nix-shared/pkgs/tb-dev-scripts/cmake_update_fetchcontent.py

"""
CMakeLists.txt FetchContent Dependency Updater

This script:
1. Searches upward for all CMakeLists.txt files containing FetchContent
2. Extracts current git hashes/tags/branches
3. Queries remote repositories for latest versions
4. Shows available updates and prompts for confirmation
5. Updates CMakeLists.txt files with new versions
"""

import re
import subprocess
import sys
import urllib.request
import json
from pathlib import Path
from typing import List, Optional, Tuple


def parse_semver(version_string: str) -> Tuple[int, int, int]:
    """
    Parse a semantic version string.

    Converts version strings like 'v1.2.3' or '1.2.3'
    into (major, minor, patch) tuple.
    """
    # Remove 'v' prefix if present
    clean = version_string.lstrip('v')

    # Split by dots and take first 3 components
    parts = clean.split('.')
    major = int(parts[0]) if len(parts) > 0 and parts[0].isdigit() else 0
    minor = int(parts[1]) if len(parts) > 1 and parts[1].isdigit() else 0

    # Handle patch with possible suffix (e.g., "3-rc1")
    patch = 0
    if len(parts) > 2:
        # Take only numeric part before dash
        patch_str = parts[2].split('-')[0]
        patch = int(patch_str) if patch_str.isdigit() else 0

    return (major, minor, patch)


class FetchContentDependency:
    """Represents a single FetchContent_Declare dependency"""

    def __init__(
        self, name: str, git_repo: str, git_tag: str,
        file_path: Path, line_start: int, line_end: int, full_text: str
    ):
        self.name = name
        self.git_repo = git_repo
        self.git_tag = git_tag
        self.file_path = file_path
        self.line_start = line_start
        self.line_end = line_end
        self.full_text = full_text
        self.latest_version: Optional[str] = None
        self.update_available = False

    def __repr__(self):
        return f"FetchContentDependency({self.name}, {self.git_tag})"


def extract_subdirectories(cmake_file: Path, content: str) -> List[Path]:
    """Extract add_subdirectory() calls from a CMakeLists.txt file content"""
    subdirs = []
    try:
        # Match add_subdirectory(path) with/without quotes
        pattern = r'add_subdirectory\s*\(\s*(["\']?)([^)"\']+)\1'
        for match in re.finditer(pattern, content, re.IGNORECASE):
            subdir_name = match.group(2).strip()
            # Resolve relative to the CMakeLists.txt directory
            subdir_path = (cmake_file.parent / subdir_name).resolve()
            subdirs.append(subdir_path)
    except Exception as e:
        print(
            f"Warning: Error parsing subdirectories in {cmake_file}: {e}",
            file=sys.stderr
        )
    return subdirs


def find_cmake_files_recursive(cmake_dir: Path, visited: set) -> List[Path]:
    """
    Recursively find CMakeLists.txt files.

    Follows add_subdirectory() calls to discover all CMake files.
    """
    cmake_files = []

    # Avoid infinite loops
    cmake_dir = cmake_dir.resolve()
    if cmake_dir in visited:
        return cmake_files
    visited.add(cmake_dir)

    # Check if this directory has a CMakeLists.txt
    cmake_path = cmake_dir / "CMakeLists.txt"
    if not cmake_path.exists():
        return cmake_files

    # Read the file once
    try:
        content = cmake_path.read_text()
    except Exception as e:
        print(f"Warning: Error reading {cmake_path}: {e}", file=sys.stderr)
        return cmake_files

    # Check if it contains FetchContent (if so, add to results)
    if "FetchContent" in content:
        cmake_files.append(cmake_path)

    # Always recurse into subdirectories
    # (subdirectories might have FetchContent)
    subdirs = extract_subdirectories(cmake_path, content)
    for subdir in subdirs:
        cmake_files.extend(find_cmake_files_recursive(subdir, visited))

    return cmake_files


def find_cmake_files(start_dir: Path = None) -> List[Path]:
    """
    Search upward for root CMakeLists.txt.

    Then recursively find all via add_subdirectory().
    """
    if start_dir is None:
        start_dir = Path.cwd()

    current = start_dir.resolve()

    # Search upward to find the root CMakeLists.txt
    root_cmake = None
    while True:
        cmake_path = current / "CMakeLists.txt"
        if cmake_path.exists():
            root_cmake = cmake_path

        # Stop at filesystem root
        if current.parent == current:
            break
        current = current.parent

    # If we found a root CMakeLists.txt, start recursive search from there
    if root_cmake:
        visited = set()
        files = find_cmake_files_recursive(root_cmake.parent, visited)
        return sorted(set(files))

    return []


def parse_fetchcontent(cmake_file: Path) -> List[FetchContentDependency]:
    """Parse FetchContent_Declare blocks from CMakeLists.txt"""
    content = cmake_file.read_text()
    lines = content.split('\n')

    dependencies = []

    # Pattern to match FetchContent_Declare blocks
    # Extracts name, GIT_REPOSITORY and GIT_TAG
    i = 0
    while i < len(lines):
        line = lines[i]

        # Look for FetchContent_Declare(
        if "FetchContent_Declare" in line:
            start_line = i

            # Find the closing parenthesis
            paren_count = line.count('(') - line.count(')')
            block_lines = [line]
            j = i + 1

            while j < len(lines) and paren_count > 0:
                block_lines.append(lines[j])
                paren_count += lines[j].count('(') - lines[j].count(')')
                j += 1

            end_line = j - 1
            full_block = '\n'.join(block_lines)

            # Extract dependency name
            name_match = re.search(
                r'FetchContent_Declare\s*\(\s*(\S+)', full_block
            )
            if not name_match:
                i = j
                continue

            dep_name = name_match.group(1)

            # Extract GIT_REPOSITORY
            repo_match = re.search(r'GIT_REPOSITORY\s+(\S+)', full_block)
            # Extract GIT_TAG
            tag_match = re.search(r'GIT_TAG\s+(\S+)', full_block)

            if repo_match and tag_match:
                git_repo = repo_match.group(1)
                git_tag = tag_match.group(1)

                dep = FetchContentDependency(
                    name=dep_name,
                    git_repo=git_repo,
                    git_tag=git_tag,
                    file_path=cmake_file,
                    line_start=start_line,
                    line_end=end_line,
                    full_text=full_block
                )
                dependencies.append(dep)

            i = j
        else:
            i += 1

    return dependencies


def run_git_command(args: List[str]) -> Optional[str]:
    """Run a git command and return output, or None on error"""
    try:
        result = subprocess.run(
            args,
            capture_output=True,
            text=True,
            timeout=30,
            check=True
        )
        return result.stdout.strip()
    except subprocess.CalledProcessError as e:
        print(
            f"Warning: Error running git command: {' '.join(args)}",
            file=sys.stderr
        )
        print(f"  {e.stderr}", file=sys.stderr)
        return None
    except subprocess.TimeoutExpired:
        print(
            f"Warning: Timeout running git command: {' '.join(args)}",
            file=sys.stderr
        )
        return None


def get_latest_commit_hash(repo_url: str, branch: str) -> Optional[str]:
    """Get the latest commit hash for a branch using git ls-remote"""
    output = run_git_command(
        ['git', 'ls-remote', repo_url, f'refs/heads/{branch}']
    )

    if not output:
        print(
            f"Warning: Branch '{branch}' not found in {repo_url}",
            file=sys.stderr
        )
        return None

    # Output format: "commit_hash\trefs/heads/branch"
    commit_hash = output.split()[0]
    return commit_hash


def extract_github_info(repo_url: str) -> Optional[Tuple[str, str]]:
    """Extract owner/repo from GitHub URL"""
    # Handle both https and git@ URLs
    patterns = [
        r'github\.com[:/]([^/]+)/([^/.]+?)(?:\.git)?$',
    ]

    for pattern in patterns:
        match = re.search(pattern, repo_url)
        if match:
            return match.group(1), match.group(2)

    return None


def get_latest_github_release(repo_url: str) -> Optional[str]:
    """Get the latest release tag from GitHub API"""
    github_info = extract_github_info(repo_url)
    if not github_info:
        return None

    owner, repo = github_info
    api_url = f"https://api.github.com/repos/{owner}/{repo}/releases/latest"

    try:
        req = urllib.request.Request(api_url)
        req.add_header('Accept', 'application/vnd.github.v3+json')

        with urllib.request.urlopen(req, timeout=10) as response:
            data = json.loads(response.read())
            return data.get('tag_name')
    except urllib.error.HTTPError as e:
        if e.code == 404:
            # No releases found, try tags
            return get_latest_github_tag(repo_url)
        print(
            f"Warning: Error fetching GitHub release for {owner}/{repo}: {e}",
            file=sys.stderr
        )
        return None
    except Exception as e:
        print(f"Warning: Error fetching GitHub release: {e}", file=sys.stderr)
        return None


def get_latest_github_tag(repo_url: str) -> Optional[str]:
    """Get the latest tag from GitHub API"""
    github_info = extract_github_info(repo_url)
    if not github_info:
        return None

    owner, repo = github_info
    api_url = f"https://api.github.com/repos/{owner}/{repo}/tags"

    try:
        req = urllib.request.Request(api_url)
        req.add_header('Accept', 'application/vnd.github.v3+json')

        with urllib.request.urlopen(req, timeout=10) as response:
            data = json.loads(response.read())
            if data:
                # Filter to semantic version tags and find the latest
                version_tags = []
                for tag in data:
                    tag_name = tag['name']
                    try:
                        # Try to parse as semantic version
                        v = parse_semver(tag_name)
                        version_tags.append((v, tag_name))
                    except Exception:
                        pass

                if version_tags:
                    # Sort by version and return the latest
                    version_tags.sort(reverse=True)
                    return version_tags[0][1]

                # If no semantic versions, return the first tag
                return data[0]['name']
            return None
    except Exception as e:
        print(f"Warning: Error fetching GitHub tags: {e}", file=sys.stderr)
        return None


def is_commit_hash(git_tag: str) -> bool:
    """Check if a git tag looks like a commit hash"""
    return bool(re.match(r'^[0-9a-f]{7,40}$', git_tag))


def is_semantic_version(git_tag: str) -> bool:
    """Check if a git tag is a semantic version"""
    try:
        parse_semver(git_tag)
        return True
    except Exception:
        return False


def compare_versions(current: str, latest: str) -> bool:
    """Compare two version strings, return True if latest is newer"""
    try:
        current_v = parse_semver(current)
        latest_v = parse_semver(latest)
        return latest_v > current_v
    except Exception:
        return current != latest


def fetch_latest_version(dep: FetchContentDependency) -> None:
    """Fetch the latest version for a dependency"""
    git_tag = dep.git_tag
    git_repo = dep.git_repo

    # Case 1: Branch name (main, master, develop, etc.)
    if git_tag in ['main', 'master', 'develop', 'trunk']:
        latest = get_latest_commit_hash(git_repo, git_tag)
        if latest:
            dep.latest_version = latest
            # Always update branches to commit hash
            dep.update_available = True
        return

    # Case 2: Commit hash - check if it's the latest on main/master
    # IMPORTANT: Check this BEFORE semantic version to avoid false positives
    if is_commit_hash(git_tag):
        # Try to get latest from main first, then master
        for branch in ['main', 'master']:
            latest = get_latest_commit_hash(git_repo, branch)
            if latest:
                dep.latest_version = latest
                dep.update_available = (git_tag != latest)
                return
        return

    # Case 3: Semantic version tag
    if is_semantic_version(git_tag):
        latest = get_latest_github_release(git_repo)
        if latest:
            dep.latest_version = latest
            dep.update_available = compare_versions(git_tag, latest)
        return

    # Case 4: Unknown format - treat as branch name
    latest = get_latest_commit_hash(git_repo, git_tag)
    if latest:
        dep.latest_version = latest
        # Convert branch to commit hash
        dep.update_available = True
    else:
        print(
            f"Warning: Could not determine version type for "
            f"{dep.name}: {git_tag}",
            file=sys.stderr
        )


def print_updates_table(dependencies: List[FetchContentDependency]) -> None:
    """Print a formatted table of available updates"""
    updates = [dep for dep in dependencies if dep.update_available]

    if not updates:
        print("\nNo updates available. All dependencies are up to date.")
        return

    print(f"\n{len(updates)} update(s) available:\n")

    # Calculate column widths
    name_width = max(len(dep.name) for dep in updates)
    current_width = max(len(dep.git_tag) for dep in updates)
    latest_width = max(len(dep.latest_version or '') for dep in updates)
    file_width = max(len(str(dep.file_path.name)) for dep in updates)

    # Print header
    header = (
        f"{'Name':<{name_width}}  {'Current':<{current_width}}  "
        f"{'Latest':<{latest_width}}  {'File':<{file_width}}"
    )
    print(header)
    print('-' * len(header))

    # Print rows
    for dep in updates:
        current = dep.git_tag
        latest = dep.latest_version or '?'
        filename = dep.file_path.name

        # Truncate long hashes for display
        if len(current) > 12:
            current_display = current[:12] + '...'
        else:
            current_display = current

        if len(latest) > 12:
            latest_display = latest[:12] + '...'
        else:
            latest_display = latest

        print(
            f"{dep.name:<{name_width}}  {current_display:<{current_width}}  "
            f"{latest_display:<{latest_width}}  {filename:<{file_width}}"
        )


def update_cmake_file(dep: FetchContentDependency) -> None:
    """Update a single dependency in its CMakeLists.txt file"""
    content = dep.file_path.read_text()

    # Create the updated version of this specific FetchContent block
    old_tag_line = f"GIT_TAG {dep.git_tag}"
    new_tag_line = f"GIT_TAG {dep.latest_version}"

    if old_tag_line not in dep.full_text:
        print(
            f"Warning: Could not find exact match for {dep.name} tag line",
            file=sys.stderr
        )
        return

    # Update only this dependency's block
    new_block = dep.full_text.replace(old_tag_line, new_tag_line, 1)

    # Replace the old block with the new block in the file content
    # Use replace with count=1 to only replace the first occurrence
    if dep.full_text not in content:
        print(
            f"Warning: Could not find block for {dep.name} in file",
            file=sys.stderr
        )
        return

    new_content = content.replace(dep.full_text, new_block, 1)

    if new_content == content:
        print(f"Warning: No changes made for {dep.name}", file=sys.stderr)
        return

    # Write back
    dep.file_path.write_text(new_content)
    print(f"Updated {dep.name} in {dep.file_path.name}")


def main():
    """Main entry point"""
    print("Searching for CMakeLists.txt files with FetchContent...")

    cmake_files = find_cmake_files()

    if not cmake_files:
        print("No CMakeLists.txt files with FetchContent found.")
        return

    print(f"Found {len(cmake_files)} CMakeLists.txt file(s):")
    for f in cmake_files:
        print(f"  - {f}")

    # Parse all dependencies
    all_deps = []
    for cmake_file in cmake_files:
        deps = parse_fetchcontent(cmake_file)
        all_deps.extend(deps)

    print(f"\nFound {len(all_deps)} FetchContent dependencies")

    # Fetch latest versions
    print("\nChecking for updates...")
    for dep in all_deps:
        print(f"  Checking {dep.name} ({dep.git_repo})...", end=' ')
        fetch_latest_version(dep)
        if dep.update_available:
            print("UPDATE AVAILABLE")
        elif dep.latest_version:
            print("up to date")
        else:
            print("SKIPPED (error)")

    # Show updates table
    print_updates_table(all_deps)

    # Get confirmation
    updates = [dep for dep in all_deps if dep.update_available]
    if not updates:
        return

    print()
    response = input(
        f"Apply {len(updates)} update(s)? [y/N]: "
    ).strip().lower()

    if response != 'y':
        print("Aborted.")
        return

    # Apply updates
    print("\nApplying updates...")
    for dep in updates:
        update_cmake_file(dep)

    print(f"\nSuccessfully updated {len(updates)} dependencies.")
    print("Please review the changes and test your build.")


if __name__ == "__main__":
    main()