Prefer to use ellipsis over pass (#2292)

stdlib/2/_json.pyi

@@ -12,8 +12,6 @@ def scanstring(a, b, *args, **kwargs) -> tuple:
     raise TypeError()
 
 
-class Encoder(object):
-    pass
+class Encoder(object): ...
 
-class Scanner(object):
-    pass
+class Scanner(object): ...

stdlib/2/io.pyi

@@ -33,13 +33,10 @@ SEEK_END = ...  # type: int
 
 class IOBase(_io._IOBase): ...
 
-class RawIOBase(_io._RawIOBase, IOBase):
-    pass
+class RawIOBase(_io._RawIOBase, IOBase): ...
 
-class BufferedIOBase(_io._BufferedIOBase, IOBase):
-    pass
+class BufferedIOBase(_io._BufferedIOBase, IOBase): ...
 
 # Note: In the actual io.py, TextIOBase subclasses IOBase.
 # (Which we don't do here because we don't want to subclass both TextIO and BinaryIO.)
-class TextIOBase(_io._TextIOBase):
-    pass
+class TextIOBase(_io._TextIOBase): ...

stdlib/2/thread.pyi

@@ -15,11 +15,8 @@ class LockType:
     def __enter__(self) -> LockType: ...
     def __exit__(self, typ: Any, value: Any, traceback: Any) -> None: ...
 
-class _local(object):
-    pass
-
-class _localdummy(object):
-    pass
+class _local(object): ...
+class _localdummy(object): ...
 
 def start_new(function: Callable[..., Any], args: Any, kwargs: Any = ...) -> int: ...
 def start_new_thread(function: Callable[..., Any], args: Any, kwargs: Any = ...) -> int: ...

stdlib/2/tokenize.pyi

@@ -122,11 +122,9 @@ def tokenize(readline: Callable[[], str], tokeneater: Callable[[Tuple[int, str,
 def tokenize_loop(readline: Callable[[], str], tokeneater: Callable[[Tuple[int, str, _Pos, _Pos, str]], None]) -> None: ...
 def untokenize(iterable: Iterable[_TokenType]) -> str: ...
 
-class StopTokenizing(Exception):
-    pass
+class StopTokenizing(Exception): ...
 
-class TokenError(Exception):
-    pass
+class TokenError(Exception): ...
 
 class Untokenizer:
     prev_col = ...  # type: int