#!/usr/bin/env bash
# test-push-guard.sh - Needle harness for push-guard.sh.
#
# Every check gets BOTH polarities:
#   NEEDLE   a deliberately-broken fixture that MUST trip the guard.
#   CONTROL  a clean fixture that MUST pass.
#
# The controls are not decoration. A guard that failed unconditionally would
# satisfy every needle and look fully covered — which is the same class of defect
# (an assertion satisfied by the null case) that push-guard.sh exists to prevent.
# A needle set without controls cannot tell "working guard" from "broken guard".

set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
GUARD="$SCRIPT_DIR/push-guard.sh"
WORK_DIR="${MOSAIC_TEST_WORK_DIR:-$SCRIPT_DIR/.work}"

PASS=0
FAIL=0

# Fresh repo + bare "remote" for each case, so no case can inherit another's state.
new_repo() {
    local name="$1"
    local dir="$WORK_DIR/$name"
    rm -rf "$dir"
    mkdir -p "$dir"
    git init -q -b main "$dir/repo"
    git init -q --bare "$dir/remote.git"
    git -C "$dir/repo" remote add origin "$dir/remote.git"
    git -C "$dir/repo" config user.name "Needle Harness"
    git -C "$dir/repo" config user.email "needle@example.invalid"
    printf 'seed\n' > "$dir/repo/seed.txt"
    git -C "$dir/repo" add seed.txt
    git -C "$dir/repo" commit -q -m "seed"
    printf '%s' "$dir/repo"
}

# write_config <repo> <json-text>
# The guard reads .push-guard.json from the WORKING TREE at the repo root, so it
# does not need to be staged — which also keeps the "N staged file(s)" counts in
# the older controls unchanged.
write_config() {
    printf '%s\n' "$2" > "$1/.push-guard.json"
}

# expect <kind> <expected_exit> <description> [--out <substring>] -- <command...>
#
# --out asserts a substring of the guard's own output. It exists because exit 0
# alone cannot distinguish "checked the files and they were fine" from "matched
# no files and had nothing to check". Two controls in an earlier revision of this
# harness were passing vacuously for exactly that reason — the pathspec silently
# matched nothing. A control that cannot fail is not a control.
expect() {
    local kind="$1" want="$2" desc="$3"; shift 3
    local need_out=""
    while (( $# )); do
        case "$1" in
            --out) need_out="$2"; shift 2 ;;
            --)    shift; break ;;
            *)     break ;;
        esac
    done
    local got=0 out
    out="$("$@" 2>&1)" || got=$?

    local why=""
    [[ "$got" == "$want" ]] || why="wanted exit $want, got $got"
    if [[ -z "$why" && -n "$need_out" && "$out" != *"$need_out"* ]]; then
        why="exit $got as expected, but output never said: $need_out"
    fi

    if [[ -z "$why" ]]; then
        printf '  PASS  [%-7s] %s (exit %s)\n' "$kind" "$desc" "$got"
        PASS=$((PASS + 1))
    else
        printf '  FAIL  [%-7s] %s — %s\n' "$kind" "$desc" "$why"
        printf '%s\n' "$out" | sed 's/^/          | /'
        FAIL=$((FAIL + 1))
    fi
}

rm -rf "$WORK_DIR"
mkdir -p "$WORK_DIR"

echo "=== (a) conflict markers / unmerged index ==="

# NEEDLE: staged content carrying real conflict markers.
R="$(new_repo a1)"
cat > "$R/conflicted.txt" <<'EOF'
intro line
<<<<<<< HEAD
ours
=======
theirs
>>>>>>> feature/x
tail line
EOF
git -C "$R" add conflicted.txt
expect NEEDLE 2 "staged conflict markers are refused" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# NEEDLE: a genuine unmerged index (merge in progress).
R="$(new_repo a2)"
git -C "$R" checkout -q -b other
printf 'from-other\n' > "$R/clash.txt"
git -C "$R" add clash.txt && git -C "$R" commit -q -m other
git -C "$R" checkout -q main
printf 'from-main\n' > "$R/clash.txt"
git -C "$R" add clash.txt && git -C "$R" commit -q -m main
git -C "$R" merge other -q >/dev/null 2>&1 || true
expect NEEDLE 2 "unmerged index paths are refused" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# CONTROL: ordinary staged content passes.
# The config is REQUIRED as of the decision gate: this fixture has no generated
# JSON, so it records that fact with a reason. This is the migration cost of the
# hard cutover, paid here first — these two controls were the only pre-existing
# cases that reached the JSON stage without nominating a path, and they are
# exactly the "repo that never wired it up" the gate now refuses.
R="$(new_repo a3)"
write_config "$R" '{"json_check": "none", "reason": "fixture has no generated JSON"}'
printf 'just some code\n' > "$R/clean.txt"
git -C "$R" add clean.txt
expect CONTROL 0 "clean staged content passes (1 file actually scanned)" \
    --out "no conflict markers in 1 staged file(s)" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# CONTROL (false-positive guard): prose using ======= as an underline must NOT trip.
# This is why the marker pattern deliberately ignores a bare '======='.
R="$(new_repo a4)"
write_config "$R" '{"json_check": "none", "reason": "fixture has no generated JSON"}'
cat > "$R/README.rst" <<'EOF'
Section Title
=============

Body text.

Another Heading
=======
EOF
git -C "$R" add README.rst
expect CONTROL 0 "prose using ======= underlines does not trip the guard" \
    --out "no conflict markers in 1 staged file(s)" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# NEEDLE (fault injection): if the scan itself ERRORS, the guard must refuse
# rather than report a clean result. git grep exits 1 for "no match" but >=2 for
# a real failure; a blanket '|| true' would collapse the two. We inject the
# failure with a PATH shim rather than trying to corrupt an index.
R="$(new_repo a5)"
SHIM="$WORK_DIR/a5/bin"
mkdir -p "$SHIM"
REAL_GIT="$(command -v git)"
cat > "$SHIM/git" <<SH
#!/usr/bin/env bash
if [[ "\$1" == "grep" ]]; then exit 2; fi
exec "$REAL_GIT" "\$@"
SH
chmod +x "$SHIM/git"
printf 'ordinary content\n' > "$R/thing.txt"
git -C "$R" add thing.txt
expect NEEDLE 2 "a conflict scan that ERRORS is refused, not reported clean" \
    --out "did not run" -- \
    bash -c "cd '$R' && export PATH=\"$SHIM:\$PATH\" && '$GUARD' check-staged"

# NEEDLE: RENAME + MODIFY. Git reports a `git mv` plus a small edit as a single
# 'R' entry, which --diff-filter=ACM does not match, making the file invisible to
# every content check. A clean file is co-staged deliberately: without it the
# file list is empty and the nothing-staged guard fires by ACCIDENT, which would
# make this needle pass for the wrong reason and hide the real defect.
R="$(new_repo a6)"
mkdir -p "$R/data"
seq 1 200 | sed 's/^/line /' > "$R/data/canon.txt"
printf 'original\n' > "$R/other.txt"
git -C "$R" add -A && git -C "$R" commit -q -m seed
git -C "$R" mv data/canon.txt data/canon2.txt
printf '<<<<<<< HEAD\nours\n=======\ntheirs\n>>>>>>> b\n' >> "$R/data/canon2.txt"
printf 'an ordinary innocent change\n' > "$R/other.txt"
git -C "$R" add -A
expect NEEDLE 2 "conflict markers in a RENAMED file are refused (rename+modify)" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# NEEDLE: a path marked binary in .gitattributes. `git grep -I` honours that
# classification and skips the blob entirely while the summary still counts the
# file as scanned — a clean pass AND a false count. Hence -a, not -I.
R="$(new_repo a7)"
printf 'notes.txt binary\n' > "$R/.gitattributes"
printf 'x\n<<<<<<< HEAD\nours\n=======\ntheirs\n>>>>>>> b\ny\n' > "$R/notes.txt"
git -C "$R" add -A
expect NEEDLE 2 "conflict markers in a .gitattributes-binary file are refused" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

echo "=== (c) staged JSON parses ==="

# NEEDLE: malformed JSON, the shape a broken generator emits.
R="$(new_repo c1)"
mkdir -p "$R/data"
printf '{"a": 1,\n' > "$R/data/broken.json"
git -C "$R" add data/broken.json
expect NEEDLE 3 "staged unparseable JSON under --json-path is refused" -- \
    bash -c "cd '$R' && '$GUARD' check-staged --json-path 'data/**/*.json'"

# NEEDLE: conflict markers inside a generated JSON file — the 70-file incident.
R="$(new_repo c2)"
mkdir -p "$R/data"
cat > "$R/data/canon.json" <<'EOF'
{
<<<<<<< HEAD
  "v": 1
=======
  "v": 2
>>>>>>> main
}
EOF
git -C "$R" add data/canon.json
expect NEEDLE 2 "conflict markers in generated JSON are refused" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# CONTROL: valid JSON passes.
R="$(new_repo c3)"
mkdir -p "$R/data"
printf '{"a": 1, "b": [2, 3]}\n' > "$R/data/good.json"
git -C "$R" add data/good.json
expect CONTROL 0 "valid staged JSON passes (and was actually parsed)" \
    --out "1 staged .json file(s)" -- \
    bash -c "cd '$R' && '$GUARD' check-staged --json-path 'data/**/*.json'"

# CONTROL: an explicit exemption works (and is reported, never silent).
R="$(new_repo c4)"
mkdir -p "$R/fixtures"
printf 'not json at all' > "$R/fixtures/malformed.json"
git -C "$R" add fixtures/malformed.json
expect CONTROL 0 "deliberate malformed fixture can be exempted (exemption announced)" \
    --out "EXEMPTED by --allow-invalid-json" -- \
    bash -c "cd '$R' && '$GUARD' check-staged --json-path 'fixtures/*' --allow-invalid-json 'fixtures/*'"

# CONTROL (false-positive regression): JSONC is the common real-world case.
# tsconfig.json legally carries comments and does NOT parse strictly. Measured on
# a real repo: 17 of 119 tracked .json files are like this. An on-by-default
# check would fire on all of them; scoping by --json-path is what prevents it.
R="$(new_repo c5)"
cat > "$R/tsconfig.json" <<'EOF'
{
  // JSONC: comments are legal here and strict json.load() rejects them.
  "compilerOptions": { "strict": true }
}
EOF
mkdir -p "$R/data"
printf '{"generated": true}\n' > "$R/data/view.json"
git -C "$R" add tsconfig.json data/view.json
expect CONTROL 0 "JSONC outside --json-path does not trip the JSON check" \
    --out "1 staged .json file(s)" -- \
    bash -c "cd '$R' && '$GUARD' check-staged --json-path 'data/**/*.json'"

# NEEDLE: the same tsconfig DOES fail if someone wrongly nominates it, proving
# the check is genuinely running and the control above is not a vacuous pass.
expect NEEDLE 3 "the same JSONC file fails when explicitly nominated" -- \
    bash -c "cd '$R' && '$GUARD' check-staged --json-path 'tsconfig.json'"

echo "=== (d) the JSON decision is MANDATORY (.push-guard.json) ==="

# NEEDLE: the fail-open this gate closes. No --json-path, no config: the previous
# release printed "JSON check NOT REQUESTED" and exited 0, leaving the repo
# unprotected forever with nothing ever failing.
R="$(new_repo d1)"
printf '{"a": 1}\n' > "$R/thing.json"
git -C "$R" add thing.json
expect NEEDLE 6 "no --json-path and no config is REFUSED (the closed fail-open)" \
    --out "NO JSON DECISION" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# NEEDLE: config nominates paths, and the check genuinely runs off it.
R="$(new_repo d2)"
write_config "$R" '{"json_paths": ["data/**/*.json"]}'
mkdir -p "$R/data"
printf '{"a": 1,\n' > "$R/data/broken.json"
git -C "$R" add data/broken.json
expect NEEDLE 3 "broken JSON is caught via config-supplied json_paths" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# NEEDLE: opt-out WITHOUT a reason. An unexplained opt-out is the absence again,
# merely relocated into a file.
R="$(new_repo d3)"
write_config "$R" '{"json_check": "none"}'
printf 'code\n' > "$R/x.txt"
git -C "$R" add x.txt
expect NEEDLE 6 "json_check:none without a reason is refused" \
    --out "REQUIRES a non-empty" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# NEEDLE: malformed config must REFUSE, never fall back to "treat as absent".
# That fallback would mean a typo silently disables the check the file enables.
R="$(new_repo d4)"
write_config "$R" '{"json_paths": ["data/**/*.json",'
printf 'code\n' > "$R/x.txt"
git -C "$R" add x.txt
expect NEEDLE 6 "an unparseable config is refused, not treated as absent" \
    --out "INVALID" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# NEEDLE: a config that parses but states NO decision. Distinct from malformed —
# this one is valid JSON and still says nothing, which is the original defect
# wearing a config file as a disguise.
R="$(new_repo d5)"
write_config "$R" '{}'
printf 'code\n' > "$R/x.txt"
git -C "$R" add x.txt
expect NEEDLE 6 "a valid config that states no decision is refused" \
    --out "states no decision" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# NEEDLE: a bogus json_check value must not be read as an opt-out.
R="$(new_repo d6)"
write_config "$R" '{"json_check": "off", "reason": "typo for none"}'
printf 'code\n' > "$R/x.txt"
git -C "$R" add x.txt
expect NEEDLE 6 'json_check with an unrecognised value is refused' \
    --out 'must be "none"' -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# CONTROL: config nominates paths, JSON is clean.
# --out is REQUIRED. Exit 0 alone cannot distinguish "read the config, resolved
# the paths, parsed the files" from "matched nothing and had nothing to do" —
# the vacuous-control trap that already caught this harness once.
R="$(new_repo d7)"
write_config "$R" '{"json_paths": ["data/**/*.json"]}'
mkdir -p "$R/data"
printf '{"generated": true}\n' > "$R/data/view.json"
git -C "$R" add data/view.json
expect CONTROL 0 "config-supplied json_paths pass and are reported as parsed" \
    --out "1 staged .json file(s)" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# CONTROL: a recorded opt-out passes AND the reason is echoed. Asserting on the
# reason is the whole point — an opt-out nobody can see is what we just removed.
R="$(new_repo d8)"
write_config "$R" '{"json_check": "none", "reason": "no generated JSON in this repo"}'
printf 'code\n' > "$R/x.txt"
git -C "$R" add x.txt
expect CONTROL 0 "a recorded opt-out passes and PRINTS its reason" \
    --out "recorded reason: no generated JSON in this repo" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# CONTROL: config-supplied exemptions still work through the config path.
R="$(new_repo d9)"
write_config "$R" '{"json_paths": ["fixtures/*"], "allow_invalid_json": ["fixtures/*"]}'
mkdir -p "$R/fixtures"
printf 'not json at all' > "$R/fixtures/malformed.json"
git -C "$R" add fixtures/malformed.json
expect CONTROL 0 "config-supplied allow_invalid_json exempts (and announces it)" \
    --out "EXEMPTED by --allow-invalid-json" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

# CONTROL: --json-path alone still satisfies the decision, with no config at all.
# This is what keeps every pre-existing caller working: nominating a path IS an
# explicit decision. Only saying nothing is refused.
R="$(new_repo d10)"
mkdir -p "$R/data"
printf '{"a": 1}\n' > "$R/data/good.json"
git -C "$R" add data/good.json
expect CONTROL 0 "--json-path alone satisfies the decision with no config present" \
    --out "1 staged .json file(s)" -- \
    bash -c "cd '$R' && '$GUARD' check-staged --json-path 'data/**/*.json'"

# NEEDLE: a CLI nomination must WIN over a config opt-out, loudly. Resolving a
# contradiction toward more checking is the only safe direction, but doing it
# silently would hide a genuine disagreement between caller and repo.
R="$(new_repo d11)"
write_config "$R" '{"json_check": "none", "reason": "claims no generated JSON"}'
mkdir -p "$R/data"
printf '{"a": 1,\n' > "$R/data/broken.json"
git -C "$R" add data/broken.json
expect NEEDLE 3 "--json-path overrides a config opt-out and still catches bad JSON" \
    --out "honouring the nomination" -- \
    bash -c "cd '$R' && '$GUARD' check-staged --json-path 'data/**/*.json'"

echo "=== (z) the guard itself emits no stray output ==="

# CONTROL: the guard must not print interpreter warnings.
#
# This case exists because a real one shipped: the config parser was an inline
# `<<'PY'` heredoc inside a $( ) command substitution, so bash printed
#   warning: command substitution: 1 unterminated here-document
# on EVERY run. Thirty needles and a clean shellcheck all missed it — the
# warning went to stderr, and no assertion in this harness looked at output it
# had not already been told to expect. It was found only by running the guard
# against a real repository.
#
# The lesson is narrow and worth encoding: asserting on what you EXPECT to see
# cannot detect what you never thought to look for. This asserts on the absence
# of a whole output class instead.
R="$(new_repo z1)"
write_config "$R" '{"json_paths": ["data/**/*.json"]}'
mkdir -p "$R/data"
printf '{"a": 1}\n' > "$R/data/view.json"
git -C "$R" add data/view.json
z_out="$(cd "$R" && "$GUARD" check-staged 2>&1)"
if grep -qiE 'warning:|unterminated|command substitution' <<<"$z_out"; then
    printf '  FAIL  [%-7s] %s\n' "CONTROL" "guard emits interpreter warnings"
    printf '%s\n' "$z_out" | sed 's/^/          | /'
    FAIL=$((FAIL + 1))
else
    printf '  PASS  [%-7s] %s\n' "CONTROL" "guard runs without emitting any interpreter warning"
    PASS=$((PASS + 1))
fi

# NEEDLE for the case above: prove the detector can actually fire, otherwise it
# is one more assertion that passes because it looked at nothing.
if grep -qiE 'warning:|unterminated|command substitution' \
        <<<"bash: warning: command substitution: 1 unterminated here-document"; then
    printf '  PASS  [%-7s] %s\n' "NEEDLE" "the warning detector fires on a known warning string"
    PASS=$((PASS + 1))
else
    printf '  FAIL  [%-7s] %s\n' "NEEDLE" "the warning detector is dead — it cannot fire"
    FAIL=$((FAIL + 1))
fi

echo "=== null case: nothing staged ==="

# NEEDLE: the index equals HEAD. Committing here produces the empty commit.
R="$(new_repo n1)"
expect NEEDLE 5 "empty index is refused before a commit happens" -- \
    bash -c "cd '$R' && '$GUARD' check-staged"

echo "=== (b) push actually happened ==="

# CONTROL: a real push that moves the remote.
R="$(new_repo b1)"
printf 'work\n' > "$R/work.txt"
git -C "$R" add work.txt && git -C "$R" commit -q -m "real work"
expect CONTROL 0 "a push that moves the remote is confirmed" \
    --out "PUSH CONFIRMED" -- \
    bash -c "cd '$R' && '$GUARD' push --remote origin --branch main"

# NEEDLE: THE ORIGINAL FALSE POSITIVE. Remote already equals local HEAD, so the
# naive 'remote == local' assertion succeeds while nothing is transferred.
R="$(new_repo b2)"
printf 'work\n' > "$R/work.txt"
git -C "$R" add work.txt && git -C "$R" commit -q -m "real work"
git -C "$R" push -q origin HEAD:refs/heads/main
expect NEEDLE 4 "second push transferring nothing is refused (remote did not move)" -- \
    bash -c "cd '$R' && '$GUARD' push --remote origin --branch main"

# NEEDLE: the commit step aborted, so HEAD never advanced.
R="$(new_repo b3)"
HEAD_BEFORE="$(git -C "$R" rev-parse HEAD)"
expect NEEDLE 5 "push after an aborted commit is refused (HEAD did not advance)" -- \
    bash -c "cd '$R' && '$GUARD' push --remote origin --branch main --since-head '$HEAD_BEFORE'"

# NEEDLE: an empty commit carrying a real message.
R="$(new_repo b4)"
git -C "$R" commit -q --allow-empty -m "feat: important-sounding message"
expect NEEDLE 5 "pushing an empty commit is refused" -- \
    bash -c "cd '$R' && '$GUARD' push --remote origin --branch main"

# CONTROL: --since-head is satisfied when a real commit was made.
R="$(new_repo b5)"
HEAD_BEFORE="$(git -C "$R" rev-parse HEAD)"
printf 'work\n' > "$R/work.txt"
git -C "$R" add work.txt && git -C "$R" commit -q -m "real work"
# --out is required here. Without it this control is VACUOUS: deleting the whole
# --since-head validation block would still yield exit 0, because the empty-commit
# and remote-moved checks independently succeed. It would prove nothing about the
# code path it names.
expect CONTROL 0 "--since-head passes when a real commit was created" \
    --out "HEAD advanced" -- \
    bash -c "cd '$R' && '$GUARD' push --remote origin --branch main --since-head '$HEAD_BEFORE'"

echo
printf 'push-guard needles: %d passed, %d failed\n' "$PASS" "$FAIL"
(( FAIL == 0 )) || exit 1