From: cloos@acsu.buffalo.edu (James H. Cloos)
Newsgroups: comp.sys.handhelds,comp.sys.hp48
Subject: Small error in RPLMAN.DOC
Date: 27 Jul 91 06:21:34 GMT
Organization: State University of New York @ Buffalo

While working on my current project, I discovered an error in the
RPLMAN.DOC file that comes with the new development system.  In the doc
file, the effect of the word NDUPN is listed as:
( ob #n --> ob ... ob )
but this does not happen.  A better description would be:
( ob #n --> ob(1) ... ob(n) #n )
( ob #0 --> #0 )
where the parenthesized numbers indicate copies of the same ob, not
different ob's.  Note that this effect can be summarized by saying:
The level 2 argument is duplicated until n copies exist (not counting any
previously in levels 3+, of course), where n is the value of the level 1
short.  Said short is left on the stack.  If the short is 0, then 0 copies
of the level 2 argument are left on the stack; ie., it is dropped.

The word looks like:

:: DUP#0=case SWAPDROP DUP #1- DUP#0=case DROP ZERO_DO (DO) OVER SWAP LOOP ;

which documents my revised description.

Note that this also affects the docs in the chapter on local variables and
local environments.  There the code example:

NULLLAM TWENTYFOUR NDUPN TWENTYFOUR {}N BIND

must be changed to:

' NULLLAM TWENTYFOUR NDUPN {}N BIND

No flames, just a correction.

-JimC
--
James H. Cloos, Jr.		Phone:  +1 716 673-1250
cloos@ACSU.Buffalo.EDU		Snail:  PersonalZipCode:  14048-0772, USA
cloos@ub.UUCP			Note:	Phone & Snail to change soon!