Currently in many functions boolean argument parsed as int, with 'i' code. When converting this to Argument Clinic we can't write just

    interactive: int = False

because C default becomes "Py_False".

Instead we need explicitly specify c_default.

    interactive: int(c_default="0") = False

It would be good if True/False default values be converted to 1/0 C default values for the int type.

I think this is enough common situation.

I think these shouldn't be "int", they should be "bool".  "bool" will allow you to use a default of False.  It maps to the "p" format unit, which was new in 3.3.

Back before 3.3, when someone wanted a boolean they just used "i", and relied on the fact that True turned into 1 and False turned into 0.  (Even more so before 2.2, when we didn't even have True and False.)  In theory it's a semantic change, because "i" only accepts ints (and True/False), whereas "p" will accept floats, lists, tuples, dicts, sets... anything with a __bool__.  But the intent of code like this is clear, it's only interested in true/false.  And Python has well-established rules for what is considered a true and false.  So I feel like this change is an improvement.