Description
Currently self.command_stack is always [] in ClusterPipeline even if there are commands I believe this bug is caused by the changes in #3611 which added in execution strategies (thanks! Excited to use transactions on clusters), so now the commands are actually stored in self._execution_strategy.command_queue instead. I don't think this is actually breaking anything, but I noticed because upgrading to redis-py 6 caused datadog tracing (and probably other tracing libraries) to not work properly since it uses self.command_stack to list the commands being run so everything was showing up blank.
I have a quick fix here https://github.com/mathewlee11/redis-py/pull/1/files but before I put up a PR I wanted to check if that's the right solution, it might make more sense to have the execution strategies manipulate pipeline.command_stack directly instead of maintaining a separate list?