1"""Tests for HTMLParser.py."""
2
3import html.parser
4import pprint
5import unittest
6
7
8class EventCollector(html.parser.HTMLParser):
9
10    def __init__(self, *args, **kw):
11        self.events = []
12        self.append = self.events.append
13        html.parser.HTMLParser.__init__(self, *args, **kw)
14
15    def get_events(self):
16        # Normalize the list of events so that buffer artefacts don't
17        # separate runs of contiguous characters.
18        L = []
19        prevtype = None
20        for event in self.events:
21            type = event[0]
22            if type == prevtype == "data":
23                L[-1] = ("data", L[-1][1] + event[1])
24            else:
25                L.append(event)
26            prevtype = type
27        self.events = L
28        return L
29
30    # structure markup
31
32    def handle_starttag(self, tag, attrs):
33        self.append(("starttag", tag, attrs))
34
35    def handle_startendtag(self, tag, attrs):
36        self.append(("startendtag", tag, attrs))
37
38    def handle_endtag(self, tag):
39        self.append(("endtag", tag))
40
41    # all other markup
42
43    def handle_comment(self, data):
44        self.append(("comment", data))
45
46    def handle_charref(self, data):
47        self.append(("charref", data))
48
49    def handle_data(self, data):
50        self.append(("data", data))
51
52    def handle_decl(self, data):
53        self.append(("decl", data))
54
55    def handle_entityref(self, data):
56        self.append(("entityref", data))
57
58    def handle_pi(self, data):
59        self.append(("pi", data))
60
61    def unknown_decl(self, decl):
62        self.append(("unknown decl", decl))
63
64
65class EventCollectorExtra(EventCollector):
66
67    def handle_starttag(self, tag, attrs):
68        EventCollector.handle_starttag(self, tag, attrs)
69        self.append(("starttag_text", self.get_starttag_text()))
70
71
72class EventCollectorCharrefs(EventCollector):
73
74    def handle_charref(self, data):
75        self.fail('This should never be called with convert_charrefs=True')
76
77    def handle_entityref(self, data):
78        self.fail('This should never be called with convert_charrefs=True')
79
80
81class TestCaseBase(unittest.TestCase):
82
83    def get_collector(self):
84        return EventCollector(convert_charrefs=False)
85
86    def _run_check(self, source, expected_events, collector=None):
87        if collector is None:
88            collector = self.get_collector()
89        parser = collector
90        for s in source:
91            parser.feed(s)
92        parser.close()
93        events = parser.get_events()
94        if events != expected_events:
95            self.fail("received events did not match expected events" +
96                      "\nSource:\n" + repr(source) +
97                      "\nExpected:\n" + pprint.pformat(expected_events) +
98                      "\nReceived:\n" + pprint.pformat(events))
99
100    def _run_check_extra(self, source, events):
101        self._run_check(source, events,
102                        EventCollectorExtra(convert_charrefs=False))
103
104
105class HTMLParserTestCase(TestCaseBase):
106
107    def test_processing_instruction_only(self):
108        self._run_check("<?processing instruction>", [
109            ("pi", "processing instruction"),
110            ])
111        self._run_check("<?processing instruction ?>", [
112            ("pi", "processing instruction ?"),
113            ])
114
115    def test_simple_html(self):
116        self._run_check("""
117<!DOCTYPE html PUBLIC 'foo'>
118<HTML>&entity;&#32;
119<!--comment1a
120-></foo><bar>&lt;<?pi?></foo<bar
121comment1b-->
122<Img sRc='Bar' isMAP>sample
123text
124&#x201C;
125<!--comment2a-- --comment2b-->
126</Html>
127""", [
128    ("data", "\n"),
129    ("decl", "DOCTYPE html PUBLIC 'foo'"),
130    ("data", "\n"),
131    ("starttag", "html", []),
132    ("entityref", "entity"),
133    ("charref", "32"),
134    ("data", "\n"),
135    ("comment", "comment1a\n-></foo><bar>&lt;<?pi?></foo<bar\ncomment1b"),
136    ("data", "\n"),
137    ("starttag", "img", [("src", "Bar"), ("ismap", None)]),
138    ("data", "sample\ntext\n"),
139    ("charref", "x201C"),
140    ("data", "\n"),
141    ("comment", "comment2a-- --comment2b"),
142    ("data", "\n"),
143    ("endtag", "html"),
144    ("data", "\n"),
145    ])
146
147    def test_malformatted_charref(self):
148        self._run_check("<p>&#bad;</p>", [
149            ("starttag", "p", []),
150            ("data", "&#bad;"),
151            ("endtag", "p"),
152        ])
153        # add the [] as a workaround to avoid buffering (see #20288)
154        self._run_check(["<div>&#bad;</div>"], [
155            ("starttag", "div", []),
156            ("data", "&#bad;"),
157            ("endtag", "div"),
158        ])
159
160    def test_unclosed_entityref(self):
161        self._run_check("&entityref foo", [
162            ("entityref", "entityref"),
163            ("data", " foo"),
164            ])
165
166    def test_bad_nesting(self):
167        # Strangely, this *is* supposed to test that overlapping
168        # elements are allowed.  HTMLParser is more geared toward
169        # lexing the input that parsing the structure.
170        self._run_check("<a><b></a></b>", [
171            ("starttag", "a", []),
172            ("starttag", "b", []),
173            ("endtag", "a"),
174            ("endtag", "b"),
175            ])
176
177    def test_bare_ampersands(self):
178        self._run_check("this text & contains & ampersands &", [
179            ("data", "this text & contains & ampersands &"),
180            ])
181
182    def test_bare_pointy_brackets(self):
183        self._run_check("this < text > contains < bare>pointy< brackets", [
184            ("data", "this < text > contains < bare>pointy< brackets"),
185            ])
186
187    def test_starttag_end_boundary(self):
188        self._run_check("""<a b='<'>""", [("starttag", "a", [("b", "<")])])
189        self._run_check("""<a b='>'>""", [("starttag", "a", [("b", ">")])])
190
191    def test_buffer_artefacts(self):
192        output = [("starttag", "a", [("b", "<")])]
193        self._run_check(["<a b='<'>"], output)
194        self._run_check(["<a ", "b='<'>"], output)
195        self._run_check(["<a b", "='<'>"], output)
196        self._run_check(["<a b=", "'<'>"], output)
197        self._run_check(["<a b='<", "'>"], output)
198        self._run_check(["<a b='<'", ">"], output)
199
200        output = [("starttag", "a", [("b", ">")])]
201        self._run_check(["<a b='>'>"], output)
202        self._run_check(["<a ", "b='>'>"], output)
203        self._run_check(["<a b", "='>'>"], output)
204        self._run_check(["<a b=", "'>'>"], output)
205        self._run_check(["<a b='>", "'>"], output)
206        self._run_check(["<a b='>'", ">"], output)
207
208        output = [("comment", "abc")]
209        self._run_check(["", "<!--abc-->"], output)
210        self._run_check(["<", "!--abc-->"], output)
211        self._run_check(["<!", "--abc-->"], output)
212        self._run_check(["<!-", "-abc-->"], output)
213        self._run_check(["<!--", "abc-->"], output)
214        self._run_check(["<!--a", "bc-->"], output)
215        self._run_check(["<!--ab", "c-->"], output)
216        self._run_check(["<!--abc", "-->"], output)
217        self._run_check(["<!--abc-", "->"], output)
218        self._run_check(["<!--abc--", ">"], output)
219        self._run_check(["<!--abc-->", ""], output)
220
221    def test_valid_doctypes(self):
222        # from http://www.w3.org/QA/2002/04/valid-dtd-list.html
223        dtds = ['HTML',  # HTML5 doctype
224                ('HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" '
225                 '"http://www.w3.org/TR/html4/strict.dtd"'),
226                ('HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" '
227                 '"http://www.w3.org/TR/html4/loose.dtd"'),
228                ('html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" '
229                 '"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"'),
230                ('html PUBLIC "-//W3C//DTD XHTML 1.0 Frameset//EN" '
231                 '"http://www.w3.org/TR/xhtml1/DTD/xhtml1-frameset.dtd"'),
232                ('math PUBLIC "-//W3C//DTD MathML 2.0//EN" '
233                 '"http://www.w3.org/Math/DTD/mathml2/mathml2.dtd"'),
234                ('html PUBLIC "-//W3C//DTD '
235                 'XHTML 1.1 plus MathML 2.0 plus SVG 1.1//EN" '
236                 '"http://www.w3.org/2002/04/xhtml-math-svg/xhtml-math-svg.dtd"'),
237                ('svg PUBLIC "-//W3C//DTD SVG 1.1//EN" '
238                 '"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd"'),
239                'html PUBLIC "-//IETF//DTD HTML 2.0//EN"',
240                'html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN"']
241        for dtd in dtds:
242            self._run_check("<!DOCTYPE %s>" % dtd,
243                            [('decl', 'DOCTYPE ' + dtd)])
244
245    def test_startendtag(self):
246        self._run_check("<p/>", [
247            ("startendtag", "p", []),
248            ])
249        self._run_check("<p></p>", [
250            ("starttag", "p", []),
251            ("endtag", "p"),
252            ])
253        self._run_check("<p><img src='foo' /></p>", [
254            ("starttag", "p", []),
255            ("startendtag", "img", [("src", "foo")]),
256            ("endtag", "p"),
257            ])
258
259    def test_get_starttag_text(self):
260        s = """<foo:bar   \n   one="1"\ttwo=2   >"""
261        self._run_check_extra(s, [
262            ("starttag", "foo:bar", [("one", "1"), ("two", "2")]),
263            ("starttag_text", s)])
264
265    def test_cdata_content(self):
266        contents = [
267            '<!-- not a comment --> &not-an-entity-ref;',
268            "<not a='start tag'>",
269            '<a href="" /> <p> <span></span>',
270            'foo = "</scr" + "ipt>";',
271            'foo = "</SCRIPT" + ">";',
272            'foo = <\n/script> ',
273            '<!-- document.write("</scr" + "ipt>"); -->',
274            ('\n//<![CDATA[\n'
275             'document.write(\'<s\'+\'cript type="text/javascript" '
276             'src="http://www.example.org/r=\'+new '
277             'Date().getTime()+\'"><\\/s\'+\'cript>\');\n//]]>'),
278            '\n<!-- //\nvar foo = 3.14;\n// -->\n',
279            'foo = "</sty" + "le>";',
280            '<!-- \u2603 -->',
281            # these two should be invalid according to the HTML 5 spec,
282            # section 8.1.2.2
283            #'foo = </\nscript>',
284            #'foo = </ script>',
285        ]
286        elements = ['script', 'style', 'SCRIPT', 'STYLE', 'Script', 'Style']
287        for content in contents:
288            for element in elements:
289                element_lower = element.lower()
290                s = '<{element}>{content}</{element}>'.format(element=element,
291                                                               content=content)
292                self._run_check(s, [("starttag", element_lower, []),
293                                    ("data", content),
294                                    ("endtag", element_lower)])
295
296    def test_cdata_with_closing_tags(self):
297        # see issue #13358
298        # make sure that HTMLParser calls handle_data only once for each CDATA.
299        # The normal event collector normalizes  the events in get_events,
300        # so we override it to return the original list of events.
301        class Collector(EventCollector):
302            def get_events(self):
303                return self.events
304
305        content = """<!-- not a comment --> &not-an-entity-ref;
306                  <a href="" /> </p><p> <span></span></style>
307                  '</script' + '>'"""
308        for element in [' script', 'script ', ' script ',
309                        '\nscript', 'script\n', '\nscript\n']:
310            element_lower = element.lower().strip()
311            s = '<script>{content}</{element}>'.format(element=element,
312                                                       content=content)
313            self._run_check(s, [("starttag", element_lower, []),
314                                ("data", content),
315                                ("endtag", element_lower)],
316                            collector=Collector(convert_charrefs=False))
317
318    def test_comments(self):
319        html = ("<!-- I'm a valid comment -->"
320                '<!--me too!-->'
321                '<!------>'
322                '<!---->'
323                '<!----I have many hyphens---->'
324                '<!-- I have a > in the middle -->'
325                '<!-- and I have -- in the middle! -->')
326        expected = [('comment', " I'm a valid comment "),
327                    ('comment', 'me too!'),
328                    ('comment', '--'),
329                    ('comment', ''),
330                    ('comment', '--I have many hyphens--'),
331                    ('comment', ' I have a > in the middle '),
332                    ('comment', ' and I have -- in the middle! ')]
333        self._run_check(html, expected)
334
335    def test_condcoms(self):
336        html = ('<!--[if IE & !(lte IE 8)]>aren\'t<![endif]-->'
337                '<!--[if IE 8]>condcoms<![endif]-->'
338                '<!--[if lte IE 7]>pretty?<![endif]-->')
339        expected = [('comment', "[if IE & !(lte IE 8)]>aren't<![endif]"),
340                    ('comment', '[if IE 8]>condcoms<![endif]'),
341                    ('comment', '[if lte IE 7]>pretty?<![endif]')]
342        self._run_check(html, expected)
343
344    def test_convert_charrefs(self):
345        # default value for convert_charrefs is now True
346        collector = lambda: EventCollectorCharrefs()
347        self.assertTrue(collector().convert_charrefs)
348        charrefs = ['&quot;', '&#34;', '&#x22;', '&quot', '&#34', '&#x22']
349        # check charrefs in the middle of the text/attributes
350        expected = [('starttag', 'a', [('href', 'foo"zar')]),
351                    ('data', 'a"z'), ('endtag', 'a')]
352        for charref in charrefs:
353            self._run_check('<a href="foo{0}zar">a{0}z</a>'.format(charref),
354                            expected, collector=collector())
355        # check charrefs at the beginning/end of the text/attributes
356        expected = [('data', '"'),
357                    ('starttag', 'a', [('x', '"'), ('y', '"X'), ('z', 'X"')]),
358                    ('data', '"'), ('endtag', 'a'), ('data', '"')]
359        for charref in charrefs:
360            self._run_check('{0}<a x="{0}" y="{0}X" z="X{0}">'
361                            '{0}</a>{0}'.format(charref),
362                            expected, collector=collector())
363        # check charrefs in <script>/<style> elements
364        for charref in charrefs:
365            text = 'X'.join([charref]*3)
366            expected = [('data', '"'),
367                        ('starttag', 'script', []), ('data', text),
368                        ('endtag', 'script'), ('data', '"'),
369                        ('starttag', 'style', []), ('data', text),
370                        ('endtag', 'style'), ('data', '"')]
371            self._run_check('{1}<script>{0}</script>{1}'
372                            '<style>{0}</style>{1}'.format(text, charref),
373                            expected, collector=collector())
374        # check truncated charrefs at the end of the file
375        html = '&quo &# &#x'
376        for x in range(1, len(html)):
377            self._run_check(html[:x], [('data', html[:x])],
378                            collector=collector())
379        # check a string with no charrefs
380        self._run_check('no charrefs here', [('data', 'no charrefs here')],
381                        collector=collector())
382
383    # the remaining tests were for the "tolerant" parser (which is now
384    # the default), and check various kind of broken markup
385    def test_tolerant_parsing(self):
386        self._run_check('<html <html>te>>xt&a<<bc</a></html>\n'
387                        '<img src="URL><//img></html</html>', [
388                            ('starttag', 'html', [('<html', None)]),
389                            ('data', 'te>>xt'),
390                            ('entityref', 'a'),
391                            ('data', '<'),
392                            ('starttag', 'bc<', [('a', None)]),
393                            ('endtag', 'html'),
394                            ('data', '\n<img src="URL>'),
395                            ('comment', '/img'),
396                            ('endtag', 'html<')])
397
398    def test_starttag_junk_chars(self):
399        self._run_check("</>", [])
400        self._run_check("</$>", [('comment', '$')])
401        self._run_check("</", [('data', '</')])
402        self._run_check("</a", [('data', '</a')])
403        self._run_check("<a<a>", [('starttag', 'a<a', [])])
404        self._run_check("</a<a>", [('endtag', 'a<a')])
405        self._run_check("<!", [('data', '<!')])
406        self._run_check("<a", [('data', '<a')])
407        self._run_check("<a foo='bar'", [('data', "<a foo='bar'")])
408        self._run_check("<a foo='bar", [('data', "<a foo='bar")])
409        self._run_check("<a foo='>'", [('data', "<a foo='>'")])
410        self._run_check("<a foo='>", [('data', "<a foo='>")])
411        self._run_check("<a$>", [('starttag', 'a$', [])])
412        self._run_check("<a$b>", [('starttag', 'a$b', [])])
413        self._run_check("<a$b/>", [('startendtag', 'a$b', [])])
414        self._run_check("<a$b  >", [('starttag', 'a$b', [])])
415        self._run_check("<a$b  />", [('startendtag', 'a$b', [])])
416
417    def test_slashes_in_starttag(self):
418        self._run_check('<a foo="var"/>', [('startendtag', 'a', [('foo', 'var')])])
419        html = ('<img width=902 height=250px '
420                'src="/sites/default/files/images/homepage/foo.jpg" '
421                '/*what am I doing here*/ />')
422        expected = [(
423            'startendtag', 'img',
424            [('width', '902'), ('height', '250px'),
425             ('src', '/sites/default/files/images/homepage/foo.jpg'),
426             ('*what', None), ('am', None), ('i', None),
427             ('doing', None), ('here*', None)]
428        )]
429        self._run_check(html, expected)
430        html = ('<a / /foo/ / /=/ / /bar/ / />'
431                '<a / /foo/ / /=/ / /bar/ / >')
432        expected = [
433            ('startendtag', 'a', [('foo', None), ('=', None), ('bar', None)]),
434            ('starttag', 'a', [('foo', None), ('=', None), ('bar', None)])
435        ]
436        self._run_check(html, expected)
437        #see issue #14538
438        html = ('<meta><meta / ><meta // ><meta / / >'
439                '<meta/><meta /><meta //><meta//>')
440        expected = [
441            ('starttag', 'meta', []), ('starttag', 'meta', []),
442            ('starttag', 'meta', []), ('starttag', 'meta', []),
443            ('startendtag', 'meta', []), ('startendtag', 'meta', []),
444            ('startendtag', 'meta', []), ('startendtag', 'meta', []),
445        ]
446        self._run_check(html, expected)
447
448    def test_declaration_junk_chars(self):
449        self._run_check("<!DOCTYPE foo $ >", [('decl', 'DOCTYPE foo $ ')])
450
451    def test_illegal_declarations(self):
452        self._run_check('<!spacer type="block" height="25">',
453                        [('comment', 'spacer type="block" height="25"')])
454
455    def test_with_unquoted_attributes(self):
456        # see #12008
457        html = ("<html><body bgcolor=d0ca90 text='181008'>"
458                "<table cellspacing=0 cellpadding=1 width=100% ><tr>"
459                "<td align=left><font size=-1>"
460                "- <a href=/rabota/><span class=en> software-and-i</span></a>"
461                "- <a href='/1/'><span class=en> library</span></a></table>")
462        expected = [
463            ('starttag', 'html', []),
464            ('starttag', 'body', [('bgcolor', 'd0ca90'), ('text', '181008')]),
465            ('starttag', 'table',
466                [('cellspacing', '0'), ('cellpadding', '1'), ('width', '100%')]),
467            ('starttag', 'tr', []),
468            ('starttag', 'td', [('align', 'left')]),
469            ('starttag', 'font', [('size', '-1')]),
470            ('data', '- '), ('starttag', 'a', [('href', '/rabota/')]),
471            ('starttag', 'span', [('class', 'en')]), ('data', ' software-and-i'),
472            ('endtag', 'span'), ('endtag', 'a'),
473            ('data', '- '), ('starttag', 'a', [('href', '/1/')]),
474            ('starttag', 'span', [('class', 'en')]), ('data', ' library'),
475            ('endtag', 'span'), ('endtag', 'a'), ('endtag', 'table')
476        ]
477        self._run_check(html, expected)
478
479    def test_comma_between_attributes(self):
480        self._run_check('<form action="/xxx.php?a=1&amp;b=2&amp", '
481                        'method="post">', [
482                            ('starttag', 'form',
483                                [('action', '/xxx.php?a=1&b=2&'),
484                                 (',', None), ('method', 'post')])])
485
486    def test_weird_chars_in_unquoted_attribute_values(self):
487        self._run_check('<form action=bogus|&#()value>', [
488                            ('starttag', 'form',
489                                [('action', 'bogus|&#()value')])])
490
491    def test_invalid_end_tags(self):
492        # A collection of broken end tags. <br> is used as separator.
493        # see http://www.w3.org/TR/html5/tokenization.html#end-tag-open-state
494        # and #13993
495        html = ('<br></label</p><br></div end tmAd-leaderBoard><br></<h4><br>'
496                '</li class="unit"><br></li\r\n\t\t\t\t\t\t</ul><br></><br>')
497        expected = [('starttag', 'br', []),
498                    # < is part of the name, / is discarded, p is an attribute
499                    ('endtag', 'label<'),
500                    ('starttag', 'br', []),
501                    # text and attributes are discarded
502                    ('endtag', 'div'),
503                    ('starttag', 'br', []),
504                    # comment because the first char after </ is not a-zA-Z
505                    ('comment', '<h4'),
506                    ('starttag', 'br', []),
507                    # attributes are discarded
508                    ('endtag', 'li'),
509                    ('starttag', 'br', []),
510                    # everything till ul (included) is discarded
511                    ('endtag', 'li'),
512                    ('starttag', 'br', []),
513                    # </> is ignored
514                    ('starttag', 'br', [])]
515        self._run_check(html, expected)
516
517    def test_broken_invalid_end_tag(self):
518        # This is technically wrong (the "> shouldn't be included in the 'data')
519        # but is probably not worth fixing it (in addition to all the cases of
520        # the previous test, it would require a full attribute parsing).
521        # see #13993
522        html = '<b>This</b attr=">"> confuses the parser'
523        expected = [('starttag', 'b', []),
524                    ('data', 'This'),
525                    ('endtag', 'b'),
526                    ('data', '"> confuses the parser')]
527        self._run_check(html, expected)
528
529    def test_correct_detection_of_start_tags(self):
530        # see #13273
531        html = ('<div style=""    ><b>The <a href="some_url">rain</a> '
532                '<br /> in <span>Spain</span></b></div>')
533        expected = [
534            ('starttag', 'div', [('style', '')]),
535            ('starttag', 'b', []),
536            ('data', 'The '),
537            ('starttag', 'a', [('href', 'some_url')]),
538            ('data', 'rain'),
539            ('endtag', 'a'),
540            ('data', ' '),
541            ('startendtag', 'br', []),
542            ('data', ' in '),
543            ('starttag', 'span', []),
544            ('data', 'Spain'),
545            ('endtag', 'span'),
546            ('endtag', 'b'),
547            ('endtag', 'div')
548        ]
549        self._run_check(html, expected)
550
551        html = '<div style="", foo = "bar" ><b>The <a href="some_url">rain</a>'
552        expected = [
553            ('starttag', 'div', [('style', ''), (',', None), ('foo', 'bar')]),
554            ('starttag', 'b', []),
555            ('data', 'The '),
556            ('starttag', 'a', [('href', 'some_url')]),
557            ('data', 'rain'),
558            ('endtag', 'a'),
559        ]
560        self._run_check(html, expected)
561
562    def test_EOF_in_charref(self):
563        # see #17802
564        # This test checks that the UnboundLocalError reported in the issue
565        # is not raised, however I'm not sure the returned values are correct.
566        # Maybe HTMLParser should use self.unescape for these
567        data = [
568            ('a&', [('data', 'a&')]),
569            ('a&b', [('data', 'ab')]),
570            ('a&b ', [('data', 'a'), ('entityref', 'b'), ('data', ' ')]),
571            ('a&b;', [('data', 'a'), ('entityref', 'b')]),
572        ]
573        for html, expected in data:
574            self._run_check(html, expected)
575
576    def test_unescape_method(self):
577        from html import unescape
578        p = self.get_collector()
579        with self.assertWarns(DeprecationWarning):
580            s = '&quot;&#34;&#x22;&quot&#34&#x22&#bad;'
581            self.assertEqual(p.unescape(s), unescape(s))
582
583    def test_broken_comments(self):
584        html = ('<! not really a comment >'
585                '<! not a comment either -->'
586                '<! -- close enough -->'
587                '<!><!<-- this was an empty comment>'
588                '<!!! another bogus comment !!!>')
589        expected = [
590            ('comment', ' not really a comment '),
591            ('comment', ' not a comment either --'),
592            ('comment', ' -- close enough --'),
593            ('comment', ''),
594            ('comment', '<-- this was an empty comment'),
595            ('comment', '!! another bogus comment !!!'),
596        ]
597        self._run_check(html, expected)
598
599    def test_broken_condcoms(self):
600        # these condcoms are missing the '--' after '<!' and before the '>'
601        html = ('<![if !(IE)]>broken condcom<![endif]>'
602                '<![if ! IE]><link href="favicon.tiff"/><![endif]>'
603                '<![if !IE 6]><img src="firefox.png" /><![endif]>'
604                '<![if !ie 6]><b>foo</b><![endif]>'
605                '<![if (!IE)|(lt IE 9)]><img src="mammoth.bmp" /><![endif]>')
606        # According to the HTML5 specs sections "8.2.4.44 Bogus comment state"
607        # and "8.2.4.45 Markup declaration open state", comment tokens should
608        # be emitted instead of 'unknown decl', but calling unknown_decl
609        # provides more flexibility.
610        # See also Lib/_markupbase.py:parse_declaration
611        expected = [
612            ('unknown decl', 'if !(IE)'),
613            ('data', 'broken condcom'),
614            ('unknown decl', 'endif'),
615            ('unknown decl', 'if ! IE'),
616            ('startendtag', 'link', [('href', 'favicon.tiff')]),
617            ('unknown decl', 'endif'),
618            ('unknown decl', 'if !IE 6'),
619            ('startendtag', 'img', [('src', 'firefox.png')]),
620            ('unknown decl', 'endif'),
621            ('unknown decl', 'if !ie 6'),
622            ('starttag', 'b', []),
623            ('data', 'foo'),
624            ('endtag', 'b'),
625            ('unknown decl', 'endif'),
626            ('unknown decl', 'if (!IE)|(lt IE 9)'),
627            ('startendtag', 'img', [('src', 'mammoth.bmp')]),
628            ('unknown decl', 'endif')
629        ]
630        self._run_check(html, expected)
631
632    def test_convert_charrefs_dropped_text(self):
633        # #23144: make sure that all the events are triggered when
634        # convert_charrefs is True, even if we don't call .close()
635        parser = EventCollector(convert_charrefs=True)
636        # before the fix, bar & baz was missing
637        parser.feed("foo <a>link</a> bar &amp; baz")
638        self.assertEqual(
639            parser.get_events(),
640            [('data', 'foo '), ('starttag', 'a', []), ('data', 'link'),
641             ('endtag', 'a'), ('data', ' bar & baz')]
642        )
643
644
645class AttributesTestCase(TestCaseBase):
646
647    def test_attr_syntax(self):
648        output = [
649          ("starttag", "a", [("b", "v"), ("c", "v"), ("d", "v"), ("e", None)])
650        ]
651        self._run_check("""<a b='v' c="v" d=v e>""", output)
652        self._run_check("""<a  b = 'v' c = "v" d = v e>""", output)
653        self._run_check("""<a\nb\n=\n'v'\nc\n=\n"v"\nd\n=\nv\ne>""", output)
654        self._run_check("""<a\tb\t=\t'v'\tc\t=\t"v"\td\t=\tv\te>""", output)
655
656    def test_attr_values(self):
657        self._run_check("""<a b='xxx\n\txxx' c="yyy\t\nyyy" d='\txyz\n'>""",
658                        [("starttag", "a", [("b", "xxx\n\txxx"),
659                                            ("c", "yyy\t\nyyy"),
660                                            ("d", "\txyz\n")])])
661        self._run_check("""<a b='' c="">""",
662                        [("starttag", "a", [("b", ""), ("c", "")])])
663        # Regression test for SF patch #669683.
664        self._run_check("<e a=rgb(1,2,3)>",
665                        [("starttag", "e", [("a", "rgb(1,2,3)")])])
666        # Regression test for SF bug #921657.
667        self._run_check(
668            "<a href=mailto:xyz@example.com>",
669            [("starttag", "a", [("href", "mailto:xyz@example.com")])])
670
671    def test_attr_nonascii(self):
672        # see issue 7311
673        self._run_check(
674            "<img src=/foo/bar.png alt=\u4e2d\u6587>",
675            [("starttag", "img", [("src", "/foo/bar.png"),
676                                  ("alt", "\u4e2d\u6587")])])
677        self._run_check(
678            "<a title='\u30c6\u30b9\u30c8' href='\u30c6\u30b9\u30c8.html'>",
679            [("starttag", "a", [("title", "\u30c6\u30b9\u30c8"),
680                                ("href", "\u30c6\u30b9\u30c8.html")])])
681        self._run_check(
682            '<a title="\u30c6\u30b9\u30c8" href="\u30c6\u30b9\u30c8.html">',
683            [("starttag", "a", [("title", "\u30c6\u30b9\u30c8"),
684                                ("href", "\u30c6\u30b9\u30c8.html")])])
685
686    def test_attr_entity_replacement(self):
687        self._run_check(
688            "<a b='&amp;&gt;&lt;&quot;&apos;'>",
689            [("starttag", "a", [("b", "&><\"'")])])
690
691    def test_attr_funky_names(self):
692        self._run_check(
693            "<a a.b='v' c:d=v e-f=v>",
694            [("starttag", "a", [("a.b", "v"), ("c:d", "v"), ("e-f", "v")])])
695
696    def test_entityrefs_in_attributes(self):
697        self._run_check(
698            "<html foo='&euro;&amp;&#97;&#x61;&unsupported;'>",
699            [("starttag", "html", [("foo", "\u20AC&aa&unsupported;")])])
700
701
702    def test_attr_funky_names2(self):
703        self._run_check(
704            r"<a $><b $=%><c \=/>",
705            [("starttag", "a", [("$", None)]),
706             ("starttag", "b", [("$", "%")]),
707             ("starttag", "c", [("\\", "/")])])
708
709    def test_entities_in_attribute_value(self):
710        # see #1200313
711        for entity in ['&', '&amp;', '&#38;', '&#x26;']:
712            self._run_check('<a href="%s">' % entity,
713                            [("starttag", "a", [("href", "&")])])
714            self._run_check("<a href='%s'>" % entity,
715                            [("starttag", "a", [("href", "&")])])
716            self._run_check("<a href=%s>" % entity,
717                            [("starttag", "a", [("href", "&")])])
718
719    def test_malformed_attributes(self):
720        # see #13357
721        html = (
722            "<a href=test'style='color:red;bad1'>test - bad1</a>"
723            "<a href=test'+style='color:red;ba2'>test - bad2</a>"
724            "<a href=test'&nbsp;style='color:red;bad3'>test - bad3</a>"
725            "<a href = test'&nbsp;style='color:red;bad4'  >test - bad4</a>"
726        )
727        expected = [
728            ('starttag', 'a', [('href', "test'style='color:red;bad1'")]),
729            ('data', 'test - bad1'), ('endtag', 'a'),
730            ('starttag', 'a', [('href', "test'+style='color:red;ba2'")]),
731            ('data', 'test - bad2'), ('endtag', 'a'),
732            ('starttag', 'a', [('href', "test'\xa0style='color:red;bad3'")]),
733            ('data', 'test - bad3'), ('endtag', 'a'),
734            ('starttag', 'a', [('href', "test'\xa0style='color:red;bad4'")]),
735            ('data', 'test - bad4'), ('endtag', 'a')
736        ]
737        self._run_check(html, expected)
738
739    def test_malformed_adjacent_attributes(self):
740        # see #12629
741        self._run_check('<x><y z=""o"" /></x>',
742                        [('starttag', 'x', []),
743                            ('startendtag', 'y', [('z', ''), ('o""', None)]),
744                            ('endtag', 'x')])
745        self._run_check('<x><y z="""" /></x>',
746                        [('starttag', 'x', []),
747                            ('startendtag', 'y', [('z', ''), ('""', None)]),
748                            ('endtag', 'x')])
749
750    # see #755670 for the following 3 tests
751    def test_adjacent_attributes(self):
752        self._run_check('<a width="100%"cellspacing=0>',
753                        [("starttag", "a",
754                          [("width", "100%"), ("cellspacing","0")])])
755
756        self._run_check('<a id="foo"class="bar">',
757                        [("starttag", "a",
758                          [("id", "foo"), ("class","bar")])])
759
760    def test_missing_attribute_value(self):
761        self._run_check('<a v=>',
762                        [("starttag", "a", [("v", "")])])
763
764    def test_javascript_attribute_value(self):
765        self._run_check("<a href=javascript:popup('/popup/help.html')>",
766                        [("starttag", "a",
767                          [("href", "javascript:popup('/popup/help.html')")])])
768
769    def test_end_tag_in_attribute_value(self):
770        # see #1745761
771        self._run_check("<a href='http://www.example.org/\">;'>spam</a>",
772                        [("starttag", "a",
773                          [("href", "http://www.example.org/\">;")]),
774                         ("data", "spam"), ("endtag", "a")])
775
776
777if __name__ == "__main__":
778    unittest.main()
779