One uses CATCH and THROW in complex programs containing parts whose correct termination is not always guaranteed. With our current knowledge, it would take us too long to develop a suitable test program just for this purpose. I've thought of a trick:
Let's consider our Forth as a complex program that allows its user to do many unexpected things, i.e. generating many THROWs. The complete Forth is put in a CATCH, around which a small program is build that lets us observe what happens.
Purpose: demonstrate THROW and CATCH usage.
Forth file: ec209.frt \ CATCH and THROW (Albert Nijhof) FORTH DEFINITIONS DECIMAL : .DEPTH ( ? -- ) DEPTH 0 .R ." : " DEPTH 0< \ When necessary: IF BEGIN 0 DEPTH 0= UNTIL \ stack repair ." Stack underflow " \ with message. THEN ; : .TOP2 ( ? -- ) \ If possible, show DEPTH 1 > IF OVER . THEN \ top two numbers DEPTH IF DUP . THEN ; \ on stack. : .TOP4 ( ? -- ) DEPTH 4 > IF ." ~ " THEN DEPTH 2 > IF 2>R .TOP2 2R> THEN .TOP2 ; : STATE-SIGN ( - char ) STATE @ IF [CHAR] ] ELSE [CHAR] [ THEN ; : .SITUATION ( -- ) CR .DEPTH .TOP4 STATE-SIGN EMIT SPACE ; CREATE INPUTBUFFER 80 ALLOT : INPUT ( -- addr len ) INPUTBUFFER DUP 80 ACCEPT SPACE ; : RISKY-PROGRAM ( -- ) .SITUATION INPUT EVALUATE ; : EX ( - ) BEGIN ['] RISKY-PROGRAM CATCH DUP IF DUP . ." THROW executes " THEN DROP AGAIN ;