bpo-42643: Add support for HTTP range requests

Doc/library/http.server.rst

@@ -392,6 +392,10 @@ instantiation, of which this module provides three different variants:
    .. versionchanged:: 3.15
       Added *extra_response_headers* parameter.
 
+   .. versionchanged:: 3.16
+      Added support for HTTP single-part range requests for files,
+      as specified in :rfc:`9110#section-14`.
+
    A lot of the work, such as parsing the request, is done by the base class
    :class:`BaseHTTPRequestHandler`.  This class implements the :func:`do_GET`
    and :func:`do_HEAD` functions.

Doc/whatsnew/3.16.rst

@@ -276,6 +276,13 @@ gzip
   which is passed on to the constructor of the :class:`~gzip.GzipFile` class.
   (Contributed by Marin Misur in :gh:`91372`.)
 
+http.server
+-----------
+
+* :class:`~http.server.SimpleHTTPRequestHandler` now supports HTTP
+  single-part range requests for files, as specified in :rfc:`9110#section-14`.
+  (Contributed by David Bord in :gh:`86809`.)
+
 io
 --
 

Lib/http/server.py

@@ -76,6 +76,7 @@
 import mimetypes
 import os
 import posixpath
+import re
 import shutil
 import socket
 import socketserver
@@ -111,6 +112,9 @@
 """
 
 DEFAULT_ERROR_CONTENT_TYPE = "text/html;charset=utf-8"
+HTTP_BYTES_RANGE_HEADER = re.compile(
+    r"bytes=(\d*)-(\d*)$", re.ASCII | re.IGNORECASE)
+
 
 class HTTPServer(socketserver.TCPServer):
 
@@ -758,7 +762,7 @@ def do_GET(self):
         f = self.send_head()
         if f:
             try:
-                self.copyfile(f, self.wfile)
+                self.copyfile(f, self.wfile, http_range=self._range)
             finally:
                 f.close()
 
@@ -791,6 +795,7 @@ def send_head(self):
         """
         path = self.translate_path(self.path)
         f = None
+        self._range = None
         if os.path.isdir(path):
             parts = urllib.parse.urlsplit(self.path)
             if not parts.path.endswith(('/', '%2f', '%2F')):
@@ -810,6 +815,7 @@ def send_head(self):
                     break
             else:
                 return self.list_directory(path)
+        self._range = self.parse_range()
         ctype = self.guess_type(path)
         # check for trailing "/" which should return 404. See Issue17324
         # The test for this was added in test_httpserver.py
@@ -854,10 +860,44 @@ def send_head(self):
                             self.end_headers()
                             f.close()
                             return None
-
-            self.send_response(HTTPStatus.OK)
+            if self._range:
+                start, end = self._range
+                if start is None and end == 0:
+                    f.close()
+                    self.send_response(HTTPStatus.REQUESTED_RANGE_NOT_SATISFIABLE)
+                    self.send_header("Content-Range", f"bytes */{fs.st_size}")
+                    self.send_header("Content-Length", "0")
+                    self.end_headers()
+                    return None
+                if start is None:
+                    start = max(0, fs.st_size - end)
+                    end = fs.st_size - 1
+                elif end is None or end >= fs.st_size:
+                    end = fs.st_size - 1
+
+                if start == 0 and end >= fs.st_size - 1:
+                    self._range = None
+                elif start >= fs.st_size:
+                    f.close()
+                    self.send_response(HTTPStatus.REQUESTED_RANGE_NOT_SATISFIABLE)
+                    self.send_header("Content-Range", f"bytes */{fs.st_size}")
+                    self.send_header("Content-Length", "0")
+                    self.end_headers()
+                    return None
+                else:
+                    self._range = (start, end)
+
+            if self._range:
+                start, end = self._range
+                self.send_response(HTTPStatus.PARTIAL_CONTENT)
+                self.send_header("Content-Range",
+                                 f"bytes {start}-{end}/{fs.st_size}")
+                self.send_header("Content-Length", str(end - start + 1))
+            else:
+                self.send_response(HTTPStatus.OK)
+                self.send_header("Accept-Ranges", "bytes")
+                self.send_header("Content-Length", str(fs.st_size))
             self.send_header("Content-type", ctype)
-            self.send_header("Content-Length", str(fs[6]))
             self.send_header("Last-Modified",
                 self.date_time_string(fs.st_mtime))
             self._send_extra_response_headers()
@@ -959,21 +999,35 @@ def translate_path(self, path):
             path += '/'
         return path
 
-    def copyfile(self, source, outputfile):
+    def copyfile(self, source, outputfile, *, http_range=None):
         """Copy all data between two file objects.
 
+        If *http_range* is provided, copy only that inclusive byte range.
+
         The SOURCE argument is a file object open for reading
-        (or anything with a read() method) and the DESTINATION
-        argument is a file object open for writing (or
-        anything with a write() method).
+        (or anything with read() and seek() methods) and the DESTINATION
+        argument is a file object open for writing (or anything with a
+        write() method).
 
         The only reason for overriding this would be to change
         the block size or perhaps to replace newlines by CRLF
         -- note however that this the default server uses this
         to copy binary data as well.
 
         """
-        shutil.copyfileobj(source, outputfile)
+        if http_range is None:
+            shutil.copyfileobj(source, outputfile)
+            return
+
+        start, end = http_range
+        length = end - start + 1
+        source.seek(start)
+        while length > 0:
+            buf = source.read(min(length, shutil.COPY_BUFSIZE))
+            if not buf:
+                raise EOFError("File shrank after size was checked")
+            length -= len(buf)
+            outputfile.write(buf)
 
     def guess_type(self, path):
         """Guess the type of a file.
@@ -1000,6 +1054,25 @@ def guess_type(self, path):
             return guess
         return self.default_content_type
 
+    def parse_range(self):
+        """Parse a single-part Range header into an inclusive byte range."""
+        range_header = self.headers.get("Range")
+        if range_header is None:
+            return None
+
+        match = HTTP_BYTES_RANGE_HEADER.fullmatch(range_header)
+        if match is None:
+            return None
+
+        start = int(match.group(1)) if match.group(1) else None
+        end = int(match.group(2)) if match.group(2) else None
+
+        if start is None and end is None:
+            return None
+        if start is not None and end is not None and start > end:
+            return None
+        return start, end
+
 
 nobody = None
 

Lib/test/test_httpservers.py

@@ -786,6 +786,7 @@ def test_get_dir_redirect_location_domain_injection_bug(self):
     def test_get(self):
         #constructs the path relative to the root directory of the HTTPServer
         response = self.request(self.base_url + '/test')
+        self.assertEqual(response.getheader('accept-ranges'), 'bytes')
         self.check_status_and_reason(response, HTTPStatus.OK, data=self.data)
         # check for trailing "/" which should return 404. See Issue17324
         response = self.request(self.base_url + '/test/')
@@ -834,6 +835,81 @@ def test_get(self):
             finally:
                 os.chmod(self.tempdir, 0o755)
 
+    @support.subTests(
+        'range_header,content_range,content_length,start,end',
+        [
+            ('bYtEs=2-5', 'bytes 2-5/30', '4', 2, 6),
+            ('bytes=3-', 'bytes 3-29/30', '27', 3, None),
+            ('bytes=-5', 'bytes 25-29/30', '5', 25, None),
+            ('bytes=29-29', 'bytes 29-29/30', '1', 29, None),
+            ('bytes=25-100', 'bytes 25-29/30', '5', 25, None),
+        ],
+    )
+    def test_single_range_get(self, range_header, content_range,
+                              content_length, start, end):
+        route = self.base_url + '/test'
+        response = self.request(route, headers={'Range': range_header})
+        self.assertEqual(response.getheader('content-range'), content_range)
+        self.assertEqual(response.getheader('content-length'), content_length)
+        self.check_status_and_reason(
+            response, HTTPStatus.PARTIAL_CONTENT, data=self.data[start:end])
+
+    def test_single_range_head(self):
+        response = self.request(
+            self.base_url + '/test', method='HEAD',
+            headers={'Range': 'bytes=2-5'})
+        self.check_status_and_reason(response, HTTPStatus.PARTIAL_CONTENT)
+        self.assertEqual(response.getheader('content-range'), 'bytes 2-5/30')
+        self.assertEqual(response.getheader('content-length'), '4')
+        self.assertEqual(response.read(), b'')
+
+    @support.subTests('range_header', [
+        'bytes=4-3',
+        'bytes=wrong format',
+        'bytes=-',
+        'bytes=--',
+        'bytes=',
+        'bytes=1-2, 4-7',
+    ])
+    def test_invalid_range_get(self, range_header):
+        route = self.base_url + '/test'
+        response = self.request(route, headers={'Range': range_header})
+        self.check_status_and_reason(response, HTTPStatus.OK, data=self.data)
+
+    def test_range_get_ignored_for_directory_listing(self):
+        response = self.request(
+            self.base_url + '/', headers={'Range': 'bytes=0-3'})
+        self.check_status_and_reason(response, HTTPStatus.OK)
+        self.assertIsNone(response.getheader('content-range'))
+
+    @support.subTests('range_header', ['bytes=100-200', 'bytes=-0'])
+    def test_unsatisfiable_range_get(self, range_header):
+        route = self.base_url + '/test'
+        response = self.request(route, headers={'Range': range_header})
+        self.assertEqual(response.getheader('content-range'), 'bytes */30')
+        self.assertEqual(response.getheader('content-length'), '0')
+        self.check_status_and_reason(
+            response, HTTPStatus.REQUESTED_RANGE_NOT_SATISFIABLE)
+
+    @support.subTests('range_header', ['bytes=0-512', 'bytes=-512'])
+    def test_single_range_get_empty(self, range_header):
+        os_helper.create_empty_file(os.path.join(self.tempdir_name, 'empty'))
+        empty_path = self.base_url + '/empty'
+
+        response = self.request(empty_path, headers={'Range': range_header})
+        self.check_status_and_reason(response, HTTPStatus.OK, data=b'')
+
+    @support.subTests('range_header', ['bytes=1-2', 'bytes=-0'])
+    def test_unsatisfiable_range_get_empty(self, range_header):
+        os_helper.create_empty_file(os.path.join(self.tempdir_name, 'empty'))
+        empty_path = self.base_url + '/empty'
+
+        response = self.request(empty_path, headers={'Range': range_header})
+        self.assertEqual(response.getheader('content-range'), 'bytes */0')
+        self.assertEqual(response.getheader('content-length'), '0')
+        self.check_status_and_reason(
+            response, HTTPStatus.REQUESTED_RANGE_NOT_SATISFIABLE)
+
     def test_head(self):
         response = self.request(
             self.base_url + '/test', method='HEAD')

Misc/NEWS.d/next/Library/2026-07-24-13-05-00.gh-issue-86809.vJk93r.rst

@@ -0,0 +1,2 @@
+SimpleHTTPRequestHandler now supports single-part HTTP range requests, as
+specified in RFC 9110 section 14.