"""Tests for HTMLParser.py.""" import html.parser import pprint import unittest from test import support class EventCollector(html.parser.HTMLParser): def __init__(self, *args, **kw): self.events = [] self.append = self.events.append html.parser.HTMLParser.__init__(self, *args, **kw) def get_events(self): # Normalize the list of events so that buffer artefacts don't # separate runs of contiguous characters. L = [] prevtype = None for event in self.events: type = event[0] if type == prevtype == "data": L[-1] = ("data", L[-1][1] + event[1]) else: L.append(event) prevtype = type self.events = L return L # structure markup def handle_starttag(self, tag, attrs): self.append(("starttag", tag, attrs)) def handle_startendtag(self, tag, attrs): self.append(("startendtag", tag, attrs)) def handle_endtag(self, tag): self.append(("endtag", tag)) # all other markup def handle_comment(self, data): self.append(("comment", data)) def handle_charref(self, data): self.append(("charref", data)) def handle_data(self, data): self.append(("data", data)) def handle_decl(self, data): self.append(("decl", data)) def handle_entityref(self, data): self.append(("entityref", data)) def handle_pi(self, data): self.append(("pi", data)) def unknown_decl(self, decl): self.append(("unknown decl", decl)) class EventCollectorExtra(EventCollector): def handle_starttag(self, tag, attrs): EventCollector.handle_starttag(self, tag, attrs) self.append(("starttag_text", self.get_starttag_text())) class EventCollectorCharrefs(EventCollector): def get_events(self): return self.events def handle_charref(self, data): self.fail('This should never be called with convert_charrefs=True') def handle_entityref(self, data): self.fail('This should never be called with convert_charrefs=True') class TestCaseBase(unittest.TestCase): def get_collector(self): raise NotImplementedError def _run_check(self, source, expected_events, collector=None): if collector is None: collector = self.get_collector() parser = collector for s in source: parser.feed(s) parser.close() events = parser.get_events() if events != expected_events: self.fail("received events did not match expected events" + "\nSource:\n" + repr(source) + "\nExpected:\n" + pprint.pformat(expected_events) + "\nReceived:\n" + pprint.pformat(events)) def _run_check_extra(self, source, events): self._run_check(source, events, EventCollectorExtra(convert_charrefs=False)) def _parse_error(self, source): def parse(source=source): parser = self.get_collector() parser.feed(source) parser.close() with self.assertRaises(html.parser.HTMLParseError): with self.assertWarns(DeprecationWarning): parse() class HTMLParserStrictTestCase(TestCaseBase): def get_collector(self): with support.check_warnings(("", DeprecationWarning), quite=False): return EventCollector(strict=True, convert_charrefs=False) def test_processing_instruction_only(self): self._run_check("", [ ("pi", "processing instruction"), ]) self._run_check("", [ ("pi", "processing instruction ?"), ]) def test_simple_html(self): self._run_check(""" &entity; sample text “ """, [ ("data", "\n"), ("decl", "DOCTYPE html PUBLIC 'foo'"), ("data", "\n"), ("starttag", "html", []), ("entityref", "entity"), ("charref", "32"), ("data", "\n"), ("comment", "comment1a\n-><&#bad;

", [ ("starttag", "p", []), ("data", "&#bad;"), ("endtag", "p"), ]) # add the [] as a workaround to avoid buffering (see #20288) self._run_check(["
&#bad;
"], [ ("starttag", "div", []), ("data", "&#bad;"), ("endtag", "div"), ]) def test_unclosed_entityref(self): self._run_check("&entityref foo", [ ("entityref", "entityref"), ("data", " foo"), ]) def test_bad_nesting(self): # Strangely, this *is* supposed to test that overlapping # elements are allowed. HTMLParser is more geared toward # lexing the input that parsing the structure. self._run_check("", [ ("starttag", "a", []), ("starttag", "b", []), ("endtag", "a"), ("endtag", "b"), ]) def test_bare_ampersands(self): self._run_check("this text & contains & ampersands &", [ ("data", "this text & contains & ampersands &"), ]) def test_bare_pointy_brackets(self): self._run_check("this < text > contains < bare>pointy< brackets", [ ("data", "this < text > contains < bare>pointy< brackets"), ]) def test_illegal_declarations(self): self._parse_error('') def test_starttag_end_boundary(self): self._run_check("""""", [("starttag", "a", [("b", "<")])]) self._run_check("""""", [("starttag", "a", [("b", ">")])]) def test_buffer_artefacts(self): output = [("starttag", "a", [("b", "<")])] self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) output = [("starttag", "a", [("b", ">")])] self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) self._run_check(["'>"], output) self._run_check([""], output) self._run_check([""], output) output = [("comment", "abc")] self._run_check(["", ""], output) self._run_check(["<", "!--abc-->"], output) self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) self._run_check([""], output) self._run_check(["", ""], output) def test_starttag_junk_chars(self): self._parse_error("") self._parse_error("") self._parse_error("") self._parse_error("") self._parse_error("'") self._parse_error("" % dtd, [('decl', 'DOCTYPE ' + dtd)]) def test_declaration_junk_chars(self): self._parse_error("") def test_startendtag(self): self._run_check("

", [ ("startendtag", "p", []), ]) self._run_check("

", [ ("starttag", "p", []), ("endtag", "p"), ]) self._run_check("

", [ ("starttag", "p", []), ("startendtag", "img", [("src", "foo")]), ("endtag", "p"), ]) def test_get_starttag_text(self): s = """""" self._run_check_extra(s, [ ("starttag", "foo:bar", [("one", "1"), ("two", "2")]), ("starttag_text", s)]) def test_cdata_content(self): contents = [ ' ¬-an-entity-ref;', "", '

', 'foo = "";', 'foo = "";', 'foo = <\n/script> ', '', ('\n//<\\/s\'+\'cript>\');\n//]]>'), '\n\n', 'foo = "";', '', # these two should be invalid according to the HTML 5 spec, # section 8.1.2.2 #'foo = ', #'foo = ', ] elements = ['script', 'style', 'SCRIPT', 'STYLE', 'Script', 'Style'] for content in contents: for element in elements: element_lower = element.lower() s = '<{element}>{content}'.format(element=element, content=content) self._run_check(s, [("starttag", element_lower, []), ("data", content), ("endtag", element_lower)]) def test_cdata_with_closing_tags(self): # see issue #13358 # make sure that HTMLParser calls handle_data only once for each CDATA. # The normal event collector normalizes the events in get_events, # so we override it to return the original list of events. class Collector(EventCollector): def get_events(self): return self.events content = """ ¬-an-entity-ref;

''""" for element in [' script', 'script ', ' script ', '\nscript', 'script\n', '\nscript\n']: element_lower = element.lower().strip() s = '{1}' '{1}'.format(text, charref), expected, collector=collector()) # check truncated charrefs at the end of the file html = '&quo &# &#x' for x in range(1, len(html)): self._run_check(html[:x], [('data', html[:x])], collector=collector()) # check a string with no charrefs self._run_check('no charrefs here', [('data', 'no charrefs here')], collector=collector()) class HTMLParserTolerantTestCase(HTMLParserStrictTestCase): def get_collector(self): return EventCollector(convert_charrefs=False) def test_deprecation_warnings(self): with self.assertWarns(DeprecationWarning): EventCollector() # convert_charrefs not passed explicitly with self.assertWarns(DeprecationWarning): EventCollector(strict=True) with self.assertWarns(DeprecationWarning): EventCollector(strict=False) with self.assertRaises(html.parser.HTMLParseError): with self.assertWarns(DeprecationWarning): EventCollector().error('test') def test_tolerant_parsing(self): self._run_check('te>>xt&a<\n' ''), ('comment', '/img'), ('endtag', 'html<')]) def test_starttag_junk_chars(self): self._run_check("", []) self._run_check("", [('comment', '$')]) self._run_check("", [('starttag', 'a", [('endtag', 'a'", [('data', "", [('starttag', 'a$b', [])]) self._run_check("", [('startendtag', 'a$b', [])]) self._run_check("", [('starttag', 'a$b', [])]) self._run_check("", [('startendtag', 'a$b', [])]) def test_slashes_in_starttag(self): self._run_check('', [('startendtag', 'a', [('foo', 'var')])]) html = ('') expected = [( 'startendtag', 'img', [('width', '902'), ('height', '250px'), ('src', '/sites/default/files/images/homepage/foo.jpg'), ('*what', None), ('am', None), ('i', None), ('doing', None), ('here*', None)] )] self._run_check(html, expected) html = ('' '') expected = [ ('startendtag', 'a', [('foo', None), ('=', None), ('bar', None)]), ('starttag', 'a', [('foo', None), ('=', None), ('bar', None)]) ] self._run_check(html, expected) #see issue #14538 html = ('' '') expected = [ ('starttag', 'meta', []), ('starttag', 'meta', []), ('starttag', 'meta', []), ('starttag', 'meta', []), ('startendtag', 'meta', []), ('startendtag', 'meta', []), ('startendtag', 'meta', []), ('startendtag', 'meta', []), ] self._run_check(html, expected) def test_declaration_junk_chars(self): self._run_check("", [('decl', 'DOCTYPE foo $ ')]) def test_illegal_declarations(self): self._run_check('', [('comment', 'spacer type="block" height="25"')]) def test_with_unquoted_attributes(self): # see #12008 html = ("" "" "
" "- software-and-i" "- library
") expected = [ ('starttag', 'html', []), ('starttag', 'body', [('bgcolor', 'd0ca90'), ('text', '181008')]), ('starttag', 'table', [('cellspacing', '0'), ('cellpadding', '1'), ('width', '100%')]), ('starttag', 'tr', []), ('starttag', 'td', [('align', 'left')]), ('starttag', 'font', [('size', '-1')]), ('data', '- '), ('starttag', 'a', [('href', '/rabota/')]), ('starttag', 'span', [('class', 'en')]), ('data', ' software-and-i'), ('endtag', 'span'), ('endtag', 'a'), ('data', '- '), ('starttag', 'a', [('href', '/1/')]), ('starttag', 'span', [('class', 'en')]), ('data', ' library'), ('endtag', 'span'), ('endtag', 'a'), ('endtag', 'table') ] self._run_check(html, expected) def test_comma_between_attributes(self): self._run_check('

', [ ('starttag', 'form', [('action', '/xxx.php?a=1&b=2&'), (',', None), ('method', 'post')])]) def test_weird_chars_in_unquoted_attribute_values(self): self._run_check('', [ ('starttag', 'form', [('action', 'bogus|&#()value')])]) def test_invalid_end_tags(self): # A collection of broken end tags.
is used as separator. # see http://www.w3.org/TR/html5/tokenization.html#end-tag-open-state # and #13993 html = ('



' '


') 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 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 = 'This 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 = ('
The rain ' '
in Spain
') expected = [ ('starttag', 'div', [('style', '')]), ('starttag', 'b', []), ('data', 'The '), ('starttag', 'a', [('href', 'some_url')]), ('data', 'rain'), ('endtag', 'a'), ('data', ' '), ('startendtag', 'br', []), ('data', ' in '), ('starttag', 'span', []), ('data', 'Spain'), ('endtag', 'span'), ('endtag', 'b'), ('endtag', 'div') ] self._run_check(html, expected) html = '
The rain' expected = [ ('starttag', 'div', [('style', ''), (',', None), ('foo', 'bar')]), ('starttag', 'b', []), ('data', 'The '), ('starttag', 'a', [('href', 'some_url')]), ('data', 'rain'), ('endtag', 'a'), ] self._run_check(html, expected) def test_EOF_in_charref(self): # see #17802 # This test checks that the UnboundLocalError reported in the issue # is not raised, however I'm not sure the returned values are correct. # Maybe HTMLParser should use self.unescape for these data = [ ('a&', [('data', 'a&')]), ('a&b', [('data', 'ab')]), ('a&b ', [('data', 'a'), ('entityref', 'b'), ('data', ' ')]), ('a&b;', [('data', 'a'), ('entityref', 'b')]), ] for html, expected in data: self._run_check(html, expected) def test_unescape_method(self): from html import unescape p = self.get_collector() with self.assertWarns(DeprecationWarning): s = '""""""&#bad;' self.assertEqual(p.unescape(s), unescape(s)) def test_broken_comments(self): html = ('' '' '' '' '') expected = [ ('comment', ' not really a comment '), ('comment', ' not a comment either --'), ('comment', ' -- close enough --'), ('comment', ''), ('comment', '<-- this was an empty comment'), ('comment', '!! another bogus comment !!!'), ] self._run_check(html, expected) def test_broken_condcoms(self): # these condcoms are missing the '--' after '' html = ('broken condcom' '' '' 'foo' '') # According to the HTML5 specs sections "8.2.4.44 Bogus comment state" # and "8.2.4.45 Markup declaration open state", comment tokens should # be emitted instead of 'unknown decl', but calling unknown_decl # provides more flexibility. # See also Lib/_markupbase.py:parse_declaration expected = [ ('unknown decl', 'if !(IE)'), ('data', 'broken condcom'), ('unknown decl', 'endif'), ('unknown decl', 'if ! IE'), ('startendtag', 'link', [('href', 'favicon.tiff')]), ('unknown decl', 'endif'), ('unknown decl', 'if !IE 6'), ('startendtag', 'img', [('src', 'firefox.png')]), ('unknown decl', 'endif'), ('unknown decl', 'if !ie 6'), ('starttag', 'b', []), ('data', 'foo'), ('endtag', 'b'), ('unknown decl', 'endif'), ('unknown decl', 'if (!IE)|(lt IE 9)'), ('startendtag', 'img', [('src', 'mammoth.bmp')]), ('unknown decl', 'endif') ] self._run_check(html, expected) class AttributesStrictTestCase(TestCaseBase): def get_collector(self): with support.check_warnings(("", DeprecationWarning), quite=False): return EventCollector(strict=True, convert_charrefs=False) def test_attr_syntax(self): output = [ ("starttag", "a", [("b", "v"), ("c", "v"), ("d", "v"), ("e", None)]) ] self._run_check("""""", output) self._run_check("""""", output) self._run_check("""""", output) self._run_check("""""", output) def test_attr_values(self): self._run_check("""""", [("starttag", "a", [("b", "xxx\n\txxx"), ("c", "yyy\t\nyyy"), ("d", "\txyz\n")])]) self._run_check("""""", [("starttag", "a", [("b", ""), ("c", "")])]) # Regression test for SF patch #669683. self._run_check("", [("starttag", "e", [("a", "rgb(1,2,3)")])]) # Regression test for SF bug #921657. self._run_check( "", [("starttag", "a", [("href", "mailto:xyz@example.com")])]) def test_attr_nonascii(self): # see issue 7311 self._run_check( "\u4e2d\u6587", [("starttag", "img", [("src", "/foo/bar.png"), ("alt", "\u4e2d\u6587")])]) self._run_check( "", [("starttag", "a", [("title", "\u30c6\u30b9\u30c8"), ("href", "\u30c6\u30b9\u30c8.html")])]) self._run_check( '', [("starttag", "a", [("title", "\u30c6\u30b9\u30c8"), ("href", "\u30c6\u30b9\u30c8.html")])]) def test_attr_entity_replacement(self): self._run_check( "", [("starttag", "a", [("b", "&><\"'")])]) def test_attr_funky_names(self): self._run_check( "", [("starttag", "a", [("a.b", "v"), ("c:d", "v"), ("e-f", "v")])]) def test_entityrefs_in_attributes(self): self._run_check( "", [("starttag", "html", [("foo", "\u20AC&aa&unsupported;")])]) class AttributesTolerantTestCase(AttributesStrictTestCase): def get_collector(self): return EventCollector(convert_charrefs=False) def test_attr_funky_names2(self): self._run_check( "", [("starttag", "a", [("$", None)]), ("starttag", "b", [("$", "%")]), ("starttag", "c", [("\\", "/")])]) def test_entities_in_attribute_value(self): # see #1200313 for entity in ['&', '&', '&', '&']: self._run_check('' % entity, [("starttag", "a", [("href", "&")])]) self._run_check("" % entity, [("starttag", "a", [("href", "&")])]) self._run_check("" % entity, [("starttag", "a", [("href", "&")])]) def test_malformed_attributes(self): # see #13357 html = ( "test - bad1" "test - bad2" "test - bad3" "test - bad4" ) expected = [ ('starttag', 'a', [('href', "test'style='color:red;bad1'")]), ('data', 'test - bad1'), ('endtag', 'a'), ('starttag', 'a', [('href', "test'+style='color:red;ba2'")]), ('data', 'test - bad2'), ('endtag', 'a'), ('starttag', 'a', [('href', "test'\xa0style='color:red;bad3'")]), ('data', 'test - bad3'), ('endtag', 'a'), ('starttag', 'a', [('href', "test'\xa0style='color:red;bad4'")]), ('data', 'test - bad4'), ('endtag', 'a') ] self._run_check(html, expected) def test_malformed_adjacent_attributes(self): # see #12629 self._run_check('', [('starttag', 'x', []), ('startendtag', 'y', [('z', ''), ('o""', None)]), ('endtag', 'x')]) self._run_check('', [('starttag', 'x', []), ('startendtag', 'y', [('z', ''), ('""', None)]), ('endtag', 'x')]) # see #755670 for the following 3 tests def test_adjacent_attributes(self): self._run_check('', [("starttag", "a", [("width", "100%"), ("cellspacing","0")])]) self._run_check('', [("starttag", "a", [("id", "foo"), ("class","bar")])]) def test_missing_attribute_value(self): self._run_check('', [("starttag", "a", [("v", "")])]) def test_javascript_attribute_value(self): self._run_check("", [("starttag", "a", [("href", "javascript:popup('/popup/help.html')")])]) def test_end_tag_in_attribute_value(self): # see #1745761 self._run_check("spam", [("starttag", "a", [("href", "http://www.example.org/\">;")]), ("data", "spam"), ("endtag", "a")]) if __name__ == "__main__": unittest.main()