runtests.py

# Copyright 2010-2020 Kurt McKee <contactme@kurtmckee.org>
# Copyright 2004-2008 Mark Pilgrim
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without modification,
# are permitted provided that the following conditions are met:
#
# * Redistributions of source code must retain the above copyright notice,
#   this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright notice,
#   this list of conditions and the following disclaimer in the documentation
#   and/or other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 'AS IS'
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.

__author__ = "Kurt McKee <contactme@kurtmckee.org>"
__license__ = "BSD 2-clause"

import codecs
import datetime
import glob
import http.server
import io
import os
import posixpath
import pprint
import re
import struct
import sys
import threading
import time
import unittest
import urllib.error
import warnings
import xml.sax
import zlib

import feedparser
import feedparser.api
import feedparser.datetimes
import feedparser.http
import feedparser.mixin
import feedparser.sanitizer
import feedparser.urls
import feedparser.util
from feedparser.datetimes.asctime import _parse_date_asctime
from feedparser.datetimes.greek import _parse_date_greek
from feedparser.datetimes.hungarian import _parse_date_hungarian
from feedparser.datetimes.iso8601 import _parse_date_iso8601
from feedparser.datetimes.korean import _parse_date_onblog, _parse_date_nate
from feedparser.datetimes.perforce import _parse_date_perforce
from feedparser.datetimes.rfc822 import _parse_date_rfc822
from feedparser.datetimes.w3dtf import _parse_date_w3dtf

# ---------- custom HTTP server (used to serve test feeds) ----------

_PORT = 8097  # not really configurable, must match hardcoded port in tests
_HOST = '127.0.0.1'  # also not really configurable


class FeedParserTestRequestHandler(http.server.SimpleHTTPRequestHandler):
    headers_re = re.compile(br"^Header:\s+([^:]+):(.+)$", re.MULTILINE)

    def send_head(self):
        """Send custom headers defined in test case

        Example:
        <!--
        Header:   Content-type: application/atom+xml
        Header:   X-Foo: bar
        -->
        """
        # Short-circuit the HTTP status test `test_redirect_to_304()`
        if self.path == '/-/return-304.xml':
            self.send_response(304)
            self.send_header('Content-type', 'text/xml')
            self.end_headers()
            return io.BytesIO(b'')
        path = self.translate_path(self.path)
        # the compression tests' filenames determine the header sent
        if self.path.startswith('/tests/compression'):
            if self.path.endswith('gz'):
                headers = {'Content-Encoding': 'gzip'}
            else:
                headers = {'Content-Encoding': 'deflate'}
            headers['Content-type'] = 'application/xml'
        else:
            with open(path, 'rb') as f:
                blob = f.read()
            headers = {
                k.decode('utf-8'): v.decode('utf-8').strip()
                for k, v in self.headers_re.findall(blob)
            }
        f = open(path, 'rb')
        if (
                self.headers.get('if-modified-since') == headers.get('Last-Modified', 'nom')
                or self.headers.get('if-none-match') == headers.get('ETag', 'nomatch')
        ):
            status = 304
        else:
            status = 200
        headers.setdefault('Status', status)
        self.send_response(int(headers['Status']))
        headers.setdefault('Content-type', self.guess_type(path))
        self.send_header("Content-type", headers['Content-type'])
        self.send_header("Content-Length", str(os.stat(f.name)[6]))
        for k, v in headers.items():
            if k not in ('Status', 'Content-type'):
                self.send_header(k, v)
        self.end_headers()
        return f

    def log_request(self, *args):
        pass


class FeedParserTestServer(threading.Thread):
    """HTTP Server that runs in a thread and handles a predetermined number of requests"""

    def __init__(self, requests):
        threading.Thread.__init__(self)
        self.requests = requests
        self.ready = threading.Event()
        self.httpd = None

    def run(self):
        self.httpd = http.server.HTTPServer((_HOST, _PORT), FeedParserTestRequestHandler)
        self.ready.set()
        while self.requests:
            self.httpd.handle_request()
            self.requests -= 1
        self.ready.clear()
        self.httpd.shutdown()


# ---------- dummy test case class (test methods are added dynamically) ----------

# _bytes is only used in everything_is_unicode().
# In Python 2 it's str, and in Python 3 it's bytes.
_bytes = type(b'')


def everything_is_unicode(d):
    """Takes a dictionary, recursively verifies that every value is unicode"""
    for k, v in d.items():
        if isinstance(v, dict) and k != 'headers':
            if not everything_is_unicode(v):
                return False
        elif isinstance(v, list):
            for i in v:
                if isinstance(i, dict) and not everything_is_unicode(i):
                    return False
                elif isinstance(i, _bytes):
                    return False
        elif isinstance(v, _bytes):
            return False
    return True


def fail_unless_eval(self, xmlfile, eval_string, msg=None):
    """Fail unless eval(eval_string, env)"""
    env = feedparser.parse(xmlfile)
    if not eval(eval_string, globals(), env):
        failure = msg or 'not eval(%s) \nWITH env(%s)' % (eval_string, pprint.pformat(env))
        raise self.failureException(failure)
    if not everything_is_unicode(env):
        raise self.failureException("not everything is unicode \nWITH env(%s)" % (pprint.pformat(env), ))


class BaseTestCase(unittest.TestCase):
    fail_unless_eval = fail_unless_eval


class TestCase(BaseTestCase):
    pass


class TestTemporaryFallbackBehavior(unittest.TestCase):
    """These tests are temporarily here because of issues 310 and 328"""
    def test_issue_328_fallback_behavior(self):
        warnings.filterwarnings('error')

        d = feedparser.util.FeedParserDict()
        d['published'] = 'pub string'
        d['published_parsed'] = 'pub tuple'
        d['updated'] = 'upd string'
        d['updated_parsed'] = 'upd tuple'
        # Ensure that `updated` doesn't map to `published` when it exists
        self.assertTrue('published' in d)
        self.assertTrue('published_parsed' in d)
        self.assertTrue('updated' in d)
        self.assertTrue('updated_parsed' in d)
        self.assertEqual(d['published'], 'pub string')
        self.assertEqual(d['published_parsed'], 'pub tuple')
        self.assertEqual(d['updated'], 'upd string')
        self.assertEqual(d['updated_parsed'], 'upd tuple')

        d = feedparser.util.FeedParserDict()
        d['published'] = 'pub string'
        d['published_parsed'] = 'pub tuple'
        # Ensure that `updated` doesn't actually exist
        self.assertTrue('updated' not in d)
        self.assertTrue('updated_parsed' not in d)
        # Ensure that accessing `updated` throws a DeprecationWarning
        try:
            d['updated']
        except DeprecationWarning:
            # Expected behavior
            pass
        else:
            # Wrong behavior
            self.assertEqual(True, False)
        try:
            d['updated_parsed']
        except DeprecationWarning:
            # Expected behavior
            pass
        else:
            # Wrong behavior
            self.assertEqual(True, False)
        # Ensure that `updated` maps to `published`
        warnings.filterwarnings('ignore')
        self.assertEqual(d['updated'], 'pub string')
        self.assertEqual(d['updated_parsed'], 'pub tuple')
        warnings.resetwarnings()


class TestEverythingIsUnicode(unittest.TestCase):
    """Ensure that `everything_is_unicode()` is working appropriately"""
    def test_everything_is_unicode(self):
        self.assertTrue(everything_is_unicode(
            {'a': 'a', 'b': ['b', {'c': 'c'}], 'd': {'e': 'e'}}
        ))

    def test_not_everything_is_unicode(self):
        self.assertFalse(everything_is_unicode({'a': b'a'}))
        self.assertFalse(everything_is_unicode({'a': [b'a']}))
        self.assertFalse(everything_is_unicode({'a': {'b': b'b'}}))
        self.assertFalse(everything_is_unicode({'a': [{'b': b'b'}]}))


class TestLooseParser(BaseTestCase):
    """Test the sgmllib-based parser by manipulating feedparser into believing no XML parsers are installed"""
    def __init__(self, arg):
        unittest.TestCase.__init__(self, arg)
        self._xml_available = feedparser.api._XML_AVAILABLE

    def setUp(self):
        feedparser.api._XML_AVAILABLE = False

    def tearDown(self):
        feedparser.api._XML_AVAILABLE = self._xml_available


class TestStrictParser(BaseTestCase):
    pass


class TestEncodings(BaseTestCase):
    def test_doctype_replacement(self):
        """Ensure that non-ASCII-compatible encodings don't hide disallowed ENTITY declarations"""
        doc = """<?xml version="1.0" encoding="utf-16be"?>
        <!DOCTYPE feed [
            <!ENTITY exponential1 "bogus ">
            <!ENTITY exponential2 "&exponential1;&exponential1;">
            <!ENTITY exponential3 "&exponential2;&exponential2;">
        ]>
        <feed><title type="html">&exponential3;</title></feed>"""
        doc = codecs.BOM_UTF16_BE + doc.encode('utf-16be')
        result = feedparser.parse(doc)
        self.assertEqual(result['feed']['title'], '&amp;exponential3')

    def test_gb2312_converted_to_gb18030_in_xml_encoding(self):
        # \u55de was chosen because it exists in gb18030 but not gb2312
        feed = '''<?xml version="1.0" encoding="gb2312"?>
                  <feed><title>\u55de</title></feed>'''
        result = feedparser.parse(feed.encode('gb18030'), response_headers={
            'Content-Type': 'text/xml'
        })
        self.assertEqual(result.encoding, 'gb18030')


class TestFeedParserDict(unittest.TestCase):
    """Ensure that FeedParserDict returns values as expected and won't crash"""

    def setUp(self):
        self.d = feedparser.util.FeedParserDict()

    def _check_key(self, k):
        self.assertTrue(k in self.d)
        self.assertTrue(hasattr(self.d, k))
        self.assertEqual(self.d[k], 1)
        self.assertEqual(getattr(self.d, k), 1)

    def _check_no_key(self, k):
        self.assertTrue(k not in self.d)
        self.assertTrue(not hasattr(self.d, k))

    def test_empty(self):
        keys = (
            'a', 'entries', 'id', 'guid', 'summary', 'subtitle', 'description',
            'category', 'enclosures', 'license', 'categories',
        )
        for k in keys:
            self._check_no_key(k)
        self.assertTrue('items' not in self.d)
        self.assertTrue(hasattr(self.d, 'items'))  # dict.items() exists

    def test_neutral(self):
        self.d['a'] = 1
        self._check_key('a')

    def test_single_mapping_target_1(self):
        self.d['id'] = 1
        self._check_key('id')
        self._check_key('guid')

    def test_single_mapping_target_2(self):
        self.d['guid'] = 1
        self._check_key('id')
        self._check_key('guid')

    def test_multiple_mapping_target_1(self):
        self.d['summary'] = 1
        self._check_key('summary')
        self._check_key('description')

    def test_multiple_mapping_target_2(self):
        self.d['subtitle'] = 1
        self._check_key('subtitle')
        self._check_key('description')

    def test_multiple_mapping_mapped_key(self):
        self.d['description'] = 1
        self._check_key('summary')
        self._check_key('description')

    def test_license(self):
        self.d['links'] = []
        self.assertNotIn('license', self.d)

        self.d['links'].append({'rel': 'license'})
        self.assertNotIn('license', self.d)

        self.d['links'].append({'rel': 'license', 'href': 'http://dom.test/'})
        self.assertIn('license', self.d)
        self.assertEqual(self.d['license'], 'http://dom.test/')

    def test_category(self):
        self.d['tags'] = []
        self.assertNotIn('category', self.d)

        self.d['tags'] = [{}]
        self.assertNotIn('category', self.d)

        self.d['tags'] = [{'term': 'cat'}]
        self.assertIn('category', self.d)
        self.assertEqual(self.d['category'], 'cat')
        self.d['tags'].append({'term': 'dog'})
        self.assertEqual(self.d['category'], 'cat')


class TestOpenResource(unittest.TestCase):
    """Ensure that `_open_resource()` interprets its arguments as URIs, file-like objects, or in-memory feeds as expected"""

    def test_fileobj(self):
        r = feedparser.api._open_resource(io.BytesIO(b''), '', '', '', '', [], {}, {})
        self.assertEqual(r, b'')

    def test_feed(self):
        f = feedparser.parse('feed://localhost:8097/tests/http/target.xml')
        self.assertEqual(f.href, 'http://localhost:8097/tests/http/target.xml')

    def test_feed_http(self):
        f = feedparser.parse('feed:http://localhost:8097/tests/http/target.xml')
        self.assertEqual(f.href, 'http://localhost:8097/tests/http/target.xml')

    def test_bytes(self):
        s = b'<feed><item><title>text</title></item></feed>'
        r = feedparser.api._open_resource(s, '', '', '', '', [], {}, {})
        self.assertEqual(s, r)

    def test_string(self):
        s = b'<feed><item><title>text</title></item></feed>'
        r = feedparser.api._open_resource(s, '', '', '', '', [], {}, {})
        self.assertEqual(s, r)

    def test_unicode_1(self):
        s = b'<feed><item><title>text</title></item></feed>'
        r = feedparser.api._open_resource(s, '', '', '', '', [], {}, {})
        self.assertEqual(s, r)

    def test_unicode_2(self):
        s = br'<feed><item><title>t\u00e9xt</title></item></feed>'
        r = feedparser.api._open_resource(s, '', '', '', '', [], {}, {})
        self.assertEqual(s, r)

    def test_http_client_ascii_unicode_encode_error(self):
        """Confirm that a Unicode character doesn't cause a UnicodeEncodeError crash."""
        url = 'https://0.0.0.0/ô'
        with self.assertRaises(urllib.error.URLError):
            feedparser.api._open_resource(url, '', '', '', '', [], {}, {})


def make_safe_uri_test(rel, expect, doc):
    def fn(self):
        value = feedparser.urls.make_safe_absolute_uri(self.base, rel)
        self.assertEqual(value, expect)

    fn.__doc__ = doc
    return fn


class TestMakeSafeAbsoluteURI(unittest.TestCase):
    """Exercise the URI joining and sanitization code"""

    base = 'http://d.test/d/f.ext'

    # make the test cases; the call signature is:
    # (relative_url, expected_return_value, test_doc_string)
    test_abs = make_safe_uri_test('https://s.test/', 'https://s.test/', 'absolute uri')
    test_rel = make_safe_uri_test('/new', 'http://d.test/new', 'relative uri')
    test_bad = make_safe_uri_test('x://bad.test/', '', 'unacceptable uri protocol')
    test_mag = make_safe_uri_test('magnet:?xt=a', 'magnet:?xt=a', 'magnet uri')

    def test_catch_ValueError(self):
        """catch ValueError in Python 2.7 and up"""
        uri = 'http://bad]test/'
        value1 = feedparser.urls.make_safe_absolute_uri(uri)
        value2 = feedparser.urls.make_safe_absolute_uri(self.base, uri)
        swap = feedparser.urls.ACCEPTABLE_URI_SCHEMES
        feedparser.urls.ACCEPTABLE_URI_SCHEMES = ()
        value3 = feedparser.urls.make_safe_absolute_uri(self.base, uri)
        feedparser.urls.ACCEPTABLE_URI_SCHEMES = swap
        # Only Python 2.7 and up throw a ValueError, otherwise uri is returned
        self.assertTrue(value1 in (uri, ''))
        self.assertTrue(value2 in (uri, ''))
        self.assertTrue(value3 in (uri, ''))


class TestConvertToIdn(unittest.TestCase):
    """Test IDN support (unavailable in Jython as of Jython 2.5.2)"""
    # this is the greek test domain
    hostname = '\u03c0\u03b1\u03c1\u03ac\u03b4\u03b5\u03b9\u03b3\u03bc\u03b1'
    hostname += '.\u03b4\u03bf\u03ba\u03b9\u03bc\u03ae'

    def test_control(self):
        r = feedparser.urls.convert_to_idn('http://example.test/')
        self.assertEqual(r, 'http://example.test/')

    def test_idn(self):
        r = feedparser.urls.convert_to_idn('http://%s/' % (self.hostname,))
        self.assertEqual(r, 'http://xn--hxajbheg2az3al.xn--jxalpdlp/')

    def test_port(self):
        r = feedparser.urls.convert_to_idn('http://%s:8080/' % (self.hostname,))
        self.assertEqual(r, 'http://xn--hxajbheg2az3al.xn--jxalpdlp:8080/')


class TestCompression(unittest.TestCase):
    """Test the gzip and deflate support in the HTTP code"""

    def test_gzip_good(self):
        f = feedparser.parse('http://localhost:8097/tests/compression/gzip.gz')
        self.assertEqual(f.version, 'atom10')

    def test_gzip_not_compressed(self):
        f = feedparser.parse('http://localhost:8097/tests/compression/gzip-not-compressed.gz')
        self.assertEqual(f.bozo, 1)
        self.assertTrue(isinstance(f.bozo_exception, IOError))
        self.assertEqual(f['feed']['title'], 'gzip')

    def test_gzip_struct_error(self):
        f = feedparser.parse('http://localhost:8097/tests/compression/gzip-struct-error.gz')
        self.assertEqual(f.bozo, 1)
        # Python 3.4 throws an EOFError that gets overwritten as xml.sax.SAXException later.
        self.assertTrue(isinstance(f.bozo_exception, (xml.sax.SAXException, struct.error)))

    def test_zlib_good(self):
        f = feedparser.parse('http://localhost:8097/tests/compression/deflate.z')
        self.assertEqual(f.version, 'atom10')

    def test_zlib_no_headers(self):
        f = feedparser.parse('http://localhost:8097/tests/compression/deflate-no-headers.z')
        self.assertEqual(f.version, 'atom10')

    def test_zlib_not_compressed(self):
        f = feedparser.parse('http://localhost:8097/tests/compression/deflate-not-compressed.z')
        self.assertEqual(f.bozo, 1)
        self.assertTrue(isinstance(f.bozo_exception, zlib.error))
        self.assertEqual(f['feed']['title'], 'deflate')


class TestHTTPStatus(unittest.TestCase):
    """Test HTTP redirection and other status codes"""

    def test_301(self):
        f = feedparser.parse('http://localhost:8097/tests/http/http_status_301.xml')
        self.assertEqual(f.status, 301)
        self.assertEqual(f.href, 'http://localhost:8097/tests/http/target.xml')
        self.assertEqual(f.entries[0].title, 'target')

    def test_302(self):
        f = feedparser.parse('http://localhost:8097/tests/http/http_status_302.xml')
        self.assertEqual(f.status, 302)
        self.assertEqual(f.href, 'http://localhost:8097/tests/http/target.xml')
        self.assertEqual(f.entries[0].title, 'target')

    def test_303(self):
        f = feedparser.parse('http://localhost:8097/tests/http/http_status_303.xml')
        self.assertEqual(f.status, 303)
        self.assertEqual(f.href, 'http://localhost:8097/tests/http/target.xml')
        self.assertEqual(f.entries[0].title, 'target')

    def test_307(self):
        f = feedparser.parse('http://localhost:8097/tests/http/http_status_307.xml')
        self.assertEqual(f.status, 307)
        self.assertEqual(f.href, 'http://localhost:8097/tests/http/target.xml')
        self.assertEqual(f.entries[0].title, 'target')

    def test_304(self):
        # first retrieve the url
        u = 'http://localhost:8097/tests/http/http_status_304.xml'
        f = feedparser.parse(u)
        self.assertEqual(f.status, 200)
        self.assertEqual(f.entries[0].title, 'title 304')
        # extract the etag and last-modified headers
        e = [v for k, v in f.headers.items() if k.lower() == 'etag'][0]
        ms = f.updated
        mt = f.updated_parsed
        md = datetime.datetime(*mt[0:7])
        self.assertTrue(isinstance(mt, time.struct_time))
        self.assertTrue(isinstance(md, datetime.datetime))
        # test that sending back the etag results in a 304
        f = feedparser.parse(u, etag=e)
        self.assertEqual(f.status, 304)
        # test that sending back last-modified (string) results in a 304
        f = feedparser.parse(u, modified=ms)
        self.assertEqual(f.status, 304)
        # test that sending back last-modified (9-tuple) results in a 304
        f = feedparser.parse(u, modified=mt)
        self.assertEqual(f.status, 304)
        # test that sending back last-modified (datetime) results in a 304
        f = feedparser.parse(u, modified=md)
        self.assertEqual(f.status, 304)

    def test_404(self):
        f = feedparser.parse('http://localhost:8097/tests/http/http_status_404.xml')
        self.assertEqual(f.status, 404)

    def test_redirect_to_304(self):
        # ensure that an http redirect to an http 304 doesn't
        # trigger a bozo_exception
        u = 'http://localhost:8097/tests/http/http_redirect_to_304.xml'
        f = feedparser.parse(u)
        self.assertTrue(f.bozo == 0)
        self.assertTrue(f.status == 302)


class TestDateParsers(unittest.TestCase):
    """Test the various date parsers; most of the test cases are constructed dynamically based on the contents of the `date_tests` dict, below"""

    def test_None(self):
        self.assertTrue(feedparser.datetimes._parse_date(None) is None)

    def _check_date(self, fn, dt, expected_value):
        try:
            parsed_value = fn(dt)
        except (OverflowError, ValueError):
            parsed_value = None
        self.assertEqual(parsed_value, expected_value)

    def test_year_10000_date(self):
        # On some systems this date string will trigger an OverflowError.
        # On Jython and x64 systems, however, it's interpreted just fine.
        try:
            date = _parse_date_rfc822('Sun, 31 Dec 9999 23:59:59 -9999')
        except OverflowError:
            date = None
        self.assertTrue(date in (None, (10000, 1, 5, 4, 38, 59, 2, 5, 0)))


date_tests = {
    _parse_date_greek: (
        ('', None),  # empty string
        ('\u039a\u03c5\u03c1, 11 \u0399\u03bf\u03cd\u03bb 2004 12:00:00 EST', (2004, 7, 11, 17, 0, 0, 6, 193, 0)),
    ),
    _parse_date_hungarian: (
        ('', None),  # empty string
        ('2004-j\u00falius-13T9:15-05:00', (2004, 7, 13, 14, 15, 0, 1, 195, 0)),
    ),
    _parse_date_iso8601: (
        ('', None),  # empty string
        ('-0312', (2003, 12, 1, 0, 0, 0, 0, 335, 0)),  # 2-digit year/month only variant
        ('031231', (2003, 12, 31, 0, 0, 0, 2, 365, 0)),  # 2-digit year/month/day only, no hyphens
        ('03-12-31', (2003, 12, 31, 0, 0, 0, 2, 365, 0)),  # 2-digit year/month/day only
        ('-03-12', (2003, 12, 1, 0, 0, 0, 0, 335, 0)),  # 2-digit year/month only
        ('03335', (2003, 12, 1, 0, 0, 0, 0, 335, 0)),  # 2-digit year/ordinal, no hyphens
        ('2003-12-31T10:14:55.1234Z', (2003, 12, 31, 10, 14, 55, 2, 365, 0)),  # fractional seconds
        # Special case for Google's extra zero in the month
        ('2003-012-31T10:14:55+00:00', (2003, 12, 31, 10, 14, 55, 2, 365, 0)),
    ),
    _parse_date_nate: (
        ('', None),  # empty string
        ('2004-05-25 \uc624\ud6c4 11:23:17', (2004, 5, 25, 14, 23, 17, 1, 146, 0)),
    ),
    _parse_date_onblog: (
        ('', None),  # empty string
        ('2004\ub144 05\uc6d4 28\uc77c  01:31:15', (2004, 5, 27, 16, 31, 15, 3, 148, 0)),
    ),
    _parse_date_perforce: (
        ('', None),  # empty string
        ('Fri, 2006/09/15 08:19:53 EDT', (2006, 9, 15, 12, 19, 53, 4, 258, 0)),
    ),
    _parse_date_rfc822: (
        ('', None),  # empty string
        ('Thu, 30 Apr 2015 08:57:00 MET', (2015, 4, 30, 7, 57, 0, 3, 120, 0)),
        ('Thu, 30 Apr 2015 08:57:00 MEST', (2015, 4, 30, 6, 57, 0, 3, 120, 0)),
        ('Thu, 01 Jan 0100 00:00:01 +0100', (99, 12, 31, 23, 0, 1, 3, 365, 0)),  # ancient date
        ('Thu, 01 Jan 04 19:48:21 GMT', (2004, 1, 1, 19, 48, 21, 3, 1, 0)),  # 2-digit year
        ('Thu, 01 Jan 2004 19:48:21 GMT', (2004, 1, 1, 19, 48, 21, 3, 1, 0)),  # 4-digit year
        ('Thu,  5 Apr 2012 10:00:00 GMT', (2012, 4, 5, 10, 0, 0, 3, 96, 0)),  # 1-digit day
        ('Wed, 19 Aug 2009 18:28:00 Etc/GMT', (2009, 8, 19, 18, 28, 0, 2, 231, 0)),  # etc/gmt timezone
        ('Wed, 19 Feb 2012 22:40:00 GMT-01:01', (2012, 2, 19, 23, 41, 0, 6, 50, 0)),  # gmt+hh:mm timezone
        ('Wed, 19 Feb 2012 22:40:00 -01:01', (2012, 2, 19, 23, 41, 0, 6, 50, 0)),  # +hh:mm timezone
        ('Mon, 13 Feb, 2012 06:28:00 UTC', (2012, 2, 13, 6, 28, 0, 0, 44, 0)),  # extraneous comma
        ('Thu, 01 Jan 2004 00:00 GMT', (2004, 1, 1, 0, 0, 0, 3, 1, 0)),  # no seconds
        ('Thu, 01 Jan 2004', (2004, 1, 1, 0, 0, 0, 3, 1, 0)),  # no time
        # Additional tests to handle Disney's long month names and invalid timezones
        ('Mon, 26 January 2004 16:31:00 AT', (2004, 1, 26, 20, 31, 0, 0, 26, 0)),
        ('Mon, 26 January 2004 16:31:00 ET', (2004, 1, 26, 21, 31, 0, 0, 26, 0)),
        ('Mon, 26 January 2004 16:31:00 CT', (2004, 1, 26, 22, 31, 0, 0, 26, 0)),
        ('Mon, 26 January 2004 16:31:00 MT', (2004, 1, 26, 23, 31, 0, 0, 26, 0)),
        ('Mon, 26 January 2004 16:31:00 PT', (2004, 1, 27, 0, 31, 0, 1, 27, 0)),
        # Swapped month and day
        ('Thu Aug 30 2012 17:26:16 +0200', (2012, 8, 30, 15, 26, 16, 3, 243, 0)),
        ('Sun, 16 Dec 2012 1:2:3:4 GMT', None),  # invalid time
        ('Sun, 16 zzz 2012 11:47:32 GMT', None),  # invalid month
        ('Sun, Dec x 2012 11:47:32 GMT', None),  # invalid day (swapped day/month)
        ('Sun, 16 Dec zz 11:47:32 GMT', None),  # invalid year
        ('Sun, 16 Dec 2012 11:47:32 +zz:00', None),  # invalid timezone hour
        ('Sun, 16 Dec 2012 11:47:32 +00:zz', None),  # invalid timezone minute
        ('Sun, 99 Jun 2009 12:00:00 GMT', None),  # out-of-range day
    ),
    _parse_date_asctime: (
        ('Sun Jan  4 16:29:06 2004', (2004, 1, 4, 16, 29, 6, 6, 4, 0)),
        ('Sun Jul 15 01:16:00 +0000 2012', (2012, 7, 15, 1, 16, 0, 6, 197, 0)),
    ),
    _parse_date_w3dtf: (
        ('', None),  # empty string
        ('2003-12-31T10:14:55Z', (2003, 12, 31, 10, 14, 55, 2, 365, 0)),  # UTC
        ('2003-12-31T10:14:55-08:00', (2003, 12, 31, 18, 14, 55, 2, 365, 0)),  # San Francisco timezone
        ('2003-12-31T18:14:55+08:00', (2003, 12, 31, 10, 14, 55, 2, 365, 0)),  # Tokyo timezone
        ('2007-04-23T23:25:47.538+10:00', (2007, 4, 23, 13, 25, 47, 0, 113, 0)),  # fractional seconds
        ('2003-12-31', (2003, 12, 31, 0, 0, 0, 2, 365, 0)),  # year/month/day only
        ('2003-12', (2003, 12, 1, 0, 0, 0, 0, 335, 0)),  # year/month only
        ('2003', (2003, 1, 1, 0, 0, 0, 2, 1, 0)),  # year only
        # Special cases for rollovers in leap years
        ('2004-02-28T18:14:55-08:00', (2004, 2, 29, 2, 14, 55, 6, 60, 0)),  # feb 28 in leap year
        ('2003-02-28T18:14:55-08:00', (2003, 3, 1, 2, 14, 55, 5, 60, 0)),  # feb 28 in non-leap year
        ('2000-02-28T18:14:55-08:00', (2000, 2, 29, 2, 14, 55, 1, 60, 0)),  # feb 28 in leap year on century divisible by 400
        # Out-of-range times
        ('9999-12-31T23:59:59-99:99', None),  # Date is out-of-range
        ('2003-12-31T25:14:55Z', None),  # invalid (25 hours)
        ('2003-12-31T10:61:55Z', None),  # invalid (61 minutes)
        ('2003-12-31T10:14:61Z', None),  # invalid (61 seconds)
        # Invalid formats
        ('22013', None),  # Year is too long
        ('013', None),  # Year is too short
        ('2013-01-27-01', None),  # Date has to many parts
        ('2013-01-28T11:30:00-06:00Textra', None),  # Too many 't's
        # Non-integer values
        ('2013-xx-27', None),  # Date
        ('2013-01-28T09:xx:00Z', None),  # Time
        ('2013-01-28T09:00:00+00:xx', None),  # Timezone
        # MSSQL-style dates
        ('2004-07-08 23:56:58 -00:20', (2004, 7, 9, 0, 16, 58, 4, 191, 0)),  # with timezone
        ('2004-07-08 23:56:58', (2004, 7, 8, 23, 56, 58, 3, 190, 0)),  # without timezone
        ('2004-07-08 23:56:58.0', (2004, 7, 8, 23, 56, 58, 3, 190, 0)),  # with fractional second
    )
}


def make_date_test(f, s, t):
    return lambda self: self._check_date(f, s, t)


for func, items in date_tests.items():
    for i, (dtstring, dttuple) in enumerate(items):
        uniqfunc = make_date_test(func, dtstring, dttuple)
        setattr(TestDateParsers, 'test_%s_%02i' % (func.__name__, i), uniqfunc)


def make_html_guess_test(text, expect, doc):
    def fn(self):
        value = bool(feedparser.mixin._FeedParserMixin.looks_like_html(text))
        self.assertEqual(value, expect)

    fn.__doc__ = doc
    return fn


class TestHTMLGuessing(unittest.TestCase):
    """Exercise the HTML sniffing code"""

    test_text_1 = make_html_guess_test('plain text', False, 'plain text')
    test_text_2 = make_html_guess_test('2 < 3', False, 'plain text with angle bracket')
    test_html_1 = make_html_guess_test('<a href="">a</a>', True, 'anchor tag')
    test_html_2 = make_html_guess_test('<i>i</i>', True, 'italics tag')
    test_html_3 = make_html_guess_test('<b>b</b>', True, 'bold tag')
    test_html_4 = make_html_guess_test('<code>', False, 'allowed tag, no end tag')
    test_html_5 = make_html_guess_test('<rss> .. </rss>', False, 'disallowed tag')
    test_entity_1 = make_html_guess_test('AT&T', False, 'corporation name')
    test_entity_2 = make_html_guess_test('&copy;', True, 'named entity reference')
    test_entity_3 = make_html_guess_test('&#169;', True, 'numeric entity reference')
    test_entity_4 = make_html_guess_test('&#xA9;', True, 'hex numeric entity reference')


# ---------- additional api unit tests, not backed by files

class TestBuildRequest(unittest.TestCase):
    """Test that HTTP request objects are created as expected"""
    def test_extra_headers(self):
        """You can pass in extra headers and they go into the request object."""

        request = feedparser.http._build_urllib2_request(
          'http://example.com/feed',
          'agent-name',
          None, None, None, None, None,
          {'Cache-Control': 'max-age=0'})
        # nb, urllib2 folds the case of the headers
        self.assertEqual(
          request.get_header('Cache-control'), 'max-age=0')


class TestLxmlBug(unittest.TestCase):
    def test_lxml_etree_bug(self):
        try:
            import lxml.etree
        except ImportError:
            pass
        else:
            doc = b"<feed>&illformed_charref</feed>"
            # Importing lxml.etree currently causes libxml2 to
            # throw SAXException instead of SAXParseException.
            feedparser.parse(io.BytesIO(doc))
        self.assertTrue(True)


class TestParseFlags(unittest.TestCase):
    feed_xml = b"""
        <rss version="2.0">
            <channel>
                <item>
                    <body><script>alert("boo!")</script></body>
                </item>
                <item>
                    <body><a href="/boo.html">boo</a></body>
                </item>
            </channel>
        </rss>
    """

    def test_sanitize_html_default(self):
        d = feedparser.parse(io.BytesIO(self.feed_xml))
        self.assertEqual(u'', d.entries[0].content[0].value)

    def test_sanitize_html_on(self):
        d = feedparser.parse(io.BytesIO(self.feed_xml), sanitize_html=True)
        self.assertEqual(u'', d.entries[0].content[0].value)

    def test_sanitize_html_off(self):
        d = feedparser.parse(io.BytesIO(self.feed_xml), sanitize_html=False)
        self.assertEqual(u'<script>alert("boo!")</script>', d.entries[0].content[0].value)

    def test_resolve_relative_uris_default(self):
        d = feedparser.parse(io.BytesIO(self.feed_xml),
                             response_headers={'content-location': 'http://example.com/feed'})
        self.assertEqual(u'<a href="http://example.com/boo.html">boo</a>', d.entries[1].content[0].value)

    def test_resolve_relative_uris_on(self):
        d = feedparser.parse(io.BytesIO(self.feed_xml),
                             response_headers={'content-location': 'http://example.com/feed'},
                             resolve_relative_uris=True)
        self.assertEqual(u'<a href="http://example.com/boo.html">boo</a>', d.entries[1].content[0].value)

    def test_resolve_relative_uris_off(self):
        d = feedparser.parse(io.BytesIO(self.feed_xml),
                             response_headers={'content-location': 'http://example.com/feed.xml'},
                             resolve_relative_uris=False)
        self.assertEqual(u'<a href="/boo.html">boo</a>', d.entries[1].content[0].value)


class TestSanitizer(unittest.TestCase):
    def test_style_attr_is_enabled(self):
        html = """<p style="margin: 15em;">example</p>"""
        result = feedparser.sanitizer._sanitize_html(html, None, 'text/html')
        self.assertEqual(result, html)

    def test_style_attr_can_be_disabled(self):
        html = """<p style="margin: 15em;">example</p>"""
        expected = """<p>example</p>"""
        original_attrs = feedparser.sanitizer._HTMLSanitizer.acceptable_attributes
        feedparser.sanitizer._HTMLSanitizer.acceptable_attributes = set()
        result = feedparser.sanitizer._sanitize_html(html, None, 'text/html')
        feedparser.sanitizer._HTMLSanitizer.acceptable_attributes = original_attrs
        self.assertEqual(result, expected)


# ---------- parse test files and create test methods ----------

def convert_to_utf8(data):
    """Identify data's encoding using its byte order mark and convert it to its utf-8 equivalent"""

    if data[:4] == b'\x4c\x6f\xa7\x94':
        return data.decode('cp037').encode('utf-8')
    elif data[:4] == b'\x00\x00\xfe\xff':
        return data.decode('utf-32be').encode('utf-8')
    elif data[:4] == b'\xff\xfe\x00\x00':
        return data.decode('utf-32le').encode('utf-8')
    elif data[:4] == b'\x00\x00\x00\x3c':
        return data.decode('utf-32be').encode('utf-8')
    elif data[:4] == b'\x3c\x00\x00\x00':
        return data.decode('utf-32le').encode('utf-8')
    elif data[:4] == b'\x00\x3c\x00\x3f':
        return data.decode('utf-16be').encode('utf-8')
    elif data[:4] == b'\x3c\x00\x3f\x00':
        return data.decode('utf-16le').encode('utf-8')
    elif (data[:2] == b'\xfe\xff') and (data[2:4] != b'\x00\x00'):
        return data[2:].decode('utf-16be').encode('utf-8')
    elif (data[:2] == b'\xff\xfe') and (data[2:4] != b'\x00\x00'):
        return data[2:].decode('utf-16le').encode('utf-8')
    elif data[:3] == b'\xef\xbb\xbf':
        return data[3:]
    # no byte order mark was found
    return data


skip_re = re.compile(br"SkipUnless:\s*(.*?)\n")
desc_re = re.compile(br"Description:\s*(.*?)\s*Expect:\s*(.*)\s*-->")


def get_description(xmlfile, data):
    """Extract test data

    Each test case is an XML file which contains not only a test feed
    but also the description of the test and the condition that we
    would expect the parser to create when it parses the feed.  Example:
    <!--
    Description: feed title
    Expect:      feed['title'] == 'Example feed'
    -->
    """
    skip_results = skip_re.search(data)
    if skip_results:
        skip_unless = skip_results.group(1).strip()
    else:
        skip_unless = '1'
    search_results = desc_re.search(data)
    if not search_results:
        raise RuntimeError("can't parse %s" % xmlfile)
    description, eval_string = [s.strip() for s in list(search_results.groups())]
    description = xmlfile + ": " + description.decode('utf8')
    return description, eval_string, skip_unless


def build_test_case(xmlfile, description, eval_string):
    def fn(self, xmlfile=xmlfile, eval_string=eval_string):
        self.fail_unless_eval(xmlfile, eval_string)

    fn.__doc__ = description
    return fn


def runtests():
    """Read the files in the tests/ directory, dynamically add tests to the TestCases above, spawn the HTTP server, and run the test suite"""

    allfiles = glob.glob(os.path.join('.', 'tests', '**', '**', '*.xml'))
    wellformedfiles = glob.glob(os.path.join('.', 'tests', 'wellformed', '**', '*.xml'))
    illformedfiles = glob.glob(os.path.join('.', 'tests', 'illformed', '*.xml'))
    encodingfiles = glob.glob(os.path.join('.', 'tests', 'encoding', '*.xml'))
    entitiesfiles = glob.glob(os.path.join('.', 'tests', 'entities', '*.xml'))

    # there are several compression test cases that must be accounted for
    # as well as a number of http status tests that redirect to a target
    # and a few `_open_resource`-related tests
    httpcount = 6 + 16 + 2
    httpcount += len([f for f in allfiles if 'http' in f])
    httpcount += len([f for f in wellformedfiles if 'http' in f])
    httpcount += len([f for f in illformedfiles if 'http' in f])
    httpcount += len([f for f in encodingfiles if 'http' in f])

    for c, xmlfile in enumerate(allfiles + encodingfiles + illformedfiles + entitiesfiles):
        add_to = TestCase
        if xmlfile in encodingfiles:
            add_to = TestEncodings
        elif xmlfile in entitiesfiles:
            add_to = (TestStrictParser, TestLooseParser)
        elif xmlfile in wellformedfiles:
            add_to = (TestStrictParser, TestLooseParser)
        f = open(xmlfile, 'rb')
        data = f.read()
        f.close()
        if 'encoding' in xmlfile:
            data = convert_to_utf8(data)
            if data is None:
                # convert_to_utf8 found a byte order mark for utf_32
                # but it's not supported in this installation of Python
                if 'http' in xmlfile:
                    httpcount -= 1 + (xmlfile in wellformedfiles)
                continue
        description, eval_string, skip_unless = get_description(xmlfile, data)
        test_name = 'test_%06d' % c
        is_http = 'http' in xmlfile
        try:
            if not eval(skip_unless):
                raise NotImplementedError
        except (ImportError, LookupError, NotImplementedError, AttributeError):
            if is_http:
                httpcount -= 1 + (xmlfile in wellformedfiles)
            continue
        if is_http:
            xmlfile = 'http://%s:%s/%s' % (_HOST, _PORT, posixpath.normpath(xmlfile.replace('\\', '/')))
        test_func = build_test_case(xmlfile, description, eval_string)
        if isinstance(add_to, tuple):
            setattr(add_to[0], test_name, test_func)
            setattr(add_to[1], test_name, test_func)
        else:
            setattr(add_to, test_name, test_func)
    if httpcount:
        httpd = FeedParserTestServer(httpcount)
        httpd.daemon = True
        httpd.start()
        httpd.ready.wait()
    testsuite = unittest.TestSuite()
    testloader = unittest.TestLoader()
    testsuite.addTest(testloader.loadTestsFromTestCase(TestCase))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestStrictParser))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestLooseParser))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestEncodings))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestDateParsers))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestHTMLGuessing))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestHTTPStatus))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestCompression))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestConvertToIdn))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestOpenResource))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestFeedParserDict))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestMakeSafeAbsoluteURI))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestEverythingIsUnicode))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestTemporaryFallbackBehavior))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestLxmlBug))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestParseFlags))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestBuildRequest))
    testsuite.addTest(testloader.loadTestsFromTestCase(TestSanitizer))
    testresults = unittest.TextTestRunner(verbosity=1).run(testsuite)

    # Return 0 if successful, 1 if there was a failure
    sys.exit(not testresults.wasSuccessful())


if __name__ == "__main__":
    runtests()