[composites] show ghost states for sequence children

Reverts the *new* behaviour introduced in #330 for sequences
with memory. This shouldn't have made it into a patch
release (surprise factor) and there is some debate on what
the best approach is. Refer to the discussion in #328.

CHANGELOG.rst

@@ -4,6 +4,7 @@ Release Notes
 Forthcoming
 -----------
 * [tests] mypy conformance,  `#327 <https://github.com/splintered-reality/py_trees/pull/327>`_
+* [composites] show ghost states for sequence children,  `#330 <https://github.com/splintered-reality/py_trees/pull/330>`_ (reverts behaviour introduced in `#325 <https://github.com/splintered-reality/py_trees/pull/325>`_) 
 
 2.1.5 (2021-05-09)
 ------------------

py_trees/composites.py

@@ -442,10 +442,6 @@ def tick(self):
             self.initialise()
         else:  # self.memory is True and status is RUNNING
             index = self.children.index(self.current_child)
-            # clear out preceding status' - not actually necessary but helps
-            # visualise the case of memory vs no memory
-            for child in itertools.islice(self.children, None, index):
-                child.stop(common.Status.INVALID)
 
         # customised work
         self.update()