gh-129005: Remove copies from _pyio using take_bytes (#141539)

Memory usage now matches that of _io for large files.

diff --git a/Lib/_pyio.py b/Lib/_pyio.py
index 423178e87a868441afc1912aafc414342b3dd227..69a088df8fc9878b5c12233fe2174e5c7046c3b6 100644 (file)
--- a/Lib/_pyio.py
+++ b/Lib/_pyio.py
@@ -546,7 +546,7 @@ class IOBase(metaclass=abc.ABCMeta):
             res += b
             if res.endswith(b"\n"):
                 break
-        return bytes(res)
+        return res.take_bytes()
 
     def __iter__(self):
         self._checkClosed()
@@ -620,7 +620,7 @@ class RawIOBase(IOBase):
         if n < 0 or n > len(b):
             raise ValueError(f"readinto returned {n} outside buffer size {len(b)}")
         del b[n:]
-        return bytes(b)
+        return b.take_bytes()
 
     def readall(self):
         """Read until EOF, using multiple read() call."""
@@ -628,7 +628,7 @@ class RawIOBase(IOBase):
         while data := self.read(DEFAULT_BUFFER_SIZE):
             res += data
         if res:
-            return bytes(res)
+            return res.take_bytes()
         else:
             # b'' or None
             return data
@@ -1738,7 +1738,7 @@ class FileIO(RawIOBase):
         assert len(result) - bytes_read >= 1, \
             "os.readinto buffer size 0 will result in erroneous EOF / returns 0"
         result.resize(bytes_read)
-        return bytes(result)
+        return result.take_bytes()
 
     def readinto(self, buffer):
         """Same as RawIOBase.readinto()."""

diff --git a/Lib/test/test_io/test_bufferedio.py b/Lib/test/test_io/test_bufferedio.py
index 30c34e818b1572845a4ad28411e370ec7b9dab63..3278665bdc9dd3d7583a395e039beda10fc7ca06 100644 (file)
--- a/Lib/test/test_io/test_bufferedio.py
+++ b/Lib/test/test_io/test_bufferedio.py
@@ -1277,7 +1277,8 @@ class BufferedRandomTest(BufferedReaderTest, BufferedWriterTest):
         def _readinto(bufio, n=-1):
             b = bytearray(n if n >= 0 else 9999)
             n = bufio.readinto(b)
-            return bytes(b[:n])
+            b.resize(n)
+            return b.take_bytes()
         self.check_flush_and_read(_readinto)
 
     def test_flush_and_peek(self):

diff --git a/Lib/test/test_io/test_largefile.py b/Lib/test/test_io/test_largefile.py
index 41f7b70e5cfe8137a8e28e2f9207966f32558af4..438a90a92ed58894692d43d11da54c0acecd7495 100644 (file)
--- a/Lib/test/test_io/test_largefile.py
+++ b/Lib/test/test_io/test_largefile.py
@@ -56,9 +56,7 @@ class TestFileMethods(LargeFileTest):
     (i.e. > 2 GiB) files.
     """
 
-    # _pyio.FileIO.readall() uses a temporary bytearray then casted to bytes,
-    # so memuse=2 is needed
-    @bigmemtest(size=size, memuse=2, dry_run=False)
+    @bigmemtest(size=size, memuse=1, dry_run=False)
     def test_large_read(self, _size):
         # bpo-24658: Test that a read greater than 2GB does not fail.
         with self.open(TESTFN, "rb") as f:
@@ -154,7 +152,7 @@ class TestFileMethods(LargeFileTest):
                 f.seek(pos)
                 self.assertTrue(f.seekable())
 
-    @bigmemtest(size=size, memuse=2, dry_run=False)
+    @bigmemtest(size=size, memuse=1, dry_run=False)
     def test_seek_readall(self, _size):
         # Seek which doesn't change position should readall successfully.
         with self.open(TESTFN, 'rb') as f: