Digest #6 2016-02-07
A.22.214.171.124 suggests an implementation of OF that does not comply with the stack effect given in the specifications of OF; we fell into this trap and used this implementation in Gforth for decades, but eventually our mistake was pointed out. You can find a new implementation that I suggest as reference implementation here; we now use it in Gforth, but it has also been tested and works on SwiftForth and VFX Forth. This implementation also implements ?OF, NEXT-CASE, and CONTOF to allow more general usage of this control structure.
I actually meant 0 ENDCASE. The stack might be empty before, so DUP could cause a stack underflow.
Thanks for the statistics.