Floating point exception in mxnet.contrib.ndarray.arange_like

[DNXie]

Description

Floating point exception in mxnet.contrib.ndarray.arange_like and mxnet.ndarray.contrib.arange_like when repeat is 0.

Error Message

In mxnet 1.6:

Floating point exception (core dumped)

In mxnet nightly version (2.0.0):

terminate called after throwing an instance of 'dmlc::Error'
  what():  [22:27:50] ../src/initialize.cc:379: FloatingPointError: Floating point exception


Aborted (core dumped)

To Reproduce

import mxnet as mx
mx.contrib.ndarray.arange_like(data=mx.nd.ones(1), repeat=0)
mx.ndarray.contrib.arange_like(data=mx.nd.ones(1), repeat=0)

Environment

OS: ubuntu 18.04
Python: 3.7.6
pip: 20.0.2
numpy: 1.18.5
mxnet: 1.6.0

[leezu]

0 shapes are unsupported in the mxnet.nd interface. Please update to MXNet 2 and use mxnet.np (or use the experimental mxnet.np in mxnet 1.x versions).

I'm closing the issue as it's unsupported.

[leezu]

You can refer to #14253 for more information

[DNXie]

@leezu This issue doesn't involve 0 in shape.

[leezu]

What do you expect repeat=0 to do?

[leezu]

We can reopen this issue and track improving the error message. It would be great to have a human-readable message instead of a crash.