Haketilo

# SPDX-License-Identifier: CC0-1.0

"""

Haketilo unit tests - URL patterns

"""

# This file is part of Haketilo

#

# Copyright (C) 2021, Wojtek Kosior

#

# This program is free software: you can redistribute it and/or modify

# it under the terms of the CC0 1.0 Universal License as published by

# the Creative Commons Corporation.

#

# This program is distributed in the hope that it will be useful,

# but WITHOUT ANY WARRANTY; without even the implied warranty of

# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

# CC0 1.0 Universal License for more details.

import pytest

from ..script_loader import load_script

@pytest.mark.get_page('https://gotmyowndoma.in')

def test_regexes(execute_in_page):

    """

    patterns.js contains regexes used for URL parsing.

    Verify they work properly.

    """

    execute_in_page(load_script('common/patterns.js'))

    valid_url = 'https://example.com/a/b?ver=1.2.3#heading2'

    valid_url_rest = 'example.com/a/b?ver=1.2.3#heading2'

    # Test matching of URL protocol.

    match = execute_in_page('returnval(proto_regex.exec(arguments[0]));',

                            valid_url)

    assert match

    assert match[1] == 'https'

    assert match[2] == valid_url_rest

    match = execute_in_page('returnval(proto_regex.exec(arguments[0]));',

                            '://bad-url.missing/protocol')

    assert match is None

    # Test matching of http(s) URLs.

    match = execute_in_page('returnval(http_regex.exec(arguments[0]));',

                            valid_url_rest)

    assert match

    assert match[1] == 'example.com'

    assert match[2] == '/a/b'

    assert match[3] == '?ver=1.2.3'

    match = execute_in_page('returnval(http_regex.exec(arguments[0]));',

                            'another.example.com')

    assert match

    assert match[1] == 'another.example.com'

    assert match[2] == ''

    assert match[3] == ''

    match = execute_in_page('returnval(http_regex.exec(arguments[0]));',

                            '/bad/http/example')

    assert match == None

    # Test matching of file URLs.

    match = execute_in_page('returnval(file_regex.exec(arguments[0]));',

                            '/good/file/example')

    assert match

    assert match[1] == '/good/file/example'

    # Test matching of ftp URLs.

    match = execute_in_page('returnval(ftp_regex.exec(arguments[0]));',

                            'example.com/a/b#heading2')

    assert match

    assert match[1] is None

    assert match[2] == 'example.com'

    assert match[3] == '/a/b'

    match = execute_in_page('returnval(ftp_regex.exec(arguments[0]));',

                            'some_user@localhost')

    assert match

    assert match[1] == 'some_user@'

    assert match[2] == 'localhost'

    assert match[3] == ''

    match = execute_in_page('returnval(ftp_regex.exec(arguments[0]));',

                            '@bad.url/')

    assert match is None

@pytest.mark.get_page('https://gotmyowndoma.in')

def test_deconstruct_url(execute_in_page):

    """

    patterns.js contains deconstruct_url() function that handles URL parsing.

    Verify it works properly.

    """

    execute_in_page(load_script('common/patterns.js'))

    deco = execute_in_page('returnval(deconstruct_url(arguments[0]));',

                           'https://eXaMpLe.com/a/b?ver=1.2.3#heading2')

    assert deco

    assert deco['trailing_slash'] == False

    assert deco['proto']          == 'https'

    assert deco['domain']         == ['example', 'com']

    assert deco['path']           == ['a', 'b']

    deco = execute_in_page('returnval(deconstruct_url(arguments[0]));',

                           'http://**.example.com/')

    assert deco

    assert deco['trailing_slash'] == True

    assert deco['proto']          == 'http'

    assert deco['domain']         == ['**', 'example', 'com']

    assert deco['path']           == []

    deco = execute_in_page('returnval(deconstruct_url(arguments[0]));',

                           'ftp://user@ftp.example.com/all///passwords.txt/')

    assert deco

    assert deco['trailing_slash'] == True

    assert deco['proto']          == 'ftp'

    assert deco['domain']         == ['ftp', 'example', 'com']

    assert deco['path']           == ['all', 'passwords.txt']

    deco = execute_in_page('returnval(deconstruct_url(arguments[0]));',

                           'ftp://mirror.edu.pl.eu.org')

    assert deco

    assert deco['trailing_slash'] == False

    assert deco['proto']          == 'ftp'

    assert deco['domain']         == ['mirror', 'edu', 'pl', 'eu', 'org']

    assert deco['path']           == []

    deco = execute_in_page('returnval(deconstruct_url(arguments[0]));',

                           'file:///mnt/parabola_chroot///etc/passwd')

    assert deco

    assert deco['trailing_slash'] == False

    assert deco['proto']          == 'file'

    assert deco['path']           == ['mnt', 'parabola_chroot', 'etc', 'passwd']

    assert 'domain' not in deco

    for bad_url in [

            '://bad-url.missing/protocol',

            'http:/example.com/a/b',

            'unknown://example.com/a/b',

            'idontfancypineapple',

            'ftp://@example.org/',

            'https:///some/path/',

            'file://non-absolute/path'

    ]:

        with pytest.raises(Exception, match=r'Error in injected script'):

            deco = execute_in_page('returnval(deconstruct_url(arguments[0]));',

                                   bad_url)

    # at some point we might also consider testing url deconstruction with

    # length limits...