BUG: optimize: using `integrality` prevents `linprog` from accepting a list for `c`

Describe your issue.

If integrality is used in linprog, then c must be a numpy array. Otherwise an exception is raised because c does not have a shape attribute.

The fix is probably as simple as passing c through np.asarray() first.

Reproducing Code Example

In [35]: from scipy.optimize import linprog

In [36]: linprog([1, 1], A_ub=[[-1, 0]], b_ub=[-2.5], integrality=[1, 1])
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In [36], line 1
----> 1 linprog([1, 1], A_ub=[[-1, 0]], b_ub=[-2.5], integrality=[1, 1])

File ~/repos/git/forks/scipy/build-install/lib/python3.10/site-packages/scipy/optimize/_linprog.py:623, in linprog(c, A_ub, b_ub, A_eq, b_eq, bounds, method, callback, options, x0, integrality)
    621     warn(warning_message, OptimizeWarning)
    622 elif np.any(integrality):
--> 623     integrality = np.broadcast_to(integrality, c.shape)
    625 lp = _LPProblem(c, A_ub, b_ub, A_eq, b_eq, bounds, x0, integrality)
    626 lp, solver_options = _parse_linprog(lp, options, meth)

AttributeError: 'list' object has no attribute 'shape'

Error message

See above.

SciPy/NumPy/Python version information

1.10.0.dev0+2184.5057eff 1.24.0.dev0+1048.g939910660 sys.version_info(major=3, minor=10, micro=8, releaselevel=‘final’, serial=0)

[jgborch]

@nboudrie and I will be taking on this issue.

[WarrenWeckesser]

This was fixed by https://github.com/scipy/scipy/pull/17513.