Fix(Py3): Use `urllib.parse.urlsplit()`

[dotfiles.git] / bin / GET.py

diff --git a/bin/GET.py b/bin/GET.py
index 0e18f5f86e5bca17984227fc8438d78ee723fd3c..2f00d89b2975180179c78681d873d7e66db3d663 100755 (executable)
--- a/bin/GET.py
+++ b/bin/GET.py
@@ -1,49 +1,21 @@
 #! /usr/bin/env python3
 
-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
-
+from urllib.parse import urlsplit, parse_qsl, quote, quote_plus, urlencode
+from urllib.request import urlretrieve
 import os
+import socket
 import sys
-import urllib
-from m_lib.defenc import default_encoding
+import urllib.request
 
 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 = splithost(request)
-if host:
-    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))
+split_results = urlsplit(url)
+protocol, netloc, path, query, tag = split_results
+user = split_results.username
+password = split_results.password
+host = split_results.hostname
+port = split_results.port
+qlist = parse_qsl(query)
 
 url = protocol + "://"
 if user:
@@ -52,11 +24,7 @@ if user:
         url += ':' + quote(password)
     url += '@'
 if host:
-    if PY2:
-        host = host.decode(default_encoding)
-    host = host.encode('idna')
-    if not PY2:
-        host = host.decode('ascii')
+    host = host.encode('idna').decode('ascii')
     url += host
     if port:
         url += ':%d' % port
@@ -74,12 +42,16 @@ if tag:
 # 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"
-if PY2:
-    urllib_version = urllib.__version__
-else:
-    urllib_version = urllib.request.__version__
+urllib_version = urllib.request.__version__
 server_version = "Python-urllib/%s" % urllib_version
 
+
+class MyURLopener(urllib.request.URLopener):
+    def open(self, fullurl, data=None, timeout=socket._GLOBAL_DEFAULT_TIMEOUT):
+        return urllib.request.URLopener.open(self, fullurl, data)
+
+
+_urlopener = urllib.request._opener = MyURLopener()
 _urlopener.addheaders[0] = ('User-agent', server_version)
 _urlopener.addheaders.append(('Accept-Charset', "koi8-r;q=1.0"))