In 17.2, we observe that some implementations may evaluate expressions lazily and others eagerly and this may have effect whether or not a test raises a dynamic error. But we don't call this out as an explicit, implementation-defined feature.
I think we should. Users should know what behavior a processor uses.
In 17.2, we observe that some implementations may evaluate expressions lazily and others eagerly and this may have effect whether or not a test raises a dynamic error. But we don't call this out as an explicit, implementation-defined feature.
I think we should. Users should know what behavior a processor uses.