel-cases`. Every geometry case the old kernel (truck) got wrong, as an issue, a PR or a crash report recorded it, is a use case in [`tests/kernel-cases/`](../tests/kernel-cases/) with its own right answer: numbers from analysis or a hand calculation (**analytic**) where the geometry has them, and otherwise the invariants (**invariants**): the operation succeeds, or refuses cleanly where that is the right answer; each body is one closed shell that passes the record's structure and geometry checks and SMLib's own check; the expected face, edge and vertex counts. Never a kernel's output, and never a commercial kernel's.\n\nThe cases run on SMLib, SolidCraft's kernel. A failing case stays failing until the kernel is fixed; there is no workaround on SolidCraft's side. A failed SMLib call leaves its inputs for replay in the run's output folder (`target/kernel-cases/dumps/<area>-<case>/`). See [`tests/kernel-cases/README.md`](../tests/kernel-cases/README.md) for the format. The fillet cases are also a regression gate: CI's Fillet rig workflow runs them on every pull request that touches fillet code and fails on a case that fails and [`fillets/KNOWN_FAILING`](../tests/kernel-cases/fillets/KNOWN_FAILING) does not list, or a listed case that passes.\n\n",
"# Kernel regression rig\n\nGenerated by `cargo xtask kernel-cases`. Every geometry case the old kernel (truck) got wrong, as an issue, a PR or a crash report recorded it, is a use case in [`tests/kernel-cases/`](../tests/kernel-cases/) with its own right answer: numbers from analysis or a hand calculation (**analytic**) where the geometry has them, and otherwise the invariants (**invariants**): the operation succeeds, or refuses cleanly where that is the right answer; each body is one closed shell that passes the record's structure and geometry checks and SMLib's own check; the expected face, edge and vertex counts. Never a kernel's output, and never a commercial kernel's.\n\nThe cases run on SMLib, SolidCraft's kernel. A failing case stays failing until the kernel is fixed; there is no workaround on SolidCraft's side. A failed SMLib call leaves its inputs for replay in the run's output folder (`target/kernel-cases/dumps/<area>-<case>/`). See [`tests/kernel-cases/README.md`](../tests/kernel-cases/README.md) for the format. Every area is also a regression gate: CI's Kernel rig workflow runs them on every pull request that touches the kernel and fails on a case that fails and its area's `KNOWN_FAILING` (such as [`fillets/KNOWN_FAILING`](../tests/kernel-cases/fillets/KNOWN_FAILING)) does not list, or a listed case that passes.\n\n",
);
md += "## Coverage\n\n| Area | Cases | Results |\n|---|---|---|\n";
for (area, (n, statuses)) in &areas {
+73 -8
View File
@@ -1,4 +1,4 @@
//! `cargo xtask kernel-cases --ladder [--record REASON]`: the kernel complexity ladder.
//! `cargo xtask kernel-cases --ladder [--record REASON | --gate]`: the kernel complexity ladder.
//!
//! `tests/kernel-cases/ladder/ladder.json` lists tiers in order of complexity and, per tier, its
//! cases in order (paths under `tests/kernel-cases/`, so a tier may also list rig cases from other
@@ -9,6 +9,8 @@
//! not shared with other cases. `--record REASON` re-measures the golden and re-sorts each tier
//! by time, fastest first; both files are committed. The run's output folder (`--out`, default
//! `target/kernel-ladder`) holds `results.json` and the SMLib dumps, as the rig's does.
//! `--gate` makes the run a regression gate instead of the report: it fails on any case whose
//! status or counts differ from the golden (times aside, which are the machine's).
//!
//! The report, `docs/kernel-ladder.md`, gives per tier the counts per status and the robustness
//! frontier: the first case, in ladder order, that fails.
@@ -103,17 +105,26 @@ fn golden_entry(r: &Value) -> Value {
/// What changed against the golden: a slowdown, a status change, other counts.
fn flags(r: &Value, g: Option<&Value>) -> Vec<String> {
let Some(g) = g else { return vec!["no golden".into()] };
let mut out = Vec::new();
if g["status"] != r["status"] {
out.push(format!("status was {}", g["status"].as_str().unwrap_or("?")));
}
if let (Some(now), Some(then)) = (r["seconds"].as_f64(), g["seconds"].as_f64())
let mut out = changes(r, g);
if let Some(g) = g
&& let (Some(now), Some(then)) = (r["seconds"].as_f64(), g["seconds"].as_f64())
&& now > then * SLOWDOWN_FACTOR
&& now - then > SLOWDOWN_MIN_S
{
out.push(format!("slower: {now:.2} s (golden {then:.2} s)"));
}
out
}
/// What changed against the golden that the machine does not explain: no golden, a status
/// change, other counts. Times are left out (they depend on the machine the golden was recorded
/// on), so these are what the gate fails on.
fn changes(r: &Value, g: Option<&Value>) -> Vec<String> {
let Some(g) = g else { return vec!["no golden".into()] };
let mut out = Vec::new();
if g["status"] != r["status"] {
out.push(format!("status was {}", g["status"].as_str().unwrap_or("?")));
}
for k in ["body_count", "faces", "edges", "vertices"] {
let (now, then) = (&r["measured"][k], &g["measured"][k]);
if !now.is_null() && !then.is_null() && now != then {
@@ -123,6 +134,35 @@ fn flags(r: &Value, g: Option<&Value>) -> Vec<String> {
out
}
/// The ladder's regression gate (`--gate`): every case's status and counts are the golden's. A
/// case that changed fails it, better or worse, as does a case with no golden: the change is
/// re-recorded with its reason (`--record`) in the same pull request.
fn gate(tiers: &[Tier], results: &BTreeMap<String, Value>, golden: &Value) -> Result<String, String> {
let mut problems = Vec::new();
let mut pass = 0;
for c in tiers.iter().flat_map(|t| &t.cases) {
let Some(r) = results.get(c) else {
problems.push(format!("{c}: no result"));
continue;
};
pass += usize::from(r["status"] == "pass");
let changed = changes(r, golden["cases"].get(c));
if !changed.is_empty() {
problems.push(format!("CHANGED ({}): {c}: {}", r["status"].as_str().unwrap_or("?"), changed.join("; ")));
}
}
let n: usize = tiers.iter().map(|t| t.cases.len()).sum();
if problems.is_empty() {
Ok(format!("kernel ladder gate: {pass}/{n} pass, every case as the golden records it"))
} else {
Err(format!(
"kernel ladder gate: {} case(s) differ from tests/kernel-cases/ladder/golden.json (re-record with --record \"why\" if the change is right):\n {}",
problems.len(),
problems.join("\n ")
))
}
}
fn secs(v: &Value) -> String {
v.as_f64().map(|s| format!("{s:.3}")).unwrap_or_else(|| "–".into())
}
@@ -218,6 +258,10 @@ pub fn run(rest: &[&str]) -> Result<(), String> {
Some(None) => return Err("--record needs a reason: --record \"why the golden changes\"".into()),
Some(Some(r)) => Some(r.to_string()),
};
let gate_run = rest.contains(&"--gate");
if gate_run && reason.is_some() {
return Err("--gate and --record: a gate run compares with the golden, a record run rewrites it".into());
}
let mut o = options(rest, "target/kernel-ladder")?;
// Times are compared with the golden: cases run one at a time unless --jobs says otherwise,
// so they don't share the machine with each other.
@@ -236,7 +280,10 @@ pub fn run(rest: &[&str]) -> Result<(), String> {
save_results(&o, &all)?;
let results: BTreeMap<String, Value> = all.into_iter().map(|r| (rel_of(&r), r)).collect();
let golden = read(&golden_path());
if let Some(reason) = &reason {
if gate_run {
// A gate run grades against the golden; the report stays the last full run's.
println!("{}", gate(&tiers, &results, &golden)?);
} else if let Some(reason) = &reason {
record(reason, &manifest, &tiers, &results, &golden)?;
let tiers = self::tiers(&read(&manifest_path()))?;
println!("{}", write_report(&tiers, &results, &read(&golden_path()))?);
@@ -264,6 +311,24 @@ mod tests {
assert_eq!(flags(&same, None), ["no golden"]);
}
#[test]
fn the_gate_fails_on_any_change_but_time() {
let t = vec![Tier { tier: "01-a".into(), what: String::new(), cases: vec!["ladder/01-a/001-x".into(), "ladder/01-a/002-y".into()] }];
let golden = json!({"cases": {
"ladder/01-a/001-x": {"status": "pass", "seconds": 1.0, "measured": {"faces": 6}},
"ladder/01-a/002-y": {"status": "fail", "seconds": 1.0}
}});
let run = |x: Value, y: Value| gate(&t, &BTreeMap::from([("ladder/01-a/001-x".to_string(), x), ("ladder/01-a/002-y".to_string(), y)]), &golden);
// Ten times slower is the machine, not a change.
assert!(run(json!({"status": "pass", "seconds": 10.0, "measured": {"faces": 6}}), json!({"status": "fail"})).is_ok());
let fixed = run(json!({"status": "pass", "measured": {"faces": 6}}), json!({"status": "pass"})).unwrap_err();
assert!(fixed.contains("CHANGED (pass): ladder/01-a/002-y: status was fail"), "{fixed}");
let counts = run(json!({"status": "pass", "measured": {"faces": 7}}), json!({"status": "fail"})).unwrap_err();
assert!(counts.contains("faces 7 (golden 6)"), "{counts}");
let unrecorded = gate(&t, &BTreeMap::new(), &golden).unwrap_err();
assert!(unrecorded.contains("ladder/01-a/001-x: no result"), "{unrecorded}");
}
#[test]
fn tiers_need_paths() {
let m = json!({"tiers": [{"tier": "01-a", "what": "w", "cases": ["ladder/01-a/001-x"]}]});
+4 -3
View File
@@ -55,12 +55,13 @@ commands:
each failing SMLib call (--cli: a solidcraft-cli built with --features smlib)
--known-failing FILE: a regression gate instead of the report: fails on a case
that fails and FILE does not list, and on a listed case that passes
(tests/kernel-cases/fillets/KNOWN_FAILING, CI's fillet rig job)
--ladder [--record REASON]: the complexity ladder instead
(tests/kernel-cases/<area>/KNOWN_FAILING, CI's Kernel rig workflow)
--ladder [--record REASON | --gate]: the complexity ladder instead
(tests/kernel-cases/ladder/ladder.json, tiers in order), graded and compared
with its timing golden (one case at a time unless --jobs), into
docs/kernel-ladder.md (DIR default target/kernel-ladder); --record re-measures
the golden and re-sorts each tier by time
the golden and re-sorts each tier by time; --gate fails on any case whose status
or counts differ from the golden, instead of the report
action-matrix [--bless] [--results FILE] [--from DIR]
try every ordinary action on every face, edge, vertex and body of the design
corpus and compare with tests/action_matrix/baseline.json (--bless rewrites it;