#!/usr/bin/env bash
# Regression harness for pr-review.sh's -r/--repo and -H/--host overrides,
# the -r repo-exists preflight, and the mosaic-pr-review User-Agent header
# (patch family 5/5c, part of #891).
#
# -r/--repo and -H/--host skip git-remote slug/host inference entirely, for
# reviewer worktrees whose origin is nonstandard, wrong, or missing (mirrors
# the -r convention of the sibling wrappers pr-view.sh/pr-diff.sh/pr-ci-wait.sh,
# #867). When -r is given, the resolved repo is preflighted (GET
# .../repos/<slug>) BEFORE any write, so a wrong-host cross-wire surfaces as a
# clear preflight error instead of an opaque write-404. Every Gitea curl call
# (preflight, write, read-back, /user, PR-head) must carry a
# `User-Agent: mosaic-pr-review` header, since some Cloudflare-fronted Gitea
# hosts intermittently reject curl's default User-Agent.
#
# Round 2 (post-review): the approve/request-changes/comment dispatch bodies
# each independently called a BARE `host=$(get_remote_host)` (only for a
# best-effort tea-login guess) that ignored -H/--host entirely and, under
# `set -e`, DIED SILENTLY (exit 1, zero stdout/stderr) whenever no git origin
# existed — exactly the case -r/-H exist to support. The "no git origin at
# all" fixture below must be a directory TRULY outside any .git ancestry (a
# subdirectory with no .git of its own still resolves `git remote get-url
# origin` UPWARD to an enclosing repo's origin — a real flaw in the first
# version of this fixture, since the test tree itself sits inside the
# mosaicstack/stack checkout). It is created via `mktemp -d` under the SYSTEM
# /tmp and additionally bounded with GIT_CEILING_DIRECTORIES, so `git rev-parse
# --show-toplevel`/`git remote get-url origin` cannot walk past it under any
# circumstance.

set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
WORK_DIR="${MOSAIC_TEST_WORK_DIR:-$PWD/.mosaic-test-work/pr-review-repo-host-override}"
BIN_DIR="$WORK_DIR/bin"
STATE_DIR="$WORK_DIR/state"
CURL_LOG="$WORK_DIR/curl.log"
UA_LOG="$WORK_DIR/ua.log"
OUTPUT_FILE="$WORK_DIR/output.log"
TMP_SCRATCH="$WORK_DIR/scratch"
CREDENTIALS_FILE="$WORK_DIR/credentials.json"
# A TRUE no-git-origin fixture: created under the system /tmp (NOT under
# $WORK_DIR, which sits inside this very checkout's .git ancestry), so that
# `git remote get-url origin` run from inside it cannot resolve upward to the
# mosaicstack/stack checkout's own real origin.
NO_GIT_DIR="$(mktemp -d "${TMPDIR:-/tmp}/pr-review-no-origin.XXXXXX")"

cleanup() {
    rm -rf "$WORK_DIR" "$NO_GIT_DIR"
}
trap cleanup EXIT

HOST_OVERRIDE_VAL="pr-review-override.test"
REPO_OVERRIDE_VAL="acme/widgets"
TOKEN_VAL="override-token-xyz"
ACTING_LOGIN="override-bot"
HEAD_SHA_VAL="feedfacecafebabe0000000000000000000000"

mkdir -p "$BIN_DIR" "$STATE_DIR" "$TMP_SCRATCH"

# tea stub: deterministic tea-absent behavior regardless of whether the host
# actually has a real `tea` binary on PATH. -r/--repo + -H/--host must not
# NEED a resolvable tea login to work — the wrapper falls back to the
# host-scoped GITEA_TOKEN credential when no tea login resolves.
cat > "$BIN_DIR/tea" <<'SH'
#!/usr/bin/env bash
exit 1
SH
chmod +x "$BIN_DIR/tea"

# curl stub: models the target Gitea instance for the OVERRIDDEN host/repo
# only. Any request to a DIFFERENT host/repo (i.e. one derived from git-remote
# inference instead of the override) is unexpected and fails the run.
cat > "$BIN_DIR/curl" <<SH
#!/usr/bin/env bash
set -euo pipefail

output_file=""
method="GET"
payload=""
url=""
config_file=""
ua_seen="0"
while [[ \$# -gt 0 ]]; do
    case "\$1" in
        -o) output_file="\$2"; shift 2 ;;
        -H)
            [[ "\$2" == "User-Agent: mosaic-pr-review" ]] && ua_seen="1"
            shift 2 ;;
        -K|--config) config_file="\$2"; shift 2 ;;
        -w) shift 2 ;;
        -X) shift 2 ;;
        -d|--data) payload="\$2"; shift 2 ;;
        -s|-S|-sS) shift ;;
        http://*|https://*) url="\$1"; shift ;;
        *) shift ;;
    esac
done

printf '%s %s\n' "\$method" "\$url" >> "$CURL_LOG"
printf '%s %s\n' "\$ua_seen" "\$url" >> "$UA_LOG"

write_response() {
    local status="\$1" body="\$2"
    [[ -n "\$output_file" ]] || exit 96
    printf '%s' "\$body" > "\$output_file"
    printf '%s' "\$status"
}

mode="\$(cat "$STATE_DIR/mode" 2>/dev/null || true)"
web_base="https://$HOST_OVERRIDE_VAL"
repo="$REPO_OVERRIDE_VAL"

case "\$url" in
    "https://$HOST_OVERRIDE_VAL/api/v1/repos/$REPO_OVERRIDE_VAL")
        if [[ "\$mode" == "repo-missing" ]]; then
            write_response 404 '{"message":"not found"}'
        else
            write_response 200 "{\\"full_name\\":\\"$REPO_OVERRIDE_VAL\\"}"
        fi
        ;;
    "https://$HOST_OVERRIDE_VAL/api/v1/user")
        write_response 200 "{\\"login\\":\\"$ACTING_LOGIN\\"}"
        ;;
    "https://$HOST_OVERRIDE_VAL/api/v1/repos/$REPO_OVERRIDE_VAL/pulls/123")
        write_response 200 "{\\"head\\":{\\"sha\\":\\"$HEAD_SHA_VAL\\"}}"
        ;;
    "https://$HOST_OVERRIDE_VAL/api/v1/repos/$REPO_OVERRIDE_VAL/pulls/123/reviews")
        PR_REVIEW_PAYLOAD="\$payload" PR_REVIEW_ACTING="$ACTING_LOGIN" \
            python3 -c '
import json, os
p = json.loads(os.environ["PR_REVIEW_PAYLOAD"])
record = {
    "id": 501,
    "state": p.get("event"),
    "commit_id": p.get("commit_id"),
    "body": p.get("body"),
    "user": {"login": os.environ["PR_REVIEW_ACTING"]},
}
with open("$STATE_DIR/review.json", "w", encoding="utf-8") as fh:
    json.dump(record, fh)
'
        write_response 201 "\$(cat "$STATE_DIR/review.json")"
        ;;
    "https://$HOST_OVERRIDE_VAL/api/v1/repos/$REPO_OVERRIDE_VAL/pulls/123/reviews/501")
        write_response 200 "\$(cat "$STATE_DIR/review.json")"
        ;;
    "https://$HOST_OVERRIDE_VAL/api/v1/repos/$REPO_OVERRIDE_VAL/issues/123/comments")
        write_response 201 "{\\"id\\":456,\\"body\\":\\"durable-body\\",\\"user\\":{\\"login\\":\\"$ACTING_LOGIN\\"},\\"issue_url\\":\\"\\",\\"pull_request_url\\":\\"\${web_base}/\${repo}/pulls/123\\"}"
        ;;
    "https://$HOST_OVERRIDE_VAL/api/v1/repos/$REPO_OVERRIDE_VAL/issues/comments/456")
        write_response 200 "{\\"id\\":456,\\"body\\":\\"durable-body\\",\\"user\\":{\\"login\\":\\"$ACTING_LOGIN\\"},\\"issue_url\\":\\"\\",\\"pull_request_url\\":\\"\${web_base}/\${repo}/pulls/123\\"}"
        ;;
    *)
        echo "Unexpected curl request (host/repo not from the override — inference leaked through?): \$method \$url" >&2
        exit 97
        ;;
esac
SH
chmod +x "$BIN_DIR/curl"

run() {
    local repo_dir="$1"; shift
    : > "$CURL_LOG"
    : > "$UA_LOG"
    : > "$OUTPUT_FILE"
    (
        cd "$repo_dir"
        PATH="$BIN_DIR:$PATH" \
        TMPDIR="$TMP_SCRATCH" \
        MOSAIC_CREDENTIALS_FILE="$CREDENTIALS_FILE" \
        GITEA_TOKEN="$TOKEN_VAL" \
        GITEA_URL="https://$HOST_OVERRIDE_VAL" \
        GIT_CEILING_DIRECTORIES="$(dirname "$NO_GIT_DIR")" \
        "$SCRIPT_DIR/pr-review.sh" "$@"
    ) > "$OUTPUT_FILE" 2>&1
}

set_mode() {
    printf '%s' "$1" > "$STATE_DIR/mode"
}

# A run must not silently die: exit 1 with a COMPLETELY EMPTY combined
# stdout+stderr is precisely the pre-fix signature of the dispatch-level bare
# `host=$(get_remote_host)` call dying under `set -e` when there was no git
# origin at all. Whatever the outcome, the wrapper must always emit SOME
# diagnostic (success line, or an "Error: ..." line).
assert_not_silently_dead() {
    local context="$1" exit_code="$2"
    if [[ "$exit_code" -ne 0 && ! -s "$OUTPUT_FILE" ]]; then
        echo "FAIL: pr-review.sh died SILENTLY (exit $exit_code, zero output) in a true no-origin dir ($context)" >&2
        exit 1
    fi
}

# Sanity-check the fixture itself BEFORE relying on it: from inside NO_GIT_DIR,
# `git remote get-url origin` must fail outright — proving this directory is
# truly outside any .git ancestry (the flaw this harness had before: a bare
# subdirectory of the checkout resolves `origin` UPWARD to the checkout's own
# real remote instead of failing).
if (cd "$NO_GIT_DIR" && GIT_CEILING_DIRECTORIES="$(dirname "$NO_GIT_DIR")" git remote get-url origin) 2>/dev/null; then
    echo "FAIL: NO_GIT_DIR fixture is not actually outside any git ancestry — 'git remote get-url origin' resolved upward" >&2
    exit 1
fi

# --- Case 1: -r/--repo and -H/--host are recognized flags (regression lock on
# the previous "Unknown option: -r" / "Unknown option: -H" failure). Use a
# bogus ACTION so the run fails cheaply, WITHOUT any network I/O, at the
# platform-dispatch unknown-action branch rather than at argument parsing —
# proving both flags were consumed as options, not rejected.
set_mode "repo-ok"
if run "$NO_GIT_DIR" -n 123 -a bogus-action -r "$REPO_OVERRIDE_VAL" -H "$HOST_OVERRIDE_VAL"; then
    echo "FAIL: bogus action unexpectedly succeeded" >&2
    cat "$OUTPUT_FILE" >&2
    exit 1
fi
if grep -q 'Unknown option' "$OUTPUT_FILE"; then
    echo "FAIL: -r/--repo or -H/--host was not recognized as a valid flag" >&2
    cat "$OUTPUT_FILE" >&2
    exit 1
fi
grep -q 'Unknown action: bogus-action' "$OUTPUT_FILE"

# --- Case 2: -h/--help documents both overrides.
HELP_TEXT="$("$SCRIPT_DIR/pr-review.sh" -h)"
echo "$HELP_TEXT" | grep -q -- '-r, --repo'
echo "$HELP_TEXT" | grep -q -- '-H, --host'

# --- Case 3 (comment): a TRUE no-git-origin dir + -r/-H must not silently die
# and must not fail with "not a git repository or no origin remote" either.
set_mode "repo-ok"
comment_rc=0
run "$NO_GIT_DIR" -n 123 -a comment -c durable-body -r "$REPO_OVERRIDE_VAL" -H "$HOST_OVERRIDE_VAL" || comment_rc=$?
assert_not_silently_dead "comment" "$comment_rc"
if grep -qi 'not a git repository or no origin remote' "$OUTPUT_FILE"; then
    echo "FAIL: -r/--repo did not tolerate a missing git origin" >&2
    cat "$OUTPUT_FILE" >&2
    exit 1
fi
grep -q 'Added and verified comment on Gitea PR #123' "$OUTPUT_FILE"

# --- Case 3b (approve): same true no-origin dir, the `approve` dispatch body
# (its OWN, previously-unguarded `host=$(get_remote_host)` call) must not
# silently die either.
set_mode "repo-ok"
approve_rc=0
run "$NO_GIT_DIR" -n 123 -a approve -r "$REPO_OVERRIDE_VAL" -H "$HOST_OVERRIDE_VAL" || approve_rc=$?
assert_not_silently_dead "approve" "$approve_rc"
if grep -qi 'not a git repository or no origin remote' "$OUTPUT_FILE"; then
    echo "FAIL: approve's -r/--repo did not tolerate a missing git origin" >&2
    cat "$OUTPUT_FILE" >&2
    exit 1
fi
grep -q 'Approved and verified Gitea PR #123 (review ID 501)' "$OUTPUT_FILE"

# --- Case 3c (request-changes): same true no-origin dir, the
# `request-changes` dispatch body's OWN previously-unguarded
# `host=$(get_remote_host)` call must not silently die either.
set_mode "repo-ok"
request_changes_rc=0
run "$NO_GIT_DIR" -n 123 -a request-changes -c please-fix -r "$REPO_OVERRIDE_VAL" -H "$HOST_OVERRIDE_VAL" || request_changes_rc=$?
assert_not_silently_dead "request-changes" "$request_changes_rc"
if grep -qi 'not a git repository or no origin remote' "$OUTPUT_FILE"; then
    echo "FAIL: request-changes's -r/--repo did not tolerate a missing git origin" >&2
    cat "$OUTPUT_FILE" >&2
    exit 1
fi
grep -q 'Requested changes and verified on Gitea PR #123 (review ID 501)' "$OUTPUT_FILE"

# --- Case 4: -r/--repo and -H/--host WIN over git-remote inference, not just
# tolerate its absence — set up a real git repo whose origin points at a
# COMPLETELY DIFFERENT host/repo and confirm the override, not the remote, is
# what gets used (the curl stub only answers for the override host/repo; any
# request derived from the wrong remote is an unexpected request and fails
# the stub with exit 97).
WRONG_REMOTE_DIR="$WORK_DIR/wrong-remote-repo"
mkdir -p "$WRONG_REMOTE_DIR"
git -C "$WRONG_REMOTE_DIR" init -q
git -C "$WRONG_REMOTE_DIR" remote add origin https://git.wrong-inferred-host.example/wrong/repo.git
set_mode "repo-ok"
run "$WRONG_REMOTE_DIR" -n 123 -a comment -c durable-body -r "$REPO_OVERRIDE_VAL" -H "$HOST_OVERRIDE_VAL"
grep -q 'Added and verified comment on Gitea PR #123' "$OUTPUT_FILE"
if grep -q 'wrong-inferred-host' "$CURL_LOG"; then
    echo "FAIL: -r/--repo + -H/--host did not override git-remote inference" >&2
    cat "$CURL_LOG" >&2
    exit 1
fi

# --- Case 5 (5c): the repo-exists preflight runs BEFORE any write and fails
# closed with a clear diagnostic when the resolved repo is unreachable at the
# resolved host — proving a wrong-host cross-wire surfaces here, not as an
# opaque write-404.
set_mode "repo-missing"
if run "$NO_GIT_DIR" -n 123 -a comment -c durable-body -r "$REPO_OVERRIDE_VAL" -H "$HOST_OVERRIDE_VAL"; then
    echo "FAIL: preflight did not fail closed on a missing repo" >&2
    cat "$OUTPUT_FILE" >&2
    exit 1
fi
grep -q "repo '$REPO_OVERRIDE_VAL' not reachable" "$OUTPUT_FILE"
if grep -q '/issues/123/comments' "$CURL_LOG"; then
    echo "FAIL: wrapper posted a comment despite a failed repo preflight" >&2
    cat "$CURL_LOG" >&2
    exit 1
fi

# --- Case 6 (5): every Gitea curl call (preflight, /user, comment write,
# comment read-back) carries the User-Agent: mosaic-pr-review header.
set_mode "repo-ok"
run "$NO_GIT_DIR" -n 123 -a comment -c durable-body -r "$REPO_OVERRIDE_VAL" -H "$HOST_OVERRIDE_VAL"
grep -q 'Added and verified comment on Gitea PR #123' "$OUTPUT_FILE"
call_count="$(wc -l < "$CURL_LOG" | tr -d ' ')"
ua_count="$(awk '$1 == "1"' "$UA_LOG" | wc -l | tr -d ' ')"
if [[ "$call_count" -lt 4 ]]; then
    echo "FAIL: expected at least 4 curl calls (preflight, /user, write, read-back), got $call_count" >&2
    cat "$CURL_LOG" >&2
    exit 1
fi
if [[ "$ua_count" != "$call_count" ]]; then
    echo "FAIL: not every curl call carried User-Agent: mosaic-pr-review ($ua_count/$call_count)" >&2
    cat "$UA_LOG" >&2
    exit 1
fi

echo "pr-review.sh -r/--repo + -H/--host override + preflight + User-Agent regression passed"