`add_constr` returns broken object when adding constraint without variables

**Describe the bug**
You can pass a trivial constraint, like `xsum([]) == 0` or `xsum([]) == 1`, to a `Model` object's `add_constr` function. This returns an object which is broken in some way (using it causes errors or even segmentation faults).

**To Reproduce**
The simplest code that produces the bug is
```
from mip import *

model = Model()
print(model.add_constr(xsum([])) == 0)
```
This produces an error:
```
Traceback (most recent call last):
File "/home/mees/minimal_example.py", line 4, in
print(model.add_constr(xsum([])) == 0)
File "/home/mees/.local/lib/python3.10/site-packages/mip/model.py", line 339, in add_constr
return self.constrs.add(lin_expr, name, priority)
File "/home/mees/.local/lib/python3.10/site-packages/mip/lists.py", line 153, in add
self.__model.solver.add_constr(lin_expr, name)
File "/home/mees/.local/lib/python3.10/site-packages/mip/cbc.py", line 1429, in add_constr
cbclib.Cbc_addRow(mp, namestr, numnz, self.iidx, self.dvec, sense, rhs)
TypeError: initializer for ctype 'char' must be a bytes of length 1, not bytes
```

Other actions on the object also cause errors. For instance, the following code also fails:
```
from mip import *

model = Model()
constraint = model.add_constr(xsum([]) == 0)
model.remove(constraint)
```
In this case, the error is a segmentation fault, although in context of a larger program I have also seen other errors.

**Expected behavior**
Ideally, these dummy constraints should just be accepted and added to the model, be printable and removable, etc.

If this is not possible due to the nature of the underlying solver(s), the `add_constr` should give a useful error when such a constraint is added to it, rather than return a broken object.

**Desktop (please complete the following information):**
- Operating System, version: Ubuntu 22.04.3 LTS
- Python version: Python 3.10.12
- Python-MIP version (we recommend you to test with the latest version): 1.15.0

**Additional context**
Of course it is unusual to add dummy constraints like the one I did above. In my use case, I am modelling a dynamic logical problem, and regularly want to check for consistency of the full model. In particular, I have lists of binary variables that may grow or shrink during the computation, and regularly want to check if it is still consistent for there to be a variable in some list with value 1. Occasionally these lists are empty, which results in a constraint like `xsum([]) == 1`. It would be convenient if that did not present a special case.