|
|
# Copyright (C) 2016-2024 RhodeCode GmbH
|
|
|
#
|
|
|
# This program is free software: you can redistribute it and/or modify
|
|
|
# it under the terms of the GNU Affero General Public License, version 3
|
|
|
# (only), as published by the Free Software Foundation.
|
|
|
#
|
|
|
# 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
|
|
|
# GNU General Public License for more details.
|
|
|
#
|
|
|
# You should have received a copy of the GNU Affero General Public License
|
|
|
# along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
|
#
|
|
|
# This program is dual-licensed. If you wish to learn more about the
|
|
|
# RhodeCode Enterprise Edition, including its added features, Support services,
|
|
|
# and proprietary license terms, please see https://rhodecode.com/licenses/
|
|
|
|
|
|
"""
|
|
|
Checking the chunked data transfer via HTTP
|
|
|
"""
|
|
|
|
|
|
import os
|
|
|
import time
|
|
|
import subprocess
|
|
|
|
|
|
import pytest
|
|
|
import requests
|
|
|
|
|
|
from rhodecode.lib.middleware.utils import scm_app_http
|
|
|
from rhodecode.tests.utils import wait_for_url
|
|
|
|
|
|
|
|
|
def test_does_chunked_end_to_end_transfer(scm_app):
|
|
|
response = requests.post(scm_app, data='', stream=True)
|
|
|
assert response.headers['Transfer-Encoding'] == 'chunked'
|
|
|
times = [time.time() for chunk in response.raw.read_chunked()]
|
|
|
assert times[1] - times[0] > 0.1, "Chunks arrived at the same time"
|
|
|
|
|
|
|
|
|
SCM_APP_URL_TMPL = 'http://0.0.0.0:{port}'
|
|
|
|
|
|
|
|
|
@pytest.fixture()
|
|
|
def echo_app_chunking(request, available_port_factory):
|
|
|
"""
|
|
|
Run the EchoApp via Waitress in a subprocess.
|
|
|
|
|
|
Return the URL endpoint to reach the app.
|
|
|
"""
|
|
|
port = available_port_factory()
|
|
|
command = (
|
|
|
'waitress-serve --send-bytes 1 --port {port} --call '
|
|
|
'rhodecode.tests.lib.middleware.utils.test_scm_app_http_chunking'
|
|
|
':create_echo_app')
|
|
|
command = command.format(port=port)
|
|
|
proc = subprocess.Popen(command.split(' '), bufsize=0)
|
|
|
echo_app_url = SCM_APP_URL_TMPL.format(port=port)
|
|
|
|
|
|
@request.addfinalizer
|
|
|
def stop_echo_app():
|
|
|
proc.kill()
|
|
|
|
|
|
return echo_app_url
|
|
|
|
|
|
|
|
|
@pytest.fixture()
|
|
|
def scm_app(request, available_port_factory, echo_app_chunking):
|
|
|
"""
|
|
|
Run the scm_app in Waitress.
|
|
|
|
|
|
Returns the URL endpoint where this app can be reached.
|
|
|
"""
|
|
|
port = available_port_factory()
|
|
|
command = (
|
|
|
'waitress-serve --send-bytes 1 --port {port} --call '
|
|
|
'rhodecode.tests.lib.middleware.utils.test_scm_app_http_chunking'
|
|
|
':create_scm_app')
|
|
|
command = command.format(port=port)
|
|
|
env = os.environ.copy()
|
|
|
env["RC_ECHO_URL"] = echo_app_chunking
|
|
|
proc = subprocess.Popen(command.split(' '), bufsize=0, env=env)
|
|
|
scm_app_url = SCM_APP_URL_TMPL.format(port=port)
|
|
|
wait_for_url(scm_app_url)
|
|
|
|
|
|
@request.addfinalizer
|
|
|
def stop_echo_app():
|
|
|
proc.kill()
|
|
|
|
|
|
return scm_app_url
|
|
|
|
|
|
|
|
|
class EchoApp(object):
|
|
|
"""
|
|
|
Stub WSGI application which returns a chunked response to every request.
|
|
|
"""
|
|
|
|
|
|
def __init__(self, repo_path, repo_name, config):
|
|
|
self._repo_path = repo_path
|
|
|
|
|
|
def __call__(self, environ, start_response):
|
|
|
environ['wsgi.input'].read()
|
|
|
status = '200 OK'
|
|
|
headers = []
|
|
|
start_response(status, headers)
|
|
|
return result_generator()
|
|
|
|
|
|
|
|
|
def result_generator():
|
|
|
"""
|
|
|
Simulate chunked results.
|
|
|
|
|
|
The intended usage is to simulate a chunked response as we would get it
|
|
|
out of a vcs operation during a call to "hg clone".
|
|
|
"""
|
|
|
yield b'waiting 2 seconds'
|
|
|
# Wait long enough so that the first chunk can go out
|
|
|
time.sleep(2)
|
|
|
yield b'final chunk'
|
|
|
# Another small wait, otherwise they go together
|
|
|
time.sleep(0.1)
|
|
|
|
|
|
|
|
|
def create_echo_app():
|
|
|
"""
|
|
|
Create EchoApp filled with stub data.
|
|
|
"""
|
|
|
return EchoApp('stub_path', 'repo_name', {})
|
|
|
|
|
|
|
|
|
def create_scm_app():
|
|
|
"""
|
|
|
Create a scm_app hooked up to speak to EchoApp.
|
|
|
"""
|
|
|
echo_app_url = os.environ["RC_ECHO_URL"]
|
|
|
return scm_app_http.VcsHttpProxy(
|
|
|
echo_app_url, 'stub_path', 'stub_name', None)
|
|
|
|