Testing the ability to include libraries with the bot

lib/bs4/tests/__init__.py

@@ -0,0 +1 @@
+"The beautifulsoup tests."

lib/bs4/tests/test_builder_registry.py

@@ -0,0 +1,141 @@
+"""Tests of the builder registry."""
+
+import unittest
+
+from bs4 import BeautifulSoup
+from bs4.builder import (
+    builder_registry as registry,
+    HTMLParserTreeBuilder,
+    TreeBuilderRegistry,
+)
+
+try:
+    from bs4.builder import HTML5TreeBuilder
+    HTML5LIB_PRESENT = True
+except ImportError:
+    HTML5LIB_PRESENT = False
+
+try:
+    from bs4.builder import (
+        LXMLTreeBuilderForXML,
+        LXMLTreeBuilder,
+        )
+    LXML_PRESENT = True
+except ImportError:
+    LXML_PRESENT = False
+
+
+class BuiltInRegistryTest(unittest.TestCase):
+    """Test the built-in registry with the default builders registered."""
+
+    def test_combination(self):
+        if LXML_PRESENT:
+            self.assertEqual(registry.lookup('fast', 'html'),
+                             LXMLTreeBuilder)
+
+        if LXML_PRESENT:
+            self.assertEqual(registry.lookup('permissive', 'xml'),
+                             LXMLTreeBuilderForXML)
+        self.assertEqual(registry.lookup('strict', 'html'),
+                          HTMLParserTreeBuilder)
+        if HTML5LIB_PRESENT:
+            self.assertEqual(registry.lookup('html5lib', 'html'),
+                              HTML5TreeBuilder)
+
+    def test_lookup_by_markup_type(self):
+        if LXML_PRESENT:
+            self.assertEqual(registry.lookup('html'), LXMLTreeBuilder)
+            self.assertEqual(registry.lookup('xml'), LXMLTreeBuilderForXML)
+        else:
+            self.assertEqual(registry.lookup('xml'), None)
+            if HTML5LIB_PRESENT:
+                self.assertEqual(registry.lookup('html'), HTML5TreeBuilder)
+            else:
+                self.assertEqual(registry.lookup('html'), HTMLParserTreeBuilder)
+
+    def test_named_library(self):
+        if LXML_PRESENT:
+            self.assertEqual(registry.lookup('lxml', 'xml'),
+                             LXMLTreeBuilderForXML)
+            self.assertEqual(registry.lookup('lxml', 'html'),
+                             LXMLTreeBuilder)
+        if HTML5LIB_PRESENT:
+            self.assertEqual(registry.lookup('html5lib'),
+                              HTML5TreeBuilder)
+
+        self.assertEqual(registry.lookup('html.parser'),
+                          HTMLParserTreeBuilder)
+
+    def test_beautifulsoup_constructor_does_lookup(self):
+        # You can pass in a string.
+        BeautifulSoup("", features="html")
+        # Or a list of strings.
+        BeautifulSoup("", features=["html", "fast"])
+
+        # You'll get an exception if BS can't find an appropriate
+        # builder.
+        self.assertRaises(ValueError, BeautifulSoup,
+                          "", features="no-such-feature")
+
+class RegistryTest(unittest.TestCase):
+    """Test the TreeBuilderRegistry class in general."""
+
+    def setUp(self):
+        self.registry = TreeBuilderRegistry()
+
+    def builder_for_features(self, *feature_list):
+        cls = type('Builder_' + '_'.join(feature_list),
+                   (object,), {'features' : feature_list})
+
+        self.registry.register(cls)
+        return cls
+
+    def test_register_with_no_features(self):
+        builder = self.builder_for_features()
+
+        # Since the builder advertises no features, you can't find it
+        # by looking up features.
+        self.assertEqual(self.registry.lookup('foo'), None)
+
+        # But you can find it by doing a lookup with no features, if
+        # this happens to be the only registered builder.
+        self.assertEqual(self.registry.lookup(), builder)
+
+    def test_register_with_features_makes_lookup_succeed(self):
+        builder = self.builder_for_features('foo', 'bar')
+        self.assertEqual(self.registry.lookup('foo'), builder)
+        self.assertEqual(self.registry.lookup('bar'), builder)
+
+    def test_lookup_fails_when_no_builder_implements_feature(self):
+        builder = self.builder_for_features('foo', 'bar')
+        self.assertEqual(self.registry.lookup('baz'), None)
+
+    def test_lookup_gets_most_recent_registration_when_no_feature_specified(self):
+        builder1 = self.builder_for_features('foo')
+        builder2 = self.builder_for_features('bar')
+        self.assertEqual(self.registry.lookup(), builder2)
+
+    def test_lookup_fails_when_no_tree_builders_registered(self):
+        self.assertEqual(self.registry.lookup(), None)
+
+    def test_lookup_gets_most_recent_builder_supporting_all_features(self):
+        has_one = self.builder_for_features('foo')
+        has_the_other = self.builder_for_features('bar')
+        has_both_early = self.builder_for_features('foo', 'bar', 'baz')
+        has_both_late = self.builder_for_features('foo', 'bar', 'quux')
+        lacks_one = self.builder_for_features('bar')
+        has_the_other = self.builder_for_features('foo')
+
+        # There are two builders featuring 'foo' and 'bar', but
+        # the one that also features 'quux' was registered later.
+        self.assertEqual(self.registry.lookup('foo', 'bar'),
+                          has_both_late)
+
+        # There is only one builder featuring 'foo', 'bar', and 'baz'.
+        self.assertEqual(self.registry.lookup('foo', 'bar', 'baz'),
+                          has_both_early)
+
+    def test_lookup_fails_when_cannot_reconcile_requested_features(self):
+        builder1 = self.builder_for_features('foo', 'bar')
+        builder2 = self.builder_for_features('foo', 'baz')
+        self.assertEqual(self.registry.lookup('bar', 'baz'), None)

lib/bs4/tests/test_docs.py

@@ -0,0 +1,36 @@
+"Test harness for doctests."
+
+# pylint: disable-msg=E0611,W0142
+
+__metaclass__ = type
+__all__ = [
+    'additional_tests',
+    ]
+
+import atexit
+import doctest
+import os
+#from pkg_resources import (
+#    resource_filename, resource_exists, resource_listdir, cleanup_resources)
+import unittest
+
+DOCTEST_FLAGS = (
+    doctest.ELLIPSIS |
+    doctest.NORMALIZE_WHITESPACE |
+    doctest.REPORT_NDIFF)
+
+
+# def additional_tests():
+#     "Run the doc tests (README.txt and docs/*, if any exist)"
+#     doctest_files = [
+#         os.path.abspath(resource_filename('bs4', 'README.txt'))]
+#     if resource_exists('bs4', 'docs'):
+#         for name in resource_listdir('bs4', 'docs'):
+#             if name.endswith('.txt'):
+#                 doctest_files.append(
+#                     os.path.abspath(
+#                         resource_filename('bs4', 'docs/%s' % name)))
+#     kwargs = dict(module_relative=False, optionflags=DOCTEST_FLAGS)
+#     atexit.register(cleanup_resources)
+#     return unittest.TestSuite((
+#         doctest.DocFileSuite(*doctest_files, **kwargs)))

lib/bs4/tests/test_html5lib.py

@@ -0,0 +1,58 @@
+"""Tests to ensure that the html5lib tree builder generates good trees."""
+
+import warnings
+
+try:
+    from bs4.builder import HTML5TreeBuilder
+    HTML5LIB_PRESENT = True
+except ImportError, e:
+    HTML5LIB_PRESENT = False
+from bs4.element import SoupStrainer
+from bs4.testing import (
+    HTML5TreeBuilderSmokeTest,
+    SoupTest,
+    skipIf,
+)
+
+@skipIf(
+    not HTML5LIB_PRESENT,
+    "html5lib seems not to be present, not testing its tree builder.")
+class HTML5LibBuilderSmokeTest(SoupTest, HTML5TreeBuilderSmokeTest):
+    """See ``HTML5TreeBuilderSmokeTest``."""
+
+    @property
+    def default_builder(self):
+        return HTML5TreeBuilder()
+
+    def test_soupstrainer(self):
+        # The html5lib tree builder does not support SoupStrainers.
+        strainer = SoupStrainer("b")
+        markup = "<p>A <b>bold</b> statement.</p>"
+        with warnings.catch_warnings(record=True) as w:
+            soup = self.soup(markup, parse_only=strainer)
+        self.assertEqual(
+            soup.decode(), self.document_for(markup))
+
+        self.assertTrue(
+            "the html5lib tree builder doesn't support parse_only" in
+            str(w[0].message))
+
+    def test_correctly_nested_tables(self):
+        """html5lib inserts <tbody> tags where other parsers don't."""
+        markup = ('<table id="1">'
+                  '<tr>'
+                  "<td>Here's another table:"
+                  '<table id="2">'
+                  '<tr><td>foo</td></tr>'
+                  '</table></td>')
+
+        self.assertSoupEquals(
+            markup,
+            '<table id="1"><tbody><tr><td>Here\'s another table:'
+            '<table id="2"><tbody><tr><td>foo</td></tr></tbody></table>'
+            '</td></tr></tbody></table>')
+
+        self.assertSoupEquals(
+            "<table><thead><tr><td>Foo</td></tr></thead>"
+            "<tbody><tr><td>Bar</td></tr></tbody>"
+            "<tfoot><tr><td>Baz</td></tr></tfoot></table>")

lib/bs4/tests/test_htmlparser.py

@@ -0,0 +1,19 @@
+"""Tests to ensure that the html.parser tree builder generates good
+trees."""
+
+from bs4.testing import SoupTest, HTMLTreeBuilderSmokeTest
+from bs4.builder import HTMLParserTreeBuilder
+
+class HTMLParserTreeBuilderSmokeTest(SoupTest, HTMLTreeBuilderSmokeTest):
+
+    @property
+    def default_builder(self):
+        return HTMLParserTreeBuilder()
+
+    def test_namespaced_system_doctype(self):
+        # html.parser can't handle namespaced doctypes, so skip this one.
+        pass
+
+    def test_namespaced_public_doctype(self):
+        # html.parser can't handle namespaced doctypes, so skip this one.
+        pass

lib/bs4/tests/test_lxml.py

@@ -0,0 +1,75 @@
+"""Tests to ensure that the lxml tree builder generates good trees."""
+
+import re
+import warnings
+
+try:
+    from bs4.builder import LXMLTreeBuilder, LXMLTreeBuilderForXML
+    LXML_PRESENT = True
+except ImportError, e:
+    LXML_PRESENT = False
+
+from bs4 import (
+    BeautifulSoup,
+    BeautifulStoneSoup,
+    )
+from bs4.element import Comment, Doctype, SoupStrainer
+from bs4.testing import skipIf
+from bs4.tests import test_htmlparser
+from bs4.testing import (
+    HTMLTreeBuilderSmokeTest,
+    XMLTreeBuilderSmokeTest,
+    SoupTest,
+    skipIf,
+)
+
+@skipIf(
+    not LXML_PRESENT,
+    "lxml seems not to be present, not testing its tree builder.")
+class LXMLTreeBuilderSmokeTest(SoupTest, HTMLTreeBuilderSmokeTest):
+    """See ``HTMLTreeBuilderSmokeTest``."""
+
+    @property
+    def default_builder(self):
+        return LXMLTreeBuilder()
+
+    def test_out_of_range_entity(self):
+        self.assertSoupEquals(
+            "<p>foo&#10000000000000;bar</p>", "<p>foobar</p>")
+        self.assertSoupEquals(
+            "<p>foo&#x10000000000000;bar</p>", "<p>foobar</p>")
+        self.assertSoupEquals(
+            "<p>foo&#1000000000;bar</p>", "<p>foobar</p>")
+
+    def test_beautifulstonesoup_is_xml_parser(self):
+        # Make sure that the deprecated BSS class uses an xml builder
+        # if one is installed.
+        with warnings.catch_warnings(record=False) as w:
+            soup = BeautifulStoneSoup("<b />")
+            self.assertEqual(u"<b/>", unicode(soup.b))
+
+    def test_real_xhtml_document(self):
+        """lxml strips the XML definition from an XHTML doc, which is fine."""
+        markup = b"""<?xml version="1.0" encoding="utf-8"?>
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN">
+<html xmlns="http://www.w3.org/1999/xhtml">
+<head><title>Hello.</title></head>
+<body>Goodbye.</body>
+</html>"""
+        soup = self.soup(markup)
+        self.assertEqual(
+            soup.encode("utf-8").replace(b"\n", b''),
+            markup.replace(b'\n', b'').replace(
+                b'<?xml version="1.0" encoding="utf-8"?>', b''))
+
+
+@skipIf(
+    not LXML_PRESENT,
+    "lxml seems not to be present, not testing its XML tree builder.")
+class LXMLXMLTreeBuilderSmokeTest(SoupTest, XMLTreeBuilderSmokeTest):
+    """See ``HTMLTreeBuilderSmokeTest``."""
+
+    @property
+    def default_builder(self):
+        return LXMLTreeBuilderForXML()
+

lib/bs4/tests/test_soup.py

@@ -0,0 +1,378 @@
+# -*- coding: utf-8 -*-
+"""Tests of Beautiful Soup as a whole."""
+
+import logging
+import unittest
+import sys
+from bs4 import (
+    BeautifulSoup,
+    BeautifulStoneSoup,
+)
+from bs4.element import (
+    CharsetMetaAttributeValue,
+    ContentMetaAttributeValue,
+    SoupStrainer,
+    NamespacedAttribute,
+    )
+import bs4.dammit
+from bs4.dammit import EntitySubstitution, UnicodeDammit
+from bs4.testing import (
+    SoupTest,
+    skipIf,
+)
+import warnings
+
+try:
+    from bs4.builder import LXMLTreeBuilder, LXMLTreeBuilderForXML
+    LXML_PRESENT = True
+except ImportError, e:
+    LXML_PRESENT = False
+
+PYTHON_2_PRE_2_7 = (sys.version_info < (2,7))
+PYTHON_3_PRE_3_2 = (sys.version_info[0] == 3 and sys.version_info < (3,2))
+
+class TestDeprecatedConstructorArguments(SoupTest):
+
+    def test_parseOnlyThese_renamed_to_parse_only(self):
+        with warnings.catch_warnings(record=True) as w:
+            soup = self.soup("<a><b></b></a>", parseOnlyThese=SoupStrainer("b"))
+        msg = str(w[0].message)
+        self.assertTrue("parseOnlyThese" in msg)
+        self.assertTrue("parse_only" in msg)
+        self.assertEqual(b"<b></b>", soup.encode())
+
+    def test_fromEncoding_renamed_to_from_encoding(self):
+        with warnings.catch_warnings(record=True) as w:
+            utf8 = b"\xc3\xa9"
+            soup = self.soup(utf8, fromEncoding="utf8")
+        msg = str(w[0].message)
+        self.assertTrue("fromEncoding" in msg)
+        self.assertTrue("from_encoding" in msg)
+        self.assertEqual("utf8", soup.original_encoding)
+
+    def test_unrecognized_keyword_argument(self):
+        self.assertRaises(
+            TypeError, self.soup, "<a>", no_such_argument=True)
+
+    @skipIf(
+        not LXML_PRESENT,
+        "lxml not present, not testing BeautifulStoneSoup.")
+    def test_beautifulstonesoup(self):
+        with warnings.catch_warnings(record=True) as w:
+            soup = BeautifulStoneSoup("<markup>")
+            self.assertTrue(isinstance(soup, BeautifulSoup))
+            self.assertTrue("BeautifulStoneSoup class is deprecated")
+
+class TestSelectiveParsing(SoupTest):
+
+    def test_parse_with_soupstrainer(self):
+        markup = "No<b>Yes</b><a>No<b>Yes <c>Yes</c></b>"
+        strainer = SoupStrainer("b")
+        soup = self.soup(markup, parse_only=strainer)
+        self.assertEqual(soup.encode(), b"<b>Yes</b><b>Yes <c>Yes</c></b>")
+
+
+class TestEntitySubstitution(unittest.TestCase):
+    """Standalone tests of the EntitySubstitution class."""
+    def setUp(self):
+        self.sub = EntitySubstitution
+
+    def test_simple_html_substitution(self):
+        # Unicode characters corresponding to named HTML entites
+        # are substituted, and no others.
+        s = u"foo\u2200\N{SNOWMAN}\u00f5bar"
+        self.assertEqual(self.sub.substitute_html(s),
+                          u"foo&forall;\N{SNOWMAN}&otilde;bar")
+
+    def test_smart_quote_substitution(self):
+        # MS smart quotes are a common source of frustration, so we
+        # give them a special test.
+        quotes = b"\x91\x92foo\x93\x94"
+        dammit = UnicodeDammit(quotes)
+        self.assertEqual(self.sub.substitute_html(dammit.markup),
+                          "&lsquo;&rsquo;foo&ldquo;&rdquo;")
+
+    def test_xml_converstion_includes_no_quotes_if_make_quoted_attribute_is_false(self):
+        s = 'Welcome to "my bar"'
+        self.assertEqual(self.sub.substitute_xml(s, False), s)
+
+    def test_xml_attribute_quoting_normally_uses_double_quotes(self):
+        self.assertEqual(self.sub.substitute_xml("Welcome", True),
+                          '"Welcome"')
+        self.assertEqual(self.sub.substitute_xml("Bob's Bar", True),
+                          '"Bob\'s Bar"')
+
+    def test_xml_attribute_quoting_uses_single_quotes_when_value_contains_double_quotes(self):
+        s = 'Welcome to "my bar"'
+        self.assertEqual(self.sub.substitute_xml(s, True),
+                          "'Welcome to \"my bar\"'")
+
+    def test_xml_attribute_quoting_escapes_single_quotes_when_value_contains_both_single_and_double_quotes(self):
+        s = 'Welcome to "Bob\'s Bar"'
+        self.assertEqual(
+            self.sub.substitute_xml(s, True),
+            '"Welcome to &quot;Bob\'s Bar&quot;"')
+
+    def test_xml_quotes_arent_escaped_when_value_is_not_being_quoted(self):
+        quoted = 'Welcome to "Bob\'s Bar"'
+        self.assertEqual(self.sub.substitute_xml(quoted), quoted)
+
+    def test_xml_quoting_handles_angle_brackets(self):
+        self.assertEqual(
+            self.sub.substitute_xml("foo<bar>"),
+            "foo&lt;bar&gt;")
+
+    def test_xml_quoting_handles_ampersands(self):
+        self.assertEqual(self.sub.substitute_xml("AT&T"), "AT&amp;T")
+
+    def test_xml_quoting_ignores_ampersands_when_they_are_part_of_an_entity(self):
+        self.assertEqual(
+            self.sub.substitute_xml("&Aacute;T&T"),
+            "&Aacute;T&amp;T")
+
+    def test_quotes_not_html_substituted(self):
+        """There's no need to do this except inside attribute values."""
+        text = 'Bob\'s "bar"'
+        self.assertEqual(self.sub.substitute_html(text), text)
+
+
+class TestEncodingConversion(SoupTest):
+    # Test Beautiful Soup's ability to decode and encode from various
+    # encodings.
+
+    def setUp(self):
+        super(TestEncodingConversion, self).setUp()
+        self.unicode_data = u'<html><head><meta charset="utf-8"/></head><body><foo>Sacr\N{LATIN SMALL LETTER E WITH ACUTE} bleu!</foo></body></html>'
+        self.utf8_data = self.unicode_data.encode("utf-8")
+        # Just so you know what it looks like.
+        self.assertEqual(
+            self.utf8_data,
+            b'<html><head><meta charset="utf-8"/></head><body><foo>Sacr\xc3\xa9 bleu!</foo></body></html>')
+
+    def test_ascii_in_unicode_out(self):
+        # ASCII input is converted to Unicode. The original_encoding
+        # attribute is set.
+        ascii = b"<foo>a</foo>"
+        soup_from_ascii = self.soup(ascii)
+        unicode_output = soup_from_ascii.decode()
+        self.assertTrue(isinstance(unicode_output, unicode))
+        self.assertEqual(unicode_output, self.document_for(ascii.decode()))
+        self.assertEqual(soup_from_ascii.original_encoding.lower(), "ascii")
+
+    def test_unicode_in_unicode_out(self):
+        # Unicode input is left alone. The original_encoding attribute
+        # is not set.
+        soup_from_unicode = self.soup(self.unicode_data)
+        self.assertEqual(soup_from_unicode.decode(), self.unicode_data)
+        self.assertEqual(soup_from_unicode.foo.string, u'Sacr\xe9 bleu!')
+        self.assertEqual(soup_from_unicode.original_encoding, None)
+
+    def test_utf8_in_unicode_out(self):
+        # UTF-8 input is converted to Unicode. The original_encoding
+        # attribute is set.
+        soup_from_utf8 = self.soup(self.utf8_data)
+        self.assertEqual(soup_from_utf8.decode(), self.unicode_data)
+        self.assertEqual(soup_from_utf8.foo.string, u'Sacr\xe9 bleu!')
+
+    def test_utf8_out(self):
+        # The internal data structures can be encoded as UTF-8.
+        soup_from_unicode = self.soup(self.unicode_data)
+        self.assertEqual(soup_from_unicode.encode('utf-8'), self.utf8_data)
+
+    @skipIf(
+        PYTHON_2_PRE_2_7 or PYTHON_3_PRE_3_2,
+        "Bad HTMLParser detected; skipping test of non-ASCII characters in attribute name.")
+    def test_attribute_name_containing_unicode_characters(self):
+        markup = u'<div><a \N{SNOWMAN}="snowman"></a></div>'
+        self.assertEqual(self.soup(markup).div.encode("utf8"), markup.encode("utf8"))
+
+class TestUnicodeDammit(unittest.TestCase):
+    """Standalone tests of Unicode, Dammit."""
+
+    def test_smart_quotes_to_unicode(self):
+        markup = b"<foo>\x91\x92\x93\x94</foo>"
+        dammit = UnicodeDammit(markup)
+        self.assertEqual(
+            dammit.unicode_markup, u"<foo>\u2018\u2019\u201c\u201d</foo>")
+
+    def test_smart_quotes_to_xml_entities(self):
+        markup = b"<foo>\x91\x92\x93\x94</foo>"
+        dammit = UnicodeDammit(markup, smart_quotes_to="xml")
+        self.assertEqual(
+            dammit.unicode_markup, "<foo>&#x2018;&#x2019;&#x201C;&#x201D;</foo>")
+
+    def test_smart_quotes_to_html_entities(self):
+        markup = b"<foo>\x91\x92\x93\x94</foo>"
+        dammit = UnicodeDammit(markup, smart_quotes_to="html")
+        self.assertEqual(
+            dammit.unicode_markup, "<foo>&lsquo;&rsquo;&ldquo;&rdquo;</foo>")
+
+    def test_smart_quotes_to_ascii(self):
+        markup = b"<foo>\x91\x92\x93\x94</foo>"
+        dammit = UnicodeDammit(markup, smart_quotes_to="ascii")
+        self.assertEqual(
+            dammit.unicode_markup, """<foo>''""</foo>""")
+
+    def test_detect_utf8(self):
+        utf8 = b"\xc3\xa9"
+        dammit = UnicodeDammit(utf8)
+        self.assertEqual(dammit.unicode_markup, u'\xe9')
+        self.assertEqual(dammit.original_encoding.lower(), 'utf-8')
+
+    def test_convert_hebrew(self):
+        hebrew = b"\xed\xe5\xec\xf9"
+        dammit = UnicodeDammit(hebrew, ["iso-8859-8"])
+        self.assertEqual(dammit.original_encoding.lower(), 'iso-8859-8')
+        self.assertEqual(dammit.unicode_markup, u'\u05dd\u05d5\u05dc\u05e9')
+
+    def test_dont_see_smart_quotes_where_there_are_none(self):
+        utf_8 = b"\343\202\261\343\203\274\343\202\277\343\202\244 Watch"
+        dammit = UnicodeDammit(utf_8)
+        self.assertEqual(dammit.original_encoding.lower(), 'utf-8')
+        self.assertEqual(dammit.unicode_markup.encode("utf-8"), utf_8)
+
+    def test_ignore_inappropriate_codecs(self):
+        utf8_data = u"Räksmörgås".encode("utf-8")
+        dammit = UnicodeDammit(utf8_data, ["iso-8859-8"])
+        self.assertEqual(dammit.original_encoding.lower(), 'utf-8')
+
+    def test_ignore_invalid_codecs(self):
+        utf8_data = u"Räksmörgås".encode("utf-8")
+        for bad_encoding in ['.utf8', '...', 'utF---16.!']:
+            dammit = UnicodeDammit(utf8_data, [bad_encoding])
+            self.assertEqual(dammit.original_encoding.lower(), 'utf-8')
+
+    def test_detect_html5_style_meta_tag(self):
+
+        for data in (
+            b'<html><meta charset="euc-jp" /></html>',
+            b"<html><meta charset='euc-jp' /></html>",
+            b"<html><meta charset=euc-jp /></html>",
+            b"<html><meta charset=euc-jp/></html>"):
+            dammit = UnicodeDammit(data, is_html=True)
+            self.assertEqual(
+                "euc-jp", dammit.original_encoding)
+
+    def test_last_ditch_entity_replacement(self):
+        # This is a UTF-8 document that contains bytestrings
+        # completely incompatible with UTF-8 (ie. encoded with some other
+        # encoding).
+        #
+        # Since there is no consistent encoding for the document,
+        # Unicode, Dammit will eventually encode the document as UTF-8
+        # and encode the incompatible characters as REPLACEMENT
+        # CHARACTER.
+        #
+        # If chardet is installed, it will detect that the document
+        # can be converted into ISO-8859-1 without errors. This happens
+        # to be the wrong encoding, but it is a consistent encoding, so the
+        # code we're testing here won't run.
+        #
+        # So we temporarily disable chardet if it's present.
+        doc = b"""\357\273\277<?xml version="1.0" encoding="UTF-8"?>
+<html><b>\330\250\330\252\330\261</b>
+<i>\310\322\321\220\312\321\355\344</i></html>"""
+        chardet = bs4.dammit.chardet_dammit
+        logging.disable(logging.WARNING)
+        try:
+            def noop(str):
+                return None
+            bs4.dammit.chardet_dammit = noop
+            dammit = UnicodeDammit(doc)
+            self.assertEqual(True, dammit.contains_replacement_characters)
+            self.assertTrue(u"\ufffd" in dammit.unicode_markup)
+
+            soup = BeautifulSoup(doc, "html.parser")
+            self.assertTrue(soup.contains_replacement_characters)
+        finally:
+            logging.disable(logging.NOTSET)
+            bs4.dammit.chardet_dammit = chardet
+
+    def test_sniffed_xml_encoding(self):
+        # A document written in UTF-16LE will be converted by a different
+        # code path that sniffs the byte order markers.
+        data = b'\xff\xfe<\x00a\x00>\x00\xe1\x00\xe9\x00<\x00/\x00a\x00>\x00'
+        dammit = UnicodeDammit(data)
+        self.assertEqual(u"<a>áé</a>", dammit.unicode_markup)
+        self.assertEqual("utf-16le", dammit.original_encoding)
+
+    def test_detwingle(self):
+        # Here's a UTF8 document.
+        utf8 = (u"\N{SNOWMAN}" * 3).encode("utf8")
+
+        # Here's a Windows-1252 document.
+        windows_1252 = (
+            u"\N{LEFT DOUBLE QUOTATION MARK}Hi, I like Windows!"
+            u"\N{RIGHT DOUBLE QUOTATION MARK}").encode("windows_1252")
+
+        # Through some unholy alchemy, they've been stuck together.
+        doc = utf8 + windows_1252 + utf8
+
+        # The document can't be turned into UTF-8:
+        self.assertRaises(UnicodeDecodeError, doc.decode, "utf8")
+
+        # Unicode, Dammit thinks the whole document is Windows-1252,
+        # and decodes it into "☃☃☃“Hi, I like Windows!”☃☃☃"
+
+        # But if we run it through fix_embedded_windows_1252, it's fixed:
+
+        fixed = UnicodeDammit.detwingle(doc)
+        self.assertEqual(
+            u"☃☃☃“Hi, I like Windows!”☃☃☃", fixed.decode("utf8"))
+
+    def test_detwingle_ignores_multibyte_characters(self):
+        # Each of these characters has a UTF-8 representation ending
+        # in \x93. \x93 is a smart quote if interpreted as
+        # Windows-1252. But our code knows to skip over multibyte
+        # UTF-8 characters, so they'll survive the process unscathed.
+        for tricky_unicode_char in (
+            u"\N{LATIN SMALL LIGATURE OE}", # 2-byte char '\xc5\x93'
+            u"\N{LATIN SUBSCRIPT SMALL LETTER X}", # 3-byte char '\xe2\x82\x93'
+            u"\xf0\x90\x90\x93", # This is a CJK character, not sure which one.
+            ):
+            input = tricky_unicode_char.encode("utf8")
+            self.assertTrue(input.endswith(b'\x93'))
+            output = UnicodeDammit.detwingle(input)
+            self.assertEqual(output, input)
+
+class TestNamedspacedAttribute(SoupTest):
+
+    def test_name_may_be_none(self):
+        a = NamespacedAttribute("xmlns", None)
+        self.assertEqual(a, "xmlns")
+
+    def test_attribute_is_equivalent_to_colon_separated_string(self):
+        a = NamespacedAttribute("a", "b")
+        self.assertEqual("a:b", a)
+
+    def test_attributes_are_equivalent_if_prefix_and_name_identical(self):
+        a = NamespacedAttribute("a", "b", "c")
+        b = NamespacedAttribute("a", "b", "c")
+        self.assertEqual(a, b)
+
+        # The actual namespace is not considered.
+        c = NamespacedAttribute("a", "b", None)
+        self.assertEqual(a, c)
+
+        # But name and prefix are important.
+        d = NamespacedAttribute("a", "z", "c")
+        self.assertNotEqual(a, d)
+
+        e = NamespacedAttribute("z", "b", "c")
+        self.assertNotEqual(a, e)
+
+
+class TestAttributeValueWithCharsetSubstitution(unittest.TestCase):
+
+    def test_content_meta_attribute_value(self):
+        value = CharsetMetaAttributeValue("euc-jp")
+        self.assertEqual("euc-jp", value)
+        self.assertEqual("euc-jp", value.original_value)
+        self.assertEqual("utf8", value.encode("utf8"))
+
+
+    def test_content_meta_attribute_value(self):
+        value = ContentMetaAttributeValue("text/html; charset=euc-jp")
+        self.assertEqual("text/html; charset=euc-jp", value)
+        self.assertEqual("text/html; charset=euc-jp", value.original_value)
+        self.assertEqual("text/html; charset=utf8", value.encode("utf8"))