You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Frame equalities: a to that uses a named set's function (phase(this) == phase(other)) still gets the frame #506
Problem. Since #498, a to can refer to a named state set through its function, for example undo(this) == undo(old(this)) or phase(this) == phase(other). AuxStateHandler.getMissingStates only counts a set as mentioned when one of its states appears in to. A set used through its function still gets the frame phase(this) == phase(old(this)). For to = "phase(this) == phase(other)" that frame contradicts the spec, unless other happened to be in the same state.
Constraint. Counting every application as a mention is wrong as well. to = "phase(this) == phase(this)" keeps the set today, because the frame is added, and must keep doing so. The rule has to count an application only when it constrains the post-state receiver. That is the same question #504 asks about states.
Split out of #502, whose phase(this) == open form becomes the state shorthand before the frame logic runs and does not depend on this.
Activity