import json
import re
import sys
from html.parser import HTMLParser
class Document(HTMLParser):
CDATA_CONTENT_ELEMENTS = HTMLParser.CDATA_CONTENT_ELEMENTS + (
'noscript', 'textarea', 'title', 'xmp', 'iframe', 'noembed', 'noframes',
)
def __init__(self):
super().__init__(convert_charrefs=False)
self.scripts = []
self.current = None
self.variants = []
self.context_free = True
self.standard_mode = False
self.doctype_allowed = True
self.template_depth = 0
self.unsupported_reason = None
def handle_starttag(self, tag, attrs):
self.doctype_allowed = False
first_attrs = {}
for name, value in attrs:
first_attrs.setdefault(name, value)
attrs = first_attrs
if not self.template_depth and tag in ('svg', 'math', 'base', 'frameset'):
self.unsupported_reason = f'<{tag}> elements require a browser adapter'
if tag == 'plaintext':
self.unsupported_reason = 'plaintext elements require a browser adapter'
if tag == 'template':
if not self.template_depth and 'shadowrootmode' in attrs:
self.unsupported_reason = 'declarative shadow roots require a browser adapter'
self.template_depth += 1
if self.template_depth:
return
if tag == 'script':
self.current = {'attrs': attrs, 'code': '', 'line': self.getpos()[0]}
if tag == 'style' or tag == 'base' or 'style' in attrs or attrs.get('id') == 'target':
self.context_free = False
if tag == 'link' and 'stylesheet' in (attrs.get('rel') or '').lower().split():
self.context_free = False
if any(name.startswith('on') for name in attrs):
self.unsupported_reason = 'event handlers require a browser adapter'
self.context_free = False
if tag == 'meta' and (attrs.get('http-equiv') or '').strip().lower() == 'content-security-policy':
self.unsupported_reason = 'meta CSP requires a browser adapter'
if tag == 'meta' and attrs.get('name') == 'variant':
self.variants.append(attrs.get('content', ''))
def handle_startendtag(self, tag, attrs):
if tag in self.CDATA_CONTENT_ELEMENTS or tag == 'template':
self.unsupported_reason = f'self-closing <{tag}> requires a browser adapter'
super().handle_startendtag(tag, attrs)
def handle_decl(self, decl):
if self.doctype_allowed:
self.standard_mode = decl.strip().lower() == 'doctype html'
self.doctype_allowed = False
def handle_data(self, data):
if data.strip(' \t\r\n\f'):
self.doctype_allowed = False
if self.current is not None:
self.current['code'] += data
def handle_entityref(self, name):
self.doctype_allowed = False
def handle_charref(self, name):
self.doctype_allowed = False
def handle_endtag(self, tag):
self.doctype_allowed = False
if tag == 'template' and self.template_depth:
self.template_depth -= 1
return
if self.template_depth:
return
if tag == 'script' and self.current is not None:
escaped = False
for marker in re.finditer(r'<!--|-->|<script(?=[\t\n\f\r />])', self.current['code'], re.I):
if marker[0] == '<!--':
escaped = True
elif marker[0] == '-->':
escaped = False
elif escaped:
self.unsupported_reason = 'HTML double-escaped scripts require a browser adapter'
break
self.scripts.append(self.current)
self.current = None
for line in sys.stdin:
source = json.loads(line)
parser = Document()
if '\x00' in source:
parser.unsupported_reason = 'literal NUL HTML source requires an encoding adapter'
parser.feed(source)
result = {'scripts': parser.scripts, 'variants': parser.variants,
'contextFree': parser.context_free and parser.standard_mode}
if parser.unsupported_reason:
result['unsupportedReason'] = parser.unsupported_reason
print(json.dumps(result))