Right now, scoped releases pick up both global (unscoped) changes and changes specific to that scope. Now I write the tests for this, it feels like unexpected behaviour: I think I'd expect specifiying a scope to filter all changes down to just changes tagged with that scope.
We should decide what the correct behaviour is, fix the script (if necessary), and document those changes.
My initial thought is that including changes from multiple scopes (including the global scope) shoudl be an explicit choice. That makes things much harder to get wrong as a user, and increases the flexibility of the tool.
Right now, scoped releases pick up both global (unscoped) changes and changes specific to that scope. Now I write the tests for this, it feels like unexpected behaviour: I think I'd expect specifiying a scope to filter all changes down to just changes tagged with that scope.
We should decide what the correct behaviour is, fix the script (if necessary), and document those changes.
My initial thought is that including changes from multiple scopes (including the global scope) shoudl be an explicit choice. That makes things much harder to get wrong as a user, and increases the flexibility of the tool.