🚀 go-pugleaf

Randy Hudson writes:

> ... end-of-line is treated as a statement
>separator if the statement could end there. That is,
>
>      a := b
>      c := d
>
>is two statements, since "a := b" is a complete statement, but
>
>      a := b +
>           c
>
>is one statement, since a statement can't end with "+".

This is similar to the way BCPL parses programs.

Let's carry it further, though. What about

       a := b c := d

A parser could read in the "c" after the "b", and, instead of issuing
an error message, assume that the previous statement has ended.

I acknowledge that debugging programs would become a lot harder if
this technique was used. Let's stick with semicolons.

Rob