#!/usr/bin/env python3
"""Prepare a writable Void build checkout outside the upstream submodule."""
import fcntl
import hashlib
import os
from pathlib import Path
import re
import shutil
import subprocess
import sys
import tempfile


def git(directory, *arguments):
    return subprocess.check_output(['git', '-C', str(directory), *arguments], text=True).strip()


def inventory(root):
    result = {}
    for path in sorted(root.rglob('*')):
        name = path.relative_to(root).as_posix()
        if path.is_symlink():
            result[name] = ('link', os.readlink(path))
        elif path.is_file():
            with path.open('rb') as stream:
                result[name] = ('file', path.stat().st_mode & 0o777, hashlib.file_digest(stream, 'sha256').hexdigest())
        elif path.is_dir():
            result[name] = ('directory',)
        else:
            raise ValueError(f'Unexpected special source file: {path}')
    return result


def prepare(project):
    source = project / 'vendor/void-packages'
    workspace = project / '.host/void-packages'
    pin = (project / 'VOID_PACKAGES_COMMIT').read_text().strip()
    if not re.fullmatch('[0-9a-f]{40}', pin) or git(source, 'rev-parse', 'HEAD') != pin:
        raise ValueError('Upstream Void checkout differs from VOID_PACKAGES_COMMIT')
    if git(source, 'diff', 'HEAD', '--'):
        raise ValueError('Upstream Void has staged or tracked changes; preserve them outside the submodule first')
    if source.is_symlink() or workspace.is_symlink():
        raise ValueError('Void source and build checkout must be real directories')

    # Migrate only known generated copies after comparing every entry. Unknown
    # or independently edited source files are never deleted or overwritten.
    overlays = {path.parent.name: path.parent for path in (project / 'packages').glob('*/template')}
    for name in ('hello', 'report'):
        example = project / f'examples/software/{name}/void'
        if (example / 'template').is_file():
            overlays[f'fds-demo-{name}'] = example
    moves = []
    duplicates = []
    known = set()
    for package, original in overlays.items():
        old = source / 'srcpkgs' / package
        if not old.exists() and not old.is_symlink():
            continue
        if git(source, 'ls-tree', 'HEAD', f'srcpkgs/{package}'):
            raise ValueError(f'FDS overlay collides with upstream package: {package}')
        if old.is_symlink() or not old.is_dir() or inventory(old) != inventory(original):
            raise ValueError(f'Preserve independent edits in {old} before migrating it')
        known.add(f'srcpkgs/{package}/')
        new = workspace / 'srcpkgs' / package
        if new.exists() or new.is_symlink():
            if new.is_symlink() or not new.is_dir() or inventory(new) != inventory(original):
                raise ValueError(f'Conflicting generated overlay: {new}')
            duplicates.append(old)
        else:
            moves.append((old, new))
    for name in git(source, 'ls-files', '--others', '--exclude-standard').splitlines():
        if not any(name.startswith(prefix) for prefix in known):
            raise ValueError(f'Untracked upstream file must be moved outside the submodule: {name}')
    for old in [source / 'hostdir', *sorted(source.glob('masterdir-*')), source / 'etc/conf']:
        if not old.exists() and not old.is_symlink():
            continue
        new = workspace / old.relative_to(source)
        if old.is_symlink() or new.exists() or new.is_symlink():
            raise ValueError(f'Cannot migrate {old}: symlink or existing destination {new}; reconcile explicitly')
        if old == source / 'etc/conf' and old.read_bytes() != (project / 'config/xbps-src.conf').read_bytes():
            raise ValueError('Old Void configuration differs from config/xbps-src.conf; reconcile explicitly')
        moves.append((old, new))

    if not workspace.exists():
        # Independent objects keep this cache usable without Git alternates or
        # hardlinks back into the submodule. No network access is needed.
        with tempfile.TemporaryDirectory(prefix='void-checkout.', dir=workspace.parent) as temporary:
            checkout = Path(temporary) / 'checkout'
            subprocess.run(['git', 'clone', '--quiet', '--no-hardlinks', '--no-checkout', str(source), str(checkout)], check=True)
            subprocess.run(['git', '-C', str(checkout), 'checkout', '--quiet', '--detach', pin], check=True)
            subprocess.run(['git', '-C', str(checkout), 'remote', 'remove', 'origin'], check=True)
            checkout.rename(workspace)
    else:
        if not (workspace / '.git').is_dir() or git(workspace, 'diff', 'HEAD', '--'):
            raise ValueError('Build checkout has tracked changes; preserve them before preparing it')
        if git(workspace, 'rev-parse', 'HEAD') != pin:
            subprocess.run(['git', '-C', str(workspace), 'fetch', '--quiet', '--update-shallow', str(source), pin], check=True)
            # Git refuses an update that would overwrite local files.
            subprocess.run(['git', '-C', str(workspace), 'checkout', '--quiet', '--detach', pin], check=True)
    for old, new in moves:
        new.parent.mkdir(parents=True, exist_ok=True)
        old.rename(new)
        print(f'Moved {old.relative_to(project)} to {new.relative_to(project)}', flush=True)
    for old in duplicates:
        shutil.rmtree(old)
    if git(source, 'status', '--porcelain', '--untracked-files=all'):
        raise ValueError('Upstream submodule is not clean after preparation')
    print('PASS: clean upstream reference; writable build checkout in .host/void-packages', flush=True)


def main():
    project = Path(__file__).resolve().parents[1]
    if len(sys.argv) != 1:
        sys.exit('Usage: tools/prepare-void-workspace')
    try:
        (project / '.host').mkdir(exist_ok=True)
        with (project / '.host/void-workspace.lock').open('a') as lock:
            fcntl.flock(lock, fcntl.LOCK_EX | fcntl.LOCK_NB)
            prepare(project)
    except (OSError, ValueError, subprocess.CalledProcessError) as error:
        sys.exit(f'ERROR: {error}')


if __name__ == '__main__':
    main()