Merged revisions 81474 via svnmerge from

svn+ssh://pythondev@svn.python.org/python/branches/py3k

................
  r81474 | victor.stinner | 2010-05-22 18:59:09 +0200 (sam., 22 mai 2010) | 20 lines

  Merged revisions 81471-81472 via svnmerge from
  svn+ssh://pythondev@svn.python.org/python/trunk

  ........
    r81471 | victor.stinner | 2010-05-22 15:37:56 +0200 (sam., 22 mai 2010) | 7 lines

    Issue #6268: More bugfixes about BOM, UTF-16 and UTF-32

     * Fix seek() method of codecs.open(), don't write the BOM twice after seek(0)
     * Fix reset() method of codecs, UTF-16, UTF-32 and StreamWriter classes
     * test_codecs: use "w+" mode instead of "wt+". "t" mode is not supported by
       Solaris or Windows, but does it really exist? I found it the in the issue.
  ........
    r81472 | victor.stinner | 2010-05-22 15:44:25 +0200 (sam., 22 mai 2010) | 4 lines

    Fix my last commit (r81471) about codecs

    Rememder: don't touch the code just before a commit
  ........
................

diff --git a/Lib/codecs.py b/Lib/codecs.py
index 9490602ad3c8520afb0b861f6989b8847b37933b..f6c24481687df4c98469e129cb4d80c9c4ddcc94 100644 (file)
--- a/Lib/codecs.py
+++ b/Lib/codecs.py
@@ -374,6 +374,11 @@ class StreamWriter(Codec):
         """
         pass
 
+    def seek(self, offset, whence=0):
+        self.stream.seek(offset, whence)
+        if whence == 0 and offset == 0:
+            self.reset()
+
     def __getattr__(self, name,
                     getattr=getattr):
 
@@ -606,8 +611,8 @@ class StreamReader(Codec):
 
             Resets the codec buffers used for keeping state.
         """
-        self.reset()
         self.stream.seek(offset, whence)
+        self.reset()
 
     def __next__(self):
 
@@ -700,8 +705,10 @@ class StreamReaderWriter:
         self.writer.reset()
 
     def seek(self, offset, whence=0):
-        self.reader.seek(offset, whence)
-        self.writer.seek(offset, whence)
+        self.stream.seek(offset, whence)
+        self.reader.reset()
+        if whence == 0 and offset == 0:
+            self.writer.reset()
 
     def __getattr__(self, name,
                     getattr=getattr):

diff --git a/Lib/encodings/utf_16.py b/Lib/encodings/utf_16.py
index 5500c0623c916b9831ac18bd57a38fee090c80d9..809bc9ab3687a2ea1a49b143b66b13126e72ca8b 100644 (file)
--- a/Lib/encodings/utf_16.py
+++ b/Lib/encodings/utf_16.py
@@ -103,17 +103,23 @@ class IncrementalDecoder(codecs.BufferedIncrementalDecoder):
 
 class StreamWriter(codecs.StreamWriter):
     def __init__(self, stream, errors='strict'):
-        self.bom_written = False
         codecs.StreamWriter.__init__(self, stream, errors)
+        self.encoder = None
+
+    def reset(self):
+        codecs.StreamWriter.reset(self)
+        self.encoder = None
 
     def encode(self, input, errors='strict'):
-        self.bom_written = True
-        result = codecs.utf_16_encode(input, errors)
-        if sys.byteorder == 'little':
-            self.encode = codecs.utf_16_le_encode
+        if self.encoder is None:
+            result = codecs.utf_16_encode(input, errors)
+            if sys.byteorder == 'little':
+                self.encoder = codecs.utf_16_le_encode
+            else:
+                self.encoder = codecs.utf_16_be_encode
+            return result
         else:
-            self.encode = codecs.utf_16_be_encode
-        return result
+            return self.encoder(input, errors)
 
 class StreamReader(codecs.StreamReader):
 

diff --git a/Lib/encodings/utf_32.py b/Lib/encodings/utf_32.py
index f0b7709a23c7efa7b4ba179b52618aa2fbff8b87..c0529285f60511dd9cfef62560c92b1266e2b3f4 100644 (file)
--- a/Lib/encodings/utf_32.py
+++ b/Lib/encodings/utf_32.py
@@ -98,17 +98,23 @@ class IncrementalDecoder(codecs.BufferedIncrementalDecoder):
 
 class StreamWriter(codecs.StreamWriter):
     def __init__(self, stream, errors='strict'):
-        self.bom_written = False
+        self.encoder = None
         codecs.StreamWriter.__init__(self, stream, errors)
 
+    def reset(self):
+        codecs.StreamWriter.reset(self)
+        self.encoder = None
+
     def encode(self, input, errors='strict'):
-        self.bom_written = True
-        result = codecs.utf_32_encode(input, errors)
-        if sys.byteorder == 'little':
-            self.encode = codecs.utf_32_le_encode
+        if self.encoder is None:
+            result = codecs.utf_32_encode(input, errors)
+            if sys.byteorder == 'little':
+                self.encoder = codecs.utf_32_le_encode
+            else:
+                self.encoder = codecs.utf_32_be_encode
+            return result
         else:
-            self.encode = codecs.utf_32_be_encode
-        return result
+            return self.encoder(input, errors)
 
 class StreamReader(codecs.StreamReader):
 

diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
index 13165723739480fe431f204814570c57b1710415..5d6b5453f550418956388ef1420084077ec0b2c8 100644 (file)
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -1602,8 +1602,8 @@ class BomTest(unittest.TestCase):
                  "utf-32-le",
                  "utf-32-be")
         for encoding in tests:
-            with codecs.open('foo', 'w+', encoding=encoding) as f:
-                # Check if the BOM is written only once
+            # Check if the BOM is written only once
+            with codecs.open(support.TESTFN, 'w+', encoding=encoding) as f:
                 f.write(data)
                 f.write(data)
                 f.seek(0)
@@ -1611,6 +1611,42 @@ class BomTest(unittest.TestCase):
                 f.seek(0)
                 self.assertEquals(f.read(), data * 2)
 
+            # Check that the BOM is written after a seek(0)
+            with codecs.open(support.TESTFN, 'w+', encoding=encoding) as f:
+                f.write(data[0])
+                self.assertNotEquals(f.tell(), 0)
+                f.seek(0)
+                f.write(data)
+                f.seek(0)
+                self.assertEquals(f.read(), data)
+
+            # (StreamWriter) Check that the BOM is written after a seek(0)
+            with codecs.open(support.TESTFN, 'w+', encoding=encoding) as f:
+                f.writer.write(data[0])
+                self.assertNotEquals(f.writer.tell(), 0)
+                f.writer.seek(0)
+                f.writer.write(data)
+                f.seek(0)
+                self.assertEquals(f.read(), data)
+
+            # Check that the BOM is not written after a seek() at a position
+            # different than the start
+            with codecs.open(support.TESTFN, 'w+', encoding=encoding) as f:
+                f.write(data)
+                f.seek(f.tell())
+                f.write(data)
+                f.seek(0)
+                self.assertEquals(f.read(), data * 2)
+
+            # (StreamWriter) Check that the BOM is not written after a seek()
+            # at a position different than the start
+            with codecs.open(support.TESTFN, 'w+', encoding=encoding) as f:
+                f.writer.write(data)
+                f.writer.seek(f.writer.tell())
+                f.writer.write(data)
+                f.seek(0)
+                self.assertEquals(f.read(), data * 2)
+
 
 def test_main():
     support.run_unittest(

diff --git a/Misc/NEWS b/Misc/NEWS
index 54ebf596d7fcec1a467f755212dc1e529ea65d0b..27d16aa42eb616640d7740e62c82a2bed9524bd7 100644 (file)
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -54,8 +54,9 @@ C-API
 Library
 -------
 
-- Issue #6268: Fix seek() method of codecs.open(), don't read the BOM twice
-  after seek(0)
+- Issue #6268: Fix seek() method of codecs.open(), don't read or write the BOM
+  twice after seek(0). Fix also reset() method of codecs, UTF-16, UTF-32 and
+  StreamWriter classes.
 
 - Issue #8782: Add a trailing newline in linecache.updatecache to the last line
   of files without one.