Good to hear.
Still the question, is there a bug in the debugger walking the stack ?
The bug is reliably reproducible, and that stack is reliably reproducible, this is the perfect chance to track this and solve for good.
And the question still there, could you (anyone with Delphi 12) track it and confirm the bug in the debugger, you could use Madshi or EurekaLog and compare the stack, also a breakpoint on TApplication.ProcessMessage and capture the stack at that point, another right before the
exception, then after (it will break on its own), compare them and confirm if there is a failure in stack walking.
Again, i am just saying this is very lucky situation to get things fixed.