#13993: HTMLParser is now able to handle broken end tags when strict=False.

This commit is contained in:
Ezio Melotti 2012-02-13 11:24:50 +02:00
parent efc66f9e36
commit 5211ffe4df
3 changed files with 71 additions and 17 deletions

View File

@ -23,6 +23,9 @@ starttagopen = re.compile('<[a-zA-Z]')
piclose = re.compile('>')
commentclose = re.compile(r'--\s*>')
tagfind = re.compile('[a-zA-Z][-.a-zA-Z0-9:_]*')
# see http://www.w3.org/TR/html5/tokenization.html#tag-open-state
# and http://www.w3.org/TR/html5/tokenization.html#tag-name-state
tagfind_tolerant = re.compile('[a-zA-Z][^\t\n\r\f />\x00]*')
# Note, the strict one of this pair isn't really strict, but we can't
# make it correctly strict without breaking backward compatibility.
attrfind = re.compile(
@ -270,7 +273,7 @@ class HTMLParser(_markupbase.ParserBase):
# see http://www.w3.org/TR/html5/tokenization.html#bogus-comment-state
def parse_bogus_comment(self, i, report=1):
rawdata = self.rawdata
if rawdata[i:i+2] != '<!':
if rawdata[i:i+2] not in ('<!', '</'):
self.error('unexpected call to parse_comment()')
pos = rawdata.find('>', i+2)
if pos == -1:
@ -398,31 +401,40 @@ class HTMLParser(_markupbase.ParserBase):
match = endendtag.search(rawdata, i+1) # >
if not match:
return -1
j = match.end()
gtpos = match.end()
match = endtagfind.match(rawdata, i) # </ + tag + >
if not match:
if self.cdata_elem is not None:
self.handle_data(rawdata[i:j])
return j
self.handle_data(rawdata[i:gtpos])
return gtpos
if self.strict:
self.error("bad end tag: %r" % (rawdata[i:j],))
k = rawdata.find('<', i + 1, j)
if k > i:
j = k
if j <= i:
j = i + 1
self.handle_data(rawdata[i:j])
return j
self.error("bad end tag: %r" % (rawdata[i:gtpos],))
# find the name: w3.org/TR/html5/tokenization.html#tag-name-state
namematch = tagfind_tolerant.match(rawdata, i+2)
if not namematch:
# w3.org/TR/html5/tokenization.html#end-tag-open-state
if rawdata[i:i+3] == '</>':
return i+3
else:
return self.parse_bogus_comment(i)
tagname = namematch.group().lower()
# consume and ignore other stuff between the name and the >
# Note: this is not 100% correct, since we might have things like
# </tag attr=">">, but looking for > after tha name should cover
# most of the cases and is much simpler
gtpos = rawdata.find('>', namematch.end())
self.handle_endtag(tagname)
return gtpos+1
elem = match.group(1).lower() # script or style
if self.cdata_elem is not None:
if elem != self.cdata_elem:
self.handle_data(rawdata[i:j])
return j
self.handle_data(rawdata[i:gtpos])
return gtpos
self.handle_endtag(elem.lower())
self.clear_cdata_mode()
return j
return gtpos
# Overridable -- finish processing of start+end tag: <tag.../>
def handle_startendtag(self, tag, attrs):

View File

@ -364,8 +364,9 @@ class HTMLParserTolerantTestCase(HTMLParserStrictTestCase):
('data', '<<bc'),
('endtag', 'a'),
('endtag', 'html'),
('data', '\n<img src="URL><//img></html'),
('endtag', 'html')])
('data', '\n<img src="URL>'),
('comment', '/img'),
('endtag', 'html<')])
def test_with_unquoted_attributes(self):
# see #12008
@ -403,6 +404,44 @@ class HTMLParserTolerantTestCase(HTMLParserStrictTestCase):
('starttag', 'form',
[('action', 'bogus|&#()value')])])
def test_invalid_end_tags(self):
# A collection of broken end tags. <br> is used as separator.
# see http://www.w3.org/TR/html5/tokenization.html#end-tag-open-state
# and #13993
html = ('<br></label</p><br></div end tmAd-leaderBoard><br></<h4><br>'
'</li class="unit"><br></li\r\n\t\t\t\t\t\t</ul><br></><br>')
expected = [('starttag', 'br', []),
# < is part of the name, / is discarded, p is an attribute
('endtag', 'label<'),
('starttag', 'br', []),
# text and attributes are discarded
('endtag', 'div'),
('starttag', 'br', []),
# comment because the first char after </ is not a-zA-Z
('comment', '<h4'),
('starttag', 'br', []),
# attributes are discarded
('endtag', 'li'),
('starttag', 'br', []),
# everything till ul (included) is discarded
('endtag', 'li'),
('starttag', 'br', []),
# </> is ignored
('starttag', 'br', [])]
self._run_check(html, expected)
def test_broken_invalid_end_tag(self):
# This is technically wrong (the "> shouldn't be included in the 'data')
# but is probably not worth fixing it (in addition to all the cases of
# the previous test, it would require a full attribute parsing).
# see #13993
html = '<b>This</b attr=">"> confuses the parser'
expected = [('starttag', 'b', []),
('data', 'This'),
('endtag', 'b'),
('data', '"> confuses the parser')]
self._run_check(html, expected)
def test_correct_detection_of_start_tags(self):
# see #13273
html = ('<div style="" ><b>The <a href="some_url">rain</a> '

View File

@ -113,6 +113,9 @@ Core and Builtins
Library
-------
- Issue #13993: HTMLParser is now able to handle broken end tags when
strict=False.
- Issue #9750: Fix sqlite3.Connection.iterdump on tables and fields
with a name that is a keyword or contains quotes. Patch by Marko
Kohtala.