mypy.nodes.PlaceholderNode

Temporary symbol node that will later become a real SymbolNode.

These are only present during semantic analysis when using the new semantic analyzer. These are created if some essential dependencies of a definition are not yet complete.

A typical use is for names imported from a module which is still incomplete (within an import cycle):

from m import f # Initially may create PlaceholderNode

This is particularly important if the imported shadows a name from an enclosing scope or builtins:

from m import int # Placeholder avoids mixups with builtins.int

Another case where this is useful is when there is another definition or assignment:

from m import f def f() -> None: ...

In the above example, the presence of PlaceholderNode allows us to handle the second definition as a redefinition.

They are also used to create PlaceholderType instances for types that refer to incomplete types. Example:

class C(Sequence[C]): ...

We create a PlaceholderNode (with becomes_typeinfo=True) for C so that the type C in Sequence[C] can be bound.

Attributes:

fullname: Full name of of the PlaceholderNode. node: AST node that contains the definition that caused this to

be created. This is useful for tracking order of incomplete definitions and for debugging.

becomes_typeinfo: If True, this refers something that could later

become a TypeInfo. It can't be used with type variables, in particular, as this would cause issues with class type variable detection.

The long-term purpose of placeholder nodes/types is to evolve into something that can support general recursive types.