check.py

#!/usr/bin/env python3
# SPDX-License-Identifier: AGPL-3.0-or-later
# TestnetHub integrity gate. Run `python3 build.py && python3 check.py` before
# deploying. Validates the built site so a future edit can't silently regress:
# broken links/anchors, heading skips, CSP leaks, oversized meta
# descriptions, sitemap drift, undefined CSS classes, unbalanced tags, and (the
# important one) donation-address consistency + QR payload. Stdlib only; the QR
# check additionally uses `segno` if it is installed. Exits non-zero on failure.

import glob
import json
import os
import re
from html.parser import HTMLParser

ROOT = os.path.dirname(os.path.abspath(__file__))
os.chdir(ROOT)

errors, warnings = [], []
def err(m): errors.append(m)
def warn(m): warnings.append(m)

pages = sorted(glob.glob("*.html"))
frags = sorted(glob.glob("content/*.html"))
built_clean = {("/" if p == "index.html" else "/" + p[:-5]) for p in pages} | {"/"}
assets = set(glob.glob("assets/**/*", recursive=True)) | {"favicon.svg", "manifest.webmanifest"}

# 1. Internal links resolve; 2. anchors resolve.
for f in pages:
    h = open(f, encoding="utf-8").read()
    for href in re.findall(r'href="(/[^"#]*)(#[^"]*)?"', h):
        base = href[0].split("?", 1)[0]  # drop cache-busting query (?v=2)
        if base.startswith("/assets/") or base in ("/favicon.svg", "/manifest.webmanifest"):
            if not os.path.exists(base.lstrip("/")):
                err(f"{f}: missing asset {base}")
        elif base not in built_clean:
            err(f"{f}: broken internal link {base}")
    for aid in re.findall(r'href="#([^"]+)"', h):
        if aid and f'id="{aid}"' not in h:
            err(f"{f}: broken anchor #{aid}")

# 3. Exactly one h1 per page, no skipped heading levels.
for f in pages:
    hs = [int(x) for x in re.findall(r"<h([1-6])", open(f, encoding="utf-8").read())]
    if hs.count(1) != 1:
        err(f"{f}: expected one <h1>, found {hs.count(1)}")
    for i in range(1, len(hs)):
        if hs[i] > hs[i - 1] + 1:
            err(f"{f}: heading level skip {hs[i-1]}->{hs[i]}")

# 4. Only expected scripts; 5. no external resource loads.
ALLOWED_SCRIPT = ("theme-init.js", "copy.js", "status.js", 'type="application/ld+json"')
for f in pages:
    h = open(f, encoding="utf-8").read()
    for s in re.findall(r"<script([^>]*)>", h):
        if not any(a in s for a in ALLOWED_SCRIPT):
            err(f"{f}: unexpected <script{s}>")
    # copy.js loads only where there is a copy button; status.js only on status.
    has_copy_trigger = bool(re.search(r'<button[^>]*\bdata-copy', h))
    if "/copy.js" in h and not has_copy_trigger:
        err(f"{f}: loads copy.js but has no copy trigger")
    if has_copy_trigger and "/copy.js" not in h:
        err(f"{f}: has a copy trigger but does not load copy.js")
    if "/status.js" in h and f != "status.html":
        err(f"{f}: status.js should load only on status.html")
    # resource-loading attrs (exclude rel=canonical / navigation <a>)
    for tag in re.findall(r"<(?:script|img)\b[^>]*?(?:src)=\"([^\"]+)\"", h):
        if tag.startswith(("http", "//")):
            err(f"{f}: external resource {tag}")
    for m in re.finditer(r"<link\b([^>]*)>", h):
        attrs = m.group(1)
        rel = (re.search(r'rel="([^"]+)"', attrs) or [None, ""])[1]
        href = (re.search(r'href="([^"]+)"', attrs) or [None, ""])[1]
        if rel in ("stylesheet", "icon", "manifest", "apple-touch-icon") and href.startswith(("http", "//")):
            err(f"{f}: external {rel} {href}")

# 6. Meta descriptions within a sane length.
for f in pages:
    m = re.search(r'name="description" content="(.*?)"', open(f, encoding="utf-8").read())
    if m and len(m.group(1)) > 160:
        warn(f"{f}: meta description {len(m.group(1))} chars (>160, may truncate)")

# 7. Sitemap matches built pages (404 excluded).
sm = set(re.findall(r"<loc>https://testnethub\.com(.*?)</loc>", open("sitemap.xml").read()))
want = {p for p in built_clean if p != "/404"}
if sm != want:
    err(f"sitemap mismatch: missing {want - sm or '{}'}, extra {sm - want or '{}'}")

# 8. manifest valid JSON.
try:
    json.load(open("manifest.webmanifest"))
except Exception as e:
    err(f"manifest.webmanifest invalid JSON: {e}")

# 9. Undefined CSS classes.
css = open("assets/style.css").read()
defined = set(re.findall(r"\.([A-Za-z][\w-]*)", css))
for f in frags:
    for attr in re.findall(r'class="([^"]+)"', open(f, encoding="utf-8").read()):
        for c in attr.split():
            if c not in defined:
                err(f"{f}: undefined CSS class .{c}")

# 10. Tag balance on built pages.
VOID = {"area", "base", "br", "col", "embed", "hr", "img", "input", "link", "meta", "param", "source", "track", "wbr"}
class Bal(HTMLParser):
    def __init__(s): super().__init__(); s.st = []; s.bad = False
    def handle_starttag(s, t, a):
        if t not in VOID: s.st.append(t)
    def handle_endtag(s, t):
        if t in VOID: return
        if s.st and s.st[-1] == t: s.st.pop()
        elif t in s.st:
            while s.st and s.st[-1] != t: s.st.pop(); s.bad = True
            if s.st: s.st.pop()
        else: s.bad = True
for f in pages:
    p = Bal(); p.feed(open(f, encoding="utf-8").read())
    if p.bad or [x for x in p.st if x not in ("html", "body")]:
        err(f"{f}: unbalanced tags")

# 11. Donation addresses: one address per coin, identical across all uses.
donate = open("donate.html", encoding="utf-8").read()
for coin, pat in {"BTC": r"bc1q[a-z0-9]{20,}", "LTC": r"ltc1q[a-z0-9]{20,}", "XMR": r"\b[48][0-9A-Za-z]{94,105}\b"}.items():
    uniq = set(re.findall(pat, donate))
    if len(uniq) != 1:
        err(f"donate.html: {coin} has {len(uniq)} distinct addresses (expected 1): {uniq}")
    elif donate.count(next(iter(uniq))) < 3:
        warn(f"donate.html: {coin} address appears < 3 times (QR/text/copy/wallet expected)")

# 12. QR codes encode the correct payment URIs (needs segno).
try:
    import io, segno
    schemes = {"btc": "bitcoin", "ltc": "litecoin", "xmr": "monero"}
    addr = {c: re.search(p, donate).group(0) for c, p in
            {"btc": r"bc1q[a-z0-9]{20,}", "ltc": r"ltc1q[a-z0-9]{20,}", "xmr": r"[48][0-9A-Za-z]{94,105}"}.items()}
    for c, scheme in schemes.items():
        buf = io.BytesIO()
        segno.make(f"{scheme}:{addr[c]}", error="m").save(buf, kind="svg", scale=8, border=4, dark="#111111", light="#ffffff")
        if buf.getvalue() != open(f"assets/donate/{c}.svg", "rb").read():
            err(f"assets/donate/{c}.svg does not encode {scheme}:{addr[c][:16]}... (regenerate)")
except ImportError:
    warn("segno not installed: skipped QR payload verification (pip install segno to enable)")

# ---- Report ----
print(f"Checked {len(pages)} pages / {len(frags)} fragments.")
for w in warnings:
    print(f"  WARN  {w}")
for e in errors:
    print(f"  FAIL  {e}")
if errors:
    print(f"\n{len(errors)} error(s), {len(warnings)} warning(s). FAILED.")
    raise SystemExit(1)
print(f"\nAll checks passed ({len(warnings)} warning(s)).")