From 4e6c9c4e935458d23add259dc63c5e0a85fba9c8 Mon Sep 17 00:00:00 2001
From: Aldo Cortesi <aldo@nullcube.com>
Date: Tue, 31 May 2016 18:42:56 +1200
Subject: Extract url functions from netlib.utils and move to netlib.http.url

---
 netlib/http/url.py | 95 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 95 insertions(+)
 create mode 100644 netlib/http/url.py

(limited to 'netlib/http/url.py')

diff --git a/netlib/http/url.py b/netlib/http/url.py
new file mode 100644
index 00000000..3961998b
--- /dev/null
+++ b/netlib/http/url.py
@@ -0,0 +1,95 @@
+import six
+from six.moves import urllib
+
+from .. import utils
+
+# PY2 workaround
+def decode_parse_result(result, enc):
+    if hasattr(result, "decode"):
+        return result.decode(enc)
+    else:
+        return urllib.parse.ParseResult(*[x.decode(enc) for x in result])
+
+
+# PY2 workaround
+def encode_parse_result(result, enc):
+    if hasattr(result, "encode"):
+        return result.encode(enc)
+    else:
+        return urllib.parse.ParseResult(*[x.encode(enc) for x in result])
+
+
+def parse_url(url):
+    """
+        URL-parsing function that checks that
+            - port is an integer 0-65535
+            - host is a valid IDNA-encoded hostname with no null-bytes
+            - path is valid ASCII
+
+        Args:
+            A URL (as bytes or as unicode)
+
+        Returns:
+            A (scheme, host, port, path) tuple
+
+        Raises:
+            ValueError, if the URL is not properly formatted.
+    """
+    parsed = urllib.parse.urlparse(url)
+
+    if not parsed.hostname:
+        raise ValueError("No hostname given")
+
+    if isinstance(url, six.binary_type):
+        host = parsed.hostname
+
+        # this should not raise a ValueError,
+        # but we try to be very forgiving here and accept just everything.
+        # decode_parse_result(parsed, "ascii")
+    else:
+        host = parsed.hostname.encode("idna")
+        parsed = encode_parse_result(parsed, "ascii")
+
+    port = parsed.port
+    if not port:
+        port = 443 if parsed.scheme == b"https" else 80
+
+    full_path = urllib.parse.urlunparse(
+        (b"", b"", parsed.path, parsed.params, parsed.query, parsed.fragment)
+    )
+    if not full_path.startswith(b"/"):
+        full_path = b"/" + full_path
+
+    if not utils.is_valid_host(host):
+        raise ValueError("Invalid Host")
+    if not utils.is_valid_port(port):
+        raise ValueError("Invalid Port")
+
+    return parsed.scheme, host, port, full_path
+
+
+def unparse_url(scheme, host, port, path=""):
+    """
+    Returns a URL string, constructed from the specified components.
+
+    Args:
+        All args must be str.
+    """
+    if path == "*":
+        path = ""
+    return "%s://%s%s" % (scheme, utils.hostport(scheme, host, port), path)
+
+
+def urlencode(s):
+    """
+        Takes a list of (key, value) tuples and returns a urlencoded string.
+    """
+    s = [tuple(i) for i in s]
+    return urllib.parse.urlencode(s, False)
+
+
+def urldecode(s):
+    """
+        Takes a urlencoded string and returns a list of (key, value) tuples.
+    """
+    return urllib.parse.parse_qsl(s, keep_blank_values=True)
-- 
cgit v1.2.3


From 6dda2b2ee544c3890f04b7bf99272998e29992b6 Mon Sep 17 00:00:00 2001
From: Aldo Cortesi <aldo@nullcube.com>
Date: Tue, 31 May 2016 18:46:19 +1200
Subject: Module is part of the name - url.decode, not url.urldecode

A pattern we need to use far more often in the codebase
---
 netlib/http/url.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

(limited to 'netlib/http/url.py')

diff --git a/netlib/http/url.py b/netlib/http/url.py
index 3961998b..147ed572 100644
--- a/netlib/http/url.py
+++ b/netlib/http/url.py
@@ -19,7 +19,7 @@ def encode_parse_result(result, enc):
         return urllib.parse.ParseResult(*[x.encode(enc) for x in result])
 
 
-def parse_url(url):
+def parse(url):
     """
         URL-parsing function that checks that
             - port is an integer 0-65535
@@ -68,7 +68,7 @@ def parse_url(url):
     return parsed.scheme, host, port, full_path
 
 
-def unparse_url(scheme, host, port, path=""):
+def unparse(scheme, host, port, path=""):
     """
     Returns a URL string, constructed from the specified components.
 
@@ -80,7 +80,7 @@ def unparse_url(scheme, host, port, path=""):
     return "%s://%s%s" % (scheme, utils.hostport(scheme, host, port), path)
 
 
-def urlencode(s):
+def encode(s):
     """
         Takes a list of (key, value) tuples and returns a urlencoded string.
     """
@@ -88,7 +88,7 @@ def urlencode(s):
     return urllib.parse.urlencode(s, False)
 
 
-def urldecode(s):
+def decode(s):
     """
         Takes a urlencoded string and returns a list of (key, value) tuples.
     """
-- 
cgit v1.2.3


From 40a030f215e1943aefdb2eb6fe2a264b9b1ee33c Mon Sep 17 00:00:00 2001
From: Aldo Cortesi <aldo@nullcube.com>
Date: Tue, 31 May 2016 19:58:28 +1200
Subject: Satisfy flake8

---
 netlib/http/url.py | 1 +
 1 file changed, 1 insertion(+)

(limited to 'netlib/http/url.py')

diff --git a/netlib/http/url.py b/netlib/http/url.py
index 147ed572..8ce28578 100644
--- a/netlib/http/url.py
+++ b/netlib/http/url.py
@@ -3,6 +3,7 @@ from six.moves import urllib
 
 from .. import utils
 
+
 # PY2 workaround
 def decode_parse_result(result, enc):
     if hasattr(result, "decode"):
-- 
cgit v1.2.3