Merge pull request #1578 from rslinckx/master

http/message: Split request line as bytes to avoid splitting on 0x0A. Fixes #1577

gunicorn/http/message.py

@@ -173,7 +173,7 @@ class Request(Message):
             buf.write(rbuf)
             line, rbuf = self.read_line(unreader, buf, self.limit_request_line)
 
-        self.parse_request_line(bytes_to_str(line))
+        self.parse_request_line(line)
         buf = BytesIO()
         buf.write(rbuf)
 
@@ -301,10 +301,10 @@ class Request(Message):
             "proxy_port": d_port
         }
 
-    def parse_request_line(self, line):
-        bits = line.split(None, 2)
+    def parse_request_line(self, line_bytes):
+        bits = [bytes_to_str(bit) for bit in line_bytes.split(None, 2)]
         if len(bits) != 3:
-            raise InvalidRequestLine(line)
+            raise InvalidRequestLine(bytes_to_str(line_bytes))
 
         # Method
         if not METH_RE.match(bits[0]):
@@ -325,7 +325,7 @@ class Request(Message):
         try:
             parts = urlsplit(self.uri)
         except ValueError:
-            raise InvalidRequestLine(line)
+            raise InvalidRequestLine(bytes_to_str(line_bytes))
         self.path = parts.path or ""
         self.query = parts.query or ""
         self.fragment = parts.fragment or ""

tests/requests/valid/027.http

@@ -0,0 +1,2 @@
+GET /à%20k HTTP/1.0\r\n
+\r\n

tests/requests/valid/027.py

@@ -0,0 +1,8 @@
+request = {
+    "method": "GET",
+    "uri": uri("/\xc3\xa0%20k"),
+    "version": (1, 0),
+    "headers": [
+    ],
+    "body": ''
+}

tests/treq.py

@@ -78,8 +78,11 @@ class request(object):
             yield lines
 
     def send_bytes(self):
-        for d in str(self.data.decode("latin1")):
-            yield bytes(d.encode("latin1"))
+        for d in self.data:
+            if six.PY3:
+                yield bytes([d])
+            else:
+                yield d
 
     def send_random(self):
         maxs = round(len(self.data) / 10)