#!/usr/bin/env bash
# repair-archive-format.sh — Normalize unversioned/old-format archives to standard format.
#
# Standard format: {cat}-{pkg}-v{version}-patched.tar.gz containing:
#   source/       — recipe source tree
#   recipe.toml   — recipe metadata
#
# Repairs archives that:
#   1. Have raw repo content at root (not inside source/ dir)
#   2. Missing recipe.toml
#   3. Include build artifacts (source/target/)
#
# Usage:
#   ./local/scripts/repair-archive-format.sh [--dry-run]

set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
PROJECT_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
POOL="$PROJECT_ROOT/sources/x86_64-unknown-redox"
WORK="/tmp/redbear-archive-repair-$$"
DRY_RUN=0

[ "${1:-}" = "--dry-run" ] && DRY_RUN=1

GREEN='\033[1;32m'
YELLOW='\033[1;33m'
NC='\033[0m'

mkdir -p "$WORK"

repaired=0
skipped=0
for archive in "$POOL"/*.tar.gz; do
    name=$(basename "$archive")
    
    # Skip already properly versioned (has -v HEX or -v DIGITS)
    if echo "$name" | grep -qE '\-v[a-f0-9]{7,}|\-v\d+\.'; then
        skipped=$((skipped + 1))
        continue
    fi
    
    # Extract category and package from filename: {cat}-{pkg}-*.tar.gz
    # For unversioned: core-base.tar.gz -> cat=core, pkg=base
    # For vunknown: core-base-vunknown-patched.tar.gz -> cat=core, pkg=base
    base="${name%.tar.gz}"
    base="${base%-patched}"
    base="${base%-vunknown}"
    cat_name=$(echo "$base" | cut -d- -f1)
    pkg_name=$(echo "$base" | cut -d- -f2-)
    
    # Find the recipe.toml
    recipe=""
    for candidate in \
        "$PROJECT_ROOT/recipes/$cat_name/$pkg_name/recipe.toml" \
        "$PROJECT_ROOT/local/recipes/$cat_name/$pkg_name/recipe.toml"; do
        [ -f "$candidate" ] && recipe="$candidate" && break
    done
    
    if [ -z "$recipe" ]; then
        echo -e "${YELLOW}SKIP${NC} $name (no recipe.toml found in recipes/)"
        continue
    fi
    
    # Determine version
    version=$(python3 -c "
import tomllib, subprocess
from pathlib import Path
with open('$recipe', 'rb') as f:
    data = tomllib.load(f)
src = data.get('source', {})
if not isinstance(src, dict):
    print('unknown')
elif 'tar' in src:
    import re
    m = re.search(r'/?v?(\d+\.\d+(?:\.\d+)?)(?:\.tar|/$)', src['tar'])
    print(m.group(1) if m else 'unknown')
elif 'git' in src:
    rev = src.get('rev', '')
    if rev:
        print(rev[:7])
    else:
        pkg_dir = Path('$recipe').parent / 'source'
        if (pkg_dir / '.git').exists():
            r = subprocess.run(['git','-C',str(pkg_dir),'rev-parse','--short','HEAD'],
                             capture_output=True, text=True)
            print(r.stdout.strip() if r.returncode == 0 else 'unknown')
        else:
            print('unknown')
else:
    print('unknown')
" 2>/dev/null || echo "unknown")
    
    new_name="${cat_name}-${pkg_name}-v${version}-patched.tar.gz"
    new_path="$POOL/$new_name"
    
    if [ -f "$new_path" ] && [ "$new_name" != "$name" ]; then
        echo -e "${YELLOW}SKIP${NC} $name (target $new_name already exists)"
        continue
    fi
    
    # Check current format
    has_source=$(tar tf "$archive" 2>/dev/null | grep -c '^source/' || true)
    has_recipe=$(tar tf "$archive" 2>/dev/null | grep -c '^recipe.toml$' || true)
    has_target=$(tar tf "$archive" 2>/dev/null | grep -c 'source/target/' || true)
    
    if [ "$has_source" -gt 0 ] && [ "$has_recipe" -gt 0 ] && [ "$has_target" -eq 0 ] && [ "$new_name" = "$name" ]; then
        skipped=$((skipped + 1))
        continue
    fi
    
    echo -e "${GREEN}REPAIR${NC} $name -> $new_name (v$version)"
    
    if [ "$DRY_RUN" -eq 1 ]; then
        echo "  [dry-run] would repackage: source/=$has_source recipe=$has_recipe target=$has_target"
        repaired=$((repaired + 1))
        continue
    fi
    
    # Create temp dir for repackaging
    tmpdir="$WORK/$base"
    rm -rf "$tmpdir"
    mkdir -p "$tmpdir/source"
    
    # Extract archive
    tar xzf "$archive" -C "$tmpdir" 2>/dev/null
    
    # If content is at root (not in source/), move it into source/
    if [ "$has_source" -eq 0 ]; then
        for item in "$tmpdir"/*; do
            [ "$item" = "$tmpdir/source" ] && continue
            mv "$item" "$tmpdir/source/" 2>/dev/null || true
        done
    fi
    
    # Add recipe.toml if missing
    if [ "$has_recipe" -eq 0 ]; then
        cp "$recipe" "$tmpdir/recipe.toml"
    fi
    
    # Strip build artifacts
    rm -rf "$tmpdir/source/target" 2>/dev/null || true
    
    # Create new archive
    tar czf "$new_path" -C "$tmpdir" source recipe.toml 2>/dev/null
    
    if [ "$new_name" != "$name" ]; then
        rm -f "$archive"
    fi
    
    repaired=$((repaired + 1))
done

# Update BLAKE3SUMS for repaired archives
if [ "$DRY_RUN" -eq 0 ] && [ "$repaired" -gt 0 ]; then
    (cd "$POOL" && b3sum *.tar.gz 2>/dev/null) > "$POOL/../redbear-0.1.0/BLAKE3SUMS"
fi

rm -rf "$WORK"

echo ""
echo "========================================="
echo "  Repair complete"
echo "  Repaired: $repaired"
echo "  Skipped:  $skipped"
echo "========================================="