pitrou commented on a change in pull request #8755:
URL: https://github.com/apache/arrow/pull/8755#discussion_r530985982
##########
File path: cpp/src/arrow/python/io.cc
##########
@@ -199,25 +219,32 @@ Result<int64_t> PyReadableFile::Read(int64_t nbytes,
void* out) {
PyObject* bytes_obj = bytes.obj();
DCHECK(bytes_obj != NULL);
- if (!PyBytes_Check(bytes_obj)) {
+ Py_buffer py_buf_;
+ if (!PyObject_GetBuffer(bytes_obj, &py_buf_, PyBUF_ANY_CONTIGUOUS)) {
+ const uint8_t* data = reinterpret_cast<const uint8_t*>(py_buf_.buf);
+ ARROW_CHECK_NE(data, nullptr) << "Null pointer in Py_buffer";
+ std::memcpy(out, data, py_buf_.len);
+ return py_buf_.len;
Review comment:
You must call `PyBuffer_Release`, otherwise you have a memory leak.
##########
File path: cpp/src/arrow/python/io.cc
##########
@@ -199,25 +219,32 @@ Result<int64_t> PyReadableFile::Read(int64_t nbytes,
void* out) {
PyObject* bytes_obj = bytes.obj();
DCHECK(bytes_obj != NULL);
- if (!PyBytes_Check(bytes_obj)) {
+ Py_buffer py_buf_;
Review comment:
Nit, but underscore-ending names are for instance members.
##########
File path: cpp/src/arrow/python/io.cc
##########
@@ -199,25 +219,32 @@ Result<int64_t> PyReadableFile::Read(int64_t nbytes,
void* out) {
PyObject* bytes_obj = bytes.obj();
DCHECK(bytes_obj != NULL);
- if (!PyBytes_Check(bytes_obj)) {
+ Py_buffer py_buf_;
+ if (!PyObject_GetBuffer(bytes_obj, &py_buf_, PyBUF_ANY_CONTIGUOUS)) {
+ const uint8_t* data = reinterpret_cast<const uint8_t*>(py_buf_.buf);
+ ARROW_CHECK_NE(data, nullptr) << "Null pointer in Py_buffer";
Review comment:
Is this check required? It should basically never happen.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
