#!/usr/bin/env python3
"""Export the isolated PDFium core changes against the verified fixed archive."""
import difflib
import hashlib
import json
from pathlib import Path
import subprocess
import tarfile
import tempfile

ROOT = Path(__file__).resolve().parents[2]


def sha(path):
    with path.open('rb') as stream:
        return hashlib.file_digest(stream, 'sha256').hexdigest()


def main():
    results = ROOT / 'tests/results/pdfium-intent-build'
    materialization = json.loads((results / 'source-materialization.json').read_text())
    repository = next(row for row in materialization['repositories'] if row['path'] == '.')
    archive = ROOT / repository['archive']
    inventory = ROOT / repository['inventory']
    assert sha(archive) == repository['archiveSha256']
    assert sha(inventory) == repository['inventorySha256']
    source = ROOT / materialization['source']
    originals = {row['path']: row for line in inventory.open()
                 if (row := json.loads(line))['archived'] and row['path'].startswith('core/')}
    changed = {}
    for name, row in originals.items():
        assert row['mode'] == '100644'
        if sha(source / name) != row['sha256']:
            changed[name] = row
    added = {str(path.relative_to(source)) for path in (source / 'core').rglob('*')
             if path.is_file() and str(path.relative_to(source)) not in originals}
    assert all(Path(name).suffix in ('.h', '.cpp') for name in added)
    chunks = []
    records = []
    with tarfile.open(archive) as tar:
        for name in sorted(changed.keys() | added):
            before = b'' if name in added else tar.extractfile('pdfium/' + name).read()
            if name in changed:
                assert hashlib.sha256(before).hexdigest() == changed[name]['sha256']
            after = (source / name).read_bytes()
            diff = ''.join(difflib.unified_diff(
                before.decode().splitlines(keepends=True),
                after.decode().splitlines(keepends=True),
                fromfile='/dev/null' if name in added else 'a/' + name,
                tofile='b/' + name))
            assert diff
            chunks.append('diff --git a/' + name + ' b/' + name + '\n' + diff)
            records.append({'path': name, 'added': name in added,
                            'beforeSha256': None if name in added else hashlib.sha256(before).hexdigest(),
                            'afterSha256': hashlib.sha256(after).hexdigest()})
    destination = ROOT / 'cmake/patches/pdfium-rendering-intent.patch'
    destination.parent.mkdir(parents=True, exist_ok=True)
    destination.write_text(''.join(chunks))
    snapshot = results / 'patches' / (sha(destination) + '.patch')
    snapshot.parent.mkdir(exist_ok=True)
    if snapshot.exists():
        assert snapshot.read_bytes() == destination.read_bytes()
    else:
        snapshot.write_bytes(destination.read_bytes())
    # Check that this reviewable patch, rather than an unrecorded local change,
    # reproduces every modified core source without fuzzy matching.
    with tempfile.TemporaryDirectory(prefix='docview-pdfium-patch-') as folder:
        check_root = Path(folder)
        with tarfile.open(archive) as tar:
            for name in changed:
                target = check_root / name
                target.parent.mkdir(parents=True, exist_ok=True)
                target.write_bytes(tar.extractfile('pdfium/' + name).read())
        result = subprocess.run(['patch', '-p1', '--batch', '--fuzz=0', '-i', str(destination)],
                                cwd=check_root, capture_output=True, text=True)
        assert result.returncode == 0, result.stdout + result.stderr
        for row in records:
            assert sha(check_root / row['path']) == row['afterSha256']
    record = {'success': True, 'exporterSha256': sha(Path(__file__)),
              'baseRevision': repository['revision'], 'archiveSha256': sha(archive),
              'patch': str(destination.relative_to(ROOT)), 'patchSha256': sha(destination),
              'patchSnapshot': str(snapshot.relative_to(ROOT)),
              'appliesWithoutFuzz': True, 'appliedSourcesByteIdentical': True,
              'files': records,
              'scope': 'Candidate core patch only. Provider shared-library/public-header patches are recorded separately. Production dependency is not replaced by this exporter.'}
    (results / 'candidate-patch.json').write_text(json.dumps(record, indent=2) + '\n')
    print(json.dumps({'files': len(records), 'patchBytes': destination.stat().st_size,
                      'patchSha256': record['patchSha256']}))


if __name__ == '__main__':
    main()