This is an archive of the discontinued Mercurial Phabricator instance.

Differential D5952

url: always use str for proxy configuration
ClosedPublic

Authored by indygreg on Feb 12 2019, 5:37 PM.

Details

Reviewers
None
Group Reviewers
hg-reviewers
Commits
rHG4028897dfa05: url: always use str for proxy configuration
Summary

Previously, proxies didn't work on Python 3 for various reasons.

First, the keys to the "proxies" dict are fed into a
setattr(self, "%s_open", ...) call and passing bytestrings
results in setting an oddly named attribute due to the b''
in %s formatting. This resulted in "http_open" and "https_open"
not being properly overridden and proxies not being used.

Second, the standard library was expecting proxy URLs to be
str. And various operations (including our custom code in
url.py) would fail to account for the str/bytes mismatch.

This commit normalizes everything to str and adjusts our
proxy code in url.py to account for the presence of str
on Python 3.

Diff Detail

Repository
rHG Mercurial
Lint
Automatic diff as part of commit; lint not applicable.
Unit
Automatic diff as part of commit; unit tests not applicable.

Event Timeline

indygreg created this revision.Feb 12 2019, 5:37 PM
Closed by commit rHG4028897dfa05: url: always use str for proxy configuration (authored by indygreg). · Explain WhyFeb 15 2019, 10:34 PM
This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

PathPackages
Mmercurial/url.py (18 lines)

Diff 14120

mercurial/url.py

p in encoding.environ.get("no_proxy", '').split(',') p in encoding.environ.get("no_proxy", '').split(',')
if p.strip()]) if p.strip()])
# "http_proxy.always" config is for running tests on localhost # "http_proxy.always" config is for running tests on localhost
if ui.configbool("http_proxy", "always"): if ui.configbool("http_proxy", "always"):
self.no_list = [] self.no_list = []
else: else:
self.no_list = no_list self.no_list = no_list
proxyurl = bytes(proxy) # Keys and values need to be str because the standard library
proxies = {'http': proxyurl, 'https': proxyurl} # expects them to be.
ui.debug('proxying through %s\n' % util.hidepassword(proxyurl)) proxyurl = str(proxy)
proxies = {r'http': proxyurl, r'https': proxyurl}
ui.debug('proxying through %s\n' % util.hidepassword(bytes(proxy)))
else: else:
proxies = {} proxies = {}
urlreq.proxyhandler.__init__(self, proxies) urlreq.proxyhandler.__init__(self, proxies)
self.ui = ui self.ui = ui
def proxy_open(self, req, proxy, type_): def proxy_open(self, req, proxy, type_):
host = urllibcompat.gethost(req).split(':')[0] host = pycompat.bytesurl(urllibcompat.gethost(req)).split(':')[0]
for e in self.no_list: for e in self.no_list:
if host == e: if host == e:
return None return None
if e.startswith('*.') and host.endswith(e[2:]): if e.startswith('*.') and host.endswith(e[2:]):
return None return None
if e.startswith('.') and host.endswith(e[1:]): if e.startswith('.') and host.endswith(e[1:]):
return None return None
return proxyres return proxyres
return keepalive.HTTPConnection.getresponse(self) return keepalive.HTTPConnection.getresponse(self)
# Large parts of this function have their origin from before Python 2.6 # Large parts of this function have their origin from before Python 2.6
# and could potentially be removed. # and could potentially be removed.
def _generic_start_transaction(handler, h, req): def _generic_start_transaction(handler, h, req):
tunnel_host = req._tunnel_host tunnel_host = req._tunnel_host
if tunnel_host: if tunnel_host:
if tunnel_host[:7] not in ['http://', 'https:/']: if tunnel_host[:7] not in [r'http://', r'https:/']:
tunnel_host = 'https://' + tunnel_host tunnel_host = r'https://' + tunnel_host
new_tunnel = True new_tunnel = True
else: else:
tunnel_host = urllibcompat.getselector(req) tunnel_host = urllibcompat.getselector(req)
new_tunnel = False new_tunnel = False
if new_tunnel or tunnel_host == urllibcompat.getfullurl(req): # has proxy if new_tunnel or tunnel_host == urllibcompat.getfullurl(req): # has proxy
u = util.url(tunnel_host) u = util.url(pycompat.bytesurl(tunnel_host))
if new_tunnel or u.scheme == 'https': # only use CONNECT for HTTPS if new_tunnel or u.scheme == 'https': # only use CONNECT for HTTPS
h.realhostport = ':'.join([u.host, (u.port or '443')]) h.realhostport = ':'.join([u.host, (u.port or '443')])
h.headers = req.headers.copy() h.headers = req.headers.copy()
h.headers.update(handler.parent.addheaders) h.headers.update(handler.parent.addheaders)
return return
h.realhostport = None h.realhostport = None
h.headers = None h.headers = None
def _generic_proxytunnel(self): def _generic_proxytunnel(self):
proxyheaders = dict( proxyheaders = dict(
[(x, self.headers[x]) for x in self.headers [(x, self.headers[x]) for x in self.headers
if x.lower().startswith('proxy-')]) if x.lower().startswith(r'proxy-')])
self.send('CONNECT %s HTTP/1.0\r\n' % self.realhostport) self.send('CONNECT %s HTTP/1.0\r\n' % self.realhostport)
for header in proxyheaders.iteritems(): for header in proxyheaders.iteritems():
self.send('%s: %s\r\n' % header) self.send('%s: %s\r\n' % header)
self.send('\r\n') self.send('\r\n')
# majority of the following code is duplicated from # majority of the following code is duplicated from
# httplib.HTTPConnection as there are no adequate places to # httplib.HTTPConnection as there are no adequate places to
# override functions to provide the needed functionality # override functions to provide the needed functionality