Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!usc!ucsd!ucbvax!MITCH.ENG.SUN.COM!wmb
From: wmb@MITCH.ENG.SUN.COM
Newsgroups: comp.lang.forth
Subject: (none)
Message-ID: <9008271932.AA00162@ucbvax.Berkeley.EDU>
Date: 27 Aug 90 18:29:48 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Reply-To: wmb%MITCH.ENG.SUN.COM@SCFVM.GSFC.NASA.GOV
Organization: The Internet
Lines: 28

Robert Berkey writes:

> However, it's not clear on how to set >IN in BASIS.
>   : <NAME>   ( addr len token -- )
>      BLK @ >R   >IN @ >R   #TIB @ >R   SOURCE-FILE @ >R   \ save-it
>      ROT ROT   ( token addr len)
>      0 BLK !                                     \ prepare BLK
>      0 SOURCE-FILE !                             \ prepare SOURCE-FILE
>      TIB ROT - DUP >IN !   ( token len tib-addr) \ prepare >IN
>      + #TIB !   ( token)                         \ prepare #TIB
>      EXECUTE                                              \ do-it!
>      R> SOURCE-FILE !   R> #TIB !   R> >IN !   R> BLK !   \ restore-it
>      ;
>
> (This is here so that Mitch or anyone else can tell me how to fix or
> improve it.)

It looks good to me.  I don't see any problem.  The basic axioms are
satisfied.  Those are:

	>IN  @ TIB +	is the start of the current input stream.
	#TIB @ TIB +	is the end   of the current input stream.

Due care has been taken to save and restore BLK , SOURCE-FILE , >IN ,
and #TIB , and to set BLK and SOURCE-FILE to 0, so I give this code
a big yes vote.

Mitch