Can you provide some more context? In which situations is type(int) treated as int?

type should act as the identity operators only when applied to typedescs (which may be surprising to some purists expecting even higher-orders types-of-types).

@zah: Interestingly, the main issue I was having (f(x: int) and f(x: typedesc[int]) being confused during overload resolution) seems to have been fixed in the last release or two.

Though the following program still compiles (and it shouldn't):

static:
  assert int is type(int)
  assert type(int) is int

see also test case here:

• type(int).name should be type int, not int, to avoid confusion with int [TODO] type(int).name should be type int, not int, to avoid confusion with int #8705

Because it is unclear if type(x) means typedesc[x] or typeof(x) the short syntax is discouraged to use. A deprecation message is not emitted yet though. But type(int) is certainly not the identity operator.