python · iii-i · Feb 28, 2024
diff --git a/Include/internal/pycore_global_objects_fini_generated.h b/Include/internal/pycore_global_objects_fini_generated.h
diff --git a/Include/internal/pycore_global_strings.h b/Include/internal/pycore_global_strings.h
@@ -463,6 +463,7 @@ struct _Py_global_strings {
         STRUCT_FOR_ID(globals)
         STRUCT_FOR_ID(groupindex)
         STRUCT_FOR_ID(groups)
+        STRUCT_FOR_ID(gzip_trailer)
         STRUCT_FOR_ID(h)
         STRUCT_FOR_ID(handle)
         STRUCT_FOR_ID(handle_seq)

diff --git a/Include/internal/pycore_runtime_init_generated.h b/Include/internal/pycore_runtime_init_generated.h
diff --git a/Include/internal/pycore_unicodeobject_generated.h b/Include/internal/pycore_unicodeobject_generated.h
diff --git a/Lib/gzip.py b/Lib/gzip.py
@@ -221,7 +221,8 @@ def __init__(self, filename=None, mode=None,
                                              zlib.DEFLATED,
                                              -zlib.MAX_WBITS,
                                              zlib.DEF_MEM_LEVEL,
-                                             0)
+                                             0,
+                                             gzip_trailer=True)
             self._write_mtime = mtime
             self._buffer_size = _WRITE_BUFFER_SIZE
             self._buffer = io.BufferedWriter(_WriteBufferStream(self),
@@ -245,8 +246,6 @@ def __repr__(self):
 
     def _init_write(self, filename):
         self.name = filename
-        self.crc = zlib.crc32(b"")
-        self.size = 0
         self.writebuf = []
         self.bufsize = 0
         self.offset = 0  # Current file offset for seek(), tell(), etc
@@ -310,8 +309,6 @@ def _write_raw(self, data):
 
         if length > 0:
             self.fileobj.write(self.compress.compress(data))
-            self.size += length
-            self.crc = zlib.crc32(data, self.crc)
             self.offset += length
 
         return length
@@ -355,9 +352,6 @@ def close(self):
             if self.mode == WRITE:
                 self._buffer.flush()
                 fileobj.write(self.compress.flush())
-                write32u(fileobj, self.crc)
-                # self.size may exceed 2 GiB, or even 4 GiB
-                write32u(fileobj, self.size & 0xffffffff)
             elif self.mode == READ:
                 self._buffer.close()
         finally:
@@ -611,10 +605,11 @@ def compress(data, compresslevel=_COMPRESS_LEVEL_BEST, *, mtime=None):
         # This is faster and with less overhead.
         return zlib.compress(data, level=compresslevel, wbits=31)
     header = _create_simple_gzip_header(compresslevel, mtime)
-    trailer = struct.pack("<LL", zlib.crc32(data), (len(data) & 0xffffffff))
-    # Wbits=-15 creates a raw deflate block.
-    return (header + zlib.compress(data, level=compresslevel, wbits=-15) +
-            trailer)
+    # Wbits=-15 creates a raw deflate block. Gzip_trailer=True computes CRC32
+    # and writes gzip trailer with zlib, which on some platforms is faster
+    # than doing it manually.
+    return (header + zlib.compress(data, level=compresslevel, wbits=-15,
+                                   gzip_trailer=True))
 
 
 def decompress(data):

@@ -370,7 +370,6 @@ def __init__(self, name, mode, comptype, fileobj, bufsize,
                 except ImportError:
                     raise CompressionError("zlib module is not available") from None
                 self.zlib = zlib
-                self.crc = zlib.crc32(b"")
                 if mode == "r":
                     self.exception = zlib.error
                     self._init_read_gz()
@@ -421,7 +420,8 @@ def _init_write_gz(self, compresslevel):
                                          self.zlib.DEFLATED,
                                          -self.zlib.MAX_WBITS,
                                          self.zlib.DEF_MEM_LEVEL,
-                                         0)
+                                         0,
+                                         gzip_trailer=True)
         timestamp = struct.pack("<L", int(time.time()))
         self.__write(b"\037\213\010\010" + timestamp + b"\002\377")
         if self.name.endswith(".gz"):
@@ -434,8 +434,6 @@ def _init_write_gz(self, compresslevel):
     def write(self, s):
         """Write string s to the stream.
         """
-        if self.comptype == "gz":
-            self.crc = self.zlib.crc32(s, self.crc)
         self.pos += len(s)
         if self.comptype != "tar":
             s = self.cmp.compress(s)
@@ -465,9 +463,6 @@ def close(self):
             if self.mode == "w" and self.buf:
                 self.fileobj.write(self.buf)
                 self.buf = b""
-                if self.comptype == "gz":
-                    self.fileobj.write(struct.pack("<L", self.crc))
-                    self.fileobj.write(struct.pack("<L", self.pos & 0xffffFFFF))
         finally:
             if not self._extfileobj:
                 self.fileobj.close()

diff --git a/Misc/NEWS.d/next/Library/2023-11-17-12-26-47.gh-issue-103477._7cTsK.rst b/Misc/NEWS.d/next/Library/2023-11-17-12-26-47.gh-issue-103477._7cTsK.rst
@@ -0,0 +1 @@
+Write gzip trailer with zlib, improving gzip compression performance on s390x by roughly 40%.
diff --git a/Modules/clinic/zlibmodule.c.h b/Modules/clinic/zlibmodule.c.h
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		Write gzip trailer with zlib, improving gzip compression performance on s390x by roughly 40%.