Let's make post_mortem support passing an exception alone, like pdb.post_mortem does. Also, add an explanation in the README.md

Author: nshepperd

README.md

@@ -100,6 +100,9 @@ while True:
     ...
 ```
 
+### `post_mortem()` function
+Debugpy supports a `post_mortem()` function similar to `pdb.post_mortem()`. Call `debugpy.post_mortem(e)` with an exception or `(type(e),e,e.__traceback__)` tuple; or with no arguments in an except block. If the debugger is attached, it will pause execution and start post-mortem debugging of the exception stack as-if an uncaught exception. This respects breakpoint filters set by the debugger by default. When resuming afterward, the program will continue executing as normal (including unwinding the stack further if post_mortem() was invoked in a context manager's `__exit__` for example, or the exception is re-raised). If there's no client attached, this function does nothing, as breakpoint().
+
 ## Debugger logging
 
 To enable debugger internal logging via CLI, the `--log-to` switch can be used:

src/debugpy/public_api.py

@@ -214,7 +214,7 @@ def trace_this_thread(__should_trace: bool):
 
 @_api()
 def post_mortem(
-    __excinfo: typing.Tuple[type, BaseException, typing.Any] | None = None,
+    __excinfo: typing.Tuple[type, BaseException, typing.Any] | BaseException | None = None,
     as_uncaught: bool = True,
 ) -> None:
     """Stops the debugger on an unhandled exception.

src/debugpy/server/api.py

@@ -372,6 +372,9 @@ def post_mortem(excinfo=None, as_uncaught=True):
     if excinfo is None:
         excinfo = sys.exc_info()
 
+    if isinstance(excinfo, BaseException):
+        excinfo = (type(excinfo), excinfo, excinfo.__traceback__)
+
     exctype, value, tb = excinfo
     if exctype is None or value is None or tb is None:
         log.debug("postmortem() ignored - no exception info")