DEP: deprecate scalar conversions for arrays with ndim > 0 (#10615)

This PR reflects some of the progress achieved in issue #10404 and is used to asses the impact of the changes. 

With the changes in this PR, `float(numpy.array([1.0])` now gives a warning; likewise some other things:
```python
import numpy

a = numpy.random.rand(10, 1)
a[0] = numpy.array([1.0])       # okay
a[0] = numpy.array(1.0)         # okay
a[0] = 1.0                      # okay

b = numpy.random.rand(10)
b[0] = numpy.array([1.0])       # ValueError: setting an array element with a sequence.
b[0, ...] = numpy.array([1.0])  # okay
b[0] = numpy.array(1.0)         # okay
b[0] = 1.0                      # okay
```
This aligns the behavior of numpy arrays with that of lists:
```python
float([3.14]) 
```
```
TypeError: float() argument must be a string or a number, not 'list'
```
```python
import numpy as np

a = np.random.rand(5)
a[0] = [3.14]
```
```
ValueError: setting an array element with a sequence.
```

Fixes #10404.

1 files changed, 1 insertions, 1 deletions

diff --git a/numpy/core/function_base.py b/numpy/core/function_base.py
index c82b495b1..00e4e6b0e 100644
--- a/numpy/core/function_base.py
+++ b/numpy/core/function_base.py
@@ -168,7 +168,7 @@ def linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None,
     y += start
 
     if endpoint and num > 1:
-        y[-1] = stop
+        y[-1, ...] = stop
 
     if axis != 0:
         y = _nx.moveaxis(y, 0, axis)