Marc-Andre Lemburg:

* New exported API PyUnicode_Resize()

* The experimental Keep-Alive optimization was turned back
  on after some tweaks to the implementation. It should now
  work without causing core dumps... this has yet to tested
  though (switching it off is easy: see the unicodeobject.c
  file for details).

* Fixed a memory leak in the Unicode freelist cleanup code.

* Added tests to correctly process the return code from
  _PyUnicode_Resize().

* Fixed a bug in the 'ignore' error handling routines
  of some builtin codecs. Added test cases for these to
  test_unicode.py.

Objects/unicodeobject.c

@@ -76,6 +76,7 @@ Unicode Integration Proposal (see file Misc/unicode.txt).
 #ifdef MS_WIN32
 #include <windows.h>
 #endif
+
 /* Limit for the Unicode object free list */
 
 #define MAX_UNICODE_FREELIST_SIZE       1024
@@ -87,18 +88,17 @@ Unicode Integration Proposal (see file Misc/unicode.txt).
    limit. This reduces malloc() overhead for small Unicode objects.  
 
    At worst this will result in MAX_UNICODE_FREELIST_SIZE *
-   (sizeof(PyUnicodeObject) + STAYALIVE_SIZE_LIMIT +
+   (sizeof(PyUnicodeObject) + KEEPALIVE_SIZE_LIMIT +
    malloc()-overhead) bytes of unused garbage.
 
    Setting the limit to 0 effectively turns the feature off.
 
-   XXX The feature is currently turned off because there are
+   Note: This is an experimental feature ! If you get core dumps when
-       apparently some lingering bugs in its implementation which I
+   using Unicode objects, turn this feature off.
-       haven't yet been able to sort out.
 
 */
 
-#define STAYALIVE_SIZE_LIMIT       0
+#define KEEPALIVE_SIZE_LIMIT       9
 
 /* Endianness switches; defaults to little endian */
 
@@ -125,9 +125,9 @@ int _PyUnicode_Resize(register PyUnicodeObject *unicode,
 {
     void *oldstr;
     
-    /* Shortcut if there's nothing to do. */
+    /* Shortcut if there's nothing much to do. */
     if (unicode->length == length)
-	return 0;
+	goto reset;
 
     /* Resizing unicode_empty is not allowed. */
     if (unicode == unicode_empty) {
@@ -148,6 +148,7 @@ int _PyUnicode_Resize(register PyUnicodeObject *unicode,
     unicode->str[length] = 0;
     unicode->length = length;
 
+ reset:
     /* Reset the object caches */
     if (unicode->utf8str) {
         Py_DECREF(unicode->utf8str);
@@ -158,6 +159,23 @@ int _PyUnicode_Resize(register PyUnicodeObject *unicode,
     return 0;
 }
 
+int PyUnicode_Resize(PyObject **unicode,
+		     int length)
+{
+    PyUnicodeObject *v;
+
+    if (unicode == NULL) {
+	PyErr_BadInternalCall();
+	return -1;
+    }
+    v = (PyUnicodeObject *)*unicode;
+    if (v == NULL || !PyUnicode_Check(v) || v->ob_refcnt != 1) {
+	PyErr_BadInternalCall();
+	return -1;
+    }
+    return _PyUnicode_Resize(v, length);
+}
+
 /* We allocate one more byte to make sure the string is
    Ux0000 terminated -- XXX is this needed ? 
 
@@ -185,7 +203,9 @@ PyUnicodeObject *_PyUnicode_New(int length)
         unicode->ob_type = &PyUnicode_Type;
         _Py_NewReference((PyObject *)unicode);
 	if (unicode->str) {
-	    if (unicode->length < length &&
+	    /* Keep-Alive optimization: we only upsize the buffer,
+	       never downsize it. */
+	    if ((unicode->length < length) &&
 		_PyUnicode_Resize(unicode, length)) {
 		free(unicode->str);
 		PyMem_DEL(unicode);
@@ -220,19 +240,25 @@ PyUnicodeObject *_PyUnicode_New(int length)
 static
 void _PyUnicode_Free(register PyUnicodeObject *unicode)
 {
-    Py_XDECREF(unicode->utf8str);
     if (unicode_freelist_size < MAX_UNICODE_FREELIST_SIZE) {
-	if (unicode->length >= STAYALIVE_SIZE_LIMIT) {
+        /* Keep-Alive optimization */
+	if (unicode->length >= KEEPALIVE_SIZE_LIMIT) {
 	    free(unicode->str);
 	    unicode->str = NULL;
 	    unicode->length = 0;
 	}
+	if (unicode->utf8str) {
+	    Py_DECREF(unicode->utf8str);
+	    unicode->utf8str = NULL;
+	}
+	/* Add to free list */
         *(PyUnicodeObject **)unicode = unicode_freelist;
         unicode_freelist = unicode;
         unicode_freelist_size++;
     }
     else {
 	free(unicode->str);
+	Py_XDECREF(unicode->utf8str);
         PyMem_DEL(unicode);
     }
 }
@@ -665,7 +691,8 @@ PyObject *PyUnicode_EncodeUTF8(const Py_UNICODE *s,
         }
     }
     *p = '\0';
-    _PyString_Resize(&v, p - q);
+    if (_PyString_Resize(&v, p - q))
+	goto onError;
 
  done:
     return v;
@@ -1047,7 +1074,8 @@ PyObject *PyUnicode_DecodeUnicodeEscape(const char *s,
             break;
         }
     }
-    _PyUnicode_Resize(v, (int)(p - buf));
+    if (_PyUnicode_Resize(v, (int)(p - buf)))
+	goto onError;
     return (PyObject *)v;
     
  onError:
@@ -1119,9 +1147,14 @@ PyObject *unicodeescape_string(const Py_UNICODE *s,
         *p++ = q[1];
 
     *p = '\0';
-    _PyString_Resize(&repr, p - q);
+    if (_PyString_Resize(&repr, p - q))
+	goto onError;
 
     return repr;
+
+ onError:
+    Py_DECREF(repr);
+    return NULL;
 }
 
 PyObject *PyUnicode_EncodeUnicodeEscape(const Py_UNICODE *s,
@@ -1209,7 +1242,8 @@ PyObject *PyUnicode_DecodeRawUnicodeEscape(const char *s,
 	s += i;
 	*p++ = x;
     }
-    _PyUnicode_Resize(v, (int)(p - buf));
+    if (_PyUnicode_Resize(v, (int)(p - buf)))
+	goto onError;
     return (PyObject *)v;
     
  onError:
@@ -1247,9 +1281,14 @@ PyObject *PyUnicode_EncodeRawUnicodeEscape(const Py_UNICODE *s,
             *p++ = (char) ch;
     }
     *p = '\0';
-    _PyString_Resize(&repr, p - q);
+    if (_PyString_Resize(&repr, p - q))
+	goto onError;
 
     return repr;
+
+ onError:
+    Py_DECREF(repr);
+    return NULL;
 }
 
 PyObject *PyUnicode_AsRawUnicodeEscapeString(PyObject *unicode)
@@ -1305,6 +1344,7 @@ int latin1_encoding_error(const Py_UNICODE **source,
     }
     else if (strcmp(errors,"replace") == 0) {
 	**dest = '?';
+	(*dest)++;
 	return 0;
     }
     else {
@@ -1321,12 +1361,13 @@ PyObject *PyUnicode_EncodeLatin1(const Py_UNICODE *p,
 				 const char *errors)
 {
     PyObject *repr;
-    char *s;
+    char *s, *start;
     repr = PyString_FromStringAndSize(NULL, size);
     if (repr == NULL)
         return NULL;
 
     s = PyString_AS_STRING(repr);
+    start = s;
     while (size-- > 0) {
         Py_UNICODE ch = *p++;
 	if (ch >= 256) {
@@ -1337,6 +1378,10 @@ PyObject *PyUnicode_EncodeLatin1(const Py_UNICODE *p,
 	else
             *s++ = (char)ch;
     }
+    /* Resize if error handling skipped some characters */
+    if (s - start < PyString_GET_SIZE(repr))
+	if (_PyString_Resize(&repr, s - start))
+	    goto onError;
     return repr;
 
  onError:
@@ -1411,8 +1456,9 @@ PyObject *PyUnicode_DecodeASCII(const char *s,
 				      "ordinal not in range(128)"))
 		goto onError;
     }
-    if (p - PyUnicode_AS_UNICODE(v) < size)
+    if (p - PyUnicode_AS_UNICODE(v) < PyString_GET_SIZE(v))
-	_PyUnicode_Resize(v, (int)(p - PyUnicode_AS_UNICODE(v)));	
+	if (_PyUnicode_Resize(v, (int)(p - PyUnicode_AS_UNICODE(v))))
+	    goto onError;
     return (PyObject *)v;
     
  onError:
@@ -1438,6 +1484,7 @@ int ascii_encoding_error(const Py_UNICODE **source,
     }
     else if (strcmp(errors,"replace") == 0) {
 	**dest = '?';
+	(*dest)++;
 	return 0;
     }
     else {
@@ -1454,12 +1501,13 @@ PyObject *PyUnicode_EncodeASCII(const Py_UNICODE *p,
 				const char *errors)
 {
     PyObject *repr;
-    char *s;
+    char *s, *start;
     repr = PyString_FromStringAndSize(NULL, size);
     if (repr == NULL)
         return NULL;
 
     s = PyString_AS_STRING(repr);
+    start = s;
     while (size-- > 0) {
         Py_UNICODE ch = *p++;
 	if (ch >= 128) {
@@ -1470,6 +1518,10 @@ PyObject *PyUnicode_EncodeASCII(const Py_UNICODE *p,
 	else
             *s++ = (char)ch;
     }
+    /* Resize if error handling skipped some characters */
+    if (s - start < PyString_GET_SIZE(repr))
+	if (_PyString_Resize(&repr, s - start))
+	    goto onError;
     return repr;
 
  onError:
@@ -1898,7 +1950,8 @@ PyObject *PyUnicode_TranslateCharmap(const Py_UNICODE *s,
 	Py_DECREF(x);
     }
     if (p - PyUnicode_AS_UNICODE(v) < PyUnicode_GET_SIZE(v))
-	_PyUnicode_Resize(v, (int)(p - PyUnicode_AS_UNICODE(v)));	
+	if (_PyUnicode_Resize(v, (int)(p - PyUnicode_AS_UNICODE(v))))
+	    goto onError;
 
  done:
     return (PyObject *)v;
@@ -1959,7 +2012,7 @@ int PyUnicode_EncodeDecimal(Py_UNICODE *s,
 	    continue;
 	}
 	if (0 < ch && ch < 256) {
-	    *output++ = (char) ch;
+	    *output++ = ch;
 	    continue;
 	}
 	/* All other characters are considered invalid */
@@ -4539,7 +4592,8 @@ PyObject *PyUnicode_Format(PyObject *format,
 	Py_DECREF(args);
     }
     Py_DECREF(uformat);
-    _PyUnicode_Resize(result, reslen - rescnt);
+    if (_PyUnicode_Resize(result, reslen - rescnt))
+	goto onError;
     return (PyObject *)result;
 
  onError:
@@ -4605,6 +4659,9 @@ _PyUnicode_Fini()
     while (u != NULL) {
 	PyUnicodeObject *v = u;
 	u = *(PyUnicodeObject **)u;
+	if (v->str)
+	    free(v->str);
+	Py_XDECREF(v->utf8str);
 	free(v);
     }
     Py_XDECREF(unicode_empty);