Add raw header info to request object (#2032)

sanic/constants.py

@@ -1 +1,2 @@
 HTTP_METHODS = ("GET", "POST", "PUT", "HEAD", "OPTIONS", "PATCH", "DELETE")
+DEFAULT_HTTP_CONTENT_TYPE = "application/octet-stream"

sanic/http.py

@@ -189,8 +189,9 @@ class Http:
 
         # Parse header content
         try:
-            raw_headers = buf[:pos].decode(errors="surrogateescape")
+            head = buf[:pos]
-            reqline, *raw_headers = raw_headers.split("\r\n")
+            raw_headers = head.decode(errors="surrogateescape")
+            reqline, *split_headers = raw_headers.split("\r\n")
             method, self.url, protocol = reqline.split(" ")
 
             if protocol == "HTTP/1.1":
@@ -204,7 +205,7 @@ class Http:
             request_body = False
             headers = []
 
-            for name, value in (h.split(":", 1) for h in raw_headers):
+            for name, value in (h.split(":", 1) for h in split_headers):
                 name, value = h = name.lower(), value.lstrip()
 
                 if name in ("content-length", "transfer-encoding"):
@@ -223,6 +224,7 @@ class Http:
         request = self.protocol.request_class(
             url_bytes=self.url.encode(),
             headers=headers_instance,
+            head=bytes(head),
             version=protocol[5:],
             method=method,
             transport=self.protocol.transport,

sanic/mixins/routes.py

@@ -11,7 +11,7 @@ from urllib.parse import unquote
 from sanic_routing.route import Route  # type: ignore
 
 from sanic.compat import stat_async
-from sanic.constants import HTTP_METHODS
+from sanic.constants import DEFAULT_HTTP_CONTENT_TYPE, HTTP_METHODS
 from sanic.exceptions import (
     ContentRangeError,
     FileNotFound,
@@ -689,7 +689,7 @@ class RouteMixin:
                 content_type = (
                     content_type
                     or guess_type(file_path)[0]
-                    or "application/octet-stream"
+                    or DEFAULT_HTTP_CONTENT_TYPE
                 )
 
                 if "charset=" not in content_type and (

sanic/request.py

@@ -31,6 +31,7 @@ from urllib.parse import parse_qs, parse_qsl, unquote, urlunparse
 from httptools import parse_url  # type: ignore
 
 from sanic.compat import CancelledErrors, Header
+from sanic.constants import DEFAULT_HTTP_CONTENT_TYPE
 from sanic.exceptions import InvalidUsage
 from sanic.headers import (
     Options,
@@ -49,12 +50,6 @@ try:
 except ImportError:
     from json import loads as json_loads  # type: ignore
 
-DEFAULT_HTTP_CONTENT_TYPE = "application/octet-stream"
-
-# HTTP/1.1: https://www.w3.org/Protocols/rfc2616/rfc2616-sec7.html#sec7.2.1
-# > If the media type remains unknown, the recipient SHOULD treat it
-# > as type "application/octet-stream"
-
 
 class RequestParameters(dict):
     """
@@ -95,6 +90,7 @@ class Request:
         "conn_info",
         "ctx",
         "endpoint",
+        "head",
         "headers",
         "method",
         "name",
@@ -121,6 +117,7 @@ class Request:
         method: str,
         transport: TransportProtocol,
         app: Sanic,
+        head: bytes = b"",
     ):
         self.raw_url = url_bytes
         # TODO: Content-Encoding detection
@@ -132,6 +129,7 @@ class Request:
         self.version = version
         self.method = method
         self.transport = transport
+        self.head = head
 
         # Init but do not inhale
         self.body = b""
@@ -207,6 +205,16 @@ class Request:
         if not self.body:
             self.body = b"".join([data async for data in self.stream])
 
+    @property
+    def raw_headers(self):
+        _, headers = self.head.split(b"\r\n", 1)
+        return bytes(headers)
+
+    @property
+    def request_line(self):
+        reqline, _ = self.head.split(b"\r\n", 1)
+        return bytes(reqline)
+
     @property
     def id(self) -> Optional[Union[uuid.UUID, str, int]]:
         """

sanic/response.py

@@ -17,6 +17,7 @@ from urllib.parse import quote_plus
 from warnings import warn
 
 from sanic.compat import Header, open_async
+from sanic.constants import DEFAULT_HTTP_CONTENT_TYPE
 from sanic.cookies import CookieJar
 from sanic.helpers import has_message_body, remove_entity_headers
 from sanic.http import Http
@@ -297,7 +298,7 @@ def raw(
     body: Optional[AnyStr],
     status: int = 200,
     headers: Optional[Dict[str, str]] = None,
-    content_type: str = "application/octet-stream",
+    content_type: str = DEFAULT_HTTP_CONTENT_TYPE,
 ) -> HTTPResponse:
     """
     Returns response object without encoding the body.

tests/test_headers.py

@@ -2,11 +2,9 @@ from unittest.mock import Mock
 
 import pytest
 
-from sanic import Sanic, headers
+from sanic import headers, text
-from sanic.compat import Header
 from sanic.exceptions import PayloadTooLarge
 from sanic.http import Http
-from sanic.request import Request
 
 
 @pytest.mark.parametrize(
@@ -85,3 +83,35 @@ async def test_header_size_exceeded():
 
     with pytest.raises(PayloadTooLarge):
         await http.http1_request_header()
+
+
+def test_raw_headers(app):
+    app.route("/")(lambda _: text(""))
+    request, _ = app.test_client.get(
+        "/",
+        headers={
+            "FOO": "bar",
+            "Host": "example.com",
+            "User-Agent": "Sanic-Testing",
+        },
+    )
+
+    assert request.raw_headers == (
+        b"Host: example.com\r\nAccept: */*\r\nAccept-Encoding: gzip, "
+        b"deflate\r\nConnection: keep-alive\r\nUser-Agent: "
+        b"Sanic-Testing\r\nFOO: bar"
+    )
+
+
+def test_request_line(app):
+    app.route("/")(lambda _: text(""))
+    request, _ = app.test_client.get(
+        "/",
+        headers={
+            "FOO": "bar",
+            "Host": "example.com",
+            "User-Agent": "Sanic-Testing",
+        },
+    )
+
+    assert request.request_line == b"GET / HTTP/1.1"