summaryrefslogtreecommitdiff
path: root/Python/codecs.c
diff options
context:
space:
mode:
authorSerhiy Storchaka <storchaka@gmail.com>2016-10-23 09:44:50 +0300
committerSerhiy Storchaka <storchaka@gmail.com>2016-10-23 09:44:50 +0300
commitcb33a01bbc933415efba1c5930ebd9237545f48a (patch)
treebb1366a2cee87b2737c54de7439620bde2e569b6 /Python/codecs.c
parent523c449ca0651ce2d06129671fe18d49946bee99 (diff)
downloadcpython-git-cb33a01bbc933415efba1c5930ebd9237545f48a.tar.gz
Issue #28510: Clean up decoding error handlers.
Since PyUnicodeDecodeError_GetObject() always returns bytes, following PyBytes_AsString() can be replaced with PyBytes_AS_STRING().
Diffstat (limited to 'Python/codecs.c')
-rw-r--r--Python/codecs.c21
1 files changed, 6 insertions, 15 deletions
diff --git a/Python/codecs.c b/Python/codecs.c
index fe57d0dc42..96b3611da2 100644
--- a/Python/codecs.c
+++ b/Python/codecs.c
@@ -867,17 +867,14 @@ PyObject *PyCodec_BackslashReplaceErrors(PyObject *exc)
Py_UCS4 c;
if (PyObject_TypeCheck(exc, (PyTypeObject *)PyExc_UnicodeDecodeError)) {
- unsigned char *p;
+ const unsigned char *p;
if (PyUnicodeDecodeError_GetStart(exc, &start))
return NULL;
if (PyUnicodeDecodeError_GetEnd(exc, &end))
return NULL;
if (!(object = PyUnicodeDecodeError_GetObject(exc)))
return NULL;
- if (!(p = (unsigned char*)PyBytes_AsString(object))) {
- Py_DECREF(object);
- return NULL;
- }
+ p = (const unsigned char*)PyBytes_AS_STRING(object);
res = PyUnicode_New(4 * (end - start), 127);
if (res == NULL) {
Py_DECREF(object);
@@ -1220,7 +1217,7 @@ PyCodec_SurrogatePassErrors(PyObject *exc)
return restuple;
}
else if (PyObject_TypeCheck(exc, (PyTypeObject *)PyExc_UnicodeDecodeError)) {
- unsigned char *p;
+ const unsigned char *p;
Py_UCS4 ch = 0;
if (PyUnicodeDecodeError_GetStart(exc, &start))
return NULL;
@@ -1228,10 +1225,7 @@ PyCodec_SurrogatePassErrors(PyObject *exc)
return NULL;
if (!(object = PyUnicodeDecodeError_GetObject(exc)))
return NULL;
- if (!(p = (unsigned char*)PyBytes_AsString(object))) {
- Py_DECREF(object);
- return NULL;
- }
+ p = (const unsigned char*)PyBytes_AS_STRING(object);
if (!(encode = PyUnicodeDecodeError_GetEncoding(exc))) {
Py_DECREF(object);
return NULL;
@@ -1338,7 +1332,7 @@ PyCodec_SurrogateEscapeErrors(PyObject *exc)
}
else if (PyObject_TypeCheck(exc, (PyTypeObject *)PyExc_UnicodeDecodeError)) {
PyObject *str;
- unsigned char *p;
+ const unsigned char *p;
Py_UCS2 ch[4]; /* decode up to 4 bad bytes. */
int consumed = 0;
if (PyUnicodeDecodeError_GetStart(exc, &start))
@@ -1347,10 +1341,7 @@ PyCodec_SurrogateEscapeErrors(PyObject *exc)
return NULL;
if (!(object = PyUnicodeDecodeError_GetObject(exc)))
return NULL;
- if (!(p = (unsigned char*)PyBytes_AsString(object))) {
- Py_DECREF(object);
- return NULL;
- }
+ p = (const unsigned char*)PyBytes_AS_STRING(object);
while (consumed < 4 && consumed < end-start) {
/* Refuse to escape ASCII bytes. */
if (p[start+consumed] < 128)