Merge branch 'feature/UpdateBS4' into develop

CHANGES.md

@@ -1,5 +1,7 @@
 ### 0.18.0 (2018-xx-xx xx:xx:xx UTC)
 
+* Update Beautiful Soup 4.6.0 (r449) to 4.6.3 (r475)
+
 
 [develop changelog]
 

lib/bs4/__init__.py

@@ -21,14 +21,15 @@ http://www.crummy.com/software/BeautifulSoup/bs4/doc/
 # found in the LICENSE file.
 
 __author__ = "Leonard Richardson (leonardr@segfault.org)"
-__version__ = "4.5.3"
+__version__ = "4.6.3"
-__copyright__ = "Copyright (c) 2004-2017 Leonard Richardson"
+__copyright__ = "Copyright (c) 2004-2018 Leonard Richardson"
 __license__ = "MIT"
 
 __all__ = ['BeautifulSoup']
 
 import os
 import re
+import sys
 import traceback
 import warnings
 
@@ -82,14 +83,46 @@ class BeautifulSoup(Tag):
 
     ASCII_SPACES = '\x20\x0a\x09\x0c\x0d'
 
-    NO_PARSER_SPECIFIED_WARNING = "No parser was explicitly specified, so I'm using the best available %(markup_type)s parser for this system (\"%(parser)s\"). This usually isn't a problem, but if you run this code on another system, or in a different virtual environment, it may use a different parser and behave differently.\n\nThe code that caused this warning is on line %(line_number)s of the file %(filename)s. To get rid of this warning, change code that looks like this:\n\n BeautifulSoup(YOUR_MARKUP})\n\nto this:\n\n BeautifulSoup(YOUR_MARKUP, \"%(parser)s\")\n"
+    NO_PARSER_SPECIFIED_WARNING = "No parser was explicitly specified, so I'm using the best available %(markup_type)s parser for this system (\"%(parser)s\"). This usually isn't a problem, but if you run this code on another system, or in a different virtual environment, it may use a different parser and behave differently.\n\nThe code that caused this warning is on line %(line_number)s of the file %(filename)s. To get rid of this warning, pass the additional argument 'features=\"%(parser)s\"' to the BeautifulSoup constructor.\n"
 
     def __init__(self, markup="", features=None, builder=None,
                  parse_only=None, from_encoding=None, exclude_encodings=None,
                  **kwargs):
-        """The Soup object is initialized as the 'root tag', and the
+        """Constructor.
-        provided markup (which can be a string or a file-like object)
+
-        is fed into the underlying parser."""
+        :param markup: A string or a file-like object representing
+        markup to be parsed.
+
+        :param features: Desirable features of the parser to be used. This
+        may be the name of a specific parser ("lxml", "lxml-xml",
+        "html.parser", or "html5lib") or it may be the type of markup
+        to be used ("html", "html5", "xml"). It's recommended that you
+        name a specific parser, so that Beautiful Soup gives you the
+        same results across platforms and virtual environments.
+
+        :param builder: A specific TreeBuilder to use instead of looking one
+        up based on `features`. You shouldn't need to use this.
+
+        :param parse_only: A SoupStrainer. Only parts of the document
+        matching the SoupStrainer will be considered. This is useful
+        when parsing part of a document that would otherwise be too
+        large to fit into memory.
+
+        :param from_encoding: A string indicating the encoding of the
+        document to be parsed. Pass this in if Beautiful Soup is
+        guessing wrongly about the document's encoding.
+
+        :param exclude_encodings: A list of strings indicating
+        encodings known to be wrong. Pass this in if you don't know
+        the document's encoding but you know Beautiful Soup's guess is
+        wrong.
+
+        :param kwargs: For backwards compatibility purposes, the
+        constructor accepts certain keyword arguments used in
+        Beautiful Soup 3. None of these arguments do anything in
+        Beautiful Soup 4 and there's no need to actually pass keyword
+        arguments into the constructor.
+        """
 
         if 'convertEntities' in kwargs:
             warnings.warn(
@@ -171,14 +204,35 @@ class BeautifulSoup(Tag):
                 else:
                     markup_type = "HTML"
 
-                caller = traceback.extract_stack()[0]
+                # This code adapted from warnings.py so that we get the same line
-                filename = caller[0]
+                # of code as our warnings.warn() call gets, even if the answer is wrong
-                line_number = caller[1]
+                # (as it may be in a multithreading situation).
-                warnings.warn(self.NO_PARSER_SPECIFIED_WARNING % dict(
+                caller = None
+                try:
+                    caller = sys._getframe(1)
+                except ValueError:
+                    pass
+                if caller:
+                    globals = caller.f_globals
+                    line_number = caller.f_lineno
+                else:
+                    globals = sys.__dict__
+                    line_number= 1                    
+                filename = globals.get('__file__')
+                if filename:
+                    fnl = filename.lower()
+                    if fnl.endswith((".pyc", ".pyo")):
+                        filename = filename[:-1]
+                if filename:
+                    # If there is no filename at all, the user is most likely in a REPL,
+                    # and the warning is not necessary.
+                    values = dict(
                         filename=filename,
                         line_number=line_number,
                         parser=builder.NAME,
-                    markup_type=markup_type))
+                        markup_type=markup_type
+                    )
+                    warnings.warn(self.NO_PARSER_SPECIFIED_WARNING % values, stacklevel=2)
 
         self.builder = builder
         self.is_xml = builder.is_xml
@@ -302,9 +356,10 @@ class BeautifulSoup(Tag):
         self.preserve_whitespace_tag_stack = []
         self.pushTag(self)
 
-    def new_tag(self, name, namespace=None, nsprefix=None, **attrs):
+    def new_tag(self, name, namespace=None, nsprefix=None, attrs={}, **kwattrs):
         """Create a new tag associated with this soup."""
-        return Tag(None, self.builder, name, namespace, nsprefix, attrs)
+        kwattrs.update(attrs)
+        return Tag(None, self.builder, name, namespace, nsprefix, kwattrs)
 
     def new_string(self, s, subclass=NavigableString):
         """Create a new NavigableString associated with this soup."""

lib/bs4/builder/__init__.py

@@ -236,10 +236,16 @@ class HTMLTreeBuilder(TreeBuilder):
         # These are from HTML5.
         'area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'keygen', 'link', 'menuitem', 'meta', 'param', 'source', 'track', 'wbr',
         
-        # These are from HTML4, removed in HTML5.
+        # These are from earlier versions of HTML and are removed in HTML5.
-        'spacer', 'frame'
+        'basefont', 'bgsound', 'command', 'frame', 'image', 'isindex', 'nextid', 'spacer'
     ])
 
+    # The HTML standard defines these as block-level elements. Beautiful
+    # Soup does not treat these elements differently from other elements,
+    # but it may do so eventually, and this information is available if
+    # you need to use it.
+    block_elements = set(["address", "article", "aside", "blockquote", "canvas", "dd", "div", "dl", "dt", "fieldset", "figcaption", "figure", "footer", "form", "h1", "h2", "h3", "h4", "h5", "h6", "header", "hr", "li", "main", "nav", "noscript", "ol", "output", "p", "pre", "section", "table", "tfoot", "ul", "video"])
+    
     # The HTML standard defines these attributes as containing a
     # space-separated list of values, not a single value. That is,
     # class="foo bar" means that the 'class' attribute has two values,

lib/bs4/builder/_html5lib.py

@@ -30,14 +30,13 @@ from bs4.element import (
     )
 
 try:
-    # 0.99999999 and up
-    from html5lib.treebuilders import base as treebuilder_base
-    old_html5lib = False
-except ImportError:
     # Pre-0.99999999
     from html5lib.treebuilders import _base as treebuilder_base
-    old_html5lib = True
+    new_html5lib = False
-
+except ImportError, e:
+    # 0.99999999 and up
+    from html5lib.treebuilders import base as treebuilder_base
+    new_html5lib = True
 
 class HTML5TreeBuilder(HTMLTreeBuilder):
     """Use html5lib to build a tree."""
@@ -66,7 +65,7 @@ class HTML5TreeBuilder(HTMLTreeBuilder):
 
         extra_kwargs = dict()
         if not isinstance(markup, unicode):
-            if not old_html5lib:
+            if new_html5lib:
                 extra_kwargs['override_encoding'] = self.user_specified_encoding
             else:
                 extra_kwargs['encoding'] = self.user_specified_encoding

lib/bs4/builder/_htmlparser.py

@@ -1,3 +1,4 @@
+# encoding: utf-8
 """Use the HTMLParser library to parse HTML files that aren't too bad."""
 
 # Use of this source code is governed by a BSD-style license that can be
@@ -65,6 +66,17 @@ class BeautifulSoupHTMLParser(HTMLParser):
         # will ignore, assuming they ever show up.
         self.already_closed_empty_element = []
 
+    def error(self, msg):
+        """In Python 3, HTMLParser subclasses must implement error(), although this
+        requirement doesn't appear to be documented.
+
+        In Python 2, HTMLParser implements error() as raising an exception.
+
+        In any event, this method is called only on very strange markup and our best strategy
+        is to pretend it didn't happen and keep going.
+        """
+        warnings.warn(msg)
+        
     def handle_startendtag(self, name, attrs):
         # This is only called when the markup looks like
         # <tag/>.
@@ -129,11 +141,26 @@ class BeautifulSoupHTMLParser(HTMLParser):
         else:
             real_name = int(name)
 
+        data = None
+        if real_name < 256:
+            # HTML numeric entities are supposed to reference Unicode
+            # code points, but sometimes they reference code points in
+            # some other encoding (ahem, Windows-1252). E.g. &#147;
+            # instead of &#201; for LEFT DOUBLE QUOTATION MARK. This
+            # code tries to detect this situation and compensate.
+            for encoding in (self.soup.original_encoding, 'windows-1252'):
+                if not encoding:
+                    continue
+                try:
+                    data = bytearray([real_name]).decode(encoding)
+                except UnicodeDecodeError, e:
+                    pass
+        if not data:
             try:
                 data = unichr(real_name)
             except (ValueError, OverflowError), e:
-            data = u"\N{REPLACEMENT CHARACTER}"
+                pass
-
+        data = data or u"\N{REPLACEMENT CHARACTER}"
         self.handle_data(data)
 
     def handle_entityref(self, name):
@@ -141,7 +168,12 @@ class BeautifulSoupHTMLParser(HTMLParser):
         if character is not None:
             data = character
         else:
-            data = "&%s;" % name
+            # If this were XML, it would be ambiguous whether "&foo"
+            # was an character entity reference with a missing
+            # semicolon or the literal string "&foo". Since this is
+            # HTML, we have a complete list of all character entity references,
+            # and this one wasn't found, so assume it's the literal string "&foo".
+            data = "&%s" % name
         self.handle_data(data)
 
     def handle_comment(self, data):
@@ -213,6 +245,7 @@ class HTMLParserTreeBuilder(HTMLTreeBuilder):
         parser.soup = self.soup
         try:
             parser.feed(markup)
+            parser.close()
         except HTMLParseError, e:
             warnings.warn(RuntimeWarning(
                 "Python's built-in HTMLParser cannot parse the given document. This is not a bug in Beautiful Soup. The best solution is to install an external parser (lxml or html5lib), and use Beautiful Soup with that parser. See http://www.crummy.com/software/BeautifulSoup/bs4/doc/#installing-a-parser for help."))

lib/bs4/builder/_lxml.py

@@ -5,9 +5,13 @@ __all__ = [
     'LXMLTreeBuilder',
     ]
 
+try:
+    from collections.abc import Callable # Python 3.6
+except ImportError , e:
+    from collections import Callable
+
 from io import BytesIO
 from StringIO import StringIO
-import collections
 from lxml import etree
 from bs4.element import (
     Comment,
@@ -58,7 +62,7 @@ class LXMLTreeBuilderForXML(TreeBuilder):
         # Use the default parser.
         parser = self.default_parser(encoding)
 
-        if isinstance(parser, collections.Callable):
+        if isinstance(parser, Callable):
             # Instantiate the parser with default arguments
             parser = parser(target=self, strip_cdata=False, encoding=encoding)
         return parser
@@ -147,7 +151,7 @@ class LXMLTreeBuilderForXML(TreeBuilder):
         attrs = dict(attrs)
         nsprefix = None
         # Invert each namespace map as it comes in.
-        if len(self.nsmaps) > 1:
+        if len(nsmap) == 0 and len(self.nsmaps) > 1:
                 # There are no new namespaces for this tag, but
                 # non-default namespaces are in play, so we need a
                 # separate tag stack to know when they end.

lib/bs4/dammit.py

@@ -46,9 +46,9 @@ except ImportError:
     pass
 
 xml_encoding_re = re.compile(
-    '^<\?.*encoding=[\'"](.*?)[\'"].*\?>'.encode(), re.I)
+    '^<\\?.*encoding=[\'"](.*?)[\'"].*\\?>'.encode(), re.I)
 html_meta_re = re.compile(
-    '<\s*meta[^>]+charset\s*=\s*["\']?([^>]*?)[ /;\'">]'.encode(), re.I)
+    '<\\s*meta[^>]+charset\\s*=\\s*["\']?([^>]*?)[ /;\'">]'.encode(), re.I)
 
 class EntitySubstitution(object):
 
@@ -82,7 +82,7 @@ class EntitySubstitution(object):
         }
 
     BARE_AMPERSAND_OR_BRACKET = re.compile("([<>]|"
-                                           "&(?!#\d+;|#x[0-9a-fA-F]+;|\w+;)"
+                                           "&(?!#\\d+;|#x[0-9a-fA-F]+;|\\w+;)"
                                            ")")
 
     AMPERSAND_OR_BRACKET = re.compile("([<>&])")

lib/bs4/diagnose.py

@@ -37,7 +37,7 @@ def diagnose(data):
                 name)
 
     if 'lxml' in basic_parsers:
-        basic_parsers.append(["lxml", "xml"])
+        basic_parsers.append("lxml-xml")
         try:
             from lxml import etree
             print "Found lxml version %s" % ".".join(map(str,etree.LXML_VERSION))
@@ -56,21 +56,27 @@ def diagnose(data):
 
     if hasattr(data, 'read'):
         data = data.read()
-    elif os.path.exists(data):
-        print '"%s" looks like a filename. Reading data from the file.' % data
-        with open(data) as fp:
-            data = fp.read()
     elif data.startswith("http:") or data.startswith("https:"):
         print '"%s" looks like a URL. Beautiful Soup is not an HTTP client.' % data
         print "You need to use some other library to get the document behind the URL, and feed that document to Beautiful Soup."
         return
+    else:
+        try:
+            if os.path.exists(data):
+                print '"%s" looks like a filename. Reading data from the file.' % data
+                with open(data) as fp:
+                    data = fp.read()
+        except ValueError:
+            # This can happen on some platforms when the 'filename' is
+            # too long. Assume it's data and not a filename.
+            pass
         print
 
     for parser in basic_parsers:
         print "Trying to parse your markup with %s" % parser
         success = False
         try:
-            soup = BeautifulSoup(data, parser)
+            soup = BeautifulSoup(data, features=parser)
             success = True
         except Exception, e:
             print "%s could not parse the markup." % parser

lib/bs4/element.py

@@ -2,7 +2,10 @@
 # found in the LICENSE file.
 __license__ = "MIT"
 
-import collections
+try:
+    from collections.abc import Callable # Python 3.6
+except ImportError , e:
+    from collections import Callable
 import re
 import shlex
 import sys
@@ -12,7 +15,7 @@ from bs4.dammit import EntitySubstitution
 DEFAULT_OUTPUT_ENCODING = "utf-8"
 PY3K = (sys.version_info[0] > 2)
 
-whitespace_re = re.compile("\s+")
+whitespace_re = re.compile(r"\s+")
 
 def _alias(attr):
     """Alias one attribute name to another for backward compatibility"""
@@ -69,7 +72,7 @@ class ContentMetaAttributeValue(AttributeValueWithCharsetSubstitution):
     The value of the 'content' attribute will be one of these objects.
     """
 
-    CHARSET_RE = re.compile("((^|;)\s*charset=)([^;]*)", re.M)
+    CHARSET_RE = re.compile(r"((^|;)\s*charset=)([^;]*)", re.M)
 
     def __new__(cls, original_value):
         match = cls.CHARSET_RE.search(original_value)
@@ -123,6 +126,41 @@ class HTMLAwareEntitySubstitution(EntitySubstitution):
         return cls._substitute_if_appropriate(
             ns, EntitySubstitution.substitute_xml)
 
+class Formatter(object):
+    """Contains information about how to format a parse tree."""
+    
+    # By default, represent void elements as <tag/> rather than <tag>
+    void_element_close_prefix = '/'
+
+    def substitute_entities(self, *args, **kwargs):
+        """Transform certain characters into named entities."""
+        raise NotImplementedError()
+
+class HTMLFormatter(Formatter):
+    """The default HTML formatter."""
+    def substitute(self, *args, **kwargs):
+        return HTMLAwareEntitySubstitution.substitute_html(*args, **kwargs)
+
+class MinimalHTMLFormatter(Formatter):
+    """A minimal HTML formatter."""
+    def substitute(self, *args, **kwargs):
+        return HTMLAwareEntitySubstitution.substitute_xml(*args, **kwargs)
+    
+class HTML5Formatter(HTMLFormatter):
+    """An HTML formatter that omits the slash in a void tag."""
+    void_element_close_prefix = None
+
+class XMLFormatter(Formatter):
+    """Substitute only the essential XML entities."""
+    def substitute(self, *args, **kwargs):
+        return EntitySubstitution.substitute_xml(*args, **kwargs)
+
+class HTMLXMLFormatter(Formatter):
+    """Format XML using HTML rules."""
+    def substitute(self, *args, **kwargs):
+        return HTMLAwareEntitySubstitution.substitute_html(*args, **kwargs)
+
+    
 class PageElement(object):
     """Contains the navigational information for some part of the page
     (either a tag or a piece of text)"""
@@ -132,39 +170,48 @@ class PageElement(object):
     #
     # "html" - All Unicode characters with corresponding HTML entities
     #   are converted to those entities on output.
+    # "html5" - The same as "html", but empty void tags are represented as
+    #   <tag> rather than <tag/>
     # "minimal" - Bare ampersands and angle brackets are converted to
     #   XML entities: &amp; &lt; &gt;
     # None - The null formatter. Unicode characters are never
     #   converted to entities.  This is not recommended, but it's
     #   faster than "minimal".
-    # A function - This function will be called on every string that
+    # A callable function - it will be called on every string that needs to undergo entity substitution.
+    # A Formatter instance - Formatter.substitute(string) will be called on every string that
     #  needs to undergo entity substitution.
     #
 
-    # In an HTML document, the default "html" and "minimal" functions
+    # In an HTML document, the default "html", "html5", and "minimal"
-    # will leave the contents of <script> and <style> tags alone. For
+    # functions will leave the contents of <script> and <style> tags
-    # an XML document, all tags will be given the same treatment.
+    # alone. For an XML document, all tags will be given the same
+    # treatment.
 
     HTML_FORMATTERS = {
-        "html" : HTMLAwareEntitySubstitution.substitute_html,
+        "html" : HTMLFormatter(),
-        "minimal" : HTMLAwareEntitySubstitution.substitute_xml,
+        "html5" : HTML5Formatter(),
+        "minimal" : MinimalHTMLFormatter(),
         None : None
         }
 
     XML_FORMATTERS = {
-        "html" : EntitySubstitution.substitute_html,
+        "html" : HTMLXMLFormatter(),
-        "minimal" : EntitySubstitution.substitute_xml,
+        "minimal" : XMLFormatter(),
         None : None
         }
 
     def format_string(self, s, formatter='minimal'):
         """Format the given string using the given formatter."""
-        if not callable(formatter):
+        if isinstance(formatter, basestring):
             formatter = self._formatter_for_name(formatter)
         if formatter is None:
             output = s
         else:
+            if callable(formatter):
+                # Backwards compatibility -- you used to pass in a formatting method.
                 output = formatter(s)
+            else:
+                output = formatter.substitute(s)
         return output
 
     @property
@@ -194,11 +241,9 @@ class PageElement(object):
     def _formatter_for_name(self, name):
         "Look up a formatter function based on its name and the tree."
         if self._is_xml:
-            return self.XML_FORMATTERS.get(
+            return self.XML_FORMATTERS.get(name, XMLFormatter())
-                name, EntitySubstitution.substitute_xml)
         else:
-            return self.HTML_FORMATTERS.get(
+            return self.HTML_FORMATTERS.get(name, HTMLFormatter())
-                name, HTMLAwareEntitySubstitution.substitute_xml)
 
     def setup(self, parent=None, previous_element=None, next_element=None,
               previous_sibling=None, next_sibling=None):
@@ -316,6 +361,14 @@ class PageElement(object):
             and not isinstance(new_child, NavigableString)):
             new_child = NavigableString(new_child)
 
+        from bs4 import BeautifulSoup
+        if isinstance(new_child, BeautifulSoup):
+            # We don't want to end up with a situation where one BeautifulSoup
+            # object contains another. Insert the children one at a time.
+            for subchild in list(new_child.contents):
+                self.insert(position, subchild)
+                position += 1
+            return
         position = min(position, len(self.contents))
         if hasattr(new_child, 'parent') and new_child.parent is not None:
             # We're 'inserting' an element that's already one
@@ -536,15 +589,22 @@ class PageElement(object):
             elif isinstance(name, basestring):
                 # Optimization to find all tags with a given name.
                 if name.count(':') == 1:
-                    # This is a name with a prefix.
+                    # This is a name with a prefix. If this is a namespace-aware document,
-                    prefix, name = name.split(':', 1)
+                    # we need to match the local name against tag.name. If not,
+                    # we need to match the fully-qualified name against tag.name.
+                    prefix, local_name = name.split(':', 1)
                 else:
                     prefix = None
+                    local_name = name
                 result = (element for element in generator
                           if isinstance(element, Tag)
-                            and element.name == name
+                          and (
+                              element.name == name
+                          ) or (
+                              element.name == local_name
                               and (prefix is None or element.prefix == prefix)
                           )
+                )
                 return ResultSet(strainer, result)
         results = ResultSet(strainer)
         while True:
@@ -1046,8 +1106,10 @@ class Tag(PageElement):
             # BS3: soup.aTag -> "soup.find("a")
             tag_name = tag[:-3]
             warnings.warn(
-                '.%sTag is deprecated, use .find("%s") instead.' % (
+                '.%(name)sTag is deprecated, use .find("%(name)s") instead. If you really were looking for a tag called %(name)sTag, use .find("%(name)sTag")' % dict(
-                    tag_name, tag_name))
+                    name=tag_name
+                )
+            )
             return self.find(tag_name)
         # We special case contents to avoid recursion.
         elif not tag.startswith("__") and not tag == "contents":
@@ -1129,11 +1191,10 @@ class Tag(PageElement):
            encoding.
         """
 
-        # First off, turn a string formatter into a function. This
+        # First off, turn a string formatter into a Formatter object. This
         # will stop the lookup from happening over and over again.
-        if not callable(formatter):
+        if not isinstance(formatter, Formatter) and not callable(formatter):
             formatter = self._formatter_for_name(formatter)
-
         attrs = []
         if self.attrs:
             for key, val in sorted(self.attrs.items()):
@@ -1162,7 +1223,9 @@ class Tag(PageElement):
             prefix = self.prefix + ":"
 
         if self.is_empty_element:
-            close = '/'
+            close = ''
+            if isinstance(formatter, Formatter):
+                close = formatter.void_element_close_prefix or close
         else:
             closeTag = '</%s%s>' % (prefix, self.name)
 
@@ -1233,9 +1296,9 @@ class Tag(PageElement):
         :param formatter: The output formatter responsible for converting
            entities to Unicode characters.
         """
-        # First off, turn a string formatter into a function. This
+        # First off, turn a string formatter into a Formatter object. This
         # will stop the lookup from happening over and over again.
-        if not callable(formatter):
+        if not isinstance(formatter, Formatter) and not callable(formatter):
             formatter = self._formatter_for_name(formatter)
 
         pretty_print = (indent_level is not None)
@@ -1348,15 +1411,29 @@ class Tag(PageElement):
         # Handle grouping selectors if ',' exists, ie: p,a
         if ',' in selector:
             context = []
-            for partial_selector in selector.split(','):
+            selectors = [x.strip() for x in selector.split(",")]
-                partial_selector = partial_selector.strip()
+
+            # If a selector is mentioned multiple times we don't want
+            # to use it more than once.
+            used_selectors = set()
+
+            # We also don't want to select the same element more than once,
+            # if it's matched by multiple selectors.
+            selected_object_ids = set()
+            for partial_selector in selectors:
                 if partial_selector == '':
                     raise ValueError('Invalid group selection syntax: %s' % selector)
+                if partial_selector in used_selectors:
+                    continue
+                used_selectors.add(partial_selector)
                 candidates = self.select(partial_selector, limit=limit)
                 for candidate in candidates:
-                    if candidate not in context:
+                    # This lets us distinguish between distinct tags that
+                    # represent the same markup.
+                    object_id = id(candidate)
+                    if object_id not in selected_object_ids:
                         context.append(candidate)
-
+                        selected_object_ids.add(object_id)
                 if limit and len(context) >= limit:
                     break
             return context
@@ -1418,7 +1495,7 @@ class Tag(PageElement):
                 if tag_name == '':
                     raise ValueError(
                         "A pseudo-class must be prefixed with a tag name.")
-                pseudo_attributes = re.match('([a-zA-Z\d-]+)\(([a-zA-Z\d]+)\)', pseudo)
+                pseudo_attributes = re.match(r'([a-zA-Z\d-]+)\(([a-zA-Z\d]+)\)', pseudo)
                 found = []
                 if pseudo_attributes is None:
                     pseudo_type = pseudo
@@ -1652,7 +1729,7 @@ class SoupStrainer(object):
             markup = markup_name
             markup_attrs = markup
         call_function_with_tag_data = (
-            isinstance(self.name, collections.Callable)
+            isinstance(self.name, Callable)
             and not isinstance(markup_name, Tag))
 
         if ((not self.name)
@@ -1732,7 +1809,7 @@ class SoupStrainer(object):
             # True matches any non-None value.
             return markup is not None
 
-        if isinstance(match_against, collections.Callable):
+        if isinstance(match_against, Callable):
             return match_against(markup)
 
         # Custom callables take the tag as an argument, but all