curl/tests/http/test_02_download.py
Stefan Eissing 4ae2d9f24d
proxy: http2 proxy tunnel implementation
- currently only on debug build and when env variable
  CURL_PROXY_TUNNEL_H2 is present.
- will ALPN negotiate with the proxy server and switch
  tunnel filter based on the protocol negotiated.
- http/1.1 tunnel code moved into cf-h1-proxy.[ch]
- http/2 tunnel code implemented in cf-h2-proxy.[ch]
- tunnel start and ALPN set remains in http_proxy.c
- moving all haproxy related code into cf-haproxy.[ch]

VTLS changes
- SSL filters rely solely on the "alpn" specification they
  are created with and no longer check conn->bits.tls_enable_alpn.
- checks on which ALPN specification to use (or none at all) are
  done in vtls.c when creating the filter.

Testing
- added a nghttpx forward proxy to the pytest setup that
  speaks HTTP/2 and forwards all requests to the Apache httpd
  forward proxy server.
- extending test coverage in test_10 cases
- adding proxy tests for direct/tunnel h1/h2 use of basic auth.
- adding test for http/1.1 and h2 proxy tunneling to pytest

Closes #10780
2023-04-06 13:04:46 +02:00

290 lines
13 KiB
Python

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
#***************************************************************************
# _ _ ____ _
# Project ___| | | | _ \| |
# / __| | | | |_) | |
# | (__| |_| | _ <| |___
# \___|\___/|_| \_\_____|
#
# Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
#
# This software is licensed as described in the file COPYING, which
# you should have received as part of this distribution. The terms
# are also available at https://curl.se/docs/copyright.html.
#
# You may opt to use, copy, modify, merge, publish, distribute and/or sell
# copies of the Software, and permit persons to whom the Software is
# furnished to do so, under the terms of the COPYING file.
#
# This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
# KIND, either express or implied.
#
# SPDX-License-Identifier: curl
#
###########################################################################
#
import difflib
import filecmp
import logging
import os
import pytest
from testenv import Env, CurlClient
log = logging.getLogger(__name__)
class TestDownload:
@pytest.fixture(autouse=True, scope='class')
def _class_scope(self, env, httpd, nghttpx):
if env.have_h3():
nghttpx.start_if_needed()
httpd.clear_extra_configs()
httpd.reload()
@pytest.fixture(autouse=True, scope='class')
def _class_scope(self, env, httpd):
env.make_data_file(indir=httpd.docs_dir, fname="data-100k", fsize=100*1024)
env.make_data_file(indir=httpd.docs_dir, fname="data-1m", fsize=1024*1024)
env.make_data_file(indir=httpd.docs_dir, fname="data-10m", fsize=10*1024*1024)
# download 1 file
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_01_download_1(self, env: Env, httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
curl = CurlClient(env=env)
url = f'https://{env.authority_for(env.domain1, proto)}/data.json'
r = curl.http_download(urls=[url], alpn_proto=proto)
r.check_response(http_status=200)
# download 2 files
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_02_download_2(self, env: Env, httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
curl = CurlClient(env=env)
url = f'https://{env.authority_for(env.domain1, proto)}/data.json?[0-1]'
r = curl.http_download(urls=[url], alpn_proto=proto)
r.check_response(http_status=200, count=2)
# download 100 files sequentially
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_03_download_100_sequential(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
curl = CurlClient(env=env)
urln = f'https://{env.authority_for(env.domain1, proto)}/data.json?[0-99]'
r = curl.http_download(urls=[urln], alpn_proto=proto)
r.check_response(http_status=200, count=100, connect_count=1)
# download 100 files parallel
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_04_download_100_parallel(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
max_parallel = 6 if proto == 'http/1.1' else 50
curl = CurlClient(env=env)
urln = f'https://{env.authority_for(env.domain1, proto)}/data.json?[0-99]'
r = curl.http_download(urls=[urln], alpn_proto=proto, extra_args=[
'--parallel', '--parallel-max', f'{max_parallel}'
])
r.check_response(http_status=200, count=100)
if proto == 'http/1.1':
# http/1.1 parallel transfers will open multiple connections
assert r.total_connects > 1, r.dump_logs()
else:
# http2 parallel transfers will use one connection (common limit is 100)
assert r.total_connects == 1, r.dump_logs()
# download 500 files sequential
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_05_download_500_sequential(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
if proto == 'h3' and env.curl_uses_lib('msh3'):
pytest.skip("msh3 shaky here")
curl = CurlClient(env=env)
urln = f'https://{env.authority_for(env.domain1, proto)}/data.json?[0-499]'
r = curl.http_download(urls=[urln], alpn_proto=proto)
r.check_response(http_status=200, count=500)
if proto == 'http/1.1':
# http/1.1 parallel transfers will open multiple connections
assert r.total_connects > 1, r.dump_logs()
else:
# http2 parallel transfers will use one connection (common limit is 100)
assert r.total_connects == 1, r.dump_logs()
# download 500 files parallel
@pytest.mark.parametrize("proto", ['h2', 'h3'])
def test_02_06_download_500_parallel(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
count = 500
max_parallel = 50
curl = CurlClient(env=env)
urln = f'https://{env.authority_for(env.domain1, proto)}/data.json?[000-{count-1}]'
r = curl.http_download(urls=[urln], alpn_proto=proto, extra_args=[
'--parallel', '--parallel-max', f'{max_parallel}'
])
r.check_response(http_status=200, count=count, connect_count=1)
# download files parallel, check connection reuse/multiplex
@pytest.mark.parametrize("proto", ['h2', 'h3'])
def test_02_07_download_reuse(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
count = 200
curl = CurlClient(env=env)
urln = f'https://{env.authority_for(env.domain1, proto)}/data.json?[0-{count-1}]'
r = curl.http_download(urls=[urln], alpn_proto=proto,
with_stats=True, extra_args=[
'--parallel', '--parallel-max', '200'
])
r.check_response(http_status=200, count=count)
# should have used at most 2 connections only (test servers allow 100 req/conn)
# it may be just 1 on slow systems where request are answered faster than
# curl can exhaust the capacity or if curl runs with address-sanitizer speed
assert r.total_connects <= 2, "h2 should use fewer connections here"
# download files parallel with http/1.1, check connection not reused
@pytest.mark.parametrize("proto", ['http/1.1'])
def test_02_07b_download_reuse(self, env: Env,
httpd, nghttpx, repeat, proto):
count = 20
curl = CurlClient(env=env)
urln = f'https://{env.authority_for(env.domain1, proto)}/data.json?[0-{count-1}]'
r = curl.http_download(urls=[urln], alpn_proto=proto,
with_stats=True, extra_args=[
'--parallel'
])
r.check_response(count=count, http_status=200)
# http/1.1 should have used count connections
assert r.total_connects == count, "http/1.1 should use this many connections"
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_08_1MB_serial(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
count = 20
urln = f'https://{env.authority_for(env.domain1, proto)}/data-1m?[0-{count-1}]'
curl = CurlClient(env=env)
r = curl.http_download(urls=[urln], alpn_proto=proto)
r.check_response(count=count, http_status=200)
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_09_1MB_parallel(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
count = 20
urln = f'https://{env.authority_for(env.domain1, proto)}/data-1m?[0-{count-1}]'
curl = CurlClient(env=env)
r = curl.http_download(urls=[urln], alpn_proto=proto, extra_args=[
'--parallel'
])
r.check_response(count=count, http_status=200)
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_10_10MB_serial(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
count = 20
urln = f'https://{env.authority_for(env.domain1, proto)}/data-10m?[0-{count-1}]'
curl = CurlClient(env=env)
r = curl.http_download(urls=[urln], alpn_proto=proto)
r.check_response(count=count, http_status=200)
@pytest.mark.parametrize("proto", ['http/1.1', 'h2', 'h3'])
def test_02_11_10MB_parallel(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
if proto == 'h3' and env.curl_uses_lib('msh3'):
pytest.skip("msh3 stalls here")
count = 20
urln = f'https://{env.authority_for(env.domain1, proto)}/data-10m?[0-{count-1}]'
curl = CurlClient(env=env)
r = curl.http_download(urls=[urln], alpn_proto=proto, extra_args=[
'--parallel'
])
r.check_response(count=count, http_status=200)
@pytest.mark.parametrize("proto", ['h2', 'h3'])
def test_02_12_head_serial_https(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
count = 100
urln = f'https://{env.authority_for(env.domain1, proto)}/data-10m?[0-{count-1}]'
curl = CurlClient(env=env)
r = curl.http_download(urls=[urln], alpn_proto=proto, extra_args=[
'--head'
])
r.check_response(count=count, http_status=200)
@pytest.mark.parametrize("proto", ['h2'])
def test_02_13_head_serial_h2c(self, env: Env,
httpd, nghttpx, repeat, proto):
if proto == 'h3' and not env.have_h3():
pytest.skip("h3 not supported")
count = 100
urln = f'http://{env.domain1}:{env.http_port}/data-10m?[0-{count-1}]'
curl = CurlClient(env=env)
r = curl.http_download(urls=[urln], alpn_proto=proto, extra_args=[
'--head', '--http2-prior-knowledge', '--fail-early'
])
r.check_response(count=count, http_status=200)
def test_02_20_h2_small_frames(self, env: Env, httpd, repeat):
# Test case to reproduce content corruption as observed in
# https://github.com/curl/curl/issues/10525
# To reliably reproduce, we need an Apache httpd that supports
# setting smaller frame sizes. This is not released yet, we
# test if it works and back out if not.
httpd.set_extra_config(env.domain1, lines=[
f'H2MaxDataFrameLen 1024',
])
assert httpd.stop()
if not httpd.start():
# no, not supported, bail out
httpd.set_extra_config(env.domain1, lines=None)
assert httpd.start()
pytest.skip(f'H2MaxDataFrameLen not supported')
# ok, make 100 downloads with 2 parallel running and they
# are expected to stumble into the issue when using `lib/http2.c`
# from curl 7.88.0
count = 100
urln = f'https://{env.authority_for(env.domain1, "h2")}/data-1m?[0-{count-1}]'
curl = CurlClient(env=env)
r = curl.http_download(urls=[urln], alpn_proto="h2", extra_args=[
'--parallel', '--parallel-max', '2'
])
r.check_response(count=count, http_status=200)
srcfile = os.path.join(httpd.docs_dir, 'data-1m')
for i in range(count):
dfile = curl.download_file(i)
assert os.path.exists(dfile)
if not filecmp.cmp(srcfile, dfile, shallow=False):
diff = "".join(difflib.unified_diff(a=open(srcfile).readlines(),
b=open(dfile).readlines(),
fromfile=srcfile,
tofile=dfile,
n=1))
assert False, f'download {dfile} differs:\n{diff}'
# restore httpd defaults
httpd.set_extra_config(env.domain1, lines=None)
assert httpd.stop()
assert httpd.start()