The approach to specifying the null forgiving operator to prevent x!!! and the like causes some problems, and might be better dealt with as "make it valid in the grammar, but prohibit it in semantic text".
TG2: see Nigel's email about "null_forgiving_expression is either wrong or a lot of work for no gain" for more details.
We agree with the approach. @Nigel-Ecma will create a PR, but may ask for suggestions for how we express "even if parenthesized, it's not valid" (e.g. ((x!))!).
The approach to specifying the null forgiving operator to prevent
x!!!
and the like causes some problems, and might be better dealt with as "make it valid in the grammar, but prohibit it in semantic text".TG2: see Nigel's email about "null_forgiving_expression is either wrong or a lot of work for no gain" for more details.