How pattern matching works for basic numeric types

[zms9110750]

Pattern matching checks the type. But in pattern matching, different types of numerical types can be mixed.
How should I understand this issue?

float f = 1f;
Console.WriteLine(f is 1);
Console.WriteLine(f is 1f);

by the way.Both of these are true.but 1st generate CA1508 [false]

[HaloFour]

You're matching values, not types, and C# has built-in numeric type coercion. 1 and 1.0 are the same value as far as C# is concerned.

[zms9110750]

How about for this?

object o = f;
Console.WriteLine(o is 1); //false
Console.WriteLine(o is 1f);//true

[HaloFour]

Since the operand is boxed and not a numeric type now it does require a type check and implicit numeric coercion is not applied.

[zms9110750]

So, if it's a value type, then 1, 1L, 1f, 1d, 1m are all the same?
If it is a reference type, will the type be checked first?
Where can I read more detailed rules for this judgment?

Also, I used INumbers for testing, but they don't match me now. Why is that

INumber<float> n = f;
Console.WriteLine(n is 1); //CS0029
Console.WriteLine(n is 1f);//CS9060

[colejohnson66]

INumber<float> is actually a boxed float.

[HaloFour]

Numeric type coercion only works for the built-in primitive numeric types, not custom types nor does it consider implicit cast operators.

[huoyaoyuan]

Pattern matching is typed. They are effectively (object)f is (int)1, (object)f is (float)1.0f, and (float)f is (float)1.

Patterns are interpreted under the type of left hand side value.