+When we find `TK_in` and `TK_out` tokens which report indents we need
+to handle them directly as the grammar cannot express what we want to
+do with them.
+
+`TK_in` tokens are easy: we simply update the `next` stack frame to
+record how many indents there are and that the next token started with
+an indent.
+
+`TK_out` tokens must either be counted off against any pending indent,
+or must force reductions until there is a pending indent which isn't
+at the start of a production.