Booleans validate as integers

[mthebridge]

Yangson considers the Boolean true and false as valid values for integer type nodes.

For example, with this minimal schema:

$ cat test.json 
{
    "ietf-yang-library:modules-state": {
      "module-set-id": "",
      "module": [
        {
          "name": "test",
          "revision": "1.0.0",
          "conformance-type": "implement"
        }
      ]
    }
  }
$ cat test.yang 
module test {
    prefix ;
    revision 1.0.0;
    description "Test schema.";
    leaf number {
        type int32;
    }
}

Then I'd expect the following code to raise an exception, but it parses happily:

from yangson import DataModel
import yaml

data = DataModel.from_file("./test.json")
yaml_input = yaml.safe_load("'test:number': true")
inst = data.from_raw(yaml_input)
inst.validate()

Replacing true with any string or non-integer value triggers a RawTypeError as expected.

[mthebridge]

Suspect

yangson/yangson/datatype.py

Line 708 in acdf40c

def __contains__(self, val: int) -> bool:

is the culprit - confusingly, Python does consider Booleans to be of type Int:

>>> isinstance(True, int)
True

So probably needs some special handling...

[llhotka]

Fixed in de4a558 -> PyPI package 1.3.48

[mthebridge]

Awesome - thanks for the speedy turnaround!