I noticed that the spec for the Annotation type qualifier has a requirement that Annotation cannot be assigned to type[T]( Type qualifiers — typing documentation )
As with most special forms,
Annotatedis not assignable totypeortype[T]
However, the spec for `type[]` ( Special types in annotations — typing documentation ) seems to disagree on “as with most“ part.
The value corresponding to
type[C]must be an actual class object that’s a subtype ofC, not a special form or other kind of type.
Digging a little more on the context of the introduction of additional explanation on Annotated, I found out this discussion thread ( Is `Annotated` compatible with `type[T]`? ), where it is suggested that a value of type[T] should be an instance of type from the builtins ( Is `Annotated` compatible with `type[T]`? - #2 by Jelle )
Out of curiosity, I checked what are instances of type from the typing module, and this is what I found what-are-instance-of-types · GitHub (running on Python 3.14). Interestingly, I found out that Generic is an instance of type, although Generic[T] isn’t. (In fact, mypy accepts Generic as a valid type[T], whereas pyright doesn’t). Similar observation applies to Union (although mypy rejects Unionas a valid type[T])
With that said, I’m seeking clarification on the expected interaction between special forms and type.
On the one hand, I don’t see a particular reason why Annotated should be a special case in the specification and the conformance test suite. It seems possible to replicate this requirement for other special forms and add corresponding conformance test cases.
Moreover, we could further clarify in the specification which special forms are permitted as assignments to type and create positive conformance test cases for them.