backup/warcprox
1
0
Fork 0
mirror of https://github.com/internetarchive/warcprox.git synced 2025-01-18 13:22:09 +01:00
Code Issues Projects Releases Activity

emoji idn fails with python 2.7, so test with a BMP unicode character

Browse Source
This commit is contained in:
Noah Levitt 2016-06-29 17:16:50 -05:00
parent 33775d360a
commit 46c24833ff
2 changed files with 12 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
setup.py
View File

@ -51,7 +51,7 @@ except:
setuptools.setup( setuptools.setup(
name='warcprox', name='warcprox',
version='2.0.dev20', version='2.0.dev21',
description='WARC writing MITM HTTP/S proxy', description='WARC writing MITM HTTP/S proxy',
url='https://github.com/internetarchive/warcprox', url='https://github.com/internetarchive/warcprox',
author='Noah Levitt', author='Noah Levitt',

22
tests/test_warcprox.py
View File

@ -1004,13 +1004,13 @@ def test_domain_data_soft_limit(
http_daemon, https_daemon, warcprox_, archiving_proxies): http_daemon, https_daemon, warcprox_, archiving_proxies):
# using idn # using idn
request_meta = { request_meta = {
"stats": {"buckets": [{"bucket":"test_domain_data_limit_bucket","tally-domains":['🎵zZ.LOCALhost']}]}, "stats": {"buckets": [{"bucket":"test_domain_data_limit_bucket","tally-domains":['♛zZ.LOCALhost']}]},
# response is 135 bytes, so 3rd novel url should be disallowed # response is 135 bytes, so 3rd novel url should be disallowed
"soft-limits": {"test_domain_data_limit_bucket:🎵ZZ.localhost/new/wire_bytes":200}, "soft-limits": {"test_domain_data_limit_bucket:♛ZZ.localhost/new/wire_bytes":200},
} }
headers = {"Warcprox-Meta": json.dumps(request_meta)} headers = {"Warcprox-Meta": json.dumps(request_meta)}
url = 'http://🎵Zz.localhost:{}/y/z'.format(http_daemon.server_port) url = 'http://♛Zz.localhost:{}/y/z'.format(http_daemon.server_port)
response = requests.get( response = requests.get(
url, proxies=archiving_proxies, headers=headers, stream=True) url, proxies=archiving_proxies, headers=headers, stream=True)
assert response.status_code == 200 assert response.status_code == 200
@ -1025,7 +1025,7 @@ def test_domain_data_soft_limit(
time.sleep(2.0) time.sleep(2.0)
# duplicate, does not count toward limit # duplicate, does not count toward limit
url = 'https://baz.🎵zz.localhost:{}/y/z'.format(https_daemon.server_port) url = 'https://baz.♛zz.localhost:{}/y/z'.format(https_daemon.server_port)
response = requests.get( response = requests.get(
url, proxies=archiving_proxies, headers=headers, stream=True, url, proxies=archiving_proxies, headers=headers, stream=True,
verify=False) verify=False)
@ -1041,7 +1041,7 @@ def test_domain_data_soft_limit(
time.sleep(2.0) time.sleep(2.0)
# novel, pushes stats over the limit # novel, pushes stats over the limit
url = 'https://muh.XN--Zz-B862a.locALHOst:{}/z/~'.format(https_daemon.server_port) url = 'https://muh.XN--Zz-xZX.locALHOst:{}/z/~'.format(https_daemon.server_port)
response = requests.get( response = requests.get(
url, proxies=archiving_proxies, headers=headers, stream=True, url, proxies=archiving_proxies, headers=headers, stream=True,
verify=False) verify=False)
@ -1065,31 +1065,31 @@ def test_domain_data_soft_limit(
assert response.content == b'I am the warcprox test payload! ~~~~~~~~~~!\n' assert response.content == b'I am the warcprox test payload! ~~~~~~~~~~!\n'
# blocked because we're over the limit now # blocked because we're over the limit now
url = 'http://lOl.wHut.🎵ZZ.lOcALHOst:{}/y/z'.format(http_daemon.server_port) url = 'http://lOl.wHut.♛ZZ.lOcALHOst:{}/y/z'.format(http_daemon.server_port)
response = requests.get( response = requests.get(
url, proxies=archiving_proxies, headers=headers, stream=True) url, proxies=archiving_proxies, headers=headers, stream=True)
assert response.status_code == 430 assert response.status_code == 430
assert response.reason == "Reached soft limit" assert response.reason == "Reached soft limit"
expected_response_meta = {'reached-soft-limit': {'test_domain_data_limit_bucket:xn--zz-b862a.localhost/new/wire_bytes': 200}, 'stats': {'test_domain_data_limit_bucket:xn--zz-b862a.localhost': {'total': {'wire_bytes': 405, 'urls': 3}, 'revisit': {'wire_bytes': 135, 'urls': 1}, 'new': {'wire_bytes': 270, 'urls': 2}, 'bucket': 'test_domain_data_limit_bucket:xn--zz-b862a.localhost'}}} expected_response_meta = {'reached-soft-limit': {'test_domain_data_limit_bucket:xn--zz-xzx.localhost/new/wire_bytes': 200}, 'stats': {'test_domain_data_limit_bucket:xn--zz-xzx.localhost': {'total': {'wire_bytes': 405, 'urls': 3}, 'revisit': {'wire_bytes': 135, 'urls': 1}, 'new': {'wire_bytes': 270, 'urls': 2}, 'bucket': 'test_domain_data_limit_bucket:xn--zz-xzx.localhost'}}}
assert json.loads(response.headers["warcprox-meta"]) == expected_response_meta assert json.loads(response.headers["warcprox-meta"]) == expected_response_meta
assert response.headers["content-type"] == "text/plain;charset=utf-8" assert response.headers["content-type"] == "text/plain;charset=utf-8"
assert response.raw.data == b"request rejected by warcprox: reached soft limit test_domain_data_limit_bucket:xn--zz-b862a.localhost/new/wire_bytes=200\n" assert response.raw.data == b"request rejected by warcprox: reached soft limit test_domain_data_limit_bucket:xn--zz-xzx.localhost/new/wire_bytes=200\n"
# XXX this check is resulting in a segfault on mac and linux, from ssl I # XXX this check is resulting in a segfault on mac and linux, from ssl I
# think, probably because of the dns resolution monkey-patching # think, probably because of the dns resolution monkey-patching
# https://travis-ci.org/internetarchive/warcprox/builds/141187342 # https://travis-ci.org/internetarchive/warcprox/builds/141187342
# #
### # https also blocked ### # https also blocked
### url = 'https://xn--zz-b862ah.loCAlhost:{}/w/x'.format(https_daemon.server_port) ### url = 'https://xn--zz-xzxh.loCAlhost:{}/w/x'.format(https_daemon.server_port)
### response = requests.get( ### response = requests.get(
### url, proxies=archiving_proxies, headers=headers, stream=True, ### url, proxies=archiving_proxies, headers=headers, stream=True,
### verify=False) ### verify=False)
### assert response.status_code == 430 ### assert response.status_code == 430
### assert response.reason == "Reached soft limit" ### assert response.reason == "Reached soft limit"
### expected_response_meta = {'reached-soft-limit': {'test_domain_data_limit_bucket:xn--zz-b862a.localhost/new/wire_bytes': 200}, 'stats': {'test_domain_data_limit_bucket:xn--zz-b862a.localhost': {'total': {'wire_bytes': 405, 'urls': 3}, 'revisit': {'wire_bytes': 135, 'urls': 1}, 'new': {'wire_bytes': 270, 'urls': 2}, 'bucket': 'test_domain_data_limit_bucket:xn--zz-b862a.localhost'}}} ### expected_response_meta = {'reached-soft-limit': {'test_domain_data_limit_bucket:xn--zz-xzx.localhost/new/wire_bytes': 200}, 'stats': {'test_domain_data_limit_bucket:xn--zz-xzx.localhost': {'total': {'wire_bytes': 405, 'urls': 3}, 'revisit': {'wire_bytes': 135, 'urls': 1}, 'new': {'wire_bytes': 270, 'urls': 2}, 'bucket': 'test_domain_data_limit_bucket:xn--zz-xzx.localhost'}}}
### assert json.loads(response.headers["warcprox-meta"]) == expected_response_meta ### assert json.loads(response.headers["warcprox-meta"]) == expected_response_meta
### assert response.headers["content-type"] == "text/plain;charset=utf-8" ### assert response.headers["content-type"] == "text/plain;charset=utf-8"
### assert response.raw.data == b"request rejected by warcprox: reached soft limit test_domain_data_limit_bucket:xn--zz-b862a.localhost/new/wire_bytes=200\n" ### assert response.raw.data == b"request rejected by warcprox: reached soft limit test_domain_data_limit_bucket:xn--zz-xzx.localhost/new/wire_bytes=200\n"
# XXX this test relies on a tor proxy running at localhost:9050 with a working # XXX this test relies on a tor proxy running at localhost:9050 with a working
# connection to the internet, and relies on a third party site (facebook) being # connection to the internet, and relies on a third party site (facebook) being