1 # -*- coding: utf-8 -*- 2 """ 3 markupsafe 4 ~~~~~~~~~~ 5 6 Implements a Markup string. 7 8 :copyright: (c) 2010 by Armin Ronacher. 9 :license: BSD, see LICENSE for more details. 10 """ 11 import re 12 from itertools import imap 13 14 15 __all__ = ['Markup', 'soft_unicode', 'escape', 'escape_silent'] 16 17 18 _striptags_re = re.compile(r'(<!--.*?-->|<[^>]*>)') 19 _entity_re = re.compile(r'&([^;]+);') 20 21 22 class Markup(unicode): 23 r"""Marks a string as being safe for inclusion in HTML/XML output without 24 needing to be escaped. This implements the `__html__` interface a couple 25 of frameworks and web applications use. :class:`Markup` is a direct 26 subclass of `unicode` and provides all the methods of `unicode` just that 27 it escapes arguments passed and always returns `Markup`. 28 29 The `escape` function returns markup objects so that double escaping can't 30 happen. 31 32 The constructor of the :class:`Markup` class can be used for three 33 different things: When passed an unicode object it's assumed to be safe, 34 when passed an object with an HTML representation (has an `__html__` 35 method) that representation is used, otherwise the object passed is 36 converted into a unicode string and then assumed to be safe: 37 38 >>> Markup("Hello <em>World</em>!") 39 Markup(u'Hello <em>World</em>!') 40 >>> class Foo(object): 41 ... def __html__(self): 42 ... return '<a href="#">foo</a>' 43 ... 44 >>> Markup(Foo()) 45 Markup(u'<a href="#">foo</a>') 46 47 If you want object passed being always treated as unsafe you can use the 48 :meth:`escape` classmethod to create a :class:`Markup` object: 49 50 >>> Markup.escape("Hello <em>World</em>!") 51 Markup(u'Hello <em>World</em>!') 52 53 Operations on a markup string are markup aware which means that all 54 arguments are passed through the :func:`escape` function: 55 56 >>> em = Markup("<em>%s</em>") 57 >>> em % "foo & bar" 58 Markup(u'<em>foo & bar</em>') 59 >>> strong = Markup("<strong>%(text)s</strong>") 60 >>> strong % {'text': '<blink>hacker here</blink>'} 61 Markup(u'<strong><blink>hacker here</blink></strong>') 62 >>> Markup("<em>Hello</em> ") + "<foo>" 63 Markup(u'<em>Hello</em> <foo>') 64 """ 65 __slots__ = () 66 67 def __new__(cls, base=u'', encoding=None, errors='strict'): 68 if hasattr(base, '__html__'): 69 base = base.__html__() 70 if encoding is None: 71 return unicode.__new__(cls, base) 72 return unicode.__new__(cls, base, encoding, errors) 73 74 def __html__(self): 75 return self 76 77 def __add__(self, other): 78 if hasattr(other, '__html__') or isinstance(other, basestring): 79 return self.__class__(unicode(self) + unicode(escape(other))) 80 return NotImplemented 81 82 def __radd__(self, other): 83 if hasattr(other, '__html__') or isinstance(other, basestring): 84 return self.__class__(unicode(escape(other)) + unicode(self)) 85 return NotImplemented 86 87 def __mul__(self, num): 88 if isinstance(num, (int, long)): 89 return self.__class__(unicode.__mul__(self, num)) 90 return NotImplemented 91 __rmul__ = __mul__ 92 93 def __mod__(self, arg): 94 if isinstance(arg, tuple): 95 arg = tuple(imap(_MarkupEscapeHelper, arg)) 96 else: 97 arg = _MarkupEscapeHelper(arg) 98 return self.__class__(unicode.__mod__(self, arg)) 99 100 def __repr__(self): 101 return '%s(%s)' % ( 102 self.__class__.__name__, 103 unicode.__repr__(self) 104 ) 105 106 def join(self, seq): 107 return self.__class__(unicode.join(self, imap(escape, seq))) 108 join.__doc__ = unicode.join.__doc__ 109 110 def split(self, *args, **kwargs): 111 return map(self.__class__, unicode.split(self, *args, **kwargs)) 112 split.__doc__ = unicode.split.__doc__ 113 114 def rsplit(self, *args, **kwargs): 115 return map(self.__class__, unicode.rsplit(self, *args, **kwargs)) 116 rsplit.__doc__ = unicode.rsplit.__doc__ 117 118 def splitlines(self, *args, **kwargs): 119 return map(self.__class__, unicode.splitlines(self, *args, **kwargs)) 120 splitlines.__doc__ = unicode.splitlines.__doc__ 121 122 def unescape(self): 123 r"""Unescape markup again into an unicode string. This also resolves 124 known HTML4 and XHTML entities: 125 126 >>> Markup("Main » <em>About</em>").unescape() 127 u'Main \xbb <em>About</em>' 128 """ 129 from jinja2._markupsafe._constants import HTML_ENTITIES 130 def handle_match(m): 131 name = m.group(1) 132 if name in HTML_ENTITIES: 133 return unichr(HTML_ENTITIES[name]) 134 try: 135 if name[:2] in ('#x', '#X'): 136 return unichr(int(name[2:], 16)) 137 elif name.startswith('#'): 138 return unichr(int(name[1:])) 139 except ValueError: 140 pass 141 return u'' 142 return _entity_re.sub(handle_match, unicode(self)) 143 144 def striptags(self): 145 r"""Unescape markup into an unicode string and strip all tags. This 146 also resolves known HTML4 and XHTML entities. Whitespace is 147 normalized to one: 148 149 >>> Markup("Main » <em>About</em>").striptags() 150 u'Main \xbb About' 151 """ 152 stripped = u' '.join(_striptags_re.sub('', self).split()) 153 return Markup(stripped).unescape() 154 155 @classmethod 156 def escape(cls, s): 157 """Escape the string. Works like :func:`escape` with the difference 158 that for subclasses of :class:`Markup` this function would return the 159 correct subclass. 160 """ 161 rv = escape(s) 162 if rv.__class__ is not cls: 163 return cls(rv) 164 return rv 165 166 def make_wrapper(name): 167 orig = getattr(unicode, name) 168 def func(self, *args, **kwargs): 169 args = _escape_argspec(list(args), enumerate(args)) 170 _escape_argspec(kwargs, kwargs.iteritems()) 171 return self.__class__(orig(self, *args, **kwargs)) 172 func.__name__ = orig.__name__ 173 func.__doc__ = orig.__doc__ 174 return func 175 176 for method in '__getitem__', 'capitalize', \ 177 'title', 'lower', 'upper', 'replace', 'ljust', \ 178 'rjust', 'lstrip', 'rstrip', 'center', 'strip', \ 179 'translate', 'expandtabs', 'swapcase', 'zfill': 180 locals()[method] = make_wrapper(method) 181 182 # new in python 2.5 183 if hasattr(unicode, 'partition'): 184 partition = make_wrapper('partition'), 185 rpartition = make_wrapper('rpartition') 186 187 # new in python 2.6 188 if hasattr(unicode, 'format'): 189 format = make_wrapper('format') 190 191 # not in python 3 192 if hasattr(unicode, '__getslice__'): 193 __getslice__ = make_wrapper('__getslice__') 194 195 del method, make_wrapper 196 197 198 def _escape_argspec(obj, iterable): 199 """Helper for various string-wrapped functions.""" 200 for key, value in iterable: 201 if hasattr(value, '__html__') or isinstance(value, basestring): 202 obj[key] = escape(value) 203 return obj 204 205 206 class _MarkupEscapeHelper(object): 207 """Helper for Markup.__mod__""" 208 209 def __init__(self, obj): 210 self.obj = obj 211 212 __getitem__ = lambda s, x: _MarkupEscapeHelper(s.obj[x]) 213 __str__ = lambda s: str(escape(s.obj)) 214 __unicode__ = lambda s: unicode(escape(s.obj)) 215 __repr__ = lambda s: str(escape(repr(s.obj))) 216 __int__ = lambda s: int(s.obj) 217 __float__ = lambda s: float(s.obj) 218 219 220 # we have to import it down here as the speedups and native 221 # modules imports the markup type which is define above. 222 try: 223 from jinja2._markupsafe._speedups import escape, escape_silent, soft_unicode 224 except ImportError: 225 from jinja2._markupsafe._native import escape, escape_silent, soft_unicode 226