Path: utzoo!utgpu!watserv1!watmath!att!dptg!ulysses!andante!mit-eddie!uw-beaver!cornell!batcomputer!rpi!zaphod.mps.ohio-state.edu!usc!ucsd!ucbvax!MITCH.ENG.SUN.COM!wmb
From: wmb@MITCH.ENG.SUN.COM
Newsgroups: comp.lang.forth
Subject: Re: Safety Nets
Message-ID: <9007161525.AA14635@ucbvax.Berkeley.EDU>
Date: 16 Jul 90 03:28:15 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Reply-To: wmb%MITCH.ENG.SUN.COM@SCFVM.GSFC.NASA.GOV
Organization: The Internet
Lines: 27

> NMORGENSTERN writes:
>
> : NET CREATE 0 , 0 , 0 , ;
> : 3!  ( n1 n2 n3 a -- ) 2DUP ! 2+ NIP 2! ;
> : 3@  ( a -- n1 n2 n3) DUP 2+ 2@ ROT @ ;
> : SET-NET ( net -- ) R@ SP@ ROT  RP@ SWAP 3! ;
> : FALL ( net -- ) 2+ RP! SWAP >R 4 + SP! ;
>
>  For example: To declare a net:
>     NET SAFETY1 To set the net
>      ...  SAFETY1 SET-NET ( point A ) FOO FAH ....

This is similar to, but slightly weaker than, C's setjmp()/longjmp() .

jmp_buf safety1;	  <==>		NET SAFETY1
setjmp(safety1)		 ~<==>		SAFETY1 SET-NET
longjmp(safety1);	  <==>		SAFETY1 FALL

The reason it is weaker is because setjmp() returns a flag indicating
whether it returned normally or returned due to a longjmp().  This
feature could be added simply by adding TRUE to the end or SET-NET
and FALSE to the end of FALL  (or vice versa; either way works).

The relative merits of CATCH/THROW vs. setjmp()/longjmp() have already
been debated on this newsgroup.

Mitch Bradley