def _really_load(self, f, filename, ignore_discard, ignore_expires):
now = time.time()
magic = f.readline()
if not re.search(self.magic_re, magic):
f.close()
raise LoadError(
"%s does not look like a Netscape format cookies file" %
filename)
try:
while 1:
line = f.readline()
if line == "": break
# last field may be absent, so keep any trailing tab
if endswith(line, "\n"): line = line[:-1]
# skip comments and blank lines XXX what is $ for?
if (startswith(string.strip(line), "#")
or startswith(string.strip(line), "$")
or string.strip(line) == ""):
continue
domain, domain_specified, path, secure, expires, name, value = \
string.split(line, "\t")
secure = (secure == "TRUE")
domain_specified = (domain_specified == "TRUE")
if name == "":
name = value
value = None
initial_dot = startswith(domain, ".")
assert domain_specified == initial_dot
discard = False
if expires == "":
expires = None
discard = True
# assume path_specified is false
c = Cookie(0, name, value, None, False, domain,
domain_specified, initial_dot, path, False, secure,
expires, discard, None, None, {})
if not ignore_discard and c.discard:
continue
if not ignore_expires and c.is_expired(now):
continue
self.set_cookie(c)
except:
reraise_unmasked_exceptions((IOError, ))
raise LoadError("invalid Netscape format file %s: %s" %
(filename, line))
def save(self, filename=None, ignore_discard=False, ignore_expires=False):
if filename is None:
if self.filename is not None: filename = self.filename
else: raise ValueError(MISSING_FILENAME_TEXT)
f = open(filename, "w")
try:
f.write(self.header)
now = time.time()
debug("Saving Netscape cookies.txt file")
for cookie in self:
if not ignore_discard and cookie.discard:
debug(" Not saving %s: marked for discard" % cookie.name)
continue
if not ignore_expires and cookie.is_expired(now):
debug(" Not saving %s: expired" % cookie.name)
continue
if cookie.secure: secure = "TRUE"
else: secure = "FALSE"
if startswith(cookie.domain, "."): initial_dot = "TRUE"
else: initial_dot = "FALSE"
if cookie.expires is not None:
expires = str(cookie.expires)
else:
expires = ""
if cookie.name is not None:
name = cookie.name
else:
name = ""
f.write(
string.join([cookie.domain, initial_dot, cookie.path,
secure, expires, name, cookie.value], "\t")+
"\n")
finally:
f.close()
def parse_ns_headers(ns_headers):
"""Ad-hoc parser for Netscape protocol cookie-attributes.
The old Netscape cookie format for Set-Cookie can for instance contain
an unquoted "," in the expires field, so we have to use this ad-hoc
parser instead of split_header_words.
XXX This may not make the best possible effort to parse all the crap
that Netscape Cookie headers contain. Ronald Tschalar's HTTPClient
parser is probably better, so could do worse than following that if
this ever gives any trouble.
Currently, this is also used for parsing RFC 2109 cookies.
"""
known_attrs = (
"expires",
"domain",
"path",
"secure",
# RFC 2109 attrs (may turn up in Netscape cookies, too)
"port",
"max-age")
result = []
for ns_header in ns_headers:
pairs = []
version_set = False
params = re.split(r";\s*", ns_header)
for ii in range(len(params)):
param = params[ii]
param = string.rstrip(param)
if param == "": continue
if "=" not in param:
k, v = param, None
else:
k, v = re.split(r"\s*=\s*", param, 1)
k = string.lstrip(k)
if ii != 0:
lc = string.lower(k)
if lc in known_attrs:
k = lc
if k == "version":
# This is an RFC 2109 cookie.
version_set = True
if k == "expires":
# convert expires date to seconds since epoch
if startswith(v, '"'): v = v[1:]
if endswith(v, '"'): v = v[:-1]
v = http2time(v) # None if invalid
pairs.append((k, v))
if pairs:
if not version_set:
pairs.append(("version", "0"))
result.append(pairs)
return result
def parse_ns_headers(ns_headers):
"""Ad-hoc parser for Netscape protocol cookie-attributes.
The old Netscape cookie format for Set-Cookie can for instance contain
an unquoted "," in the expires field, so we have to use this ad-hoc
parser instead of split_header_words.
XXX This may not make the best possible effort to parse all the crap
that Netscape Cookie headers contain. Ronald Tschalar's HTTPClient
parser is probably better, so could do worse than following that if
this ever gives any trouble.
Currently, this is also used for parsing RFC 2109 cookies.
"""
known_attrs = ("expires", "domain", "path", "secure",
# RFC 2109 attrs (may turn up in Netscape cookies, too)
"port", "max-age")
result = []
for ns_header in ns_headers:
pairs = []
version_set = False
for param in re.split(r";\s*", ns_header):
param = string.rstrip(param)
if param == "": continue
if "=" not in param:
if string.lower(param) in known_attrs:
k, v = param, None
else:
# cookie with missing name
k, v = None, param
else:
k, v = re.split(r"\s*=\s*", param, 1)
k = string.lstrip(k)
if k is not None:
lc = string.lower(k)
if lc in known_attrs:
k = lc
if k == "version":
# This is an RFC 2109 cookie. Will be treated as RFC 2965
# cookie in rest of code.
# Probably it should be parsed with split_header_words, but
# that's too much hassle.
version_set = True
if k == "expires":
# convert expires date to seconds since epoch
if startswith(v, '"'): v = v[1:]
if endswith(v, '"'): v = v[:-1]
v = http2time(v) # None if invalid
pairs.append((k, v))
if pairs:
if not version_set:
pairs.append(("version", "0"))
result.append(pairs)
return result
def load_cookie_data(self,
filename,
ignore_discard=False,
ignore_expires=False):
"""Load cookies from file containing actual cookie data.
Old cookies are kept unless overwritten by newly loaded ones.
You should not call this method if the delayload attribute is set.
I think each of these files contain all cookies for one user, domain,
and path.
filename: file containing cookies -- usually found in a file like
C:\WINNT\Profiles\joe\Cookies\joe@blah[1].txt
"""
now = int(time.time())
cookie_data = self._load_cookies_from_file(filename)
for cookie in cookie_data:
flags = cookie["FLAGS"]
secure = ((flags & 0x2000) != 0)
filetime = (cookie["HIXP"] << 32) + cookie["LOXP"]
expires = epoch_time_offset_from_win32_filetime(filetime)
if expires < now:
discard = True
else:
discard = False
domain = cookie["DOMAIN"]
initial_dot = startswith(domain, ".")
if initial_dot:
domain_specified = True
else:
# MSIE 5 does not record whether the domain cookie-attribute
# was specified.
# Assuming it wasn't is conservative, because with strict
# domain matching this will match less frequently; with regular
# Netscape tail-matching, this will match at exactly the same
# times that domain_specified = True would. It also means we
# don't have to prepend a dot to achieve consistency with our
# own & Mozilla's domain-munging scheme.
domain_specified = False
# assume path_specified is false
# XXX is there other stuff in here? -- eg. comment, commentURL?
c = Cookie(0, cookie["KEY"], cookie["VALUE"], None, False, domain,
domain_specified, initial_dot, cookie["PATH"], False,
secure, expires, discard, None, None, {"flags": flags})
if not ignore_discard and c.discard:
continue
if not ignore_expires and c.is_expired(now):
continue
self.set_cookie(c)
def add_handler(self, handler):
added = False
for meth in methnames(handler):
i = string.find(meth, "_")
protocol = meth[:i]
condition = meth[i + 1:]
if startswith(condition, "error"):
j = string.find(meth[i + 1:], "_") + i + 1
kind = meth[j + 1:]
try:
kind = int(kind)
except ValueError:
pass
lookup = self.handle_error.get(protocol, {})
self.handle_error[protocol] = lookup
elif (condition == "open"
and protocol not in ["do", "proxy"]): # hack -- see below
kind = protocol
lookup = self.handle_open
elif (condition in ["response", "request"]
and protocol != "redirect"): # yucky hack
# hack above is to fix HTTPRedirectHandler problem, which
# appears to above line to be a processor because of the
# redirect_request method :-((
kind = protocol
lookup = getattr(self, "process_" + condition)
else:
continue
if lookup.has_key(kind):
bisect.insort(lookup[kind], handler)
else:
lookup[kind] = [handler]
added = True
continue
if added:
# XXX why does self.handlers need to be sorted?
bisect.insort(self.handlers, handler)
handler.add_parent(self)
def add_handler(self, handler):
added = False
for meth in methnames(handler):
i = string.find(meth, "_")
protocol = meth[:i]
condition = meth[i+1:]
if startswith(condition, "error"):
j = string.find(meth[i+1:], "_") + i + 1
kind = meth[j+1:]
try:
kind = int(kind)
except ValueError:
pass
lookup = self.handle_error.get(protocol, {})
self.handle_error[protocol] = lookup
elif (condition == "open" and
protocol not in ["do", "proxy"]): # hack -- see below
kind = protocol
lookup = self.handle_open
elif (condition in ["response", "request"] and
protocol != "redirect"): # yucky hack
# hack above is to fix HTTPRedirectHandler problem, which
# appears to above line to be a processor because of the
# redirect_request method :-((
kind = protocol
lookup = getattr(self, "process_"+condition)
else:
continue
if lookup.has_key(kind):
bisect.insort(lookup[kind], handler)
else:
lookup[kind] = [handler]
added = True
continue
if added:
# XXX why does self.handlers need to be sorted?
bisect.insort(self.handlers, handler)
handler.add_parent(self)
def _really_load(self, f, filename, ignore_discard, ignore_expires):
magic = f.readline()
if not re.search(self.magic_re, magic):
msg = "%s does not seem to contain cookies" % filename
raise LoadError(msg)
now = time.time()
header = "Set-Cookie3:"
boolean_attrs = ("port_spec", "path_spec", "domain_dot",
"secure", "discard", "rfc2109")
value_attrs = ("version",
"port", "path", "domain",
"expires",
"comment", "commenturl")
try:
while 1:
line = f.readline()
if line == "": break
if not startswith(line, header):
continue
line = string.strip(line[len(header):])
for data in split_header_words([line]):
name, value = data[0]
standard = {}
rest = {}
for k in boolean_attrs:
standard[k] = False
for k, v in data[1:]:
if k is not None:
lc = string.lower(k)
else:
lc = None
# don't lose case distinction for unknown fields
if (lc in value_attrs) or (lc in boolean_attrs):
k = lc
if k in boolean_attrs:
if v is None: v = True
standard[k] = v
elif k in value_attrs:
standard[k] = v
else:
rest[k] = v
h = standard.get
expires = h("expires")
discard = h("discard")
if expires is not None:
expires = iso2time(expires)
if expires is None:
discard = True
domain = h("domain")
domain_specified = startswith(domain, ".")
c = Cookie(h("version"), name, value,
h("port"), h("port_spec"),
domain, domain_specified, h("domain_dot"),
h("path"), h("path_spec"),
h("secure"),
expires,
discard,
h("comment"),
h("commenturl"),
rest,
h("rfc2109"),
)
if not ignore_discard and c.discard:
continue
if not ignore_expires and c.is_expired(now):
continue
self.set_cookie(c)
except:
reraise_unmasked_exceptions((IOError,))
raise LoadError("invalid Set-Cookie3 format file %s" % filename)
def split_header_words(header_values):
r"""Parse header values into a list of lists containing key,value pairs.
The function knows how to deal with ",", ";" and "=" as well as quoted
values after "=". A list of space separated tokens are parsed as if they
were separated by ";".
If the header_values passed as argument contains multiple values, then they
are treated as if they were a single value separated by comma ",".
This means that this function is useful for parsing header fields that
follow this syntax (BNF as from the HTTP/1.1 specification, but we relax
the requirement for tokens).
headers = #header
header = (token | parameter) *( [";"] (token | parameter))
token = 1*<any CHAR except CTLs or separators>
separators = "(" | ")" | "<" | ">" | "@"
| "," | ";" | ":" | "\" | <">
| "/" | "[" | "]" | "?" | "="
| "{" | "}" | SP | HT
quoted-string = ( <"> *(qdtext | quoted-pair ) <"> )
qdtext = <any TEXT except <">>
quoted-pair = "\" CHAR
parameter = attribute "=" value
attribute = token
value = token | quoted-string
Each header is represented by a list of key/value pairs. The value for a
simple token (not part of a parameter) is None. Syntactically incorrect
headers will not necessarily be parsed as you would want.
This is easier to describe with some examples:
>>> split_header_words(['foo="bar"; port="80,81"; discard, bar=baz'])
[[('foo', 'bar'), ('port', '80,81'), ('discard', None)], [('bar', 'baz')]]
>>> split_header_words(['text/html; charset="iso-8859-1"'])
[[('text/html', None), ('charset', 'iso-8859-1')]]
>>> split_header_words([r'Basic realm="\"foo\bar\""'])
[[('Basic', None), ('realm', '"foobar"')]]
"""
assert type(header_values) not in STRING_TYPES
result = []
for text in header_values:
orig_text = text
pairs = []
while text:
m = token_re.search(text)
if m:
text = unmatched(m)
name = m.group(1)
m = quoted_value_re.search(text)
if m: # quoted value
text = unmatched(m)
value = m.group(1)
value = escape_re.sub(r"\1", value)
else:
m = value_re.search(text)
if m: # unquoted value
text = unmatched(m)
value = m.group(1)
value = string.rstrip(value)
else:
# no value, a lone token
value = None
pairs.append((name, value))
elif startswith(string.lstrip(text), ","):
# concatenated headers, as per RFC 2616 section 4.2
text = string.lstrip(text)[1:]
if pairs: result.append(pairs)
pairs = []
else:
# skip junk
non_junk, nr_junk_chars = re.subn("^[=\s;]*", "", text)
assert nr_junk_chars > 0, (
"split_header_words bug: '%s', '%s', %s" %
(orig_text, text, pairs))
text = non_junk
if pairs: result.append(pairs)
return result
def split_header_words(header_values):
r"""Parse header values into a list of lists containing key,value pairs.
The function knows how to deal with ",", ";" and "=" as well as quoted
values after "=". A list of space separated tokens are parsed as if they
were separated by ";".
If the header_values passed as argument contains multiple values, then they
are treated as if they were a single value separated by comma ",".
This means that this function is useful for parsing header fields that
follow this syntax (BNF as from the HTTP/1.1 specification, but we relax
the requirement for tokens).
headers = #header
header = (token | parameter) *( [";"] (token | parameter))
token = 1*<any CHAR except CTLs or separators>
separators = "(" | ")" | "<" | ">" | "@"
| "," | ";" | ":" | "\" | <">
| "/" | "[" | "]" | "?" | "="
| "{" | "}" | SP | HT
quoted-string = ( <"> *(qdtext | quoted-pair ) <"> )
qdtext = <any TEXT except <">>
quoted-pair = "\" CHAR
parameter = attribute "=" value
attribute = token
value = token | quoted-string
Each header is represented by a list of key/value pairs. The value for a
simple token (not part of a parameter) is None. Syntactically incorrect
headers will not necessarily be parsed as you would want.
This is easier to describe with some examples:
>>> split_header_words(['foo="bar"; port="80,81"; discard, bar=baz'])
[[('foo', 'bar'), ('port', '80,81'), ('discard', None)], [('bar', 'baz')]]
>>> split_header_words(['text/html; charset="iso-8859-1"'])
[[('text/html', None), ('charset', 'iso-8859-1')]]
>>> split_header_words([r'Basic realm="\"foo\bar\""'])
[[('Basic', None), ('realm', '"foobar"')]]
"""
assert type(header_values) not in STRING_TYPES
result = []
for text in header_values:
orig_text = text
pairs = []
while text:
m = token_re.search(text)
if m:
text = unmatched(m)
name = m.group(1)
m = quoted_value_re.search(text)
if m: # quoted value
text = unmatched(m)
value = m.group(1)
value = escape_re.sub(r"\1", value)
else:
m = value_re.search(text)
if m: # unquoted value
text = unmatched(m)
value = m.group(1)
value = string.rstrip(value)
else:
# no value, a lone token
value = None
pairs.append((name, value))
elif startswith(string.lstrip(text), ","):
# concatenated headers, as per RFC 2616 section 4.2
text = string.lstrip(text)[1:]
if pairs: result.append(pairs)
pairs = []
else:
# skip junk
non_junk, nr_junk_chars = re.subn("^[=\s;]*", "", text)
assert nr_junk_chars > 0, (
"split_header_words bug: '%s', '%s', %s" %
(orig_text, text, pairs))
text = non_junk
if pairs: result.append(pairs)
return result
def load_cookie_data(self, filename,
ignore_discard=False, ignore_expires=False):
"""Load cookies from file containing actual cookie data.
Old cookies are kept unless overwritten by newly loaded ones.
You should not call this method if the delayload attribute is set.
I think each of these files contain all cookies for one user, domain,
and path.
filename: file containing cookies -- usually found in a file like
C:\WINNT\Profiles\joe\Cookies\joe@blah[1].txt
"""
now = int(time.time())
cookie_data = self._load_cookies_from_file(filename)
for cookie in cookie_data:
flags = cookie["FLAGS"]
secure = ((flags & 0x2000) != 0)
filetime = (cookie["HIXP"] << 32) + cookie["LOXP"]
expires = epoch_time_offset_from_win32_filetime(filetime)
if expires < now:
discard = True
else:
discard = False
domain = cookie["DOMAIN"]
initial_dot = startswith(domain, ".")
if initial_dot:
domain_specified = True
else:
# MSIE 5 does not record whether the domain cookie-attribute
# was specified.
# Assuming it wasn't is conservative, because with strict
# domain matching this will match less frequently; with regular
# Netscape tail-matching, this will match at exactly the same
# times that domain_specified = True would. It also means we
# don't have to prepend a dot to achieve consistency with our
# own & Mozilla's domain-munging scheme.
domain_specified = False
# assume path_specified is false
# XXX is there other stuff in here? -- eg. comment, commentURL?
c = Cookie(0,
cookie["KEY"], cookie["VALUE"],
None, False,
domain, domain_specified, initial_dot,
cookie["PATH"], False,
secure,
expires,
discard,
None,
None,
{"flags": flags})
if not ignore_discard and c.discard:
continue
if not ignore_expires and c.is_expired(now):
continue
CookieJar.set_cookie(self, c)
def _really_load(self, f, filename, ignore_discard, ignore_expires):
now = time.time()
magic = f.readline()
if not re.search(self.magic_re, magic):
f.close()
raise LoadError(
"%s does not look like a Netscape format cookies file" %
filename)
try:
while 1:
line = f.readline()
if line == "": break
# last field may be absent, so keep any trailing tab
if endswith(line, "\n"): line = line[:-1]
# skip comments and blank lines XXX what is $ for?
if (startswith(string.strip(line), "#") or
startswith(string.strip(line), "$") or
string.strip(line) == ""):
continue
domain, domain_specified, path, secure, expires, name, value = \
string.split(line, "\t")
secure = (secure == "TRUE")
domain_specified = (domain_specified == "TRUE")
if name == "":
name = value
value = None
initial_dot = startswith(domain, ".")
assert domain_specified == initial_dot
discard = False
if expires == "":
expires = None
discard = True
# assume path_specified is false
c = Cookie(0, name, value,
None, False,
domain, domain_specified, initial_dot,
path, False,
secure,
expires,
discard,
None,
None,
{})
if not ignore_discard and c.discard:
continue
if not ignore_expires and c.is_expired(now):
continue
self.set_cookie(c)
except:
reraise_unmasked_exceptions((IOError,))
raise LoadError("invalid Netscape format file %s: %s" %
(filename, line))
