#!/usr/bin/env bash
# check-cargo-patches.sh — verify all [patch.crates-io] and [patch."<URL>"]
# sections in local sources' Cargo.toml resolve to the expected paths.
#
# Scans local/sources/*/Cargo.toml for [patch.*] sections and runs
# `cargo metadata` on each source to confirm the patch is applied
# (i.e. the resolved source URL matches the expected local path).
#
# This is the Phase J / Improvement C verification step. The cookbook's
# own patch validation handles file-level patches; this script handles
# the Cargo-level [patch] sections which are needed for transitive
# deps like redox_syscall, libredox, etc.
#
# Usage: ./local/scripts/check-cargo-patches.sh [--strict]
#   --strict: exit non-zero if any patch fails verification (for CI)

set -euo pipefail

ROOT="$(cd "$(dirname "$0")/../.." && pwd)"
STRICT=false
[[ "${1:-}" == "--strict" ]] && STRICT=true

PASSED=0
FAILED=0
SKIPPED=0

echo "=== Cargo [patch] Verification ==="

# Find all Cargo.toml files in local sources
# Skip relibc explicitly — its [patch] section is only the
# cc-rs git branch override (no `path` patches), and
# `cargo metadata` on relibc takes minutes (hundreds of
# deps) which would hang the script.
while IFS= read -r -d '' cargo_toml; do
    case "$cargo_toml" in
        *local/sources/relibc/Cargo.toml) continue ;;
    esac
    source_dir="$(dirname "$cargo_toml")"
    source_name="$(basename "$source_dir")"
    relative_to_root="${cargo_toml#$ROOT/}"

    # Find [patch.crates-io] and [patch."<URL>"] sections
    patch_lines=$(grep -E '^\[patch\.' "$cargo_toml" 2>/dev/null || true)
    if [[ -z "$patch_lines" ]]; then
        continue
    fi

    echo ""
    echo "--- $relative_to_root ---"
    echo "  $patch_lines" | sed 's/^/    /'

    # Check if the source is a workspace member
    if ! grep -q '^\[workspace\]' "$cargo_toml" 2>/dev/null; then
        # Single-package — [patch] sections need [workspace] to apply
        # (cargo treats them as ignored). This is a common bug.
        if grep -q '^\[patch\.' "$cargo_toml" 2>/dev/null; then
            echo "  ⚠️  WARN: source has [patch] sections but no [workspace]"
            echo "      cargo silently ignores [patch] in non-workspace manifests"
            SKIPPED=$((SKIPPED + 1))
        fi
        continue
    fi

    # Check if cargo is available
    if ! command -v cargo >/dev/null 2>&1; then
        echo "  SKIP: cargo not available"
        SKIPPED=$((SKIPPED + 1))
        continue
    fi

    # Run cargo metadata and parse the resolved source URLs.
    # Use a 30-second timeout to prevent hangs on large
    # workspaces like relibc (which has hundreds of deps).
    if ! timeout 30 cargo metadata --format-version 1 --manifest-path "$cargo_toml" >/dev/null 2>&1; then
        echo "  SKIP: cargo metadata timed out (30s)"
        SKIPPED=$((SKIPPED + 1))
        continue
    fi

    # Extract the resolved source URLs and check against the
    # expected [patch] paths. The expected [patch] entries have
    # a `path = "..."` field; the resolved source URL from
    # cargo metadata should match the local path.
    expected_paths=$(grep -A1 '^\[patch\.' "$cargo_toml" 2>/dev/null | grep -oP 'path\s*=\s*"\K[^"]+' | sort -u)
    if [[ -z "$expected_paths" ]]; then
        # No `path` entries in the [patch] sections. This is
        # common for git-branch patches (e.g. relibc's cc-rs).
        # The verification only applies to `path` patches.
        # Continue to next file.
        continue
    fi

    resolved_urls=$(cargo metadata --format-version 1 --manifest-path "$cargo_toml" 2>/dev/null \
        | python3 -c '
import json, sys
data = json.load(sys.stdin)
# For path-deps, source is null but manifest_path points to
# the local fork. Print both forms.
for pkg in data.get("packages", []):
    src = pkg.get("source", "")
    if src:
        print(src)
    mp = pkg.get("manifest_path", "")
    if mp:
        print(mp)
' 2>/dev/null | sort -u)

    good=0
    bad=0
    # Skip the resolution loop if there are no `path` entries.
    # (Empty `expected_paths` would loop with empty lines.)
    if [[ -n "$expected_paths" ]]; then
        while IFS= read -r expected; do
            # Skip empty lines (the heredoc may emit one)
            if [[ -z "$expected" ]]; then
                continue
            fi
            # expected is a path like "../syscall" or "../libredox"
            # Resolve to absolute
            expected_abs="$(cd "$source_dir" && realpath "$expected" 2>/dev/null || echo "")"
            if [[ -z "$expected_abs" ]]; then
                continue
            fi
            if echo "$resolved_urls" | grep -qF "$expected_abs"; then
                good=$((good + 1))
            else
                # Check if the expected path is the source itself
                if [[ "$expected_abs" == "$source_dir" ]]; then
                    # self-patch is OK
                    good=$((good + 1))
                else
                    bad=$((bad + 1))
                    echo "  ⚠️  [patch] path $expected → $expected_abs NOT in resolved sources"
                fi
            fi
        done <<< "$expected_paths"
    fi

    if [[ $bad -eq 0 ]]; then
        echo "  ✅ All $good [patch] entries resolved correctly"
        PASSED=$((PASSED + 1))
    else
        echo "  ❌ $bad [patch] entries did not resolve"
        FAILED=$((FAILED + 1))
    fi
done < <(find "$ROOT/local/sources" -name 'Cargo.toml' -print0 2>/dev/null)

echo ""
echo "=== Summary ==="
echo "  Passed: $PASSED"
echo "  Failed: $FAILED"
echo "  Skipped: $SKIPPED"
if [[ $FAILED -gt 0 ]]; then
    echo ""
    echo "❌ Cargo [patch] verification FAILED"
    $STRICT && exit 1
    exit 1
fi
echo ""
echo "✅ Cargo [patch] verification PASSED"
exit 0