]>
Commit | Line | Data |
---|---|---|
cc52de43 | 1 | #!/usr/bin/env python3 |
54007a45 | 2 | |
83fda3c0 PH |
3 | # Allow direct execution |
4 | import os | |
5 | import sys | |
227bf1a3 | 6 | |
7 | import pytest | |
f8271158 | 8 | |
83fda3c0 PH |
9 | sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__)))) |
10 | ||
227bf1a3 | 11 | import functools |
08916a49 | 12 | import gzip |
227bf1a3 | 13 | import http.client |
08916a49 | 14 | import http.cookiejar |
54007a45 | 15 | import http.server |
227bf1a3 | 16 | import inspect |
08916a49 | 17 | import io |
18 | import pathlib | |
227bf1a3 | 19 | import random |
f8271158 | 20 | import ssl |
08916a49 | 21 | import tempfile |
f8271158 | 22 | import threading |
227bf1a3 | 23 | import time |
08916a49 | 24 | import urllib.error |
ac668111 | 25 | import urllib.request |
227bf1a3 | 26 | import warnings |
daafbf49 | 27 | import zlib |
227bf1a3 | 28 | from email.message import Message |
29 | from http.cookiejar import CookieJar | |
f8271158 | 30 | |
227bf1a3 | 31 | from test.helper import FakeYDL, http_server_port |
daafbf49 | 32 | from yt_dlp.dependencies import brotli |
227bf1a3 | 33 | from yt_dlp.networking import ( |
34 | HEADRequest, | |
35 | PUTRequest, | |
36 | Request, | |
37 | RequestDirector, | |
38 | RequestHandler, | |
39 | Response, | |
40 | ) | |
41 | from yt_dlp.networking._urllib import UrllibRH | |
42 | from yt_dlp.networking.common import _REQUEST_HANDLERS | |
43 | from yt_dlp.networking.exceptions import ( | |
44 | CertificateVerifyError, | |
45 | HTTPError, | |
46 | IncompleteRead, | |
47 | NoSupportingHandlers, | |
48 | RequestError, | |
49 | SSLError, | |
50 | TransportError, | |
51 | UnsupportedRequest, | |
52 | ) | |
53 | from yt_dlp.utils._utils import _YDLLogger as FakeLogger | |
54 | from yt_dlp.utils.networking import HTTPHeaderDict | |
83fda3c0 PH |
55 | |
56 | TEST_DIR = os.path.dirname(os.path.abspath(__file__)) | |
57 | ||
03d8d4df | 58 | |
227bf1a3 | 59 | def _build_proxy_handler(name): |
60 | class HTTPTestRequestHandler(http.server.BaseHTTPRequestHandler): | |
61 | proxy_name = name | |
62 | ||
63 | def log_message(self, format, *args): | |
64 | pass | |
65 | ||
66 | def do_GET(self): | |
67 | self.send_response(200) | |
68 | self.send_header('Content-Type', 'text/plain; charset=utf-8') | |
69 | self.end_headers() | |
70 | self.wfile.write('{self.proxy_name}: {self.path}'.format(self=self).encode()) | |
71 | return HTTPTestRequestHandler | |
72 | ||
73 | ||
ac668111 | 74 | class HTTPTestRequestHandler(http.server.BaseHTTPRequestHandler): |
08916a49 | 75 | protocol_version = 'HTTP/1.1' |
76 | ||
83fda3c0 PH |
77 | def log_message(self, format, *args): |
78 | pass | |
79 | ||
08916a49 | 80 | def _headers(self): |
227bf1a3 | 81 | payload = str(self.headers).encode() |
08916a49 | 82 | self.send_response(200) |
83 | self.send_header('Content-Type', 'application/json') | |
84 | self.send_header('Content-Length', str(len(payload))) | |
85 | self.end_headers() | |
86 | self.wfile.write(payload) | |
87 | ||
88 | def _redirect(self): | |
89 | self.send_response(int(self.path[len('/redirect_'):])) | |
90 | self.send_header('Location', '/method') | |
91 | self.send_header('Content-Length', '0') | |
92 | self.end_headers() | |
93 | ||
94 | def _method(self, method, payload=None): | |
95 | self.send_response(200) | |
96 | self.send_header('Content-Length', str(len(payload or ''))) | |
97 | self.send_header('Method', method) | |
98 | self.end_headers() | |
99 | if payload: | |
100 | self.wfile.write(payload) | |
101 | ||
102 | def _status(self, status): | |
103 | payload = f'<html>{status} NOT FOUND</html>'.encode() | |
104 | self.send_response(int(status)) | |
105 | self.send_header('Content-Type', 'text/html; charset=utf-8') | |
106 | self.send_header('Content-Length', str(len(payload))) | |
107 | self.end_headers() | |
108 | self.wfile.write(payload) | |
109 | ||
110 | def _read_data(self): | |
111 | if 'Content-Length' in self.headers: | |
112 | return self.rfile.read(int(self.headers['Content-Length'])) | |
113 | ||
114 | def do_POST(self): | |
227bf1a3 | 115 | data = self._read_data() + str(self.headers).encode() |
08916a49 | 116 | if self.path.startswith('/redirect_'): |
117 | self._redirect() | |
118 | elif self.path.startswith('/method'): | |
119 | self._method('POST', data) | |
120 | elif self.path.startswith('/headers'): | |
121 | self._headers() | |
122 | else: | |
123 | self._status(404) | |
124 | ||
125 | def do_HEAD(self): | |
126 | if self.path.startswith('/redirect_'): | |
127 | self._redirect() | |
128 | elif self.path.startswith('/method'): | |
129 | self._method('HEAD') | |
130 | else: | |
131 | self._status(404) | |
132 | ||
133 | def do_PUT(self): | |
227bf1a3 | 134 | data = self._read_data() + str(self.headers).encode() |
08916a49 | 135 | if self.path.startswith('/redirect_'): |
136 | self._redirect() | |
137 | elif self.path.startswith('/method'): | |
138 | self._method('PUT', data) | |
139 | else: | |
140 | self._status(404) | |
141 | ||
83fda3c0 PH |
142 | def do_GET(self): |
143 | if self.path == '/video.html': | |
08916a49 | 144 | payload = b'<html><video src="/vid.mp4" /></html>' |
83fda3c0 PH |
145 | self.send_response(200) |
146 | self.send_header('Content-Type', 'text/html; charset=utf-8') | |
227bf1a3 | 147 | self.send_header('Content-Length', str(len(payload))) |
83fda3c0 | 148 | self.end_headers() |
08916a49 | 149 | self.wfile.write(payload) |
83fda3c0 | 150 | elif self.path == '/vid.mp4': |
08916a49 | 151 | payload = b'\x00\x00\x00\x00\x20\x66\x74[video]' |
83fda3c0 PH |
152 | self.send_response(200) |
153 | self.send_header('Content-Type', 'video/mp4') | |
08916a49 | 154 | self.send_header('Content-Length', str(len(payload))) |
83fda3c0 | 155 | self.end_headers() |
08916a49 | 156 | self.wfile.write(payload) |
8c32e5dc | 157 | elif self.path == '/%E4%B8%AD%E6%96%87.html': |
08916a49 | 158 | payload = b'<html><video src="/vid.mp4" /></html>' |
8c32e5dc YCH |
159 | self.send_response(200) |
160 | self.send_header('Content-Type', 'text/html; charset=utf-8') | |
08916a49 | 161 | self.send_header('Content-Length', str(len(payload))) |
162 | self.end_headers() | |
163 | self.wfile.write(payload) | |
164 | elif self.path == '/%c7%9f': | |
165 | payload = b'<html><video src="/vid.mp4" /></html>' | |
166 | self.send_response(200) | |
167 | self.send_header('Content-Type', 'text/html; charset=utf-8') | |
168 | self.send_header('Content-Length', str(len(payload))) | |
169 | self.end_headers() | |
170 | self.wfile.write(payload) | |
227bf1a3 | 171 | elif self.path.startswith('/redirect_loop'): |
172 | self.send_response(301) | |
173 | self.send_header('Location', self.path) | |
174 | self.send_header('Content-Length', '0') | |
175 | self.end_headers() | |
08916a49 | 176 | elif self.path.startswith('/redirect_'): |
177 | self._redirect() | |
178 | elif self.path.startswith('/method'): | |
227bf1a3 | 179 | self._method('GET', str(self.headers).encode()) |
08916a49 | 180 | elif self.path.startswith('/headers'): |
181 | self._headers() | |
f8b4bcc0 | 182 | elif self.path.startswith('/308-to-headers'): |
183 | self.send_response(308) | |
184 | self.send_header('Location', '/headers') | |
185 | self.send_header('Content-Length', '0') | |
186 | self.end_headers() | |
08916a49 | 187 | elif self.path == '/trailing_garbage': |
188 | payload = b'<html><video src="/vid.mp4" /></html>' | |
189 | self.send_response(200) | |
190 | self.send_header('Content-Type', 'text/html; charset=utf-8') | |
191 | self.send_header('Content-Encoding', 'gzip') | |
192 | buf = io.BytesIO() | |
193 | with gzip.GzipFile(fileobj=buf, mode='wb') as f: | |
194 | f.write(payload) | |
195 | compressed = buf.getvalue() + b'trailing garbage' | |
196 | self.send_header('Content-Length', str(len(compressed))) | |
197 | self.end_headers() | |
198 | self.wfile.write(compressed) | |
199 | elif self.path == '/302-non-ascii-redirect': | |
200 | new_url = f'http://127.0.0.1:{http_server_port(self.server)}/中文.html' | |
201 | self.send_response(301) | |
202 | self.send_header('Location', new_url) | |
203 | self.send_header('Content-Length', '0') | |
8c32e5dc | 204 | self.end_headers() |
daafbf49 | 205 | elif self.path == '/content-encoding': |
206 | encodings = self.headers.get('ytdl-encoding', '') | |
207 | payload = b'<html><video src="/vid.mp4" /></html>' | |
208 | for encoding in filter(None, (e.strip() for e in encodings.split(','))): | |
209 | if encoding == 'br' and brotli: | |
210 | payload = brotli.compress(payload) | |
211 | elif encoding == 'gzip': | |
212 | buf = io.BytesIO() | |
213 | with gzip.GzipFile(fileobj=buf, mode='wb') as f: | |
214 | f.write(payload) | |
215 | payload = buf.getvalue() | |
216 | elif encoding == 'deflate': | |
217 | payload = zlib.compress(payload) | |
218 | elif encoding == 'unsupported': | |
219 | payload = b'raw' | |
220 | break | |
221 | else: | |
222 | self._status(415) | |
223 | return | |
224 | self.send_response(200) | |
225 | self.send_header('Content-Encoding', encodings) | |
226 | self.send_header('Content-Length', str(len(payload))) | |
227 | self.end_headers() | |
228 | self.wfile.write(payload) | |
227bf1a3 | 229 | elif self.path.startswith('/gen_'): |
230 | payload = b'<html></html>' | |
231 | self.send_response(int(self.path[len('/gen_'):])) | |
232 | self.send_header('Content-Type', 'text/html; charset=utf-8') | |
233 | self.send_header('Content-Length', str(len(payload))) | |
234 | self.end_headers() | |
235 | self.wfile.write(payload) | |
236 | elif self.path.startswith('/incompleteread'): | |
237 | payload = b'<html></html>' | |
238 | self.send_response(200) | |
239 | self.send_header('Content-Type', 'text/html; charset=utf-8') | |
240 | self.send_header('Content-Length', '234234') | |
241 | self.end_headers() | |
242 | self.wfile.write(payload) | |
243 | self.finish() | |
244 | elif self.path.startswith('/timeout_'): | |
245 | time.sleep(int(self.path[len('/timeout_'):])) | |
246 | self._headers() | |
247 | elif self.path == '/source_address': | |
248 | payload = str(self.client_address[0]).encode() | |
249 | self.send_response(200) | |
250 | self.send_header('Content-Type', 'text/html; charset=utf-8') | |
251 | self.send_header('Content-Length', str(len(payload))) | |
252 | self.end_headers() | |
253 | self.wfile.write(payload) | |
254 | self.finish() | |
83fda3c0 | 255 | else: |
08916a49 | 256 | self._status(404) |
257 | ||
258 | def send_header(self, keyword, value): | |
259 | """ | |
260 | Forcibly allow HTTP server to send non percent-encoded non-ASCII characters in headers. | |
261 | This is against what is defined in RFC 3986, however we need to test we support this | |
262 | since some sites incorrectly do this. | |
263 | """ | |
264 | if keyword.lower() == 'connection': | |
265 | return super().send_header(keyword, value) | |
266 | ||
267 | if not hasattr(self, '_headers_buffer'): | |
268 | self._headers_buffer = [] | |
269 | ||
270 | self._headers_buffer.append(f'{keyword}: {value}\r\n'.encode()) | |
83fda3c0 PH |
271 | |
272 | ||
227bf1a3 | 273 | def validate_and_send(rh, req): |
274 | rh.validate(req) | |
275 | return rh.send(req) | |
83fda3c0 | 276 | |
83fda3c0 | 277 | |
227bf1a3 | 278 | class TestRequestHandlerBase: |
279 | @classmethod | |
280 | def setup_class(cls): | |
281 | cls.http_httpd = http.server.ThreadingHTTPServer( | |
f19eae42 | 282 | ('127.0.0.1', 0), HTTPTestRequestHandler) |
227bf1a3 | 283 | cls.http_port = http_server_port(cls.http_httpd) |
284 | cls.http_server_thread = threading.Thread(target=cls.http_httpd.serve_forever) | |
08916a49 | 285 | # FIXME: we should probably stop the http server thread after each test |
286 | # See: https://github.com/yt-dlp/yt-dlp/pull/7094#discussion_r1199746041 | |
227bf1a3 | 287 | cls.http_server_thread.daemon = True |
288 | cls.http_server_thread.start() | |
08916a49 | 289 | |
290 | # HTTPS server | |
83fda3c0 | 291 | certfn = os.path.join(TEST_DIR, 'testcert.pem') |
227bf1a3 | 292 | cls.https_httpd = http.server.ThreadingHTTPServer( |
f19eae42 | 293 | ('127.0.0.1', 0), HTTPTestRequestHandler) |
b6dc37fe | 294 | sslctx = ssl.SSLContext(ssl.PROTOCOL_TLS_SERVER) |
295 | sslctx.load_cert_chain(certfn, None) | |
227bf1a3 | 296 | cls.https_httpd.socket = sslctx.wrap_socket(cls.https_httpd.socket, server_side=True) |
297 | cls.https_port = http_server_port(cls.https_httpd) | |
298 | cls.https_server_thread = threading.Thread(target=cls.https_httpd.serve_forever) | |
299 | cls.https_server_thread.daemon = True | |
300 | cls.https_server_thread.start() | |
301 | ||
302 | ||
303 | @pytest.fixture | |
304 | def handler(request): | |
305 | RH_KEY = request.param | |
306 | if inspect.isclass(RH_KEY) and issubclass(RH_KEY, RequestHandler): | |
307 | handler = RH_KEY | |
308 | elif RH_KEY in _REQUEST_HANDLERS: | |
309 | handler = _REQUEST_HANDLERS[RH_KEY] | |
310 | else: | |
311 | pytest.skip(f'{RH_KEY} request handler is not available') | |
312 | ||
313 | return functools.partial(handler, logger=FakeLogger) | |
314 | ||
315 | ||
316 | class TestHTTPRequestHandler(TestRequestHandlerBase): | |
317 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
318 | def test_verify_cert(self, handler): | |
319 | with handler() as rh: | |
320 | with pytest.raises(CertificateVerifyError): | |
321 | validate_and_send(rh, Request(f'https://127.0.0.1:{self.https_port}/headers')) | |
322 | ||
323 | with handler(verify=False) as rh: | |
324 | r = validate_and_send(rh, Request(f'https://127.0.0.1:{self.https_port}/headers')) | |
325 | assert r.status == 200 | |
08916a49 | 326 | r.close() |
327 | ||
227bf1a3 | 328 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
329 | def test_ssl_error(self, handler): | |
330 | # HTTPS server with too old TLS version | |
331 | # XXX: is there a better way to test this than to create a new server? | |
332 | https_httpd = http.server.ThreadingHTTPServer( | |
333 | ('127.0.0.1', 0), HTTPTestRequestHandler) | |
334 | sslctx = ssl.SSLContext(ssl.PROTOCOL_TLS_SERVER) | |
335 | https_httpd.socket = sslctx.wrap_socket(https_httpd.socket, server_side=True) | |
336 | https_port = http_server_port(https_httpd) | |
337 | https_server_thread = threading.Thread(target=https_httpd.serve_forever) | |
338 | https_server_thread.daemon = True | |
339 | https_server_thread.start() | |
340 | ||
341 | with handler(verify=False) as rh: | |
342 | with pytest.raises(SSLError, match='sslv3 alert handshake failure') as exc_info: | |
343 | validate_and_send(rh, Request(f'https://127.0.0.1:{https_port}/headers')) | |
344 | assert not issubclass(exc_info.type, CertificateVerifyError) | |
345 | ||
346 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
347 | def test_percent_encode(self, handler): | |
348 | with handler() as rh: | |
08916a49 | 349 | # Unicode characters should be encoded with uppercase percent-encoding |
227bf1a3 | 350 | res = validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/中文.html')) |
351 | assert res.status == 200 | |
08916a49 | 352 | res.close() |
353 | # don't normalize existing percent encodings | |
227bf1a3 | 354 | res = validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/%c7%9f')) |
355 | assert res.status == 200 | |
08916a49 | 356 | res.close() |
357 | ||
227bf1a3 | 358 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
359 | def test_unicode_path_redirection(self, handler): | |
360 | with handler() as rh: | |
361 | r = validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/302-non-ascii-redirect')) | |
362 | assert r.url == f'http://127.0.0.1:{self.http_port}/%E4%B8%AD%E6%96%87.html' | |
08916a49 | 363 | r.close() |
364 | ||
227bf1a3 | 365 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
366 | def test_raise_http_error(self, handler): | |
367 | with handler() as rh: | |
368 | for bad_status in (400, 500, 599, 302): | |
369 | with pytest.raises(HTTPError): | |
370 | validate_and_send(rh, Request('http://127.0.0.1:%d/gen_%d' % (self.http_port, bad_status))) | |
371 | ||
372 | # Should not raise an error | |
373 | validate_and_send(rh, Request('http://127.0.0.1:%d/gen_200' % self.http_port)).close() | |
374 | ||
375 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
376 | def test_response_url(self, handler): | |
377 | with handler() as rh: | |
378 | # Response url should be that of the last url in redirect chain | |
379 | res = validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/redirect_301')) | |
380 | assert res.url == f'http://127.0.0.1:{self.http_port}/method' | |
381 | res.close() | |
382 | res2 = validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/gen_200')) | |
383 | assert res2.url == f'http://127.0.0.1:{self.http_port}/gen_200' | |
384 | res2.close() | |
385 | ||
386 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
387 | def test_redirect(self, handler): | |
388 | with handler() as rh: | |
389 | def do_req(redirect_status, method, assert_no_content=False): | |
08916a49 | 390 | data = b'testdata' if method in ('POST', 'PUT') else None |
227bf1a3 | 391 | res = validate_and_send( |
392 | rh, Request(f'http://127.0.0.1:{self.http_port}/redirect_{redirect_status}', method=method, data=data)) | |
393 | ||
394 | headers = b'' | |
395 | data_sent = b'' | |
396 | if data is not None: | |
397 | data_sent += res.read(len(data)) | |
398 | if data_sent != data: | |
399 | headers += data_sent | |
400 | data_sent = b'' | |
401 | ||
402 | headers += res.read() | |
403 | ||
404 | if assert_no_content or data is None: | |
405 | assert b'Content-Type' not in headers | |
406 | assert b'Content-Length' not in headers | |
407 | else: | |
408 | assert b'Content-Type' in headers | |
409 | assert b'Content-Length' in headers | |
410 | ||
411 | return data_sent.decode(), res.headers.get('method', '') | |
08916a49 | 412 | |
413 | # A 303 must either use GET or HEAD for subsequent request | |
227bf1a3 | 414 | assert do_req(303, 'POST', True) == ('', 'GET') |
415 | assert do_req(303, 'HEAD') == ('', 'HEAD') | |
08916a49 | 416 | |
227bf1a3 | 417 | assert do_req(303, 'PUT', True) == ('', 'GET') |
08916a49 | 418 | |
419 | # 301 and 302 turn POST only into a GET | |
227bf1a3 | 420 | assert do_req(301, 'POST', True) == ('', 'GET') |
421 | assert do_req(301, 'HEAD') == ('', 'HEAD') | |
422 | assert do_req(302, 'POST', True) == ('', 'GET') | |
423 | assert do_req(302, 'HEAD') == ('', 'HEAD') | |
08916a49 | 424 | |
227bf1a3 | 425 | assert do_req(301, 'PUT') == ('testdata', 'PUT') |
426 | assert do_req(302, 'PUT') == ('testdata', 'PUT') | |
08916a49 | 427 | |
428 | # 307 and 308 should not change method | |
429 | for m in ('POST', 'PUT'): | |
227bf1a3 | 430 | assert do_req(307, m) == ('testdata', m) |
431 | assert do_req(308, m) == ('testdata', m) | |
08916a49 | 432 | |
227bf1a3 | 433 | assert do_req(307, 'HEAD') == ('', 'HEAD') |
434 | assert do_req(308, 'HEAD') == ('', 'HEAD') | |
08916a49 | 435 | |
436 | # These should not redirect and instead raise an HTTPError | |
437 | for code in (300, 304, 305, 306): | |
227bf1a3 | 438 | with pytest.raises(HTTPError): |
08916a49 | 439 | do_req(code, 'GET') |
440 | ||
227bf1a3 | 441 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
442 | def test_request_cookie_header(self, handler): | |
f8b4bcc0 | 443 | # We should accept a Cookie header being passed as in normal headers and handle it appropriately. |
227bf1a3 | 444 | with handler() as rh: |
f8b4bcc0 | 445 | # Specified Cookie header should be used |
227bf1a3 | 446 | res = validate_and_send( |
447 | rh, Request( | |
08916a49 | 448 | f'http://127.0.0.1:{self.http_port}/headers', |
227bf1a3 | 449 | headers={'Cookie': 'test=test'})).read().decode() |
450 | assert 'Cookie: test=test' in res | |
08916a49 | 451 | |
227bf1a3 | 452 | # Specified Cookie header should be removed on any redirect |
453 | res = validate_and_send( | |
454 | rh, Request( | |
455 | f'http://127.0.0.1:{self.http_port}/308-to-headers', | |
456 | headers={'Cookie': 'test=test'})).read().decode() | |
457 | assert 'Cookie: test=test' not in res | |
458 | ||
459 | # Specified Cookie header should override global cookiejar for that request | |
460 | cookiejar = http.cookiejar.CookieJar() | |
461 | cookiejar.set_cookie(http.cookiejar.Cookie( | |
462 | version=0, name='test', value='ytdlp', port=None, port_specified=False, | |
463 | domain='127.0.0.1', domain_specified=True, domain_initial_dot=False, path='/', | |
464 | path_specified=True, secure=False, expires=None, discard=False, comment=None, | |
465 | comment_url=None, rest={})) | |
466 | ||
467 | with handler(cookiejar=cookiejar) as rh: | |
468 | data = validate_and_send( | |
469 | rh, Request(f'http://127.0.0.1:{self.http_port}/headers', headers={'cookie': 'test=test'})).read() | |
470 | assert b'Cookie: test=ytdlp' not in data | |
471 | assert b'Cookie: test=test' in data | |
472 | ||
473 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
474 | def test_redirect_loop(self, handler): | |
475 | with handler() as rh: | |
476 | with pytest.raises(HTTPError, match='redirect loop'): | |
477 | validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/redirect_loop')) | |
478 | ||
479 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
480 | def test_incompleteread(self, handler): | |
481 | with handler(timeout=2) as rh: | |
482 | with pytest.raises(IncompleteRead): | |
483 | validate_and_send(rh, Request('http://127.0.0.1:%d/incompleteread' % self.http_port)).read() | |
484 | ||
485 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
486 | def test_cookies(self, handler): | |
487 | cookiejar = http.cookiejar.CookieJar() | |
488 | cookiejar.set_cookie(http.cookiejar.Cookie( | |
489 | 0, 'test', 'ytdlp', None, False, '127.0.0.1', True, | |
490 | False, '/headers', True, False, None, False, None, None, {})) | |
491 | ||
492 | with handler(cookiejar=cookiejar) as rh: | |
493 | data = validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/headers')).read() | |
494 | assert b'Cookie: test=ytdlp' in data | |
495 | ||
496 | # Per request | |
497 | with handler() as rh: | |
498 | data = validate_and_send( | |
499 | rh, Request(f'http://127.0.0.1:{self.http_port}/headers', extensions={'cookiejar': cookiejar})).read() | |
500 | assert b'Cookie: test=ytdlp' in data | |
501 | ||
502 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
503 | def test_headers(self, handler): | |
504 | ||
505 | with handler(headers=HTTPHeaderDict({'test1': 'test', 'test2': 'test2'})) as rh: | |
506 | # Global Headers | |
507 | data = validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/headers')).read() | |
508 | assert b'Test1: test' in data | |
509 | ||
510 | # Per request headers, merged with global | |
511 | data = validate_and_send(rh, Request( | |
512 | f'http://127.0.0.1:{self.http_port}/headers', headers={'test2': 'changed', 'test3': 'test3'})).read() | |
513 | assert b'Test1: test' in data | |
514 | assert b'Test2: changed' in data | |
515 | assert b'Test2: test2' not in data | |
516 | assert b'Test3: test3' in data | |
517 | ||
518 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
519 | def test_timeout(self, handler): | |
520 | with handler() as rh: | |
521 | # Default timeout is 20 seconds, so this should go through | |
522 | validate_and_send( | |
523 | rh, Request(f'http://127.0.0.1:{self.http_port}/timeout_3')) | |
524 | ||
525 | with handler(timeout=0.5) as rh: | |
526 | with pytest.raises(TransportError): | |
527 | validate_and_send( | |
528 | rh, Request(f'http://127.0.0.1:{self.http_port}/timeout_1')) | |
529 | ||
530 | # Per request timeout, should override handler timeout | |
531 | validate_and_send( | |
532 | rh, Request(f'http://127.0.0.1:{self.http_port}/timeout_1', extensions={'timeout': 4})) | |
533 | ||
534 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
535 | def test_source_address(self, handler): | |
536 | source_address = f'127.0.0.{random.randint(5, 255)}' | |
537 | with handler(source_address=source_address) as rh: | |
538 | data = validate_and_send( | |
539 | rh, Request(f'http://127.0.0.1:{self.http_port}/source_address')).read().decode() | |
540 | assert source_address == data | |
541 | ||
542 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
543 | def test_gzip_trailing_garbage(self, handler): | |
544 | with handler() as rh: | |
545 | data = validate_and_send(rh, Request(f'http://localhost:{self.http_port}/trailing_garbage')).read().decode() | |
546 | assert data == '<html><video src="/vid.mp4" /></html>' | |
547 | ||
548 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
549 | @pytest.mark.skipif(not brotli, reason='brotli support is not installed') | |
550 | def test_brotli(self, handler): | |
551 | with handler() as rh: | |
552 | res = validate_and_send( | |
553 | rh, Request( | |
daafbf49 | 554 | f'http://127.0.0.1:{self.http_port}/content-encoding', |
555 | headers={'ytdl-encoding': 'br'})) | |
227bf1a3 | 556 | assert res.headers.get('Content-Encoding') == 'br' |
557 | assert res.read() == b'<html><video src="/vid.mp4" /></html>' | |
558 | ||
559 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
560 | def test_deflate(self, handler): | |
561 | with handler() as rh: | |
562 | res = validate_and_send( | |
563 | rh, Request( | |
daafbf49 | 564 | f'http://127.0.0.1:{self.http_port}/content-encoding', |
565 | headers={'ytdl-encoding': 'deflate'})) | |
227bf1a3 | 566 | assert res.headers.get('Content-Encoding') == 'deflate' |
567 | assert res.read() == b'<html><video src="/vid.mp4" /></html>' | |
568 | ||
569 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
570 | def test_gzip(self, handler): | |
571 | with handler() as rh: | |
572 | res = validate_and_send( | |
573 | rh, Request( | |
daafbf49 | 574 | f'http://127.0.0.1:{self.http_port}/content-encoding', |
575 | headers={'ytdl-encoding': 'gzip'})) | |
227bf1a3 | 576 | assert res.headers.get('Content-Encoding') == 'gzip' |
577 | assert res.read() == b'<html><video src="/vid.mp4" /></html>' | |
daafbf49 | 578 | |
227bf1a3 | 579 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
580 | def test_multiple_encodings(self, handler): | |
581 | with handler() as rh: | |
daafbf49 | 582 | for pair in ('gzip,deflate', 'deflate, gzip', 'gzip, gzip', 'deflate, deflate'): |
227bf1a3 | 583 | res = validate_and_send( |
584 | rh, Request( | |
daafbf49 | 585 | f'http://127.0.0.1:{self.http_port}/content-encoding', |
586 | headers={'ytdl-encoding': pair})) | |
227bf1a3 | 587 | assert res.headers.get('Content-Encoding') == pair |
588 | assert res.read() == b'<html><video src="/vid.mp4" /></html>' | |
589 | ||
590 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
591 | def test_unsupported_encoding(self, handler): | |
592 | with handler() as rh: | |
593 | res = validate_and_send( | |
594 | rh, Request( | |
daafbf49 | 595 | f'http://127.0.0.1:{self.http_port}/content-encoding', |
596 | headers={'ytdl-encoding': 'unsupported'})) | |
227bf1a3 | 597 | assert res.headers.get('Content-Encoding') == 'unsupported' |
598 | assert res.read() == b'raw' | |
599 | ||
600 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
601 | def test_read(self, handler): | |
602 | with handler() as rh: | |
603 | res = validate_and_send( | |
604 | rh, Request(f'http://127.0.0.1:{self.http_port}/headers')) | |
605 | assert res.readable() | |
606 | assert res.read(1) == b'H' | |
607 | assert res.read(3) == b'ost' | |
608 | ||
609 | ||
610 | class TestHTTPProxy(TestRequestHandlerBase): | |
611 | @classmethod | |
612 | def setup_class(cls): | |
613 | super().setup_class() | |
614 | # HTTP Proxy server | |
615 | cls.proxy = http.server.ThreadingHTTPServer( | |
616 | ('127.0.0.1', 0), _build_proxy_handler('normal')) | |
617 | cls.proxy_port = http_server_port(cls.proxy) | |
618 | cls.proxy_thread = threading.Thread(target=cls.proxy.serve_forever) | |
619 | cls.proxy_thread.daemon = True | |
620 | cls.proxy_thread.start() | |
daafbf49 | 621 | |
227bf1a3 | 622 | # Geo proxy server |
623 | cls.geo_proxy = http.server.ThreadingHTTPServer( | |
624 | ('127.0.0.1', 0), _build_proxy_handler('geo')) | |
625 | cls.geo_port = http_server_port(cls.geo_proxy) | |
626 | cls.geo_proxy_thread = threading.Thread(target=cls.geo_proxy.serve_forever) | |
627 | cls.geo_proxy_thread.daemon = True | |
628 | cls.geo_proxy_thread.start() | |
629 | ||
630 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
631 | def test_http_proxy(self, handler): | |
632 | http_proxy = f'http://127.0.0.1:{self.proxy_port}' | |
633 | geo_proxy = f'http://127.0.0.1:{self.geo_port}' | |
634 | ||
635 | # Test global http proxy | |
636 | # Test per request http proxy | |
637 | # Test per request http proxy disables proxy | |
638 | url = 'http://foo.com/bar' | |
01218f91 | 639 | |
227bf1a3 | 640 | # Global HTTP proxy |
641 | with handler(proxies={'http': http_proxy}) as rh: | |
642 | res = validate_and_send(rh, Request(url)).read().decode() | |
643 | assert res == f'normal: {url}' | |
644 | ||
645 | # Per request proxy overrides global | |
646 | res = validate_and_send(rh, Request(url, proxies={'http': geo_proxy})).read().decode() | |
647 | assert res == f'geo: {url}' | |
648 | ||
649 | # and setting to None disables all proxies for that request | |
650 | real_url = f'http://127.0.0.1:{self.http_port}/headers' | |
651 | res = validate_and_send( | |
652 | rh, Request(real_url, proxies={'http': None})).read().decode() | |
653 | assert res != f'normal: {real_url}' | |
654 | assert 'Accept' in res | |
655 | ||
656 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
657 | def test_noproxy(self, handler): | |
658 | with handler(proxies={'proxy': f'http://127.0.0.1:{self.proxy_port}'}) as rh: | |
659 | # NO_PROXY | |
660 | for no_proxy in (f'127.0.0.1:{self.http_port}', '127.0.0.1', 'localhost'): | |
661 | nop_response = validate_and_send( | |
662 | rh, Request(f'http://127.0.0.1:{self.http_port}/headers', proxies={'no': no_proxy})).read().decode( | |
663 | 'utf-8') | |
664 | assert 'Accept' in nop_response | |
665 | ||
666 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
667 | def test_allproxy(self, handler): | |
668 | url = 'http://foo.com/bar' | |
669 | with handler() as rh: | |
670 | response = validate_and_send(rh, Request(url, proxies={'all': f'http://127.0.0.1:{self.proxy_port}'})).read().decode( | |
671 | 'utf-8') | |
672 | assert response == f'normal: {url}' | |
673 | ||
674 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
675 | def test_http_proxy_with_idn(self, handler): | |
676 | with handler(proxies={ | |
677 | 'http': f'http://127.0.0.1:{self.proxy_port}', | |
678 | }) as rh: | |
679 | url = 'http://中文.tw/' | |
680 | response = rh.send(Request(url)).read().decode() | |
681 | # b'xn--fiq228c' is '中文'.encode('idna') | |
682 | assert response == 'normal: http://xn--fiq228c.tw/' | |
683 | ||
684 | ||
685 | class TestClientCertificate: | |
686 | ||
687 | @classmethod | |
688 | def setup_class(cls): | |
bb58c9ed | 689 | certfn = os.path.join(TEST_DIR, 'testcert.pem') |
227bf1a3 | 690 | cls.certdir = os.path.join(TEST_DIR, 'testdata', 'certificate') |
691 | cacertfn = os.path.join(cls.certdir, 'ca.crt') | |
692 | cls.httpd = http.server.ThreadingHTTPServer(('127.0.0.1', 0), HTTPTestRequestHandler) | |
bb58c9ed | 693 | sslctx = ssl.SSLContext(ssl.PROTOCOL_TLS_SERVER) |
694 | sslctx.verify_mode = ssl.CERT_REQUIRED | |
695 | sslctx.load_verify_locations(cafile=cacertfn) | |
696 | sslctx.load_cert_chain(certfn, None) | |
227bf1a3 | 697 | cls.httpd.socket = sslctx.wrap_socket(cls.httpd.socket, server_side=True) |
698 | cls.port = http_server_port(cls.httpd) | |
699 | cls.server_thread = threading.Thread(target=cls.httpd.serve_forever) | |
700 | cls.server_thread.daemon = True | |
701 | cls.server_thread.start() | |
702 | ||
703 | def _run_test(self, handler, **handler_kwargs): | |
704 | with handler( | |
bb58c9ed | 705 | # Disable client-side validation of unacceptable self-signed testcert.pem |
706 | # The test is of a check on the server side, so unaffected | |
227bf1a3 | 707 | verify=False, |
708 | **handler_kwargs, | |
709 | ) as rh: | |
710 | validate_and_send(rh, Request(f'https://127.0.0.1:{self.port}/video.html')).read().decode() | |
711 | ||
712 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
713 | def test_certificate_combined_nopass(self, handler): | |
714 | self._run_test(handler, client_cert={ | |
715 | 'client_certificate': os.path.join(self.certdir, 'clientwithkey.crt'), | |
bb58c9ed | 716 | }) |
bb58c9ed | 717 | |
227bf1a3 | 718 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
719 | def test_certificate_nocombined_nopass(self, handler): | |
720 | self._run_test(handler, client_cert={ | |
721 | 'client_certificate': os.path.join(self.certdir, 'client.crt'), | |
722 | 'client_certificate_key': os.path.join(self.certdir, 'client.key'), | |
723 | }) | |
bb58c9ed | 724 | |
227bf1a3 | 725 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
726 | def test_certificate_combined_pass(self, handler): | |
727 | self._run_test(handler, client_cert={ | |
728 | 'client_certificate': os.path.join(self.certdir, 'clientwithencryptedkey.crt'), | |
729 | 'client_certificate_password': 'foobar', | |
730 | }) | |
bb58c9ed | 731 | |
227bf1a3 | 732 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
733 | def test_certificate_nocombined_pass(self, handler): | |
734 | self._run_test(handler, client_cert={ | |
735 | 'client_certificate': os.path.join(self.certdir, 'client.crt'), | |
736 | 'client_certificate_key': os.path.join(self.certdir, 'clientencrypted.key'), | |
737 | 'client_certificate_password': 'foobar', | |
738 | }) | |
bb58c9ed | 739 | |
bb58c9ed | 740 | |
227bf1a3 | 741 | class TestUrllibRequestHandler(TestRequestHandlerBase): |
742 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
743 | def test_file_urls(self, handler): | |
744 | # See https://github.com/ytdl-org/youtube-dl/issues/8227 | |
745 | tf = tempfile.NamedTemporaryFile(delete=False) | |
746 | tf.write(b'foobar') | |
747 | tf.close() | |
748 | req = Request(pathlib.Path(tf.name).as_uri()) | |
749 | with handler() as rh: | |
750 | with pytest.raises(UnsupportedRequest): | |
751 | rh.validate(req) | |
752 | ||
753 | # Test that urllib never loaded FileHandler | |
754 | with pytest.raises(TransportError): | |
755 | rh.send(req) | |
756 | ||
757 | with handler(enable_file_urls=True) as rh: | |
758 | res = validate_and_send(rh, req) | |
759 | assert res.read() == b'foobar' | |
760 | res.close() | |
bb58c9ed | 761 | |
227bf1a3 | 762 | os.unlink(tf.name) |
01218f91 | 763 | |
227bf1a3 | 764 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) |
765 | def test_http_error_returns_content(self, handler): | |
766 | # urllib HTTPError will try close the underlying response if reference to the HTTPError object is lost | |
767 | def get_response(): | |
768 | with handler() as rh: | |
769 | # headers url | |
770 | try: | |
771 | validate_and_send(rh, Request(f'http://127.0.0.1:{self.http_port}/gen_404')) | |
772 | except HTTPError as e: | |
773 | return e.response | |
774 | ||
775 | assert get_response().read() == b'<html></html>' | |
776 | ||
777 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
778 | def test_verify_cert_error_text(self, handler): | |
779 | # Check the output of the error message | |
780 | with handler() as rh: | |
781 | with pytest.raises( | |
782 | CertificateVerifyError, | |
783 | match=r'\[SSL: CERTIFICATE_VERIFY_FAILED\] certificate verify failed: self.signed certificate' | |
784 | ): | |
785 | validate_and_send(rh, Request(f'https://127.0.0.1:{self.https_port}/headers')) | |
786 | ||
787 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
788 | def test_httplib_validation_errors(self, handler): | |
789 | with handler() as rh: | |
790 | ||
791 | # https://github.com/python/cpython/blob/987b712b4aeeece336eed24fcc87a950a756c3e2/Lib/http/client.py#L1256 | |
792 | with pytest.raises(RequestError, match='method can\'t contain control characters') as exc_info: | |
793 | validate_and_send(rh, Request('http://127.0.0.1', method='GET\n')) | |
794 | assert not isinstance(exc_info.value, TransportError) | |
795 | ||
796 | # https://github.com/python/cpython/blob/987b712b4aeeece336eed24fcc87a950a756c3e2/Lib/http/client.py#L1265 | |
797 | with pytest.raises(RequestError, match='URL can\'t contain control characters') as exc_info: | |
798 | validate_and_send(rh, Request('http://127.0.0. 1', method='GET\n')) | |
799 | assert not isinstance(exc_info.value, TransportError) | |
800 | ||
801 | # https://github.com/python/cpython/blob/987b712b4aeeece336eed24fcc87a950a756c3e2/Lib/http/client.py#L1288C31-L1288C50 | |
802 | with pytest.raises(RequestError, match='Invalid header name') as exc_info: | |
803 | validate_and_send(rh, Request('http://127.0.0.1', headers={'foo\n': 'bar'})) | |
804 | assert not isinstance(exc_info.value, TransportError) | |
805 | ||
806 | ||
807 | def run_validation(handler, fail, req, **handler_kwargs): | |
808 | with handler(**handler_kwargs) as rh: | |
809 | if fail: | |
810 | with pytest.raises(UnsupportedRequest): | |
811 | rh.validate(req) | |
812 | else: | |
813 | rh.validate(req) | |
814 | ||
815 | ||
816 | class TestRequestHandlerValidation: | |
817 | ||
818 | class ValidationRH(RequestHandler): | |
819 | def _send(self, request): | |
820 | raise RequestError('test') | |
821 | ||
822 | class NoCheckRH(ValidationRH): | |
823 | _SUPPORTED_FEATURES = None | |
824 | _SUPPORTED_PROXY_SCHEMES = None | |
825 | _SUPPORTED_URL_SCHEMES = None | |
826 | ||
827 | class HTTPSupportedRH(ValidationRH): | |
828 | _SUPPORTED_URL_SCHEMES = ('http',) | |
829 | ||
830 | URL_SCHEME_TESTS = [ | |
831 | # scheme, expected to fail, handler kwargs | |
832 | ('Urllib', [ | |
833 | ('http', False, {}), | |
834 | ('https', False, {}), | |
835 | ('data', False, {}), | |
836 | ('ftp', False, {}), | |
837 | ('file', True, {}), | |
838 | ('file', False, {'enable_file_urls': True}), | |
839 | ]), | |
840 | (NoCheckRH, [('http', False, {})]), | |
841 | (ValidationRH, [('http', True, {})]) | |
842 | ] | |
843 | ||
844 | PROXY_SCHEME_TESTS = [ | |
845 | # scheme, expected to fail | |
846 | ('Urllib', [ | |
847 | ('http', False), | |
848 | ('https', True), | |
849 | ('socks4', False), | |
850 | ('socks4a', False), | |
851 | ('socks5', False), | |
852 | ('socks5h', False), | |
853 | ('socks', True), | |
854 | ]), | |
855 | (NoCheckRH, [('http', False)]), | |
856 | (HTTPSupportedRH, [('http', True)]), | |
857 | ] | |
858 | ||
859 | PROXY_KEY_TESTS = [ | |
860 | # key, expected to fail | |
861 | ('Urllib', [ | |
862 | ('all', False), | |
863 | ('unrelated', False), | |
864 | ]), | |
865 | (NoCheckRH, [('all', False)]), | |
866 | (HTTPSupportedRH, [('all', True)]), | |
867 | (HTTPSupportedRH, [('no', True)]), | |
868 | ] | |
869 | ||
870 | @pytest.mark.parametrize('handler,scheme,fail,handler_kwargs', [ | |
871 | (handler_tests[0], scheme, fail, handler_kwargs) | |
872 | for handler_tests in URL_SCHEME_TESTS | |
873 | for scheme, fail, handler_kwargs in handler_tests[1] | |
874 | ||
875 | ], indirect=['handler']) | |
876 | def test_url_scheme(self, handler, scheme, fail, handler_kwargs): | |
877 | run_validation(handler, fail, Request(f'{scheme}://'), **(handler_kwargs or {})) | |
878 | ||
879 | @pytest.mark.parametrize('handler,fail', [('Urllib', False)], indirect=['handler']) | |
880 | def test_no_proxy(self, handler, fail): | |
881 | run_validation(handler, fail, Request('http://', proxies={'no': '127.0.0.1,github.com'})) | |
882 | run_validation(handler, fail, Request('http://'), proxies={'no': '127.0.0.1,github.com'}) | |
883 | ||
884 | @pytest.mark.parametrize('handler,proxy_key,fail', [ | |
885 | (handler_tests[0], proxy_key, fail) | |
886 | for handler_tests in PROXY_KEY_TESTS | |
887 | for proxy_key, fail in handler_tests[1] | |
888 | ], indirect=['handler']) | |
889 | def test_proxy_key(self, handler, proxy_key, fail): | |
890 | run_validation(handler, fail, Request('http://', proxies={proxy_key: 'http://example.com'})) | |
891 | run_validation(handler, fail, Request('http://'), proxies={proxy_key: 'http://example.com'}) | |
892 | ||
893 | @pytest.mark.parametrize('handler,scheme,fail', [ | |
894 | (handler_tests[0], scheme, fail) | |
895 | for handler_tests in PROXY_SCHEME_TESTS | |
896 | for scheme, fail in handler_tests[1] | |
897 | ], indirect=['handler']) | |
898 | def test_proxy_scheme(self, handler, scheme, fail): | |
899 | run_validation(handler, fail, Request('http://', proxies={'http': f'{scheme}://example.com'})) | |
900 | run_validation(handler, fail, Request('http://'), proxies={'http': f'{scheme}://example.com'}) | |
901 | ||
902 | @pytest.mark.parametrize('handler', ['Urllib', HTTPSupportedRH], indirect=True) | |
903 | def test_empty_proxy(self, handler): | |
904 | run_validation(handler, False, Request('http://', proxies={'http': None})) | |
905 | run_validation(handler, False, Request('http://'), proxies={'http': None}) | |
906 | ||
907 | @pytest.mark.parametrize('proxy_url', ['//example.com', 'example.com', '127.0.0.1']) | |
908 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
909 | def test_missing_proxy_scheme(self, handler, proxy_url): | |
910 | run_validation(handler, True, Request('http://', proxies={'http': 'example.com'})) | |
911 | ||
912 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
913 | def test_cookiejar_extension(self, handler): | |
914 | run_validation(handler, True, Request('http://', extensions={'cookiejar': 'notacookiejar'})) | |
915 | ||
916 | @pytest.mark.parametrize('handler', ['Urllib'], indirect=True) | |
917 | def test_timeout_extension(self, handler): | |
918 | run_validation(handler, True, Request('http://', extensions={'timeout': 'notavalidtimeout'})) | |
919 | ||
920 | def test_invalid_request_type(self): | |
921 | rh = self.ValidationRH(logger=FakeLogger()) | |
922 | for method in (rh.validate, rh.send): | |
923 | with pytest.raises(TypeError, match='Expected an instance of Request'): | |
924 | method('not a request') | |
925 | ||
926 | ||
927 | class FakeResponse(Response): | |
928 | def __init__(self, request): | |
929 | # XXX: we could make request part of standard response interface | |
930 | self.request = request | |
931 | super().__init__(fp=io.BytesIO(b''), headers={}, url=request.url) | |
932 | ||
933 | ||
934 | class FakeRH(RequestHandler): | |
935 | ||
936 | def _validate(self, request): | |
937 | return | |
938 | ||
939 | def _send(self, request: Request): | |
940 | if request.url.startswith('ssl://'): | |
941 | raise SSLError(request.url[len('ssl://'):]) | |
942 | return FakeResponse(request) | |
943 | ||
944 | ||
945 | class FakeRHYDL(FakeYDL): | |
946 | def __init__(self, *args, **kwargs): | |
947 | super().__init__(*args, **kwargs) | |
948 | self._request_director = self.build_request_director([FakeRH]) | |
949 | ||
950 | ||
951 | class TestRequestDirector: | |
952 | ||
953 | def test_handler_operations(self): | |
954 | director = RequestDirector(logger=FakeLogger()) | |
955 | handler = FakeRH(logger=FakeLogger()) | |
956 | director.add_handler(handler) | |
957 | assert director.handlers.get(FakeRH.RH_KEY) is handler | |
958 | ||
959 | # Handler should overwrite | |
960 | handler2 = FakeRH(logger=FakeLogger()) | |
961 | director.add_handler(handler2) | |
962 | assert director.handlers.get(FakeRH.RH_KEY) is not handler | |
963 | assert director.handlers.get(FakeRH.RH_KEY) is handler2 | |
964 | assert len(director.handlers) == 1 | |
965 | ||
966 | class AnotherFakeRH(FakeRH): | |
01218f91 | 967 | pass |
227bf1a3 | 968 | director.add_handler(AnotherFakeRH(logger=FakeLogger())) |
969 | assert len(director.handlers) == 2 | |
970 | assert director.handlers.get(AnotherFakeRH.RH_KEY).RH_KEY == AnotherFakeRH.RH_KEY | |
01218f91 | 971 | |
227bf1a3 | 972 | director.handlers.pop(FakeRH.RH_KEY, None) |
973 | assert director.handlers.get(FakeRH.RH_KEY) is None | |
974 | assert len(director.handlers) == 1 | |
01218f91 | 975 | |
227bf1a3 | 976 | # RequestErrors should passthrough |
977 | with pytest.raises(SSLError): | |
978 | director.send(Request('ssl://something')) | |
01218f91 | 979 | |
227bf1a3 | 980 | def test_send(self): |
981 | director = RequestDirector(logger=FakeLogger()) | |
982 | with pytest.raises(RequestError): | |
983 | director.send(Request('any://')) | |
984 | director.add_handler(FakeRH(logger=FakeLogger())) | |
985 | assert isinstance(director.send(Request('http://')), FakeResponse) | |
01218f91 | 986 | |
227bf1a3 | 987 | def test_unsupported_handlers(self): |
988 | director = RequestDirector(logger=FakeLogger()) | |
989 | director.add_handler(FakeRH(logger=FakeLogger())) | |
01218f91 | 990 | |
227bf1a3 | 991 | class SupportedRH(RequestHandler): |
992 | _SUPPORTED_URL_SCHEMES = ['http'] | |
01218f91 | 993 | |
227bf1a3 | 994 | def _send(self, request: Request): |
995 | return Response(fp=io.BytesIO(b'supported'), headers={}, url=request.url) | |
efbed08d | 996 | |
227bf1a3 | 997 | # This handler should by default take preference over FakeRH |
998 | director.add_handler(SupportedRH(logger=FakeLogger())) | |
999 | assert director.send(Request('http://')).read() == b'supported' | |
1000 | assert director.send(Request('any://')).read() == b'' | |
582be358 | 1001 | |
227bf1a3 | 1002 | director.handlers.pop(FakeRH.RH_KEY) |
1003 | with pytest.raises(NoSupportingHandlers): | |
1004 | director.send(Request('any://')) | |
1005 | ||
1006 | def test_unexpected_error(self): | |
1007 | director = RequestDirector(logger=FakeLogger()) | |
1008 | ||
1009 | class UnexpectedRH(FakeRH): | |
1010 | def _send(self, request: Request): | |
1011 | raise TypeError('something') | |
1012 | ||
1013 | director.add_handler(UnexpectedRH(logger=FakeLogger)) | |
1014 | with pytest.raises(NoSupportingHandlers, match=r'1 unexpected error'): | |
1015 | director.send(Request('any://')) | |
1016 | ||
1017 | director.handlers.clear() | |
1018 | assert len(director.handlers) == 0 | |
1019 | ||
1020 | # Should not be fatal | |
1021 | director.add_handler(FakeRH(logger=FakeLogger())) | |
1022 | director.add_handler(UnexpectedRH(logger=FakeLogger)) | |
1023 | assert director.send(Request('any://')) | |
1024 | ||
1025 | ||
1026 | # XXX: do we want to move this to test_YoutubeDL.py? | |
1027 | class TestYoutubeDLNetworking: | |
1028 | ||
1029 | @staticmethod | |
1030 | def build_handler(ydl, handler: RequestHandler = FakeRH): | |
1031 | return ydl.build_request_director([handler]).handlers.get(handler.RH_KEY) | |
1032 | ||
1033 | def test_compat_opener(self): | |
08916a49 | 1034 | with FakeYDL() as ydl: |
227bf1a3 | 1035 | with warnings.catch_warnings(): |
1036 | warnings.simplefilter('ignore', category=DeprecationWarning) | |
1037 | assert isinstance(ydl._opener, urllib.request.OpenerDirector) | |
1038 | ||
1039 | @pytest.mark.parametrize('proxy,expected', [ | |
1040 | ('http://127.0.0.1:8080', {'all': 'http://127.0.0.1:8080'}), | |
1041 | ('', {'all': '__noproxy__'}), | |
1042 | (None, {'http': 'http://127.0.0.1:8081', 'https': 'http://127.0.0.1:8081'}) # env, set https | |
1043 | ]) | |
1044 | def test_proxy(self, proxy, expected): | |
1045 | old_http_proxy = os.environ.get('HTTP_PROXY') | |
1046 | try: | |
1047 | os.environ['HTTP_PROXY'] = 'http://127.0.0.1:8081' # ensure that provided proxies override env | |
1048 | with FakeYDL({'proxy': proxy}) as ydl: | |
1049 | assert ydl.proxies == expected | |
1050 | finally: | |
1051 | if old_http_proxy: | |
1052 | os.environ['HTTP_PROXY'] = old_http_proxy | |
1053 | ||
1054 | def test_compat_request(self): | |
1055 | with FakeRHYDL() as ydl: | |
1056 | assert ydl.urlopen('test://') | |
1057 | urllib_req = urllib.request.Request('http://foo.bar', data=b'test', method='PUT', headers={'X-Test': '1'}) | |
1058 | urllib_req.add_unredirected_header('Cookie', 'bob=bob') | |
1059 | urllib_req.timeout = 2 | |
3d2623a8 | 1060 | with warnings.catch_warnings(): |
1061 | warnings.simplefilter('ignore', category=DeprecationWarning) | |
1062 | req = ydl.urlopen(urllib_req).request | |
1063 | assert req.url == urllib_req.get_full_url() | |
1064 | assert req.data == urllib_req.data | |
1065 | assert req.method == urllib_req.get_method() | |
1066 | assert 'X-Test' in req.headers | |
1067 | assert 'Cookie' in req.headers | |
1068 | assert req.extensions.get('timeout') == 2 | |
227bf1a3 | 1069 | |
1070 | with pytest.raises(AssertionError): | |
1071 | ydl.urlopen(None) | |
1072 | ||
1073 | def test_extract_basic_auth(self): | |
1074 | with FakeRHYDL() as ydl: | |
1075 | res = ydl.urlopen(Request('http://user:pass@foo.bar')) | |
1076 | assert res.request.headers['Authorization'] == 'Basic dXNlcjpwYXNz' | |
1077 | ||
1078 | def test_sanitize_url(self): | |
1079 | with FakeRHYDL() as ydl: | |
1080 | res = ydl.urlopen(Request('httpss://foo.bar')) | |
1081 | assert res.request.url == 'https://foo.bar' | |
1082 | ||
1083 | def test_file_urls_error(self): | |
1084 | # use urllib handler | |
1085 | with FakeYDL() as ydl: | |
1086 | with pytest.raises(RequestError, match=r'file:// URLs are disabled by default'): | |
1087 | ydl.urlopen('file://') | |
1088 | ||
1089 | def test_legacy_server_connect_error(self): | |
1090 | with FakeRHYDL() as ydl: | |
1091 | for error in ('UNSAFE_LEGACY_RENEGOTIATION_DISABLED', 'SSLV3_ALERT_HANDSHAKE_FAILURE'): | |
1092 | with pytest.raises(RequestError, match=r'Try using --legacy-server-connect'): | |
1093 | ydl.urlopen(f'ssl://{error}') | |
1094 | ||
1095 | with pytest.raises(SSLError, match='testerror'): | |
1096 | ydl.urlopen('ssl://testerror') | |
1097 | ||
1098 | @pytest.mark.parametrize('proxy_key,proxy_url,expected', [ | |
1099 | ('http', '__noproxy__', None), | |
1100 | ('no', '127.0.0.1,foo.bar', '127.0.0.1,foo.bar'), | |
1101 | ('https', 'example.com', 'http://example.com'), | |
1102 | ('https', 'socks5://example.com', 'socks5h://example.com'), | |
1103 | ('http', 'socks://example.com', 'socks4://example.com'), | |
1104 | ('http', 'socks4://example.com', 'socks4://example.com'), | |
1105 | ]) | |
1106 | def test_clean_proxy(self, proxy_key, proxy_url, expected): | |
1107 | # proxies should be cleaned in urlopen() | |
1108 | with FakeRHYDL() as ydl: | |
1109 | req = ydl.urlopen(Request('test://', proxies={proxy_key: proxy_url})).request | |
1110 | assert req.proxies[proxy_key] == expected | |
1111 | ||
1112 | # and should also be cleaned when building the handler | |
1113 | env_key = f'{proxy_key.upper()}_PROXY' | |
1114 | old_env_proxy = os.environ.get(env_key) | |
1115 | try: | |
1116 | os.environ[env_key] = proxy_url # ensure that provided proxies override env | |
1117 | with FakeYDL() as ydl: | |
1118 | rh = self.build_handler(ydl) | |
1119 | assert rh.proxies[proxy_key] == expected | |
1120 | finally: | |
1121 | if old_env_proxy: | |
1122 | os.environ[env_key] = old_env_proxy | |
1123 | ||
1124 | def test_clean_proxy_header(self): | |
1125 | with FakeRHYDL() as ydl: | |
1126 | req = ydl.urlopen(Request('test://', headers={'ytdl-request-proxy': '//foo.bar'})).request | |
1127 | assert 'ytdl-request-proxy' not in req.headers | |
1128 | assert req.proxies == {'all': 'http://foo.bar'} | |
1129 | ||
1130 | with FakeYDL({'http_headers': {'ytdl-request-proxy': '//foo.bar'}}) as ydl: | |
1131 | rh = self.build_handler(ydl) | |
1132 | assert 'ytdl-request-proxy' not in rh.headers | |
1133 | assert rh.proxies == {'all': 'http://foo.bar'} | |
1134 | ||
1135 | def test_clean_header(self): | |
1136 | with FakeRHYDL() as ydl: | |
1137 | res = ydl.urlopen(Request('test://', headers={'Youtubedl-no-compression': True})) | |
1138 | assert 'Youtubedl-no-compression' not in res.request.headers | |
1139 | assert res.request.headers.get('Accept-Encoding') == 'identity' | |
1140 | ||
1141 | with FakeYDL({'http_headers': {'Youtubedl-no-compression': True}}) as ydl: | |
1142 | rh = self.build_handler(ydl) | |
1143 | assert 'Youtubedl-no-compression' not in rh.headers | |
1144 | assert rh.headers.get('Accept-Encoding') == 'identity' | |
1145 | ||
1146 | def test_build_handler_params(self): | |
1147 | with FakeYDL({ | |
1148 | 'http_headers': {'test': 'testtest'}, | |
1149 | 'socket_timeout': 2, | |
1150 | 'proxy': 'http://127.0.0.1:8080', | |
1151 | 'source_address': '127.0.0.45', | |
1152 | 'debug_printtraffic': True, | |
1153 | 'compat_opts': ['no-certifi'], | |
1154 | 'nocheckcertificate': True, | |
75dc8e67 | 1155 | 'legacyserverconnect': True, |
227bf1a3 | 1156 | }) as ydl: |
1157 | rh = self.build_handler(ydl) | |
1158 | assert rh.headers.get('test') == 'testtest' | |
1159 | assert 'Accept' in rh.headers # ensure std_headers are still there | |
1160 | assert rh.timeout == 2 | |
1161 | assert rh.proxies.get('all') == 'http://127.0.0.1:8080' | |
1162 | assert rh.source_address == '127.0.0.45' | |
1163 | assert rh.verbose is True | |
1164 | assert rh.prefer_system_certs is True | |
1165 | assert rh.verify is False | |
1166 | assert rh.legacy_ssl_support is True | |
1167 | ||
1168 | @pytest.mark.parametrize('ydl_params', [ | |
1169 | {'client_certificate': 'fakecert.crt'}, | |
1170 | {'client_certificate': 'fakecert.crt', 'client_certificate_key': 'fakekey.key'}, | |
1171 | {'client_certificate': 'fakecert.crt', 'client_certificate_key': 'fakekey.key', 'client_certificate_password': 'foobar'}, | |
1172 | {'client_certificate_key': 'fakekey.key', 'client_certificate_password': 'foobar'}, | |
1173 | ]) | |
1174 | def test_client_certificate(self, ydl_params): | |
1175 | with FakeYDL(ydl_params) as ydl: | |
1176 | rh = self.build_handler(ydl) | |
1177 | assert rh._client_cert == ydl_params # XXX: Too bound to implementation | |
1178 | ||
1179 | def test_urllib_file_urls(self): | |
1180 | with FakeYDL({'enable_file_urls': False}) as ydl: | |
1181 | rh = self.build_handler(ydl, UrllibRH) | |
1182 | assert rh.enable_file_urls is False | |
08916a49 | 1183 | |
227bf1a3 | 1184 | with FakeYDL({'enable_file_urls': True}) as ydl: |
1185 | rh = self.build_handler(ydl, UrllibRH) | |
1186 | assert rh.enable_file_urls is True | |
1187 | ||
1188 | ||
1189 | class TestRequest: | |
1190 | ||
1191 | def test_query(self): | |
1192 | req = Request('http://example.com?q=something', query={'v': 'xyz'}) | |
1193 | assert req.url == 'http://example.com?q=something&v=xyz' | |
1194 | ||
1195 | req.update(query={'v': '123'}) | |
1196 | assert req.url == 'http://example.com?q=something&v=123' | |
1197 | req.update(url='http://example.com', query={'v': 'xyz'}) | |
1198 | assert req.url == 'http://example.com?v=xyz' | |
1199 | ||
1200 | def test_method(self): | |
1201 | req = Request('http://example.com') | |
1202 | assert req.method == 'GET' | |
1203 | req.data = b'test' | |
1204 | assert req.method == 'POST' | |
1205 | req.data = None | |
1206 | assert req.method == 'GET' | |
1207 | req.data = b'test2' | |
1208 | req.method = 'PUT' | |
1209 | assert req.method == 'PUT' | |
1210 | req.data = None | |
1211 | assert req.method == 'PUT' | |
1212 | with pytest.raises(TypeError): | |
1213 | req.method = 1 | |
1214 | ||
1215 | def test_request_helpers(self): | |
1216 | assert HEADRequest('http://example.com').method == 'HEAD' | |
1217 | assert PUTRequest('http://example.com').method == 'PUT' | |
1218 | ||
1219 | def test_headers(self): | |
1220 | req = Request('http://example.com', headers={'tesT': 'test'}) | |
1221 | assert req.headers == HTTPHeaderDict({'test': 'test'}) | |
1222 | req.update(headers={'teSt2': 'test2'}) | |
1223 | assert req.headers == HTTPHeaderDict({'test': 'test', 'test2': 'test2'}) | |
1224 | ||
1225 | req.headers = new_headers = HTTPHeaderDict({'test': 'test'}) | |
1226 | assert req.headers == HTTPHeaderDict({'test': 'test'}) | |
1227 | assert req.headers is new_headers | |
1228 | ||
1229 | # test converts dict to case insensitive dict | |
1230 | req.headers = new_headers = {'test2': 'test2'} | |
1231 | assert isinstance(req.headers, HTTPHeaderDict) | |
1232 | assert req.headers is not new_headers | |
1233 | ||
1234 | with pytest.raises(TypeError): | |
1235 | req.headers = None | |
1236 | ||
1237 | def test_data_type(self): | |
1238 | req = Request('http://example.com') | |
1239 | assert req.data is None | |
1240 | # test bytes is allowed | |
1241 | req.data = b'test' | |
1242 | assert req.data == b'test' | |
1243 | # test iterable of bytes is allowed | |
1244 | i = [b'test', b'test2'] | |
1245 | req.data = i | |
1246 | assert req.data == i | |
1247 | ||
1248 | # test file-like object is allowed | |
1249 | f = io.BytesIO(b'test') | |
1250 | req.data = f | |
1251 | assert req.data == f | |
1252 | ||
1253 | # common mistake: test str not allowed | |
1254 | with pytest.raises(TypeError): | |
1255 | req.data = 'test' | |
1256 | assert req.data != 'test' | |
1257 | ||
1258 | # common mistake: test dict is not allowed | |
1259 | with pytest.raises(TypeError): | |
1260 | req.data = {'test': 'test'} | |
1261 | assert req.data != {'test': 'test'} | |
1262 | ||
1263 | def test_content_length_header(self): | |
1264 | req = Request('http://example.com', headers={'Content-Length': '0'}, data=b'') | |
1265 | assert req.headers.get('Content-Length') == '0' | |
1266 | ||
1267 | req.data = b'test' | |
1268 | assert 'Content-Length' not in req.headers | |
1269 | ||
1270 | req = Request('http://example.com', headers={'Content-Length': '10'}) | |
1271 | assert 'Content-Length' not in req.headers | |
1272 | ||
1273 | def test_content_type_header(self): | |
1274 | req = Request('http://example.com', headers={'Content-Type': 'test'}, data=b'test') | |
1275 | assert req.headers.get('Content-Type') == 'test' | |
1276 | req.data = b'test2' | |
1277 | assert req.headers.get('Content-Type') == 'test' | |
1278 | req.data = None | |
1279 | assert 'Content-Type' not in req.headers | |
1280 | req.data = b'test3' | |
1281 | assert req.headers.get('Content-Type') == 'application/x-www-form-urlencoded' | |
1282 | ||
71baa490 | 1283 | def test_update_req(self): |
1284 | req = Request('http://example.com') | |
1285 | assert req.data is None | |
1286 | assert req.method == 'GET' | |
1287 | assert 'Content-Type' not in req.headers | |
1288 | # Test that zero-byte payloads will be sent | |
1289 | req.update(data=b'') | |
1290 | assert req.data == b'' | |
1291 | assert req.method == 'POST' | |
1292 | assert req.headers.get('Content-Type') == 'application/x-www-form-urlencoded' | |
1293 | ||
227bf1a3 | 1294 | def test_proxies(self): |
1295 | req = Request(url='http://example.com', proxies={'http': 'http://127.0.0.1:8080'}) | |
1296 | assert req.proxies == {'http': 'http://127.0.0.1:8080'} | |
1297 | ||
1298 | def test_extensions(self): | |
1299 | req = Request(url='http://example.com', extensions={'timeout': 2}) | |
1300 | assert req.extensions == {'timeout': 2} | |
1301 | ||
1302 | def test_copy(self): | |
1303 | req = Request( | |
1304 | url='http://example.com', | |
1305 | extensions={'cookiejar': CookieJar()}, | |
1306 | headers={'Accept-Encoding': 'br'}, | |
1307 | proxies={'http': 'http://127.0.0.1'}, | |
1308 | data=[b'123'] | |
1309 | ) | |
1310 | req_copy = req.copy() | |
1311 | assert req_copy is not req | |
1312 | assert req_copy.url == req.url | |
1313 | assert req_copy.headers == req.headers | |
1314 | assert req_copy.headers is not req.headers | |
1315 | assert req_copy.proxies == req.proxies | |
1316 | assert req_copy.proxies is not req.proxies | |
1317 | ||
1318 | # Data is not able to be copied | |
1319 | assert req_copy.data == req.data | |
1320 | assert req_copy.data is req.data | |
1321 | ||
1322 | # Shallow copy extensions | |
1323 | assert req_copy.extensions is not req.extensions | |
1324 | assert req_copy.extensions['cookiejar'] == req.extensions['cookiejar'] | |
1325 | ||
1326 | # Subclasses are copied by default | |
1327 | class AnotherRequest(Request): | |
1328 | pass | |
08916a49 | 1329 | |
227bf1a3 | 1330 | req = AnotherRequest(url='http://127.0.0.1') |
1331 | assert isinstance(req.copy(), AnotherRequest) | |
1332 | ||
1333 | def test_url(self): | |
1334 | req = Request(url='https://фtest.example.com/ some spaceв?ä=c',) | |
1335 | assert req.url == 'https://xn--test-z6d.example.com/%20some%20space%D0%B2?%C3%A4=c' | |
1336 | ||
1337 | assert Request(url='//example.com').url == 'http://example.com' | |
1338 | ||
1339 | with pytest.raises(TypeError): | |
1340 | Request(url='https://').url = None | |
1341 | ||
1342 | ||
1343 | class TestResponse: | |
1344 | ||
1345 | @pytest.mark.parametrize('reason,status,expected', [ | |
1346 | ('custom', 200, 'custom'), | |
1347 | (None, 404, 'Not Found'), # fallback status | |
1348 | ('', 403, 'Forbidden'), | |
1349 | (None, 999, None) | |
1350 | ]) | |
1351 | def test_reason(self, reason, status, expected): | |
1352 | res = Response(io.BytesIO(b''), url='test://', headers={}, status=status, reason=reason) | |
1353 | assert res.reason == expected | |
1354 | ||
1355 | def test_headers(self): | |
1356 | headers = Message() | |
1357 | headers.add_header('Test', 'test') | |
1358 | headers.add_header('Test', 'test2') | |
1359 | headers.add_header('content-encoding', 'br') | |
1360 | res = Response(io.BytesIO(b''), headers=headers, url='test://') | |
1361 | assert res.headers.get_all('test') == ['test', 'test2'] | |
1362 | assert 'Content-Encoding' in res.headers | |
1363 | ||
1364 | def test_get_header(self): | |
1365 | headers = Message() | |
1366 | headers.add_header('Set-Cookie', 'cookie1') | |
1367 | headers.add_header('Set-cookie', 'cookie2') | |
1368 | headers.add_header('Test', 'test') | |
1369 | headers.add_header('Test', 'test2') | |
1370 | res = Response(io.BytesIO(b''), headers=headers, url='test://') | |
1371 | assert res.get_header('test') == 'test, test2' | |
1372 | assert res.get_header('set-Cookie') == 'cookie1' | |
1373 | assert res.get_header('notexist', 'default') == 'default' | |
1374 | ||
1375 | def test_compat(self): | |
1376 | res = Response(io.BytesIO(b''), url='test://', status=404, headers={'test': 'test'}) | |
3d2623a8 | 1377 | with warnings.catch_warnings(): |
1378 | warnings.simplefilter('ignore', category=DeprecationWarning) | |
1379 | assert res.code == res.getcode() == res.status | |
1380 | assert res.geturl() == res.url | |
1381 | assert res.info() is res.headers | |
1382 | assert res.getheader('test') == res.get_header('test') |