tree 7fe8cc974fb7de4cb32c8098d62d7dd373fb4d27
parent 733afaabd512cbf396d3edf09805d1054a6ec5f8
author C Freeman <cecille@google.com> 1669387594 -0500
committer GitHub <noreply@github.com> 1669387594 -0500
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsBcBAABCAAQBQJjgNVKCRBK7hj4Ov3rIwAAvyQIAHGlnaBIn/DyIVeAzq1DOgQk
 09FetNNN2o1nLT2H0B0vxaq4SZoxd6FM0MjFR1q1f0xz/3mArdlWg92OfkikPzvv
 n/ltnKK8YsiQuVidqgRzqFiiWp4JtzcIbdSrEQd5g5XGgLFBsWyg5ZJrfOmCoCMi
 n9tKRMKCanXcgHhw+yA7Somg9vv8wI94ptyfjuR0cRU+BwRheBMRoGIjdDQu9vAw
 SEJEkczSLTXt9fFhPAL3URpwa86V2dqxKnl9YccRRSaP8fRSGUoizz9sfMrLbrxQ
 hXum6oLJCnxYSPvbIZncn5ZHk9zZkcxdFjB1o/ZbLiWQHfWPCJ0WIZdX5vTBIQM=
 =mOTE
 -----END PGP SIGNATURE-----
 

Fix warning on TestStateMachine (#23744)

This warning was bugging me, so I fixed it. That being said,
this warning does point to a strange circular dependency between
the transitions and the state machine where the state machine owns
transitions, but transitions can post events to the state machine.
It was abstracted out somewhat by the virtual Context interface,
but that still leaves us with this circular dependency. A better
solution would be to fix this circular dependency.

One solution would be to have the transitions table either return
a state or an event, and have the state machine dispatch to itself
rather than having the transitions hold a reference back to the
state machine. Then the transitions would only need to know about
states and events.