Include stddef.h and use size_t in dpctl4pybind11

dpctl/apis/include/dpctl4pybind11.hpp

@@ -27,9 +27,9 @@
 
 #include "dpctl_capi.h"
 #include <complex>
-#include <cstddef>
 #include <memory>
 #include <pybind11/pybind11.h>
+#include <stddef.h>
 #include <stdexcept>
 #include <sycl/sycl.hpp>
 #include <utility>
@@ -94,9 +94,9 @@ class dpctl_capi
     void *(*Memory_GetOpaquePointer_)(Py_MemoryObject *);
     DPCTLSyclContextRef (*Memory_GetContextRef_)(Py_MemoryObject *);
     DPCTLSyclQueueRef (*Memory_GetQueueRef_)(Py_MemoryObject *);
-    std::size_t (*Memory_GetNumBytes_)(Py_MemoryObject *);
+    size_t (*Memory_GetNumBytes_)(Py_MemoryObject *);
     PyObject *(*Memory_Make_)(DPCTLSyclUSMRef,
-                              std::size_t,
+                              size_t,
                               DPCTLSyclQueueRef,
                               PyObject *);
 
@@ -126,7 +126,7 @@ class dpctl_capi
                                                   Py_MemoryObject *,
                                                   py::ssize_t,
                                                   char);
-    PyObject *(*UsmNDArray_MakeSimpleFromPtr_)(std::size_t,
+    PyObject *(*UsmNDArray_MakeSimpleFromPtr_)(size_t,
                                                int,
                                                DPCTLSyclUSMRef,
                                                DPCTLSyclQueueRef,
@@ -760,7 +760,7 @@ class usm_memory : public py::object
      *  lifetime of the USM allocation.
      */
     usm_memory(void *usm_ptr,
-               std::size_t nbytes,
+               size_t nbytes,
                const sycl::queue &q,
                std::shared_ptr<void> shptr)
     {
@@ -820,7 +820,7 @@ class usm_memory : public py::object
         return reinterpret_cast<char *>(MRef);
     }
 
-    std::size_t get_nbytes() const
+    size_t get_nbytes() const
     {
         auto const &api = ::dpctl::detail::dpctl_capi::get();
         Py_MemoryObject *mem_obj = reinterpret_cast<Py_MemoryObject *>(m_ptr);
@@ -1259,18 +1259,18 @@ struct ManagedMemory
 
 } // end of namespace detail
 
-template <std::size_t num>
+template <size_t num>
 sycl::event keep_args_alive(sycl::queue &q,
                             const py::object (&py_objs)[num],
                             const std::vector<sycl::event> &depends = {})
 {
-    std::size_t n_objects_held = 0;
+    size_t n_objects_held = 0;
     std::array<std::shared_ptr<py::handle>, num> shp_arr{};
 
-    std::size_t n_usm_owners_held = 0;
+    size_t n_usm_owners_held = 0;
     std::array<std::shared_ptr<void>, num> shp_usm{};
 
-    for (std::size_t i = 0; i < num; ++i) {
+    for (size_t i = 0; i < num; ++i) {
         const auto &py_obj_i = py_objs[i];
         if (detail::ManagedMemory::is_usm_managed_by_shared_ptr(py_obj_i)) {
             const auto &shp =
@@ -1317,7 +1317,7 @@ sycl::event keep_args_alive(sycl::queue &q,
             cgh.host_task([n_objects_held, shp_arr = std::move(shp_arr)]() {
                 py::gil_scoped_acquire acquire;
 
-                for (std::size_t i = 0; i < n_objects_held; ++i) {
+                for (size_t i = 0; i < n_objects_held; ++i) {
                     shp_arr[i]->dec_ref();
                 }
             });
@@ -1329,11 +1329,11 @@ sycl::event keep_args_alive(sycl::queue &q,
 
 /*! @brief Check if all allocation queues are the same as the
     execution queue */
-template <std::size_t num>
+template <size_t num>
 bool queues_are_compatible(const sycl::queue &exec_q,
                            const sycl::queue (&alloc_qs)[num])
 {
-    for (std::size_t i = 0; i < num; ++i) {
+    for (size_t i = 0; i < num; ++i) {
 
         if (exec_q != alloc_qs[i]) {
             return false;
@@ -1344,11 +1344,11 @@ bool queues_are_compatible(const sycl::queue &exec_q,
 
 /*! @brief Check if all allocation queues of usm_ndarays are the same as
     the execution queue */
-template <std::size_t num>
+template <size_t num>
 bool queues_are_compatible(const sycl::queue &exec_q,
                            const ::dpctl::tensor::usm_ndarray (&arrs)[num])
 {
-    for (std::size_t i = 0; i < num; ++i) {
+    for (size_t i = 0; i < num; ++i) {
 
         if (exec_q != arrs[i].get_queue()) {
             return false;