Bug/6294 zero bytes files are chunked

src/requests/models.py

@@ -545,7 +545,7 @@ def prepare_body(self, data, files, json=None):
                     "Streamed bodies and files are mutually exclusive."
                 )
 
-            if length:
+            if length is not None:
                 self.headers["Content-Length"] = builtin_str(length)
             else:
                 self.headers["Transfer-Encoding"] = "chunked"
@@ -573,7 +573,7 @@ def prepare_content_length(self, body):
         """Prepare Content-Length header based on request method and body"""
         if body is not None:
             length = super_len(body)
-            if length:
+            if length is not None:
                 # If length exists, set it. Otherwise, we fallback
                 # to Transfer-Encoding: chunked.
                 self.headers["Content-Length"] = builtin_str(length)

src/requests/utils.py

@@ -12,6 +12,7 @@
 import os
 import re
 import socket
+import stat
 import struct
 import sys
 import tempfile
@@ -131,7 +132,14 @@ def dict_to_sequence(d):
 
 
 def super_len(o):
+    """Returns the length of the object or None if the length cannot be measured.
+
+    Tries looking for length attributes, file handles and seek/tell in order
+    to figure out the object length.  If the length cannot be found, None
+    is returned.
+    """
     total_length = None
+    length_undefined = False
     current_position = 0
 
     if isinstance(o, str):
@@ -146,11 +154,18 @@ def super_len(o):
     elif hasattr(o, "fileno"):
         try:
             fileno = o.fileno()
+            if not stat.S_ISREG(os.fstat(fileno).st_mode):
+                raise BufferError("Cannot tell size of non regular file")
         except (io.UnsupportedOperation, AttributeError):
             # AttributeError is a surprising exception, seeing as how we've just checked
             # that `hasattr(o, 'fileno')`.  It happens for objects obtained via
             # `Tarfile.extractfile()`, per issue 5229.
             pass
+        except BufferError:
+            # Telling size of non regular files does is not realiable
+            # if it is a socket or a pipe, they need to be handled as
+            # streams with no known end.
+            length_undefined = True
         else:
             total_length = os.fstat(fileno).st_size
 
@@ -169,7 +184,13 @@ def super_len(o):
                     FileModeWarning,
                 )
 
-    if hasattr(o, "tell"):
+    if hasattr(o, "tell") and not length_undefined:
+        # pipes have tell/seek, tell() will fail on Linux/MacOS,
+        # but not in windows, causing super_len to report the
+        # length as the number of bytes currently written to
+        # the pipe, which will be wrong if the pipe is continuously
+        # being written to. The length_underfined check will
+        # prevent trying to use tell/seek for a pipe.
         try:
             current_position = o.tell()
         except OSError:
@@ -191,10 +212,10 @@ def super_len(o):
                     # partially read file-like objects
                     o.seek(current_position or 0)
                 except OSError:
-                    total_length = 0
+                    total_length = None
 
     if total_length is None:
-        total_length = 0
+        return total_length
 
     return max(0, total_length - current_position)
 

tests/test_requests.py

@@ -130,6 +130,22 @@ def test_empty_content_length(self, httpbin, method):
         req = requests.Request(method, httpbin(method.lower()), data="").prepare()
         assert req.headers["Content-Length"] == "0"
 
+    @pytest.mark.parametrize("method", ("POST", "PUT", "PATCH", "OPTIONS"))
+    def test_empty_file_content_length(self, httpbin, method):
+        data = io.BytesIO(b"")
+        req = requests.Request(method, httpbin(method.lower()), data=data).prepare()
+        assert req.headers["Content-Length"] == "0"
+
+    @pytest.mark.parametrize("method", ("POST", "PUT", "PATCH", "OPTIONS"))
+    def test_not_readable_content_length_pipe(self, httpbin, method):
+        pipe_r, pipe_w = os.pipe()
+        pipe_rf = os.fdopen(pipe_r, "rb")
+        pipe_wf = os.fdopen(pipe_w, "wb")
+        pipe_wf.write(b"hello")
+        req = requests.Request(method, httpbin(method.lower()), data=pipe_rf).prepare()
+        assert req.headers.get("Content-Length") is None
+        assert req.headers["Transfer-Encoding"] == "chunked"
+
     def test_override_content_length(self, httpbin):
         headers = {"Content-Length": "not zero"}
         r = requests.Request("POST", httpbin("post"), headers=headers).prepare()
@@ -2151,7 +2167,9 @@ def test_response_without_release_conn(self):
         resp.close()
         assert resp.raw.closed
 
-    def test_empty_stream_with_auth_does_not_set_content_length_header(self, httpbin):
+    def test_empty_stream_with_auth_does_not_set_transfer_encoding_header(
+        self, httpbin
+    ):
         """Ensure that a byte stream with size 0 will not set both a Content-Length
         and Transfer-Encoding header.
         """
@@ -2160,8 +2178,8 @@ def test_empty_stream_with_auth_does_not_set_content_length_header(self, httpbin
         file_obj = io.BytesIO(b"")
         r = requests.Request("POST", url, auth=auth, data=file_obj)
         prepared_request = r.prepare()
-        assert "Transfer-Encoding" in prepared_request.headers
-        assert "Content-Length" not in prepared_request.headers
+        assert "Transfer-Encoding" not in prepared_request.headers
+        assert "Content-Length" in prepared_request.headers
 
     def test_stream_with_auth_does_not_set_transfer_encoding_header(self, httpbin):
         """Ensure that a byte stream with size > 0 will not set both a Content-Length

tests/test_utils.py

@@ -92,7 +92,7 @@ def tell(self):
             def seek(self, offset, whence):
                 pass
 
-        assert super_len(NoLenBoomFile()) == 0
+        assert super_len(NoLenBoomFile()) is None
 
     def test_string(self):
         assert super_len("Test") == 4
@@ -148,8 +148,14 @@ def test_super_len_with_fileno(self):
         assert length == len(file_data)
 
     def test_super_len_with_no_matches(self):
-        """Ensure that objects without any length methods default to 0"""
-        assert super_len(object()) == 0
+        """Ensure that objects without any length methods default to None"""
+        assert super_len(object()) is None
+
+    def test_super_len_with_pipe(self):
+        """Ensure that ojects with a fileno that are not regular files default to length None"""
+        r, w = os.pipe()
+        rf = os.fdopen(r, "rb")
+        assert super_len(rf) is None
 
 
 class TestToKeyValList: