-#! /usr/bin/env python
+#! /usr/bin/env python3
-import sys
-url = sys.argv[1]
+try:
+ PY2 = False
+ from urllib.parse import parse_qsl, urlencode, \
+ quote, quote_plus, unquote, unquote_plus, \
+ splittype, splithost, splituser, splitpasswd, \
+ splitport, splittag, splitquery
+ from urllib.request import urlretrieve
+ import urllib.request
+except ImportError:
+ PY2 = True
+ from cgi import parse_qsl
+ from urllib import urlencode, quote, quote_plus, unquote, unquote_plus, \
+ splittype, splithost, splituser, splitpasswd, \
+ splitport, splittag, splitquery, urlretrieve
+import os
+import sys
import urllib
-urllib._urlopener = urllib.FancyURLopener()
-
-from cgi import parse_qsl
from m_lib.defenc import default_encoding
-protocol, request = urllib.splittype(url)
+url = sys.argv[1]
+if PY2:
+ _urlopener = urllib._urlopener = urllib.FancyURLopener()
+else:
+ _urlopener = urllib.request._opener = urllib.request.FancyURLopener()
+
+protocol, request = splittype(url)
user, password, port = None, None, None
-host, path = urllib.splithost(request)
+host, path = splithost(request)
if host:
- user, host = urllib.splituser(host)
- if user:
- user, password = urllib.splitpasswd(user)
- host, port = urllib.splitport(host)
- if port: port = int(port)
-path, tag = urllib.splittag(path)
-path, query = urllib.splitquery(path)
-path = urllib.unquote(path)
-if tag: tag = urllib.unquote_plus(tag)
+ user, host = splituser(host)
+ if user:
+ user, password = splitpasswd(user)
+ host, port = splitport(host)
+ if port: port = int(port)
+path, tag = splittag(path)
+path, query = splitquery(path)
+path = unquote(path)
+if tag: tag = unquote_plus(tag)
if query:
- qlist = []
- for name, value in parse_qsl(query):
- qlist.append((name, value))
+ qlist = []
+ for name, value in parse_qsl(query):
+ qlist.append((name, value))
url = protocol + "://"
if user:
- url += urllib.quote(user)
- if password:
- url += ':' + urllib.quote(password)
- url += '@'
+ url += quote(user)
+ if password:
+ url += ':' + quote(password)
+ url += '@'
if host:
- url += host.decode(default_encoding).encode('idna')
- if port:
- url += ':%d' % port
+ if PY2:
+ host = host.decode(default_encoding)
+ host = host.encode('idna')
+ if not PY2:
+ host = host.decode('ascii')
+ url += host
+ if port:
+ url += ':%d' % port
if path:
- if protocol == "file":
- url += urllib.quote(path)
- else:
- url += urllib.quote(path)
+ if protocol == "file":
+ url += quote(path)
+ else:
+ url += quote(path)
if query:
- url += '?' + urllib.urlencode(qlist)
+ url += '?' + urlencode(qlist)
if tag:
- url += '#' + urllib.quote_plus(tag)
+ url += '#' + quote_plus(tag)
# I remember seeing some sites that return broken HTML or even HTTP response
# without "compatible" user agent; I don't know if such sites are still around,
# but this header doesn't cause any harm so I'd better continue to use it.
# UPDATE: I saw a number of sites that forbid "Mozilla compatible"
-server_version = "Python-urllib/%s" % urllib.__version__
-urllib._urlopener.addheaders[0] = ('User-agent', server_version)
-urllib._urlopener.addheaders.append(('Accept-Charset', "koi8-r;q=1.0"))
+if PY2:
+ urllib_version = urllib.__version__
+else:
+ urllib_version = urllib.request.__version__
+server_version = "Python-urllib/%s" % urllib_version
-import os
-dest_file = os.path.basename(url)
+_urlopener.addheaders[0] = ('User-agent', server_version)
+_urlopener.addheaders.append(('Accept-Charset', "koi8-r;q=1.0"))
+dest_file = os.path.basename(url)
if not dest_file:
- dest_file = "_index.html"
+ dest_file = "_index.html"
-filename, headers = urllib.urlretrieve(url, dest_file)
+filename, headers = urlretrieve(url, filename=dest_file)
-if headers.has_key("last-modified"):
- from m_lib.net.www.util import parse_time
- last_modified = parse_time(headers["last-modified"])
- if last_modified:
- os.utime(dest_file, (last_modified, last_modified))
+if "last-modified" in headers:
+ from m_lib.net.www.util import parse_time
+ last_modified = parse_time(headers["last-modified"])
+ if last_modified:
+ os.utime(dest_file, (last_modified, last_modified))