dpctl.tensor.astype raises ValueError with empty list as input

[LukichevaPolina]

Reproducer:

import dpctl.tensor as dpt

a = dpt.asarray([])
dpt.astype(a, float)

Output:

Traceback (most recent call last):
  File "/localdisk/work/plukiche/dpnp_fork/dpctl_empty_astype_repro.py", line 4, in <module>
    dpt.astype(a, float)
  File "/localdisk/work/plukiche/miniconda/miniconda3/envs/dpnptc4_internal/lib/python3.8/site-packages/dpctl/tensor/_copy_utils.py", line 447, in astype
    _copy_from_usm_ndarray_to_usm_ndarray(R, usm_ary)
  File "/localdisk/work/plukiche/miniconda/miniconda3/envs/dpnptc4_internal/lib/python3.8/site-packages/dpctl/tensor/_copy_utils.py", line 282, in _copy_from_usm_ndarray_to_usm_ndarray
    copy_same_shape(dst, src)
  File "/localdisk/work/plukiche/miniconda/miniconda3/envs/dpnptc4_internal/lib/python3.8/site-packages/dpctl/tensor/_copy_utils.py", line 239, in copy_same_shape
    copy_same_dtype(dst, src)
  File "/localdisk/work/plukiche/miniconda/miniconda3/envs/dpnptc4_internal/lib/python3.8/site-packages/dpctl/tensor/_copy_utils.py", line 197, in copy_same_dtype
    if _has_memory_overlap(dst, src):
  File "/localdisk/work/plukiche/miniconda/miniconda3/envs/dpnptc4_internal/lib/python3.8/site-packages/dpctl/tensor/_copy_utils.py", line 69, in _has_memory_overlap
    m1 = dpm.as_usm_memory(x1)
  File "dpctl/memory/_memory.pyx", line 736, in dpctl.memory._memory.as_usm_memory
  File "dpctl/memory/_memory.pyx", line 203, in dpctl.memory._memory._Memory._cinit_other
  File "dpctl/memory/_sycl_usm_array_interface_utils.pxi", line 185, in dpctl.memory._memory._USMBufferData.from_sycl_usm_ary_iface
  File "dpctl/memory/_sycl_usm_array_interface_utils.pxi", line 83, in dpctl.memory._memory._pointers_from_shape_and_stride
ValueError: Array shape elements need to be positive

dpctl version is 0.12.0dev1+104.gccd21b68

[LukichevaPolina]

@oleksandr-pavlyk, Could you please look at this problem?

[densmirn]

@oleksandr-pavlyk, could you please either leave a comment or prepare a fix on this issue? The issue is blocking DPNP testing in external CI.

[oleksandr-pavlyk]

The issue is with the logic of copying:

In [5]: import dpctl.tensor as dpt
   ...:
   ...: a = dpt.asarray([])

In [6]: b = dpt.asarray([], 'd')

In [7]: b[...] = a
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)