# -*- coding: utf-8 -*-
import os, re
targets = [
    r'E:\mcsd\MCStudio',
    r'E:\MCStudioDownload\game',
]
needles = ['缺少Manifest', 'Manifest配置', 'ManifestConfig', 'missing manifest', 'no manifest']
exts = ('.py', '.txt', '.json', '.lang', '.lua', '.js', '.xml', '.ini', '.cfg', '.md')
encs = ['utf-8', 'gbk', 'utf-16']
hits = []
for root in targets:
    for dp, dn, fn in os.walk(root):
        # skip heavy dirs
        if any(s in dp.lower() for s in ('\\temp', '\\cache', '\\__pycache__')):
            continue
        for n in fn:
            if not n.lower().endswith(exts):
                continue
            p = os.path.join(dp, n)
            try:
                if os.path.getsize(p) > 5_000_000:
                    continue
                raw = open(p, 'rb').read()
            except Exception:
                continue
            for enc in encs:
                try:
                    t = raw.decode(enc)
                except Exception:
                    continue
                for nd in needles:
                    if nd in t:
                        hits.append('%s [%s] %s' % (p, enc, nd))
                        break
                break
with open(r'e:\MCStudioDownload\work\3027246179@qq.com\Cpp\AddOn\0aaaa4f2f16c46e8964025a5ab0ecc02\_sa_hanhua_out\_tmp\grep_manifest.txt', 'w', encoding='utf-8') as f:
    f.write('\n'.join(hits) if hits else 'NO HITS')
print('hits=%d' % len(hits))