diff options
Diffstat (limited to 'languages/elisp')
| -rwxr-xr-x | languages/elisp/claude/hooks/validate-el.sh | 4 | ||||
| -rwxr-xr-x | languages/elisp/githooks/pre-commit | 23 | ||||
| -rw-r--r-- | languages/elisp/tests/test-pre-commit-hook.bats | 126 | ||||
| -rw-r--r-- | languages/elisp/tests/test-validate-el-hook.bats | 100 |
4 files changed, 247 insertions, 6 deletions
diff --git a/languages/elisp/claude/hooks/validate-el.sh b/languages/elisp/claude/hooks/validate-el.sh index d028789..870eefe 100755 --- a/languages/elisp/claude/hooks/validate-el.sh +++ b/languages/elisp/claude/hooks/validate-el.sh @@ -39,8 +39,6 @@ f="$(jq -r '.tool_input.file_path // .tool_response.filePath // empty')" [ -z "$f" ] && exit 0 [ "${f##*.}" = "el" ] || exit 0 -MAX_AUTO_TEST_FILES=20 # skip if more matches than this (large test suites) - # --- Phase 1: syntax + byte-compile --- case "$f" in */init.el|*/early-init.el) @@ -96,7 +94,7 @@ case "$f" in esac count="${#tests[@]}" -if [ "$count" -ge 1 ] && [ "$count" -le "$MAX_AUTO_TEST_FILES" ]; then +if [ "$count" -ge 1 ]; then load_args=() for t in "${tests[@]}"; do load_args+=("-l" "$t"); done if ! output="$(emacs --batch --no-site-file --no-site-lisp \ diff --git a/languages/elisp/githooks/pre-commit b/languages/elisp/githooks/pre-commit index 2c0cf0b..a87bedf 100755 --- a/languages/elisp/githooks/pre-commit +++ b/languages/elisp/githooks/pre-commit @@ -18,8 +18,18 @@ cd "$REPO_ROOT" || exit 1 SECRET_PATTERNS_CS='(AKIA[0-9A-Z]{16}|sk-[a-zA-Z0-9_-]{20,}|-----BEGIN (RSA|DSA|EC|OPENSSH|PGP)( PRIVATE)?( KEY| KEY BLOCK)?-----)' SECRET_PATTERNS_CI='(api[_-]?key|api[_-]?secret|auth[_-]?token|secret[_-]?key|bearer[_-]?token|access[_-]?token|password)[[:space:]]*[:=][[:space:]]*["'"'"'][^"'"'"']{16,}["'"'"']' -added_lines="$(git diff --cached -U0 --diff-filter=AM \ - | grep '^+' | grep -v '^+++' || true)" +# Read the diff on its own so a git failure is distinguishable from "grep +# matched nothing". Both end in a non-zero status, but only one of them means +# there is nothing to scan; piping them together and swallowing the result with +# `|| true` made a broken git look like a clean commit — the scan searched an +# empty string, found nothing, and the secret went in. +if ! staged_diff="$(git diff --cached -U0 --diff-filter=AM)"; then + echo "pre-commit: cannot read the staged diff — refusing to skip the secret scan" >&2 + exit 1 +fi + +# The greps keep their `|| true`: exiting 1 on no match is their normal result. +added_lines="$(printf '%s\n' "$staged_diff" | grep '^+' | grep -v '^+++' || true)" cs_hits="$(printf '%s\n' "$added_lines" | grep -nE "$SECRET_PATTERNS_CS" || true)" ci_hits="$(printf '%s\n' "$added_lines" | grep -niE "$SECRET_PATTERNS_CI" || true)" @@ -37,7 +47,14 @@ if [ -n "$secret_hits" ]; then fi # --- 2. Paren check on staged .el files --- -staged_el="$(git diff --cached --name-only --diff-filter=AM | grep '\.el$' || true)" +# Same split as the secret scan above: a git failure must not read as "no files +# staged", which would skip the language check silently. +if ! staged_names="$(git diff --cached --name-only --diff-filter=AM)"; then + echo "pre-commit: cannot read the staged file list — refusing to skip the check" >&2 + exit 1 +fi + +staged_el="$(printf '%s\n' "$staged_names" | grep '\.el$' || true)" if [ -n "$staged_el" ]; then paren_fail="" diff --git a/languages/elisp/tests/test-pre-commit-hook.bats b/languages/elisp/tests/test-pre-commit-hook.bats new file mode 100644 index 0000000..413c71d --- /dev/null +++ b/languages/elisp/tests/test-pre-commit-hook.bats @@ -0,0 +1,126 @@ +#!/usr/bin/env bats +# Tests for githooks/pre-commit — the secret scan and paren check. +# +# The scan reads its input through a pipeline: +# +# added_lines="$(git diff --cached ... | grep '^+' | grep -v '^+++' || true)" +# +# `grep` exits 1 when it matches nothing, which is the ordinary case, so the +# `|| true` has to stay. But with no `pipefail` it also swallows a failure of +# `git diff` itself, and an empty `added_lines` makes the scan search nothing, +# find nothing, and report clean. A gate that passes without looking is the +# failure this file exists to pin: the fail-open test drives a broken `git diff` +# and asserts the hook refuses rather than exiting 0. +# +# Each test builds a throwaway git repo in BATS_TEST_TMPDIR, so nothing touches +# the real repository or its hooks. + +setup() { + HOOK="${BATS_TEST_DIRNAME}/../githooks/pre-commit" + REPO="${BATS_TEST_TMPDIR}/repo" + mkdir -p "$REPO" + cd "$REPO" || return 1 + git init -q . + git config user.email t@example.com + git config user.name Test + # Split so the fixtures never appear as credential-shaped literals here. + AWS_TAIL="IOSFODNN7EXAMPLE" + WORD_TAIL="word" +} + +# Put a stub `git` ahead of the real one that fails for the staged-diff call +# and delegates everything else, so only the pipeline under test breaks. +break_staged_diff() { + mkdir -p "${BATS_TEST_TMPDIR}/bin" + cat > "${BATS_TEST_TMPDIR}/bin/git" <<'STUB' +#!/usr/bin/env bash +if [ "${1:-}" = "diff" ] && [ "${2:-}" = "--cached" ] && [ "${3:-}" = "-U0" ]; then + echo "simulated git failure" >&2 + exit 128 +fi +exec /usr/bin/git "$@" +STUB + chmod +x "${BATS_TEST_TMPDIR}/bin/git" + PATH="${BATS_TEST_TMPDIR}/bin:$PATH" +} + +# ------------------------------- Normal cases ------------------------------- + +@test "secret scan: blocks a staged AWS key" { + # Assembled at runtime: a literal key-shaped string in this file would trip + # the very hook under test on every commit that touches it, and this repo + # mirrors to a public remote. + printf 'aws = "%s"\n' "AKIA${AWS_TAIL}" > creds.txt + git add creds.txt + run "$HOOK" + [ "$status" -eq 1 ] + [[ "$output" == *"potential secret"* ]] +} + +@test "secret scan: blocks a staged keyword=value password" { + printf '%s = "%s"\n' "pass${WORD_TAIL}" "correcthorsebatterystaple" > conf.txt + git add conf.txt + run "$HOOK" + [ "$status" -eq 1 ] + [[ "$output" == *"potential secret"* ]] +} + +@test "secret scan: allows an ordinary staged file" { + printf 'just some prose\n' > notes.txt + git add notes.txt + run "$HOOK" + [ "$status" -eq 0 ] +} + +# ------------------------------ Boundary cases ------------------------------ + +@test "secret scan: allows a commit with nothing staged" { + run "$HOOK" + [ "$status" -eq 0 ] +} + +@test "paren check: blocks an unbalanced staged .el file" { + printf '(defun broken ()\n (message "no close"\n' > bad.el + git add bad.el + run "$HOOK" + [ "$status" -eq 1 ] + [[ "$output" == *"paren check failed"* ]] +} + +@test "paren check: allows a balanced staged .el file" { + printf '(defun fine ()\n (message "ok"))\n' > good.el + git add good.el + run "$HOOK" + [ "$status" -eq 0 ] +} + +# -------------------------------- Error cases ------------------------------- + +@test "secret scan: refuses to pass when the staged diff cannot be read" { + # The scan must not report clean after searching nothing. Without a + # pipefail-aware guard the broken diff yields an empty added_lines and the + # hook exits 0, letting a real secret through unscanned. + printf 'aws = "%s"\n' "AKIA${AWS_TAIL}" > creds.txt + git add creds.txt + break_staged_diff + run "$HOOK" + [ "$status" -ne 0 ] +} + +@test "paren check: refuses to pass when the staged file list cannot be read" { + printf '(defun broken ()\n (message "no close"\n' > bad.el + git add bad.el + mkdir -p "${BATS_TEST_TMPDIR}/bin2" + cat > "${BATS_TEST_TMPDIR}/bin2/git" <<'STUB' +#!/usr/bin/env bash +if [ "${1:-}" = "diff" ] && [ "${2:-}" = "--cached" ] && [ "${3:-}" = "--name-only" ]; then + echo "simulated git failure" >&2 + exit 128 +fi +exec /usr/bin/git "$@" +STUB + chmod +x "${BATS_TEST_TMPDIR}/bin2/git" + PATH="${BATS_TEST_TMPDIR}/bin2:$PATH" + run "$HOOK" + [ "$status" -ne 0 ] +} diff --git a/languages/elisp/tests/test-validate-el-hook.bats b/languages/elisp/tests/test-validate-el-hook.bats new file mode 100644 index 0000000..d4d6f23 --- /dev/null +++ b/languages/elisp/tests/test-validate-el-hook.bats @@ -0,0 +1,100 @@ +#!/usr/bin/env bats +# Tests for .claude/hooks/validate-el.sh — the auto-test runner. +# +# The runner used to skip entirely above MAX_AUTO_TEST_FILES=20, with no else +# branch: nothing printed, exit 0, indistinguishable from a passing run. That +# was live for the three largest families here (calendar-sync 63 test files, +# music 45, ai-term 35), so every edit to those ran parens and byte-compile and +# zero tests, silently. +# +# The cap was removed rather than made loud, because its premise did not hold. +# Measured on this machine, running a whole family takes about a second: +# ai-term 208 tests in 1.0s, music 403 in 1.7s, calendar-sync 633 in 0.9s. It +# was also concealing a real cross-test pollution bug in calendar-sync that +# only appears when that family runs in one process. +# +# These tests pin that no file count is skipped. Each builds a synthetic +# project in BATS_TEST_TMPDIR and points CLAUDE_PROJECT_DIR at it, so nothing +# runs against the real tree. + +setup() { + # Bundle layout: the hook ships at claude/hooks/ here and installs to + # .claude/hooks/ in a consuming project. This test was written against the + # installed layout, so re-homing it needed the path adjusted. + HOOK="${BATS_TEST_DIRNAME}/../claude/hooks/validate-el.sh" + PROJ="${BATS_TEST_TMPDIR}/proj" + mkdir -p "$PROJ/modules" "$PROJ/tests" + export CLAUDE_PROJECT_DIR="$PROJ" + printf '(provide (quote widget))\n' > "$PROJ/modules/widget.el" +} + +# N green test files matching the widget stem. +make_tests() { + local n="$1" i + for ((i = 1; i <= n; i++)); do + printf '(require (quote ert))\n(ert-deftest test-widget-%d () (should t))\n' \ + "$i" > "$PROJ/tests/test-widget-${i}.el" + done +} + +# One failing test file, to prove the run is real rather than merely quiet. +make_failing_test() { + printf '(require (quote ert))\n(ert-deftest test-widget-bad () (should nil))\n' \ + > "$PROJ/tests/test-widget-bad.el" +} + +hook_input() { + printf '{"tool_input":{"file_path":"%s"}}' "$PROJ/modules/widget.el" +} + +run_hook() { + run bash -c "$(printf '%q' "$HOOK") <<< '$(hook_input)'" +} + +# ------------------------------- Normal cases ------------------------------- + +@test "a small family runs and passes quietly" { + make_tests 3 + run_hook + [ "$status" -eq 0 ] +} + +@test "a failing test blocks, so a quiet pass means the tests really ran" { + make_tests 3 + make_failing_test + run_hook + [ "$status" -eq 2 ] + [[ "$output" == *"TESTS FAILED"* ]] +} + +# ------------------------------ Boundary cases ------------------------------ + +@test "at the old cap of 20 files: runs" { + make_tests 20 + run_hook + [ "$status" -eq 0 ] +} + +@test "past the old cap: still runs, no longer skipped" { + make_tests 21 + run_hook + [ "$status" -eq 0 ] + [[ "${output,,}" != *"skipped"* ]] +} + +@test "well past the old cap: a failure in file 63 is still caught" { + # The regression this guards: at 63 files the runner used to skip, so a red + # test in a big family reported clean. calendar-sync is exactly this size. + make_tests 63 + make_failing_test + run_hook + [ "$status" -eq 2 ] + [[ "$output" == *"TESTS FAILED"* ]] +} + +# -------------------------------- Error cases ------------------------------- + +@test "no matching tests: exits clean without running anything" { + run_hook + [ "$status" -eq 0 ] +} |