iframe-proxy | Sunbelt Computer Software

History

197 lines (157 loc) · 7.88 KB

Raw

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

#!/usr/bin/env python

"""

See the file 'LICENSE' for copying permission

Unit coverage for PURE functions in lib/request/basic.py.

These exercise getHeuristicCharEncoding (with its kb.cache.encoding memoization)

and decodePage's charset + HTML-entity decoding branches, in isolation - WITHOUT

touching the network, the DBMS or any interactive prompt.

stdlib unittest only (no pytest / no pip); works on Python 2.7 and 3.x.

"""

import os

import sys

import unittest

sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))

from _testutils import bootstrap

bootstrap()

from lib.core.data import conf, kb

class TestBasicHeuristicCharEncoding(unittest.TestCase):

def test_ascii(self):

from lib.request.basic import getHeuristicCharEncoding

self.assertEqual(getHeuristicCharEncoding(b"<html></html>"), "ascii")

def test_cache_hit_returns_same(self):

from lib.request.basic import getHeuristicCharEncoding

page = b"<html>hello world</html>"

first = getHeuristicCharEncoding(page)

# second call for identical page must come back identical (and from cache)

self.assertEqual(getHeuristicCharEncoding(page), first)

key = (len(page), hash(page))

self.assertEqual(kb.cache.encoding.get(key), first)

class TestBasicDecodePage(unittest.TestCase):

"""decodePage charset + HTML-entity decoding branches."""

def setUp(self):

self._old_encoding = conf.encoding

self._old_null = conf.nullConnection

conf.nullConnection = False

def tearDown(self):

conf.encoding = self._old_encoding

conf.nullConnection = self._old_null

def test_html_entity_amp(self):

from lib.request.basic import decodePage

from lib.core.common import getText

conf.encoding = None

self.assertEqual(

getText(decodePage(b"<html>foo&bar</html>", None, "text/html; charset=utf-8")),

"<html>foo&bar</html>",

)

def test_numeric_hex_entity_tab(self):

from lib.request.basic import decodePage

from lib.core.common import getText

conf.encoding = None

self.assertEqual(getText(decodePage(b"	", None, "text/html; charset=utf-8")), "\t")

def test_numeric_hex_entity_letter(self):

from lib.request.basic import decodePage

from lib.core.common import getText

conf.encoding = None

self.assertEqual(getText(decodePage(b"J", None, "text/html; charset=utf-8")), "J")

def test_unicode_entity(self):

from lib.request.basic import decodePage

conf.encoding = None

self.assertEqual(decodePage(b"™", None, "text/html; charset=utf-8"), u"\u2122")

def test_empty_page(self):

from lib.request.basic import decodePage

from lib.core.common import getText

# empty page short-circuits to getUnicode(page)

self.assertEqual(getText(decodePage(b"", None, "text/html")), "")

class TestRedirectSetCookieMerge(unittest.TestCase):

"""A 302 that sets more than one cookie sends SEPARATE Set-Cookie headers (RFC-6265 forbids

comma-folding them). The handler must merge ALL of them into the follow-up request's Cookie

header; a __getitem__ read returns only the first, silently dropping the 2nd+ (e.g. a CSRF token)."""

_CONF = ("cookieDel", "scope")

def setUp(self):

self._c = dict((k, conf.get(k)) for k in self._CONF)

self._redirect = kb.choices.get("redirect") if kb.get("choices") else None

def tearDown(self):

for k, v in self._c.items():

conf[k] = v

if kb.get("choices"):

kb.choices.redirect = self._redirect

def test_all_set_cookies_merged_across_redirect(self):

from lib.core.enums import HTTP_HEADER, REDIRECTION

from thirdparty import six

from thirdparty.six.moves import urllib as _urllib

import lib.request.redirecthandler as rh

conf.cookieDel = None

conf.scope = None

kb.choices.redirect = REDIRECTION.YES

raw = ("Location: http://example.com/home\r\n"

"Set-Cookie: sid=NEW; Path=/; HttpOnly\r\n"

"Set-Cookie: csrf=XYZ; Path=/\r\n\r\n")

# build the same headers object type the redirect handler receives on this interpreter:

# py3 http.client.HTTPMessage (email.message.Message, get_all) / py2 mimetools.Message (getheaders)

if six.PY2:

import mimetools

headers = mimetools.Message(six.StringIO(raw))

else:

from email import message_from_string

from thirdparty.six.moves.http_client import HTTPMessage

headers = message_from_string(raw, _class=HTTPMessage)

# stub the network-following parent so the test touches no socket

saved = _urllib.request.HTTPRedirectHandler.http_error_302

_urllib.request.HTTPRedirectHandler.http_error_302 = lambda self, req, fp, code, msg, headers: fp

try:

req = _urllib.request.Request("http://example.com/login", headers={"Cookie": "sid=OLD"})

fp = _urllib.response.addinfourl(six.BytesIO(b""), headers, req.get_full_url())

rh.SmartRedirectHandler().http_error_302(req, fp, 302, "Found", headers)

finally:

_urllib.request.HTTPRedirectHandler.http_error_302 = saved

merged = req.headers.get(HTTP_HEADER.COOKIE) or req.headers.get("Cookie") or ""

self.assertIn("sid=NEW", merged)

self.assertIn("csrf=XYZ", merged) # the 2nd Set-Cookie must survive the redirect (order-independent)

class TestForgeHeadersCookieMerge(unittest.TestCase):

"""A domain-scoped jar cookie (Domain=example.com -> '.example.com') must merge into the

request for the apex host, not be dropped by a naive endswith() domain check."""

_CONF = ("cj", "hostname", "httpHeaders", "loadCookies", "cookieDel", "parameters", "csrfToken", "safeUrl")

_KB = ("mergeCookies", "testMode", "injection")

def setUp(self):

self._c = dict((k, conf.get(k)) for k in self._CONF)

self._k = dict((k, kb.get(k)) for k in self._KB)

def tearDown(self):

for k, v in self._c.items():

conf[k] = v

for k, v in self._k.items():

kb[k] = v

def _jar_with_domain_cookie(self):

try:

from http.cookiejar import CookieJar, Cookie

except ImportError:

from cookielib import CookieJar, Cookie

# a domain-scoped cookie the jar stores as '.example.com' (domain_specified=True),

# exactly as it would after Set-Cookie: sid=NEW; Domain=example.com

cookie = Cookie(version=0, name="sid", value="NEW", port=None, port_specified=False,

domain=".example.com", domain_specified=True, domain_initial_dot=True,

path="/", path_specified=True, secure=False, expires=None, discard=True,

comment=None, comment_url=None, rest={})

cj = CookieJar()

cj.set_cookie(cookie)

return cj

def test_domain_cookie_merged_on_apex_host(self):

from lib.request.basic import forgeHeaders

from lib.core.enums import PLACE, HTTP_HEADER

from lib.core.datatype import AttribDict

conf.cj = self._jar_with_domain_cookie()

conf.hostname = "example.com" # apex host == cookie domain

conf.httpHeaders = [(HTTP_HEADER.COOKIE, "sid=OLD")]

conf.loadCookies = False

conf.cookieDel = None

conf.parameters = {}

conf.csrfToken = conf.safeUrl = None

kb.mergeCookies = True

kb.testMode = False

kb.injection = AttribDict()

kb.injection.place = PLACE.GET

headers = forgeHeaders()

# before the fix the domain cookie was skipped for the apex host, leaving 'sid=OLD'

self.assertEqual(headers.get(HTTP_HEADER.COOKIE), "sid=NEW")

if __name__ == "__main__":

unittest.main(verbosity=2)

Sunbelt Computer Software

PL/B Language Development and Support

Navigation Menu

Uh oh!

Uh oh!

Sunbelt Computer Software

PL/B Language Development and Support

Uh oh!

Uh oh!

FilesExpand file tree

test_request_basic.py

Latest commit

History

test_request_basic.py

File metadata and controls