ADDITIONAL SYSTEM INFORMATION :
macOS Monterey, 12.3.1

A DESCRIPTION OF THE PROBLEM :
when calling a static method foo(..) parameterized by a type parameter U, if the method foo takes two formal parameters of generic type U, one can call foo("2", 1);

I was under the impression one of the main reasons for adding generic types was to preclude behavior like this?

REGRESSION : Last worked in version 8

STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
Here is a minimal example:
    public static <U> U foo(U x, U y) {
        return y != null ? y : x;
    }

    public static void main(String[] args) {
        Integer a = 1;
        String b = "x";
        foo(a, b); //<-- U first gets instantiated by Integer, but then a String is supplied.
    }


EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
Shouldn't the call to foo(..) in main represent a compile time bug caught by the typechecker? Shouldn't both arguments to foo be required to be the same type (since both parameters were given U as their type)?
ACTUAL -
program compiles fine and without issue. It's only when one does this:
Integer m = foo(a, b);

does the typechecker pick it up as an issue. But if foo(..) returned a boolean or void, then the code would compile just fine and the bug might go unseen.

Instantiating the type parameter manually makes the typechecker catch it: Integer m = Main.<Integer>foo(a, b);

---------- BEGIN SOURCE ----------
see steps to reproduce
---------- END SOURCE ----------

FREQUENCY : rarely