Issue #22982: Improve BOM handling when seeking to multiple positions of a writable text file.

2 files changed, 28 insertions, 11 deletions

diff --git a/Lib/_pyio.py b/Lib/_pyio.py
index 3ed02e4..c0b5fd1 100644
--- a/Lib/_pyio.py
+++ b/Lib/_pyio.py
@@ -1865,6 +1865,19 @@ class TextIOWrapper(TextIOBase):
         return buffer
 
     def seek(self, cookie, whence=0):
+        def _reset_encoder(position):
+            """Reset the encoder (merely useful for proper BOM handling)"""
+            try:
+                encoder = self._encoder or self._get_encoder()
+            except LookupError:
+                # Sometimes the encoder doesn't exist
+                pass
+            else:
+                if position != 0:
+                    encoder.setstate(0)
+                else:
+                    encoder.reset()
+
         if self.closed:
             raise ValueError("tell on closed file")
         if not self._seekable:
@@ -1885,6 +1898,7 @@ class TextIOWrapper(TextIOBase):
             self._snapshot = None
             if self._decoder:
                 self._decoder.reset()
+            _reset_encoder(position)
             return position
         if whence != 0:
             raise ValueError("unsupported whence (%r)" % (whence,))
@@ -1922,17 +1936,7 @@ class TextIOWrapper(TextIOBase):
                 raise OSError("can't restore logical file position")
             self._decoded_chars_used = chars_to_skip
 
-        # Finally, reset the encoder (merely useful for proper BOM handling)
-        try:
-            encoder = self._encoder or self._get_encoder()
-        except LookupError:
-            # Sometimes the encoder doesn't exist
-            pass
-        else:
-            if cookie != 0:
-                encoder.setstate(0)
-            else:
-                encoder.reset()
+        _reset_encoder(cookie)
         return cookie
 
     def read(self, size=None):
diff --git a/Lib/test/test_io.py b/Lib/test/test_io.py
index dfa3d77..ea109ac 100644
--- a/Lib/test/test_io.py
+++ b/Lib/test/test_io.py
@@ -2669,6 +2669,19 @@ class TextIOWrapperTest(unittest.TestCase):
             with self.open(filename, 'rb') as f:
                 self.assertEqual(f.read(), 'bbbzzz'.encode(charset))
 
+    def test_seek_append_bom(self):
+        # Same test, but first seek to the start and then to the end
+        filename = support.TESTFN
+        for charset in ('utf-8-sig', 'utf-16', 'utf-32'):
+            with self.open(filename, 'w', encoding=charset) as f:
+                f.write('aaa')
+            with self.open(filename, 'a', encoding=charset) as f:
+                f.seek(0)
+                f.seek(0, self.SEEK_END)
+                f.write('xxx')
+            with self.open(filename, 'rb') as f:
+                self.assertEqual(f.read(), 'aaaxxx'.encode(charset))
+
     def test_errors_property(self):
         with self.open(support.TESTFN, "w") as f:
             self.assertEqual(f.errors, "strict")