Describe the bug

After deploying a stack that uses a stage as it's scope (and not the App), so that it is not defined at the top level of the construct tree, a subsequent deploy of that same stack with all resources removed (including analytics_reporting=False) will recognize the stack as empty and existing, but does not delete it.

Expected Behavior

A previously existing stack deployed with no resources should delete the stack.

Current Behavior

The previously existing stack is not deleted.

Reproduction Steps

#app.py
from aws_cdk import App, Stage, Stack
app = App()
stage = Stage(app, 'Stage')
stack = Stack(stage, 'Stack')

cdk deploy '**'

#changed app.py
from aws_cdk import App, Stage, Stack
app = App()
stage = Stage(app, 'Stage')
stack = Stack(stage, 'Stack', analytics_reporting=False)

cdk deploy '**'

Possible Solution

I suspect the stack.stackName reference during the deploy of an existing empty stack should be stack.node.path instead.

>>> app = App()
>>> top_level_stack = Stack(app, 'TopLevelStack')
>>> stage = Stage(app, 'Stage')
>>> staged_level_stack = Stack(stage, 'StagedLevelStack')
>>> top_level_stack.stack_name; top_level_stack.node.path
'TopLevelStack'
'TopLevelStack'
>>> staged_level_stack.stack_name; staged_level_stack.node.path
'Stage-StagedLevelStack'
'Stage/StagedLevelStack'

Additional Information/Context

No response

CDK CLI Version

2.28.1 (build d035432)

Framework Version

2.28.1

Node.js Version

14.16.0

OS

AL2

Language

Python

Language Version

Python 3.8

Other information

No response