#!/usr/bin/env python3
"""Cek status setiap tautan di naskah baru #2908 (HTTP 200) sebelum ditulis."""
import re, urllib.request, urllib.error

F = "/home/mantapman/.hermes/profiles/atun/workspace/tslaw_pemulihan/perbaikan_20261019/post_2908_SESUDAH_dryrun.html"
UA = {"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/124 Safari/537.36"}
urls = sorted(set(re.findall(r'https://www\.tslaw\.id/[^"\'\s<]+', open(F).read())))
print("tautan:", len(urls))
bad = []
for u in urls:
    try:
        with urllib.request.urlopen(urllib.request.Request(u, headers=UA), timeout=60) as x:
            st = x.status
    except urllib.error.HTTPError as e:
        st = e.code
    except Exception as e:
        st = str(e)[:40]
    print(f"  {st}  {u}")
    if st != 200:
        bad.append((st, u))
print("mati:", bad)