def __init__(self, codecs=None):
"""
Instantiates a new DefaultEncoder.
@param codecs: : a list of codec instances to use for canonicalization
"""
Encoder.__init__(self)
self.html_codec = HTMLEntityCodec()
self.percent_codec = PercentCodec()
self.javascript_codec = JavascriptCodec()
self.vbscript_codec = VBScriptCodec()
self.css_codec = CSSCodec()
self.ldap_codec = LDAPCodec()
self.ldap_dn_codec = LDAPDNCodec()
self.logger = ESAPI.logger("Encoder")
# Used for canonicalization
self.codecs = []
if codecs is None:
self.codecs.append(self.html_codec)
self.codecs.append(self.percent_codec)
self.codecs.append(self.javascript_codec)
# Leaving out css_codec because it eats / characters
# Leaving out vbscript_codec because it eats " characters
else:
for codec in codecs:
if not isinstance(codec, Codec):
raise TypeError(
_("Codecs in list must be instances of children of Codec"
))
self.codecs.append(codec)
def test_codec_for_css(self):
instance = ESAPI.encoder()
### High level
self.assertEquals(None, instance.encode_for_css(None))
self.assertEquals("\\3c script\\3e ", instance.encode_for_css("<script>"))
self.assertEquals("\\21 \\40 \\24 \\25 \\28 \\29 \\3d \\2b \\7b \\7d \\5b \\5d ", instance.encode_for_css("!@$%()=+{}[]"))
# Unicode
self.assertEquals(unichr(12345), instance.encode_for_css(unichr(12345)))
### Low level
codec = CSSCodec()
cases = (
# PLAIN - ENCODED - ALT_ENCODINGS
('', '', ()), # 0 length string
('t','t', ()),
('test', 'test', ()),
('<script>', '\\3c script\\3e ', ()),
('!@#$%^&*(){}[]?+/=|\\', '\\21 \\40 \\23 \\24 \\25 \\5e \\26 \\2a \\28 \\29 \\7b \\7d \\5b \\5d \\3f \\2b \\2f \\3d \\7c \\5c ', ()),
('"`~1234_-', '\\22 \\60 \\7e 1234\\5f \\2d ', ()),
(unichr(9), "\\9 ", ()),
(unichr(12345), unichr(12345), ()),
('\\', '\\5c ', ('\\')),
('\\2aq', None, ("\\2aq",)), # Malformed hex
('\\2aq ', None, ("\\2aq ",)), # Malformed hex
('\\q ', None, ('\\q ',)), # Malformed hex
)
for case in cases:
if case[ENCODED] is not None:
self.assertEquals(case[ENCODED], codec.encode('', case[PLAIN]))
self.assertEquals(case[PLAIN], codec.decode(case[ENCODED]))
for encoding in case[ALT_ENCODINGS]:
print "encoding=",encoding
self.assertEquals(case[PLAIN], codec.decode(encoding))
def __init__(self, logger, extra={}):
self.logger = logger
self.extra = extra
# Enable code for html, JS, url and CSS
codeclist = [HTMLEntityCodec(), JavascriptCodec(), PercentCodec(), CSSCodec()]
self.encoder = SecurityEncoder(codeclist)
class DefaultEncoder(Encoder):
"""
Reference implementation of the Encoder interface. This implementation
takes a whitelist approach to encoding, meaning that everything not
specifically identified in a list of "immune" characters is encoded.
@author: Craig Younkins ([email protected])
"""
IMMUNE_HTML = ',.-_ '
IMMUNE_HTMLATTR = ',.-_'
IMMUNE_CSS = ''
IMMUNE_JAVASCRIPT = ',._'
IMMUNE_VBSCRIPT = ',._'
IMMUNE_XML = ',.-_ '
IMMUNE_SQL = ' '
IMMUNE_OS = '-'
IMMUNE_XMLATTR = ',.-_'
IMMUNE_XPATH = ',.-_ '
IMMUNE_LDAP = ''
IMMUNE_LDAP_DN = ''
# Unreserved characters as specified in RFC 3986
IMMUNE_URL = '-_.~'
def __init__(self, codecs=None):
"""
Instantiates a new DefaultEncoder.
@param codecs: : a list of codec instances to use for canonicalization
"""
Encoder.__init__(self)
self.html_codec = HTMLEntityCodec()
self.percent_codec = PercentCodec()
self.javascript_codec = JavascriptCodec()
self.vbscript_codec = VBScriptCodec()
self.css_codec = CSSCodec()
self.ldap_codec = LDAPCodec()
self.ldap_dn_codec = LDAPDNCodec()
self.logger = ESAPI.logger("Encoder")
# Used for canonicalization
self.codecs = []
if codecs is None:
self.codecs.append(self.html_codec)
self.codecs.append(self.percent_codec)
self.codecs.append(self.javascript_codec)
# Leaving out css_codec because it eats / characters
# Leaving out vbscript_codec because it eats " characters
else:
for codec in codecs:
if not isinstance(codec, Codec):
raise TypeError(
_("Codecs in list must be instances of children of Codec"
))
self.codecs.append(codec)
def canonicalize(self, input_, strict=True):
if input_ is None:
return None
working = input_[:]
codecs_found = []
found_count = 0
clean = False
while not clean:
clean = True
# Try each codec and keep track of which ones work
for codec in self.codecs:
old = working[:]
working = codec.decode(working)
if old != working:
if codec.__class__.__name__ not in codecs_found:
codecs_found.append(codec.__class__.__name__)
if clean:
found_count += 1
clean = False
if found_count >= 2 and len(codecs_found) > 1:
if strict:
raise IntrusionException(
_("Input validation failure"),
_("Multiple (%(times_encoded)sx) and mixed encoding (%(codecs_found)s) detected in %(input)s"
) % {
'times_encoded': found_count,
'codecs_found': str(codecs_found),
'input': input_
})
else:
self.logger.warning(
Logger.SECURITY_FAILURE,
_("Multiple (%s(times_encoded)x) and mixed encoding (%(codecs_found)s) detected in %(input)s"
) % {
'times_encoded': found_count,
'codecs_found': str(codecs_found),
'input': input_
})
elif found_count >= 2:
if strict:
raise IntrusionException(
_("Input validation failure"),
_("Multiple (%s(times_encoded)x) encoding detected in %(input)s"
) % {
'times_encoded': found_count,
'input': input_
})
else:
self.logger.warning(
Logger.SECURITY_FAILURE,
_("Multiple (%s(times_encoded)x) encoding detected in %(input)s"
) % {
'times_encoded': found_count,
'input': input_
})
elif len(codecs_found) > 1:
if strict:
raise IntrusionException(
_("Input validation failure"),
_("Mixed encoding (%(codecs_found)s) detected in %(input)s"
) % {
'codecs_found': str(codecs_found),
'input': input_
})
else:
self.logger.warning(
Logger.SECURITY_FAILURE,
_("Mixed encoding (%(codecs_found)s) detected in %(input)s"
) % {
'codecs_found': str(codecs_found),
'input': input_
})
return working
def encode_for_css(self, input_):
return self.css_codec.encode(DefaultEncoder.IMMUNE_CSS, input_)
def encode_for_html(self, input_):
return self.html_codec.encode(DefaultEncoder.IMMUNE_HTML, input_)
def encode_for_html_attribute(self, input_):
return self.html_codec.encode(DefaultEncoder.IMMUNE_HTMLATTR, input_)
def encode_for_javascript(self, input_):
return self.javascript_codec.encode(DefaultEncoder.IMMUNE_JAVASCRIPT,
input_)
def encode_for_vbscript(self, input_):
return self.vbscript_codec.encode(DefaultEncoder.IMMUNE_VBSCRIPT,
input_)
def encode_for_sql(self, codec, input_):
return codec.encode(DefaultEncoder.IMMUNE_SQL, input_)
def encode_for_os(self, codec, input_):
return codec.encode(DefaultEncoder.IMMUNE_OS, input_)
def encode_for_ldap(self, input_):
return self.ldap_codec.encode(DefaultEncoder.IMMUNE_LDAP, input_)
def encode_for_dn(self, input_):
return self.ldap_dn_codec.encode(DefaultEncoder.IMMUNE_LDAP_DN, input_)
def encode_for_xpath(self, input_):
return self.html_codec.encode(DefaultEncoder.IMMUNE_XPATH, input_)
def encode_for_xml(self, input_):
return self.html_codec.encode(DefaultEncoder.IMMUNE_XML, input_)
def encode_for_xml_attribute(self, input_):
return self.html_codec.encode(DefaultEncoder.IMMUNE_XMLATTR, input_)
def encode_for_url(self, input_):
return self.percent_codec.encode(DefaultEncoder.IMMUNE_URL, input_)
def decode_from_url(self, input_):
if input_ is None:
return None
canonical = self.canonicalize(input_)
return self.percent_codec.decode(canonical)
def encode_for_base64(self, input_):
try:
return base64.b64encode(input_)
except:
return None
def decode_from_base64(self, input_):
try:
return base64.b64decode(input_)
except:
return None
def test_canonicalize(self):
codecs = [HTMLEntityCodec(), PercentCodec()]
encoder_class = ESAPI.security_configuration().get_class_for_interface('encoder')
instance = encoder_class(codecs)
# Test None paths
self.assertEquals( None, instance.canonicalize(None))
self.assertEquals( None, instance.canonicalize(None, True))
self.assertEquals( None, instance.canonicalize(None, False))
# test exception paths
self.assertEquals( "%", instance.canonicalize("%25", True))
self.assertEquals( "%", instance.canonicalize("%25", False))
self.assertEquals( "%", instance.canonicalize("%25"))
self.assertEquals( "%F", instance.canonicalize("%25F"))
self.assertEquals( "<", instance.canonicalize("%3c"))
self.assertEquals( "<", instance.canonicalize("%3C"))
self.assertEquals( "%X1", instance.canonicalize("%X1"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "%", instance.canonicalize("%"))
self.assertEquals( "%", instance.canonicalize("%"))
self.assertEquals( "%b", instance.canonicalize("%b"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
# percent encoding
self.assertEquals( "<", instance.canonicalize("%3c"))
self.assertEquals( "<", instance.canonicalize("%3C"))
# html entity encoding
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("&lT"))
self.assertEquals( "<", instance.canonicalize("&Lt"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<", instance.canonicalize("&lT;"))
self.assertEquals( "<", instance.canonicalize("≪"))
self.assertEquals( "<", instance.canonicalize("<"))
self.assertEquals( "<script>alert(\"hello\");</script>", instance.canonicalize("%3Cscript%3Ealert%28%22hello%22%29%3B%3C%2Fscript%3E") )
self.assertEquals( "<script>alert(\"hello\");</script>", instance.canonicalize("%3Cscript>alert%28%22hello"%29%3B%3C%2Fscript%3E", False) )
# javascript escape syntax
js = [JavascriptCodec()]
instance = encoder_class( js )
self.assertEquals( "\0", instance.canonicalize("\\0"))
self.assertEquals( "\b", instance.canonicalize("\\b"))
self.assertEquals( "\t", instance.canonicalize("\\t"))
self.assertEquals( "\n", instance.canonicalize("\\n"))
self.assertEquals( unichr(0x0b), instance.canonicalize("\\v"))
self.assertEquals( "\f", instance.canonicalize("\\f"))
self.assertEquals( "\r", instance.canonicalize("\\r"))
self.assertEquals( "\'", instance.canonicalize("\\'"))
self.assertEquals( "\"", instance.canonicalize("\\\""))
self.assertEquals( "\\", instance.canonicalize("\\\\"))
self.assertEquals( "<", instance.canonicalize("\\<"))
self.assertEquals( "<", instance.canonicalize("\\u003c"))
self.assertEquals( "<", instance.canonicalize("\\U003c"))
self.assertEquals( "<", instance.canonicalize("\\u003C"))
self.assertEquals( "<", instance.canonicalize("\\U003C"))
self.assertEquals( "<", instance.canonicalize("\\x3c"))
self.assertEquals( "<", instance.canonicalize("\\X3c"))
self.assertEquals( "<", instance.canonicalize("\\x3C"))
self.assertEquals( "<", instance.canonicalize("\\X3C"))
# css escape syntax
# be careful because some codecs see \0 as null byte
css = [CSSCodec()]
instance = encoder_class( css )
self.assertEquals( "<", instance.canonicalize("\\3c")); # add strings to prevent null byte
self.assertEquals( "<", instance.canonicalize("\\03c"))
self.assertEquals( "<", instance.canonicalize("\\003c"))
self.assertEquals( "<", instance.canonicalize("\\0003c"))
self.assertEquals( "<", instance.canonicalize("\\00003c"))
self.assertEquals( "<", instance.canonicalize("\\3C"))
self.assertEquals( "<", instance.canonicalize("\\03C"))
self.assertEquals( "<", instance.canonicalize("\\003C"))
self.assertEquals( "<", instance.canonicalize("\\0003C"))
self.assertEquals( "<", instance.canonicalize("\\00003C"))
def test_encoder_constructor_exception(self):
"""
Checks that only valid codecs are allowed.
"""
codecs = [CSSCodec(), str]
self.assertRaises(TypeError, ESAPI.encoder, codecs)