fix(3597): clear residual Windows test failures + add ratchet lint guard
Two more diagnostic passes (clusters J: residuals in already-touched files,
K: 12 untouched files) plus a production-code path fix and a new
ratchet-style lint guard.
## Test-only fixes (14 files)
bug-1736, bug-2248, bug-2698 — replace inline 1s-budget rmSync with the
shared cleanup() helper (5s budget, 20×250ms retries). The earlier
inline maxRetries:10 / retryDelay:100 wasn't enough to absorb Windows
Defender's deferred-scan handle hold on cold runners.
bug-2256, skill-manifest — also override USERPROFILE alongside HOME in
beforeEach/runGsdTools calls. os.homedir() reads USERPROFILE on win32,
so HOME-only stubs leak the runner's real home into the SUT.
bug-2784, bug-3608, enh-2500, enh-2790, few-shot-calibration,
gsd-settings-advanced — CRLF tolerance: literal \n in regexes against
file content (frontmatter anchors, bash-fence regex, multi-line
numbered-list captures, awk-block extractors) becomes \r?\n; split('\n')
becomes split(/\r?\n/). Windows checkout with autocrlf=true puts \r
before every \n; .+ doesn't match \r in JS regex by default.
bug-2966 — three-part fix to extractStepRun (CRLF split), awk regex
(\r?\n), and conflict-marker parser (rawLine + \r$ strip).
bug-2969, config — normalize separators on test assertions where the
SUT correctly emits \ on win32 but the test compares against /.
prompt-injection-scan — normalize relPath via replace(/\\/g, '/') before
ALLOWLIST.has() lookup. ALLOWLIST keys are POSIX; path.relative returns
backslashes on win32 → falsely scans allowlisted security module → trips
the boundary-tag detector on its own legitimate detection code.
prune-orphaned-worktrees — use the existing canonicalPath +
listedWorktreePaths(repoDir).has(...) helpers instead of substring
matching the raw path. git stores long-form canonical paths
(runneradmin), but mkdtempSync returns 8.3 short-form (RUNNER~1) on
Windows runners; plain string compare misses every entry.
## Production-code fix (1 file)
get-shit-done/bin/lib/init.cjs — bug-3491 in_nested_subdir computation
canonicalizes both worktreeRoot and cwd via fs.realpathSync.native +
path.relative before declaring "nested." Windows runner cwd (8.3 short
name) vs git's --show-toplevel (long form, forward slashes) made the
raw string compare always say true even at the worktree root, breaking
the "init new-project at worktree root" subtest.
## New ratchet lint guard
tests/windows-test-parity-guard.test.cjs — scans tests/ for 7
anti-patterns that drove the Windows failure clusters. Each rule has a
baseline count snapshot from this PR; the test fails when a new
occurrence appears (count grows above baseline), ratcheting down as
existing offenders are fixed. Patterns covered:
G1 split('\n') after readFileSync (use /\r?\n/)
G2 ```bash\n fence regex (use ```bash\r?\n)
G3 ^---\n frontmatter anchor (use ^---\r?\n)
G4 hardcoded "/tmp/..." literal passed to fs.* (use os.tmpdir())
G5 bare 'npm' to execFileSync without {shell:true} on win32
G6 process.env.HOME stub with no USERPROFILE
G7 fs.rmSync({recursive,force}) without maxRetries
Future Windows-parity regressions get caught at PR time rather than
five iterations into a CI loop.
Validated: holodeck (ubuntu docker) 11232/0 pass (count +8 = the 7
new ratchet tests + parent describe).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Tom Boucher
@@ -463,7 +463,22 @@ function cmdInitNewProject(cwd, raw) {
|
||||
...(() => {
|
||||
const info = gitWorktreeInfoInternal(cwd);
|
||||
const worktreeRoot = info.worktreeRoot;
|
||||
const inNestedSubdir = info.inside && worktreeRoot !== null && worktreeRoot !== cwd;
|
||||
// Canonicalize both sides before comparing: on Windows the runner's
|
||||
// cwd may be the 8.3 short-name form (RUNNER~1) while git's
|
||||
// --show-toplevel emits the long-form path with forward slashes.
|
||||
// Without canonicalization, in_nested_subdir is `true` even at the
|
||||
// worktree root (bug #3491). realpathSync.native handles 8.3→long
|
||||
// expansion; path.resolve normalizes separators. Wrap in try so
|
||||
// a missing path falls back to the original string compare.
|
||||
let inNestedSubdir = info.inside && worktreeRoot !== null && worktreeRoot !== cwd;
|
||||
if (inNestedSubdir) {
|
||||
try {
|
||||
const canonRoot = fs.realpathSync.native(worktreeRoot);
|
||||
const canonCwd = fs.realpathSync.native(cwd);
|
||||
const rel = path.relative(canonRoot, canonCwd);
|
||||
inNestedSubdir = rel !== '' && !rel.startsWith('..');
|
||||
} catch { /* keep raw-string compare result */ }
|
||||
}
|
||||
return {
|
||||
has_git: info.inside,
|
||||
git_worktree_root: worktreeRoot,
|
||||
@@ -607,7 +622,22 @@ function cmdInitIngestDocs(cwd, raw) {
|
||||
// Bug #3491 — see cmdInitNewProject above. Same shallow-check bug.
|
||||
const info = gitWorktreeInfoInternal(cwd);
|
||||
const worktreeRoot = info.worktreeRoot;
|
||||
const inNestedSubdir = info.inside && worktreeRoot !== null && worktreeRoot !== cwd;
|
||||
// Canonicalize both sides before comparing: on Windows the runner's
|
||||
// cwd may be the 8.3 short-name form (RUNNER~1) while git's
|
||||
// --show-toplevel emits the long-form path with forward slashes.
|
||||
// Without canonicalization, in_nested_subdir is `true` even at the
|
||||
// worktree root (bug #3491). realpathSync.native handles 8.3→long
|
||||
// expansion; path.resolve normalizes separators. Wrap in try so
|
||||
// a missing path falls back to the original string compare.
|
||||
let inNestedSubdir = info.inside && worktreeRoot !== null && worktreeRoot !== cwd;
|
||||
if (inNestedSubdir) {
|
||||
try {
|
||||
const canonRoot = fs.realpathSync.native(worktreeRoot);
|
||||
const canonCwd = fs.realpathSync.native(cwd);
|
||||
const rel = path.relative(canonRoot, canonCwd);
|
||||
inNestedSubdir = rel !== '' && !rel.startsWith('..');
|
||||
} catch { /* keep raw-string compare result */ }
|
||||
}
|
||||
return {
|
||||
has_git: info.inside,
|
||||
git_worktree_root: worktreeRoot,
|
||||
|
||||
@@ -22,6 +22,7 @@ const { execFileSync } = require('child_process');
|
||||
const INSTALL_SRC = path.join(__dirname, '..', 'bin', 'install.js');
|
||||
const BUILD_SCRIPT = path.join(__dirname, '..', 'scripts', 'build-hooks.js');
|
||||
const { install, copyCommandsAsClaudeSkills } = require(INSTALL_SRC);
|
||||
const { cleanup } = require('./helpers.cjs');
|
||||
|
||||
// ─── Ensure hooks/dist/ is populated before install tests ────────────────────
|
||||
// With --test-concurrency=4, other install tests (bug-1834, bug-1924) run
|
||||
@@ -46,7 +47,9 @@ describe('#1736: local Claude install populates .claude/commands/gsd/', () => {
|
||||
});
|
||||
|
||||
afterEach(() => {
|
||||
fs.rmSync(tmpDir, { recursive: true, force: true, maxRetries: 10, retryDelay: 100 });
|
||||
// Use the shared helper which has a 5s Windows-EBUSY retry budget
|
||||
// (20×250ms). The inline 1s budget here was insufficient on cold runners.
|
||||
cleanup(tmpDir);
|
||||
});
|
||||
|
||||
test('local install creates .claude/commands/gsd/ directory', (t) => {
|
||||
|
||||
@@ -28,6 +28,7 @@ const { execFileSync } = require('child_process');
|
||||
const INSTALL_SRC = path.join(__dirname, '..', 'bin', 'install.js');
|
||||
const BUILD_SCRIPT = path.join(__dirname, '..', 'scripts', 'build-hooks.js');
|
||||
const { install, finishInstall } = require(INSTALL_SRC);
|
||||
const { cleanup } = require('./helpers.cjs');
|
||||
|
||||
// ─── Ensure hooks/dist/ is populated before install tests ────────────────────
|
||||
before(() => {
|
||||
@@ -47,7 +48,8 @@ describe('#2248: local Claude install does not clobber profile-level statusLine'
|
||||
});
|
||||
|
||||
afterEach(() => {
|
||||
fs.rmSync(tmpDir, { recursive: true, force: true, maxRetries: 10, retryDelay: 100 });
|
||||
// Use the shared 5s Windows-EBUSY retry budget instead of inline 1s.
|
||||
cleanup(tmpDir);
|
||||
});
|
||||
|
||||
test('local install does not write statusLine to .claude/settings.json', (t) => {
|
||||
|
||||
@@ -19,6 +19,8 @@ const fs = require('fs');
|
||||
const path = require('path');
|
||||
const os = require('os');
|
||||
|
||||
const isWindows = process.platform === 'win32';
|
||||
|
||||
const {
|
||||
readGsdEffectiveModelOverrides,
|
||||
generateCodexAgentToml,
|
||||
@@ -43,17 +45,27 @@ describe('bug #2256 — readGsdEffectiveModelOverrides', () => {
|
||||
let projectDir;
|
||||
let homeDir;
|
||||
let origHome;
|
||||
let origUserProfile;
|
||||
|
||||
beforeEach(() => {
|
||||
projectDir = makeTmp('proj');
|
||||
homeDir = makeTmp('home');
|
||||
origHome = process.env.HOME;
|
||||
// On Windows, os.homedir() reads USERPROFILE (not HOME). Tests that
|
||||
// need to redirect ~ must override both — otherwise the SUT reads
|
||||
// the real user's home and the fixture is invisible.
|
||||
origUserProfile = process.env.USERPROFILE;
|
||||
process.env.HOME = homeDir;
|
||||
if (isWindows) process.env.USERPROFILE = homeDir;
|
||||
});
|
||||
|
||||
afterEach(() => {
|
||||
if (origHome === undefined) delete process.env.HOME;
|
||||
else process.env.HOME = origHome;
|
||||
if (isWindows) {
|
||||
if (origUserProfile === undefined) delete process.env.USERPROFILE;
|
||||
else process.env.USERPROFILE = origUserProfile;
|
||||
}
|
||||
rmr(projectDir);
|
||||
rmr(homeDir);
|
||||
});
|
||||
|
||||
@@ -43,6 +43,7 @@ const { execFileSync } = require('child_process');
|
||||
const INSTALL_SRC = path.join(__dirname, '..', 'bin', 'install.js');
|
||||
const BUILD_SCRIPT = path.join(__dirname, '..', 'scripts', 'build-hooks.js');
|
||||
const { install, GSD_CODEX_MARKER } = require(INSTALL_SRC);
|
||||
const { cleanup } = require('./helpers.cjs');
|
||||
|
||||
// Ensure hooks/dist/ is populated before install tests
|
||||
before(() => {
|
||||
@@ -60,7 +61,8 @@ describe('#2698: CRLF stale gsd-update-check block is removed on Codex reinstall
|
||||
});
|
||||
|
||||
afterEach(() => {
|
||||
fs.rmSync(tmpDir, { recursive: true, force: true, maxRetries: 10, retryDelay: 100 });
|
||||
// Use the shared 5s Windows-EBUSY retry budget instead of inline 1s.
|
||||
cleanup(tmpDir);
|
||||
});
|
||||
|
||||
// Helper: pre-populate .codex/config.toml with a GSD marker + stale hooks block
|
||||
|
||||
@@ -61,10 +61,10 @@ describe('bug-2784: update.md cache-clear covers shared cache path', () => {
|
||||
const stepContent = stepMatch[0];
|
||||
|
||||
const bashLines = [];
|
||||
const fenceRe = /```(?:bash|sh)\n([\s\S]*?)```/g;
|
||||
const fenceRe = /```(?:bash|sh)\r?\n([\s\S]*?)```/g;
|
||||
let m;
|
||||
while ((m = fenceRe.exec(stepContent)) !== null) {
|
||||
for (const line of m[1].split('\n')) {
|
||||
for (const line of m[1].split(/\r?\n/)) {
|
||||
const trimmed = line.trim();
|
||||
if (trimmed) bashLines.push(trimmed);
|
||||
}
|
||||
|
||||
@@ -58,7 +58,9 @@ const WORKFLOW_PATH = path.join(__dirname, '..', '.github', 'workflows', 'releas
|
||||
* one for a single test isn't justified.
|
||||
*/
|
||||
function extractStepRun(workflowText, stepName) {
|
||||
const lines = workflowText.split('\n');
|
||||
// CRLF-tolerant split: Windows checkout (autocrlf=true) leaves trailing
|
||||
// \r on every line which downstream regex anchors don't tolerate.
|
||||
const lines = workflowText.split(/\r?\n/);
|
||||
for (let i = 0; i < lines.length; i++) {
|
||||
const m = lines[i].match(/^(\s*)- name:\s*(.+?)\s*$/);
|
||||
if (!m || m[2] !== stepName) continue;
|
||||
@@ -197,7 +199,9 @@ describe('bug-2966: release-sdk hotfix cherry-pick classifies context-missing vs
|
||||
const blocks = [];
|
||||
let inHead = false;
|
||||
let head = '';
|
||||
for (const line of conflicted.split('\n')) {
|
||||
for (const rawLine of conflicted.split(/\r?\n/)) {
|
||||
// Strip residual \r so /^=======$/ matches even on CRLF content.
|
||||
const line = rawLine.replace(/\r$/, '');
|
||||
if (/^<<<<<<< /.test(line)) { inHead = true; head = ''; continue; }
|
||||
if (/^=======$/.test(line) && inHead) { inHead = false; continue; }
|
||||
if (/^>>>>>>> /.test(line)) { blocks.push(head); head = ''; continue; }
|
||||
@@ -217,7 +221,7 @@ describe('bug-2966: release-sdk hotfix cherry-pick classifies context-missing vs
|
||||
// exercises the exact predicate that runs in CI — not a copy.
|
||||
const yaml = fs.readFileSync(WORKFLOW_PATH, 'utf8');
|
||||
const script = extractStepRun(yaml, 'Prepare hotfix branch');
|
||||
const awkMatch = script.match(/awk '\n([\s\S]+?)' "\$CONFLICTED"/);
|
||||
const awkMatch = script.match(/awk '\r?\n([\s\S]+?)' "\$CONFLICTED"/);
|
||||
assert.ok(awkMatch, 'expected to find the conflict-classifying awk script in the workflow');
|
||||
const awkProgram = awkMatch[1];
|
||||
|
||||
|
||||
@@ -128,7 +128,8 @@ describe('Bug #2969: deterministic Step 5 verification gate', () => {
|
||||
assert.equal(status, 1);
|
||||
assert.equal(report.failures, 1);
|
||||
const r0 = report.results[0];
|
||||
assert.equal(r0.file, 'skills/discuss-phase/SKILL.md');
|
||||
// Normalize separators: on Windows the SUT emits 'skills\discuss-phase\SKILL.md'.
|
||||
assert.equal(r0.file.replace(/\\/g, '/'), 'skills/discuss-phase/SKILL.md');
|
||||
assert.equal(r0.status, 'fail');
|
||||
assert.equal(r0.reason, REASON.FAIL_USER_LINES_MISSING);
|
||||
assert.ok(
|
||||
|
||||
@@ -89,7 +89,7 @@ describe('bug #3608: update.md models Antigravity as a first-class runtime', ()
|
||||
// Extract the inference block — the if/elif ladder that maps env vars to runtime.
|
||||
// Match from the comment marker through the closing `fi` of the inference block.
|
||||
const blockMatch = content.match(
|
||||
/If runtime is still unknown, infer from runtime env vars[\s\S]*?\nfi\n/,
|
||||
/If runtime is still unknown, infer from runtime env vars[\s\S]*?\r?\nfi\r?\n/,
|
||||
);
|
||||
assert.ok(blockMatch, 'env-var inference block not found');
|
||||
|
||||
|
||||
@@ -953,14 +953,15 @@ describe('config-path command (#2282)', () => {
|
||||
test('returns root config path when no workstream is active', () => {
|
||||
const result = runGsdTools('config-path', tmpDir);
|
||||
assert.ok(result.success, `config-path failed: ${result.error}`);
|
||||
assert.ok(result.output.trim().endsWith('.planning/config.json'), `expected root config path, got: ${result.output}`);
|
||||
// Normalize separators: Windows emits backslashes in the resolved path.
|
||||
assert.ok(result.output.trim().replace(/\\/g, '/').endsWith('.planning/config.json'), `expected root config path, got: ${result.output}`);
|
||||
assert.ok(!result.output.includes('workstreams'), 'should not include workstreams in path');
|
||||
});
|
||||
|
||||
test('returns workstream config path when GSD_WORKSTREAM is set', () => {
|
||||
const result = runGsdTools('config-path', tmpDir, { GSD_WORKSTREAM: 'my-stream' });
|
||||
assert.ok(result.success, `config-path failed: ${result.error}`);
|
||||
assert.ok(result.output.trim().includes('workstreams/my-stream/config.json'), `expected workstream config path, got: ${result.output}`);
|
||||
assert.ok(result.output.trim().replace(/\\/g, '/').includes('workstreams/my-stream/config.json'), `expected workstream config path, got: ${result.output}`);
|
||||
});
|
||||
|
||||
test('config-path and config-get agree on the active path', () => {
|
||||
|
||||
@@ -99,7 +99,9 @@ describe('enh-2500: gsd-codebase-mapper arch focus — rich architecture output'
|
||||
|
||||
test('template includes data flow traces with numbered steps', () => {
|
||||
const hasPrimaryRequestPath = /###\s+Primary Request Path/i.test(archTemplate);
|
||||
const hasThreeNumberedSteps = /^\s*1\..+\n\s*2\..+\n\s*3\./m.test(archTemplate);
|
||||
// [^\n]+ + \r?\n is CRLF-tolerant: .+ doesn't match \r in JS regex by
|
||||
// default, so \r before the literal \n in CRLF content kills the match.
|
||||
const hasThreeNumberedSteps = /^\s*1\.[^\n]+\r?\n\s*2\.[^\n]+\r?\n\s*3\./m.test(archTemplate);
|
||||
const hasFileLineRefs = /\(`\[.*:(?:line|\d+)\]`\)/.test(archTemplate);
|
||||
|
||||
assert.ok(
|
||||
|
||||
@@ -18,7 +18,9 @@ const COMMANDS_DIR = path.join(__dirname, '..', 'commands', 'gsd');
|
||||
*/
|
||||
function parseFrontmatter(filePath) {
|
||||
const raw = fs.readFileSync(filePath, 'utf8');
|
||||
const lines = raw.split('\n');
|
||||
// CRLF-tolerant: Windows checkouts leave \r on every line. lines.indexOf('---', 1)
|
||||
// would never match because elements would be '---\r' instead of '---'.
|
||||
const lines = raw.split(/\r?\n/);
|
||||
if (lines[0].trim() !== '---') return {};
|
||||
const endIdx = lines.indexOf('---', 1);
|
||||
if (endIdx === -1) return {};
|
||||
|
||||
@@ -32,7 +32,7 @@ describe('few-shot calibration examples', () => {
|
||||
describe('frontmatter metadata', () => {
|
||||
test('plan-checker.md has version and component in frontmatter', () => {
|
||||
const content = readFile(path.join(REFS_DIR, 'plan-checker.md'));
|
||||
assert.match(content, /^---\n/);
|
||||
assert.match(content, /^---\r?\n/);
|
||||
assert.match(content, /component:\s*plan-checker/);
|
||||
assert.match(content, /version:\s*\d+/);
|
||||
assert.match(content, /last_calibrated:\s*\d{4}-\d{2}-\d{2}/);
|
||||
@@ -40,7 +40,7 @@ describe('few-shot calibration examples', () => {
|
||||
|
||||
test('verifier.md has version and component in frontmatter', () => {
|
||||
const content = readFile(path.join(REFS_DIR, 'verifier.md'));
|
||||
assert.match(content, /^---\n/);
|
||||
assert.match(content, /^---\r?\n/);
|
||||
assert.match(content, /component:\s*verifier/);
|
||||
assert.match(content, /version:\s*\d+/);
|
||||
assert.match(content, /last_calibrated:\s*\d{4}-\d{2}-\d{2}/);
|
||||
|
||||
@@ -85,7 +85,7 @@ describe('gsd-settings-advanced — file scaffolding', () => {
|
||||
|
||||
test('command frontmatter has name, description, allowed-tools', () => {
|
||||
const text = fs.readFileSync(COMMAND_PATH, 'utf-8');
|
||||
const fmMatch = text.match(/^---\n([\s\S]*?)\n---/);
|
||||
const fmMatch = text.match(/^---\r?\n([\s\S]*?)\r?\n---/);
|
||||
assert.ok(fmMatch, 'command file missing frontmatter block');
|
||||
const fm = fmMatch[1];
|
||||
assert.match(fm, /name:\s*gsd:config/, 'frontmatter missing name (gsd:config)');
|
||||
|
||||
@@ -101,7 +101,10 @@ describe('codebase prompt injection scan', () => {
|
||||
const findings = [];
|
||||
|
||||
for (const file of agentFiles) {
|
||||
const relPath = path.relative(PROJECT_ROOT, file);
|
||||
// Normalize to POSIX separators so ALLOWLIST.has() works on Windows
|
||||
// (path.relative returns 'get-shit-done\bin\...' on win32; allowlist
|
||||
// keys are POSIX 'get-shit-done/bin/...').
|
||||
const relPath = path.relative(PROJECT_ROOT, file).replace(/\\/g, '/');
|
||||
if (ALLOWLIST.has(relPath)) continue;
|
||||
|
||||
const content = fs.readFileSync(file, 'utf-8');
|
||||
@@ -131,7 +134,10 @@ describe('codebase prompt injection scan', () => {
|
||||
const oversized = [];
|
||||
|
||||
for (const file of agentFiles) {
|
||||
const relPath = path.relative(PROJECT_ROOT, file);
|
||||
// Normalize to POSIX separators so ALLOWLIST.has() works on Windows
|
||||
// (path.relative returns 'get-shit-done\bin\...' on win32; allowlist
|
||||
// keys are POSIX 'get-shit-done/bin/...').
|
||||
const relPath = path.relative(PROJECT_ROOT, file).replace(/\\/g, '/');
|
||||
if (ALLOWLIST.has(relPath)) continue;
|
||||
|
||||
const content = fs.readFileSync(file, 'utf-8');
|
||||
@@ -152,7 +158,10 @@ describe('codebase prompt injection scan', () => {
|
||||
const findings = [];
|
||||
|
||||
for (const file of workflowFiles) {
|
||||
const relPath = path.relative(PROJECT_ROOT, file);
|
||||
// Normalize to POSIX separators so ALLOWLIST.has() works on Windows
|
||||
// (path.relative returns 'get-shit-done\bin\...' on win32; allowlist
|
||||
// keys are POSIX 'get-shit-done/bin/...').
|
||||
const relPath = path.relative(PROJECT_ROOT, file).replace(/\\/g, '/');
|
||||
if (ALLOWLIST.has(relPath)) continue;
|
||||
|
||||
const content = fs.readFileSync(file, 'utf-8');
|
||||
@@ -175,7 +184,10 @@ describe('codebase prompt injection scan', () => {
|
||||
const findings = [];
|
||||
|
||||
for (const file of commandFiles) {
|
||||
const relPath = path.relative(PROJECT_ROOT, file);
|
||||
// Normalize to POSIX separators so ALLOWLIST.has() works on Windows
|
||||
// (path.relative returns 'get-shit-done\bin\...' on win32; allowlist
|
||||
// keys are POSIX 'get-shit-done/bin/...').
|
||||
const relPath = path.relative(PROJECT_ROOT, file).replace(/\\/g, '/');
|
||||
if (ALLOWLIST.has(relPath)) continue;
|
||||
|
||||
const content = fs.readFileSync(file, 'utf-8');
|
||||
@@ -198,7 +210,10 @@ describe('codebase prompt injection scan', () => {
|
||||
const findings = [];
|
||||
|
||||
for (const file of hookFiles) {
|
||||
const relPath = path.relative(PROJECT_ROOT, file);
|
||||
// Normalize to POSIX separators so ALLOWLIST.has() works on Windows
|
||||
// (path.relative returns 'get-shit-done\bin\...' on win32; allowlist
|
||||
// keys are POSIX 'get-shit-done/bin/...').
|
||||
const relPath = path.relative(PROJECT_ROOT, file).replace(/\\/g, '/');
|
||||
if (ALLOWLIST.has(relPath)) continue;
|
||||
|
||||
const content = fs.readFileSync(file, 'utf-8');
|
||||
@@ -221,7 +236,10 @@ describe('codebase prompt injection scan', () => {
|
||||
const findings = [];
|
||||
|
||||
for (const file of libFiles) {
|
||||
const relPath = path.relative(PROJECT_ROOT, file);
|
||||
// Normalize to POSIX separators so ALLOWLIST.has() works on Windows
|
||||
// (path.relative returns 'get-shit-done\bin\...' on win32; allowlist
|
||||
// keys are POSIX 'get-shit-done/bin/...').
|
||||
const relPath = path.relative(PROJECT_ROOT, file).replace(/\\/g, '/');
|
||||
if (ALLOWLIST.has(relPath)) continue;
|
||||
|
||||
const content = fs.readFileSync(file, 'utf-8');
|
||||
@@ -244,7 +262,10 @@ describe('codebase prompt injection scan', () => {
|
||||
const invisiblePattern = /[\u200B-\u200F\u2028-\u202F\uFEFF\u00AD]/;
|
||||
|
||||
for (const file of allFiles) {
|
||||
const relPath = path.relative(PROJECT_ROOT, file);
|
||||
// Normalize to POSIX separators so ALLOWLIST.has() works on Windows
|
||||
// (path.relative returns 'get-shit-done\bin\...' on win32; allowlist
|
||||
// keys are POSIX 'get-shit-done/bin/...').
|
||||
const relPath = path.relative(PROJECT_ROOT, file).replace(/\\/g, '/');
|
||||
if (ALLOWLIST.has(relPath)) continue;
|
||||
|
||||
const content = fs.readFileSync(file, 'utf-8');
|
||||
@@ -273,7 +294,10 @@ describe('codebase prompt injection scan', () => {
|
||||
const boundaryPattern = /<\/?(?:system|assistant|human)>/i;
|
||||
|
||||
for (const file of allFiles) {
|
||||
const relPath = path.relative(PROJECT_ROOT, file);
|
||||
// Normalize to POSIX separators so ALLOWLIST.has() works on Windows
|
||||
// (path.relative returns 'get-shit-done\bin\...' on win32; allowlist
|
||||
// keys are POSIX 'get-shit-done/bin/...').
|
||||
const relPath = path.relative(PROJECT_ROOT, file).replace(/\\/g, '/');
|
||||
if (ALLOWLIST.has(relPath)) continue;
|
||||
// Allow .md files to use common tags in examples/docs
|
||||
// But flag .js/.cjs files that embed these
|
||||
|
||||
@@ -171,12 +171,14 @@ describe('pruneOrphanedWorktrees', () => {
|
||||
execSync('git worktree add "' + worktreeDir + '" -b fix/stale-ref', { cwd: repoDir, stdio: 'pipe' });
|
||||
assert.ok(fs.existsSync(worktreeDir), 'worktree dir should exist before manual deletion');
|
||||
|
||||
// Verify it appears in git worktree list
|
||||
const beforeList = execSync('git worktree list --porcelain', { cwd: repoDir, encoding: 'utf8' });
|
||||
// git worktree list --porcelain emits forward slashes on Windows even
|
||||
// when path.join produced backslashes; normalize both sides for compare.
|
||||
const normalizeSlashes = (p) => p.replace(/\\/g, '/');
|
||||
assert.ok(normalizeSlashes(beforeList).includes(normalizeSlashes(worktreeDir)), 'worktree should appear in list before deletion');
|
||||
// Use the canonicalPath helper so Windows 8.3 short-name (RUNNER~1) vs
|
||||
// long-form (runneradmin) and slash-direction differences both collapse
|
||||
// to the same key before comparison. git stores the long-form path in
|
||||
// its administrative files; substring matching on the raw path fails.
|
||||
// Capture the canonical key BEFORE deletion since canonicalPath calls
|
||||
// realpathSync.native which fails on missing paths.
|
||||
const wantedKey = canonicalPath(worktreeDir);
|
||||
assert.ok(listedWorktreePaths(repoDir).has(wantedKey), 'worktree should appear in list before deletion');
|
||||
|
||||
// Manually delete the worktree directory (simulate orphan)
|
||||
fs.rmSync(worktreeDir, { recursive: true, force: true });
|
||||
@@ -185,11 +187,10 @@ describe('pruneOrphanedWorktrees', () => {
|
||||
const pruneOrphanedWorktrees = getPruneOrphanedWorktrees();
|
||||
pruneOrphanedWorktrees(repoDir);
|
||||
|
||||
// Assert: git worktree list no longer shows the stale entry
|
||||
const afterList = execSync('git worktree list --porcelain', { cwd: repoDir, encoding: 'utf8' });
|
||||
// Assert: git worktree list no longer shows the stale entry.
|
||||
assert.ok(
|
||||
!normalizeSlashes(afterList).includes(normalizeSlashes(worktreeDir)),
|
||||
'git worktree list still shows stale entry after prune:\n' + afterList
|
||||
!listedWorktreePaths(repoDir).has(wantedKey),
|
||||
'git worktree list still shows stale entry after prune'
|
||||
);
|
||||
});
|
||||
});
|
||||
|
||||
@@ -52,7 +52,10 @@ describe('skill-manifest', () => {
|
||||
});
|
||||
|
||||
test('returns normalized inventory across canonical roots', () => {
|
||||
const result = runGsdTools(['skill-manifest'], tmpDir, { HOME: homeDir });
|
||||
// On Windows, os.homedir() reads USERPROFILE (not HOME). The SUT scans
|
||||
// global skill roots via os.homedir(), so the test must also override
|
||||
// USERPROFILE to keep the fixture's homeDir visible.
|
||||
const result = runGsdTools(['skill-manifest'], tmpDir, { HOME: homeDir, USERPROFILE: homeDir });
|
||||
assert.ok(result.success, `Command should succeed: ${result.error || result.output}`);
|
||||
|
||||
const manifest = JSON.parse(result.output);
|
||||
@@ -117,7 +120,7 @@ describe('skill-manifest', () => {
|
||||
});
|
||||
|
||||
test('writes manifest to .planning/skill-manifest.json when --write flag is used', () => {
|
||||
const result = runGsdTools(['skill-manifest', '--write'], tmpDir, { HOME: homeDir });
|
||||
const result = runGsdTools(['skill-manifest', '--write'], tmpDir, { HOME: homeDir, USERPROFILE: homeDir });
|
||||
assert.ok(result.success, `Command should succeed: ${result.error || result.output}`);
|
||||
|
||||
const manifestPath = path.join(tmpDir, '.planning', 'skill-manifest.json');
|
||||
@@ -131,6 +134,7 @@ describe('skill-manifest', () => {
|
||||
test('global roots honor runtime-home env overrides instead of hardcoded home paths', () => {
|
||||
const result = runGsdTools(['skill-manifest'], tmpDir, {
|
||||
HOME: homeDir,
|
||||
USERPROFILE: homeDir,
|
||||
CLAUDE_CONFIG_DIR: path.join(homeDir, 'claude-custom'),
|
||||
CODEX_HOME: path.join(homeDir, 'codex-custom'),
|
||||
});
|
||||
|
||||
183
tests/windows-test-parity-guard.test.cjs
Normal file
183
tests/windows-test-parity-guard.test.cjs
Normal file
@@ -0,0 +1,183 @@
|
||||
'use strict';
|
||||
|
||||
process.env.GSD_TEST_MODE = '1';
|
||||
|
||||
/**
|
||||
* Ratchet-style lint guard against Windows-test-parity regressions.
|
||||
*
|
||||
* PR #3649 cleared ~270 Windows-only test failures from the chunking fix
|
||||
* in #3597 surfaced. Each cluster reduced to a handful of repeating
|
||||
* patterns. This guard prevents the patterns from being re-introduced.
|
||||
*
|
||||
* Strategy: per-pattern offender list is snapshotted at the count present
|
||||
* at the time of PR #3649. The test fails if a NEW file is added that
|
||||
* matches the anti-pattern (count grows above the baseline). Existing
|
||||
* offenders are acknowledged as technical debt that can be cleared
|
||||
* incrementally without blocking this PR.
|
||||
*
|
||||
* When you fix an existing offender, lower the corresponding BASELINE
|
||||
* count by 1. When CI breaks because BASELINE is set higher than the
|
||||
* actual offender count, lower BASELINE to match (one-way ratchet down).
|
||||
*
|
||||
* Scope: tests/ only. Production-code Windows-compat is enforced via
|
||||
* behavioural tests (see no-unconditional-win32-skip.test.cjs).
|
||||
*/
|
||||
|
||||
const { test, describe } = require('node:test');
|
||||
const assert = require('node:assert/strict');
|
||||
const fs = require('node:fs');
|
||||
const path = require('node:path');
|
||||
|
||||
const TESTS_DIR = path.join(__dirname);
|
||||
const SELF = path.basename(__filename);
|
||||
|
||||
// ── Baseline counts after PR #3649 batch ─────────────────────────────────
|
||||
// Set these to the exact number of offending files at the time of merge.
|
||||
// Each rule must not exceed its baseline; CI fails when a new offender appears.
|
||||
// Decrement when an existing offender is fixed.
|
||||
const BASELINE = {
|
||||
splitNewlineOnFileContent: 3,
|
||||
fenceRegexLiteralNewline: 2,
|
||||
frontmatterAnchorLiteralNewline: 5,
|
||||
hardcodedTmpToFsCall: 0,
|
||||
bareNpmExecWithoutShell: 0,
|
||||
stubsHomeNoUserProfile: 8,
|
||||
rmSyncNoMaxRetries: 95,
|
||||
};
|
||||
|
||||
function listTestFiles() {
|
||||
return fs.readdirSync(TESTS_DIR)
|
||||
.filter((f) => /\.(test|spec)\.cjs$/.test(f))
|
||||
.filter((f) => f !== SELF)
|
||||
.map((f) => path.join(TESTS_DIR, f));
|
||||
}
|
||||
|
||||
function readFileText(filePath) {
|
||||
return fs.readFileSync(filePath, 'utf8');
|
||||
}
|
||||
|
||||
// Strip line comments and block comments before pattern matching to avoid
|
||||
// false-positives in commentary describing the very pattern we forbid.
|
||||
function stripComments(text) {
|
||||
return text
|
||||
.replace(/\/\*[\s\S]*?\*\//g, '')
|
||||
.replace(/(^|[^:])\/\/[^\n]*/g, '$1');
|
||||
}
|
||||
|
||||
function countMatchingFiles(predicate) {
|
||||
let count = 0;
|
||||
const offenders = [];
|
||||
for (const file of listTestFiles()) {
|
||||
const text = stripComments(readFileText(file));
|
||||
if (predicate(text, file)) {
|
||||
count += 1;
|
||||
offenders.push(path.basename(file));
|
||||
}
|
||||
}
|
||||
return { count, offenders };
|
||||
}
|
||||
|
||||
function ratchetAssert(rule, actualCount, baselineCount, offenders, guidance) {
|
||||
if (actualCount > baselineCount) {
|
||||
const newCount = actualCount - baselineCount;
|
||||
assert.fail(
|
||||
`Windows-parity guard "${rule}": ${actualCount} offenders, baseline is ${baselineCount} ` +
|
||||
`(+${newCount} new). New occurrences of this anti-pattern were added. ` +
|
||||
`${guidance}\n\nFull offender list (${actualCount}):\n ` +
|
||||
offenders.join('\n '),
|
||||
);
|
||||
}
|
||||
}
|
||||
|
||||
describe('Windows test-parity lint guards (ratchet baseline: PR #3649)', () => {
|
||||
// ── G1 — CRLF: file-content split on literal '\n' ─────────────────────
|
||||
test('split-on-newline after readFileSync (use /\\r?\\n/)', () => {
|
||||
const { count, offenders } = countMatchingFiles((text) => {
|
||||
return /\.readFileSync\s*\([^)]*\)[^;]*\.split\(\s*['"]\\n['"]\s*\)/.test(text);
|
||||
});
|
||||
ratchetAssert(
|
||||
'splitNewlineOnFileContent', count, BASELINE.splitNewlineOnFileContent, offenders,
|
||||
"Replace .split('\\n') with .split(/\\r?\\n/) so the test tolerates CRLF " +
|
||||
"checkout (autocrlf=true on Windows leaves trailing \\r on every line).",
|
||||
);
|
||||
});
|
||||
|
||||
// ── G2 — CRLF: ```bash|sh\n fence regex on file content ──────────────
|
||||
test('markdown-fence regex with literal \\n after ```bash/sh', () => {
|
||||
const { count, offenders } = countMatchingFiles((text) => {
|
||||
return /\/[^/]*```(?:bash|sh)\\n[^/]*\//.test(text);
|
||||
});
|
||||
ratchetAssert(
|
||||
'fenceRegexLiteralNewline', count, BASELINE.fenceRegexLiteralNewline, offenders,
|
||||
"Use /```(?:bash|sh)\\r?\\n([\\s\\S]*?)```/g — Windows CRLF makes the byte after " +
|
||||
"`bash` be \\r, the regex never matches, and bash-block extraction returns empty.",
|
||||
);
|
||||
});
|
||||
|
||||
// ── G3 — CRLF: frontmatter regex with literal '\n' ────────────────────
|
||||
test('frontmatter regex anchors on /^---\\n/', () => {
|
||||
const { count, offenders } = countMatchingFiles((text) => {
|
||||
return /\/\^---\\n/.test(text);
|
||||
});
|
||||
ratchetAssert(
|
||||
'frontmatterAnchorLiteralNewline', count, BASELINE.frontmatterAnchorLiteralNewline, offenders,
|
||||
"Use /^---\\r?\\n/ — on Windows the byte after --- is \\r, not \\n, so the " +
|
||||
"anchor fails to match and parseFrontmatter returns null/{}.",
|
||||
);
|
||||
});
|
||||
|
||||
// ── G4 — POSIX-tmp: hardcoded '/tmp/' literal passed to fs.* ─────────
|
||||
test('fs.* call receives a hardcoded "/tmp/..." literal', () => {
|
||||
const { count, offenders } = countMatchingFiles((text) => {
|
||||
return /\bfs\.[A-Za-z]+\s*\([^)]*['"]\/tmp\/[^'"]+['"][^)]*\)/.test(text);
|
||||
});
|
||||
ratchetAssert(
|
||||
'hardcodedTmpToFsCall', count, BASELINE.hardcodedTmpToFsCall, offenders,
|
||||
"Use os.tmpdir() — on Windows '/tmp/foo' becomes 'D:\\tmp\\foo' where D:\\tmp " +
|
||||
"doesn't exist by default → ENOENT.",
|
||||
);
|
||||
});
|
||||
|
||||
// ── G5 — npm.cmd: bare 'npm' to exec*Sync without shell:true ─────────
|
||||
test('bare npm exec without shell-true Windows fallback', () => {
|
||||
const { count, offenders } = countMatchingFiles((text) => {
|
||||
const re = /\b(?:execFileSync|spawnSync)\s*\(\s*['"]npm['"]\s*,[^)]*\)/g;
|
||||
const matches = text.match(re) || [];
|
||||
return matches.some((m) =>
|
||||
!/shell\s*:\s*true/.test(m) && !/shell\s*:\s*isWindows/.test(m),
|
||||
);
|
||||
});
|
||||
ratchetAssert(
|
||||
'bareNpmExecWithoutShell', count, BASELINE.bareNpmExecWithoutShell, offenders,
|
||||
"On Windows npm is npm.cmd — pass {shell: process.platform === 'win32'} or " +
|
||||
"use npm.cmd directly, otherwise execFileSync errors ENOENT.",
|
||||
);
|
||||
});
|
||||
|
||||
// ── G6 — Test stubs HOME without USERPROFILE ─────────────────────────
|
||||
test('test stubs process.env.HOME but never references USERPROFILE', () => {
|
||||
const { count, offenders } = countMatchingFiles((text) => {
|
||||
return /process\.env\.HOME\s*=\s*/.test(text) && !/USERPROFILE/.test(text);
|
||||
});
|
||||
ratchetAssert(
|
||||
'stubsHomeNoUserProfile', count, BASELINE.stubsHomeNoUserProfile, offenders,
|
||||
"On Windows os.homedir() reads USERPROFILE (not HOME). Tests redirecting ~ " +
|
||||
"must override both, or the SUT sees the real user's home.",
|
||||
);
|
||||
});
|
||||
|
||||
// ── G7 — rmSync cleanup without retry budget ─────────────────────────
|
||||
test('test teardown rmSync without maxRetries', () => {
|
||||
const { count, offenders } = countMatchingFiles((text) => {
|
||||
const re = /fs\.rmSync\s*\([^)]*recursive\s*:\s*true[^)]*force\s*:\s*true[^)]*\)/g;
|
||||
const matches = text.match(re) || [];
|
||||
return matches.some((m) => !/maxRetries/.test(m));
|
||||
});
|
||||
ratchetAssert(
|
||||
'rmSyncNoMaxRetries', count, BASELINE.rmSyncNoMaxRetries, offenders,
|
||||
"Use helpers.cleanup() (shared 5s retry budget) or pass " +
|
||||
"{maxRetries: 10, retryDelay: 100} — Windows AV scanners can hold handles for " +
|
||||
"seconds after a process exits, surfacing as flaky EBUSY teardown failures.",
|
||||
);
|
||||
});
|
||||
});
|
||||
Reference in New Issue
Block a user