# -*- coding: latin-1 -*-

"""Tests for cookielib.py."""

import re, os, time

from unittest import TestCase

from test import test_support

class DateTimeTests(TestCase):

    def test_time2isoz(self):

        from cookielib import time2isoz

        base = 1019227000

        day = 24*3600

        self.assertEquals(time2isoz(base), "2002-04-19 14:36:40Z")

        self.assertEquals(time2isoz(base+day), "2002-04-20 14:36:40Z")

        self.assertEquals(time2isoz(base+2*day), "2002-04-21 14:36:40Z")

        self.assertEquals(time2isoz(base+3*day), "2002-04-22 14:36:40Z")

        az = time2isoz()

        bz = time2isoz(500000)

        for text in (az, bz):

            self.assert_(re.search(r"^\d{4}-\d\d-\d\d \d\d:\d\d:\d\dZ$", text),

                         "bad time2isoz format: %s %s" % (az, bz))

    def test_http2time(self):

        from cookielib import http2time

        def parse_date(text):

            return time.gmtime(http2time(text))[:6]

        self.assertEquals(parse_date("01 Jan 2001"), (2001, 1, 1, 0, 0, 0.0))

        # this test will break around year 2070

        self.assertEquals(parse_date("03-Feb-20"), (2020, 2, 3, 0, 0, 0.0))

        # this test will break around year 2048

        self.assertEquals(parse_date("03-Feb-98"), (1998, 2, 3, 0, 0, 0.0))

    def test_http2time_formats(self):

        from cookielib import http2time, time2isoz

        # test http2time for supported dates.  Test cases with 2 digit year

        # will probably break in year 2044.

        tests = [

         'Thu, 03 Feb 1994 00:00:00 GMT',  # proposed new HTTP format

         'Thursday, 03-Feb-94 00:00:00 GMT',  # old rfc850 HTTP format

         'Thursday, 03-Feb-1994 00:00:00 GMT',  # broken rfc850 HTTP format

         '03 Feb 1994 00:00:00 GMT',  # HTTP format (no weekday)

         '03-Feb-94 00:00:00 GMT',  # old rfc850 (no weekday)

         '03-Feb-1994 00:00:00 GMT',  # broken rfc850 (no weekday)

         '03-Feb-1994 00:00 GMT',  # broken rfc850 (no weekday, no seconds)

         '03-Feb-1994 00:00',  # broken rfc850 (no weekday, no seconds, no tz)

         '03-Feb-94',  # old rfc850 HTTP format (no weekday, no time)

         '03-Feb-1994',  # broken rfc850 HTTP format (no weekday, no time)

         '03 Feb 1994',  # proposed new HTTP format (no weekday, no time)

         # A few tests with extra space at various places

         '  03   Feb   1994  0:00  ',

         '  03-Feb-1994  ',

        ]

        test_t = 760233600  # assume broken POSIX counting of seconds

        result = time2isoz(test_t)

        expected = "1994-02-03 00:00:00Z"

        self.assertEquals(result, expected,

                          "%s  =>  '%s' (%s)" % (test_t, result, expected))

        for s in tests:

            t = http2time(s)

            t2 = http2time(s.lower())

            t3 = http2time(s.upper())

            self.assert_(t == t2 == t3 == test_t,

                         "'%s'  =>  %s, %s, %s (%s)" % (s, t, t2, t3, test_t))

    def test_http2time_garbage(self):

        from cookielib import http2time

        for test in [

            '',

            'Garbage',

            'Mandag 16. September 1996',

            '01-00-1980',

            '01-13-1980',

            '00-01-1980',

            '32-01-1980',

            '01-01-1980 25:00:00',

            '01-01-1980 00:61:00',

            '01-01-1980 00:00:62',

            ]:

            self.assert_(http2time(test) is None,

                         "http2time(%s) is not None\n"

                         "http2time(test) %s" % (test, http2time(test))

                         )

class HeaderTests(TestCase):

    def test_parse_ns_headers(self):

        from cookielib import parse_ns_headers

        # quotes should be stripped

        expected = [[('foo', 'bar'), ('expires', 2209069412L), ('version', '0')]]

        for hdr in [

            'foo=bar; expires=01 Jan 2040 22:23:32 GMT',

            'foo=bar; expires="01 Jan 2040 22:23:32 GMT"',

            ]:

            self.assertEquals(parse_ns_headers([hdr]), expected)

    def test_parse_ns_headers_special_names(self):

        # names such as 'expires' are not special in first name=value pair

        # of Set-Cookie: header

        from cookielib import parse_ns_headers

        # Cookie with name 'expires'

        hdr = 'expires=01 Jan 2040 22:23:32 GMT'

        expected = [[("expires", "01 Jan 2040 22:23:32 GMT"), ("version", "0")]]

        self.assertEquals(parse_ns_headers([hdr]), expected)

    def test_join_header_words(self):

        from cookielib import join_header_words

        joined = join_header_words([[("foo", None), ("bar", "baz")]])

        self.assertEquals(joined, "foo; bar=baz")

        self.assertEquals(join_header_words([[]]), "")

    def test_split_header_words(self):

        from cookielib import split_header_words

        tests = [

            ("foo", [[("foo", None)]]),

            ("foo=bar", [[("foo", "bar")]]),

            ("   foo   ", [[("foo", None)]]),

            ("   foo=   ", [[("foo", "")]]),

            ("   foo=", [[("foo", "")]]),

            ("   foo=   ; ", [[("foo", "")]]),

            ("   foo=   ; bar= baz ", [[("foo", ""), ("bar", "baz")]]),

            ("foo=bar bar=baz", [[("foo", "bar"), ("bar", "baz")]]),

            # doesn't really matter if this next fails, but it works ATM

            ("foo= bar=baz", [[("foo", "bar=baz")]]),

            ("foo=bar;bar=baz", [[("foo", "bar"), ("bar", "baz")]]),

            ('foo bar baz', [[("foo", None), ("bar", None), ("baz", None)]]),

            ("a, b, c", [[("a", None)], [("b", None)], [("c", None)]]),

            (r'foo; bar=baz, spam=, foo="\,\;\"", bar= ',

             [[("foo", None), ("bar", "baz")],

              [("spam", "")], [("foo", ',;"')], [("bar", "")]]),

            ]

        for arg, expect in tests:

            try:

                result = split_header_words([arg])

            except:

                import traceback, StringIO

                f = StringIO.StringIO()

                traceback.print_exc(None, f)

                result = "(error -- traceback follows)\n\n%s" % f.getvalue()

            self.assertEquals(result,  expect, """

When parsing: '%s'

Expected:     '%s'

Got:          '%s'

""" % (arg, expect, result))

    def test_roundtrip(self):

        from cookielib import split_header_words, join_header_words

        tests = [

            ("foo", "foo"),

            ("foo=bar", "foo=bar"),

            ("   foo   ", "foo"),

            ("foo=", 'foo=""'),

            ("foo=bar bar=baz", "foo=bar; bar=baz"),

            ("foo=bar;bar=baz", "foo=bar; bar=baz"),

            ('foo bar baz', "foo; bar; baz"),

            (r'foo="\"" bar="\\"', r'foo="\""; bar="\\"'),

            ('foo,,,bar', 'foo, bar'),

            ('foo=bar,bar=baz', 'foo=bar, bar=baz'),

            ('text/html; charset=iso-8859-1',

             'text/html; charset="iso-8859-1"'),

            ('foo="bar"; port="80,81"; discard, bar=baz',

             'foo=bar; port="80,81"; discard, bar=baz'),

            (r'Basic realm="\"foo\\\\bar\""',

             r'Basic; realm="\"foo\\\\bar\""')

            ]

        for arg, expect in tests:

            input = split_header_words([arg])

            res = join_header_words(input)

            self.assertEquals(res, expect, """

When parsing: '%s'

Expected:     '%s'

Got:          '%s'

Input was:    '%s'

""" % (arg, expect, res, input))

class FakeResponse:

    def __init__(self, headers=[], url=None):

        """

        headers: list of RFC822-style 'Key: value' strings

        """

        import mimetools, StringIO

        f = StringIO.StringIO("\n".join(headers))

        self._headers = mimetools.Message(f)

        self._url = url

    def info(self): return self._headers

def interact_2965(cookiejar, url, *set_cookie_hdrs):

    return _interact(cookiejar, url, set_cookie_hdrs, "Set-Cookie2")

def interact_netscape(cookiejar, url, *set_cookie_hdrs):

    return _interact(cookiejar, url, set_cookie_hdrs, "Set-Cookie")

def _interact(cookiejar, url, set_cookie_hdrs, hdr_name):

    """Perform a single request / response cycle, returning Cookie: header."""

    from urllib2 import Request

    req = Request(url)

    cookiejar.add_cookie_header(req)

    cookie_hdr = req.get_header("Cookie", "")

    headers = []

    for hdr in set_cookie_hdrs:

        headers.append("%s: %s" % (hdr_name, hdr))

    res = FakeResponse(headers, url)

    cookiejar.extract_cookies(res, req)

    return cookie_hdr

class FileCookieJarTests(TestCase):

    def test_lwp_valueless_cookie(self):

        # cookies with no value should be saved and loaded consistently

        from cookielib import LWPCookieJar

        filename = test_support.TESTFN

        c = LWPCookieJar()

        interact_netscape(c, "http://www.acme.com/", 'boo')

        self.assertEqual(c._cookies["www.acme.com"]["/"]["boo"].value, None)

        try:

            c.save(filename, ignore_discard=True)

            c = LWPCookieJar()

            c.load(filename, ignore_discard=True)

        finally:

            try: os.unlink(filename)

            except OSError: pass

        self.assertEqual(c._cookies["www.acme.com"]["/"]["boo"].value, None)

    def test_bad_magic(self):

        from cookielib import LWPCookieJar, MozillaCookieJar, LoadError

        # IOErrors (eg. file doesn't exist) are allowed to propagate

        filename = test_support.TESTFN

        for cookiejar_class in LWPCookieJar, MozillaCookieJar:

            c = cookiejar_class()

            try:

                c.load(filename="for this test to work, a file with this "

                                "filename should not exist")

            except IOError, exc:

                # exactly IOError, not LoadError

                self.assertEqual(exc.__class__, IOError)

            else:

                self.fail("expected IOError for invalid filename")

        # Invalid contents of cookies file (eg. bad magic string)

        # causes a LoadError.

        try:

            f = open(filename, "w")

            f.write("oops\n")

            for cookiejar_class in LWPCookieJar, MozillaCookieJar:

                c = cookiejar_class()

                self.assertRaises(LoadError, c.load, filename)

        finally:

            try: os.unlink(filename)

            except OSError: pass

class CookieTests(TestCase):

    # XXX

    # Get rid of string comparisons where not actually testing str / repr.

    # .clear() etc.

    # IP addresses like 50 (single number, no dot) and domain-matching

    #  functions (and is_HDN)?  See draft RFC 2965 errata.

    # Strictness switches

    # is_third_party()

    # unverifiability / third-party blocking

    # Netscape cookies work the same as RFC 2965 with regard to port.

    # Set-Cookie with negative max age.

    # If turn RFC 2965 handling off, Set-Cookie2 cookies should not clobber

    #  Set-Cookie cookies.

    # Cookie2 should be sent if *any* cookies are not V1 (ie. V0 OR V2 etc.).

    # Cookies (V1 and V0) with no expiry date should be set to be discarded.

    # RFC 2965 Quoting:

    #  Should accept unquoted cookie-attribute values?  check errata draft.

    #   Which are required on the way in and out?

    #  Should always return quoted cookie-attribute values?

    # Proper testing of when RFC 2965 clobbers Netscape (waiting for errata).

    # Path-match on return (same for V0 and V1).

    # RFC 2965 acceptance and returning rules

    #  Set-Cookie2 without version attribute is rejected.

    # Netscape peculiarities list from Ronald Tschalar.

    # The first two still need tests, the rest are covered.

## - Quoting: only quotes around the expires value are recognized as such

##   (and yes, some folks quote the expires value); quotes around any other

##   value are treated as part of the value.

## - White space: white space around names and values is ignored

## - Default path: if no path parameter is given, the path defaults to the

##   path in the request-uri up to, but not including, the last '/'. Note

##   that this is entirely different from what the spec says.

## - Commas and other delimiters: Netscape just parses until the next ';'.

##   This means it will allow commas etc inside values (and yes, both

##   commas and equals are commonly appear in the cookie value). This also

##   means that if you fold multiple Set-Cookie header fields into one,

##   comma-separated list, it'll be a headache to parse (at least my head

##   starts hurting everytime I think of that code).

## - Expires: You'll get all sorts of date formats in the expires,

##   including emtpy expires attributes ("expires="). Be as flexible as you

##   can, and certainly don't expect the weekday to be there; if you can't

##   parse it, just ignore it and pretend it's a session cookie.

## - Domain-matching: Netscape uses the 2-dot rule for _all_ domains, not

##   just the 7 special TLD's listed in their spec. And folks rely on

##   that...

    def test_domain_return_ok(self):

        # test optimization: .domain_return_ok() should filter out most

        # domains in the CookieJar before we try to access them (because that

        # may require disk access -- in particular, with MSIECookieJar)

        # This is only a rough check for performance reasons, so it's not too

        # critical as long as it's sufficiently liberal.

        import cookielib, urllib2

        pol = cookielib.DefaultCookiePolicy()

        for url, domain, ok in [

            ("http://foo.bar.com/", "blah.com", False),

            ("http://foo.bar.com/", "rhubarb.blah.com", False),

            ("http://foo.bar.com/", "rhubarb.foo.bar.com", False),

            ("http://foo.bar.com/", ".foo.bar.com", True),

            ("http://foo.bar.com/", "foo.bar.com", True),

            ("http://foo.bar.com/", ".bar.com", True),

            ("http://foo.bar.com/", "com", True),

            ("http://foo.com/", "rhubarb.foo.com", False),

            ("http://foo.com/", ".foo.com", True),

            ("http://foo.com/", "foo.com", True),

            ("http://foo.com/", "com", True),

            ("http://foo/", "rhubarb.foo", False),

            ("http://foo/", ".foo", True),

            ("http://foo/", "foo", True),

            ("http://foo/", "foo.local", True),

            ("http://foo/", ".local", True),

            ]:

            request = urllib2.Request(url)

            r = pol.domain_return_ok(domain, request)

            if ok: self.assert_(r)

            else: self.assert_(not r)

    def test_missing_value(self):

        from cookielib import MozillaCookieJar, lwp_cookie_str

        # missing = sign in Cookie: header is regarded by Mozilla as a missing

        # name, and by cookielib as a missing value

        filename = test_support.TESTFN

        c = MozillaCookieJar(filename)

        interact_netscape(c, "http://www.acme.com/", 'eggs')

        interact_netscape(c, "http://www.acme.com/", '"spam"; path=/foo/')

        cookie = c._cookies["www.acme.com"]["/"]["eggs"]

        self.assert_(cookie.value is None)

        self.assertEquals(cookie.name, "eggs")

        cookie = c._cookies["www.acme.com"]['/foo/']['"spam"']

        self.assert_(cookie.value is None)

        self.assertEquals(cookie.name, '"spam"')

        self.assertEquals(lwp_cookie_str(cookie), (

            r'"spam"; path="/foo/"; domain="www.acme.com"; '

            'path_spec; discard; version=0'))

        old_str = repr(c)

        c.save(ignore_expires=True, ignore_discard=True)

        try:

            c = MozillaCookieJar(filename)

            c.revert(ignore_expires=True, ignore_discard=True)

        finally:

            os.unlink(c.filename)

        # cookies unchanged apart from lost info re. whether path was specified

        self.assertEquals(

            repr(c),

            re.sub("path_specified=%s" % True, "path_specified=%s" % False,

                   old_str)

            )

        self.assertEquals(interact_netscape(c, "http://www.acme.com/foo/"),

                          '"spam"; eggs')

    def test_rfc2109_handling(self):

        # RFC 2109 cookies are handled as RFC 2965 or Netscape cookies,

        # dependent on policy settings

        from cookielib import CookieJar, DefaultCookiePolicy

        for rfc2109_as_netscape, rfc2965, version in [

            # default according to rfc2965 if not explicitly specified

            (None, False, 0),

            (None, True, 1),

            # explicit rfc2109_as_netscape

            (False, False, None),  # version None here means no cookie stored

            (False, True, 1),

            (True, False, 0),

            (True, True, 0),

            ]:

            policy = DefaultCookiePolicy(

                rfc2109_as_netscape=rfc2109_as_netscape,

                rfc2965=rfc2965)

            c = CookieJar(policy)

            interact_netscape(c, "http://www.example.com/", "ni=ni; Version=1")

            try:

                cookie = c._cookies["www.example.com"]["/"]["ni"]

            except KeyError:

                self.assert_(version is None)  # didn't expect a stored cookie

            else:

                self.assertEqual(cookie.version, version)

                # 2965 cookies are unaffected

                interact_2965(c, "http://www.example.com/",

                              "foo=bar; Version=1")

                if rfc2965:

                    cookie2965 = c._cookies["www.example.com"]["/"]["foo"]

                    self.assertEqual(cookie2965.version, 1)

    def test_ns_parser(self):

        from cookielib import CookieJar, DEFAULT_HTTP_PORT

        c = CookieJar()

        interact_netscape(c, "http://www.acme.com/",

                          'spam=eggs; DoMain=.acme.com; port; blArgh="feep"')

        interact_netscape(c, "http://www.acme.com/", 'ni=ni; port=80,8080')

        interact_netscape(c, "http://www.acme.com:80/", 'nini=ni')

        interact_netscape(c, "http://www.acme.com:80/", 'foo=bar; expires=')

        interact_netscape(c, "http://www.acme.com:80/", 'spam=eggs; '

                          'expires="Foo Bar 25 33:22:11 3022"')

        cookie = c._cookies[".acme.com"]["/"]["spam"]

        self.assertEquals(cookie.domain, ".acme.com")

        self.assert_(cookie.domain_specified)

        self.assertEquals(cookie.port, DEFAULT_HTTP_PORT)

        self.assert_(not cookie.port_specified)

        # case is preserved

        self.assert_(cookie.has_nonstandard_attr("blArgh") and

                     not cookie.has_nonstandard_attr("blargh"))

        cookie = c._cookies["www.acme.com"]["/"]["ni"]

        self.assertEquals(cookie.domain, "www.acme.com")

        self.assert_(not cookie.domain_specified)

        self.assertEquals(cookie.port, "80,8080")

        self.assert_(cookie.port_specified)

        cookie = c._cookies["www.acme.com"]["/"]["nini"]

        self.assert_(cookie.port is None)

        self.assert_(not cookie.port_specified)

        # invalid expires should not cause cookie to be dropped

        foo = c._cookies["www.acme.com"]["/"]["foo"]

        spam = c._cookies["www.acme.com"]["/"]["foo"]

        self.assert_(foo.expires is None)

        self.assert_(spam.expires is None)

    def test_ns_parser_special_names(self):

        # names such as 'expires' are not special in first name=value pair

        # of Set-Cookie: header

        from cookielib import CookieJar

        c = CookieJar()

        interact_netscape(c, "http://www.acme.com/", 'expires=eggs')

        interact_netscape(c, "http://www.acme.com/", 'version=eggs; spam=eggs')

        cookies = c._cookies["www.acme.com"]["/"]

        self.assert_('expires' in cookies)

        self.assert_('version' in cookies)

    def test_expires(self):

        from cookielib import time2netscape, CookieJar

        # if expires is in future, keep cookie...

        c = CookieJar()

        future = time2netscape(time.time()+3600)

        interact_netscape(c, "http://www.acme.com/", 'spam="bar"; expires=%s' %

                          future)

        self.assertEquals(len(c), 1)

        now = time2netscape(time.time()-1)

        # ... and if in past or present, discard it

        interact_netscape(c, "http://www.acme.com/", 'foo="eggs"; expires=%s' %

                          now)

        h = interact_netscape(c, "http://www.acme.com/")

        self.assertEquals(len(c), 1)

        self.assert_('spam="bar"' in h and "foo" not in h)

        # max-age takes precedence over expires, and zero max-age is request to

        # delete both new cookie and any old matching cookie

        interact_netscape(c, "http://www.acme.com/", 'eggs="bar"; expires=%s' %

                          future)

        interact_netscape(c, "http://www.acme.com/", 'bar="bar"; expires=%s' %

                          future)

        self.assertEquals(len(c), 3)

        interact_netscape(c, "http://www.acme.com/", 'eggs="bar"; '

                          'expires=%s; max-age=0' % future)

        interact_netscape(c, "http://www.acme.com/", 'bar="bar"; '

                          'max-age=0; expires=%s' % future)

        h = interact_netscape(c, "http://www.acme.com/")

        self.assertEquals(len(c), 1)

        # test expiry at end of session for cookies with no expires attribute

        interact_netscape(c, "http://www.rhubarb.net/", 'whum="fizz"')

        self.assertEquals(len(c), 2)

        c.clear_session_cookies()

        self.assertEquals(len(c), 1)

        self.assert_('spam="bar"' in h)

        # XXX RFC 2965 expiry rules (some apply to V0 too)

    def test_default_path(self):

        from cookielib import CookieJar, DefaultCookiePolicy

        # RFC 2965

        pol = DefaultCookiePolicy(rfc2965=True)

        c = CookieJar(pol)

        interact_2965(c, "http://www.acme.com/", 'spam="bar"; Version="1"')

        self.assert_("/" in c._cookies["www.acme.com"])

        c = CookieJar(pol)

        interact_2965(c, "http://www.acme.com/blah", 'eggs="bar"; Version="1"')

        self.assert_("/" in c._cookies["www.acme.com"])

        c = CookieJar(pol)

        interact_2965(c, "http://www.acme.com/blah/rhubarb",

                      'eggs="bar"; Version="1"')