From @mail.uunet.ca:beejay@micor Wed Aug  5 14:15:03 1992
Received: from calvin.sfasu.edu by umaxc.weeg.uiowa.edu (5.61.jnf/920629)
	  on Wed, 5 Aug 92 14:14:59 -0500 id AA22447 with SMTP 
Received: from mail.uunet.ca (uunet.ca) by calvin.sfasu.EDU with SMTP (5.59/25-eef)
	id AA03754; Wed, 5 Aug 92 13:52:09 CDT
Return-Path: <@mail.uunet.ca:beejay@micor>
Received: from ocunix by mail.uunet.ca with UUCP id <10563>; Wed, 5 Aug 1992 15:01:34 -0400
Received: by micor.OCUnix.On.Ca (smail2.5)
	id AA28005; 5 Aug 92 11:12:21 EDT (Wed)
To: hyperc-l@calvin.sfasu.edu
Subject: HyperC hint
Date: 	Wed, 5 Aug 1992 07:12:21 -0400
Message-Id: <9208051112.AA28003@micor.OCUnix.On.Ca>
From: beejay@micor.OCUnix.On.Ca (Basil Johnson)
Status: R

 
A HyperC hint
-------------
There have been a few occasions in the hyperc-l mail where people have 
said that they are aware of the routine 'setup', but see no reason for 
using it.  None of the source code of applications I have seen use it 
either; everybody seems to follow the Andy Werner model of taking 
parameters directly off the stack and eschew the use of 'setup'.   Allow 
me to introduce you to Mr. setup.
 
Consider the C function 'peekb'
char peekb(address) /* address is any location 0x0000 - 0xffff 
                       [excluding video page 0x400 - 0x7ff]. 
 
peekb calls the assembly language program _peekb and returns the value 
of the byte at address in r1.   
 
 
Below are two alternate approaches using the Andy Werner model and 
'setup'.  You'll see that even though there's only one parm for this 
function, the most efficient version of the Werner model, including 
the use of 65c02 code, is 42% larger than the version that uses 'setup'.  
In a situation with more parameters, the efficiency in transferring the 
parm bytes to usable addresses would be greater.  The number of bytes 
for the version using 'setup' would not change, while the other would 
with every additional parameter.
 
This doesn't mean that setup is the best approach in _every_ situation.  
If _all_ the parms are to be used as self-modifying code and that's the 
only way to accomplish your task, then direct picks from the stack is very 
appropriate.  But if at least one parm is a pointer, an address or a data 
byte that you have to make reference to more than once; if indirect access 
using the y index is needed, then you are much better off using 'setup'.
 
Now, go do your setups. ;-)
 
Basil Johnson
=======================================
;This version of peekb uses the Werner model
; - it picks parameters from the stack.
;
; peekb.a
; Callable from C - Syntax: peekb(addr);
; Returns the value of the byte at memory address 'addr'.
;
.entry _peekb
_peekb:
    ldy #0
    sty r1h        ; returning 8 bit value so zero high byte
    lda [sp],y     ; get lsb of address into ./batchmail0
    sta 0
    iny
    lda [sp],y     ;     msb of address into ./batchmail1
    sta 1
    lda (0)        ; get the byte ( 65c02 code )
    sta r1         ; return 8 bits in r1
    rts
;
=======================================
; This version of pekb uses setup
;
; peekb.a
;
; by Basil Johnson 92/08/04
;
; Callable from C - Syntax: peekb(addr);
; Returns the value of the byte at memory address 'addr'.
;
.entry _peekb
_peekb:
    ldy #1     ; get 2 bytes
    jsr setup  ; y = 0xff on completion
    iny
    lda [0],y  ; get lsb of addr into r1
    sta r1
    sty r1h    ; returning 8 bit value so zero high byte
    rts
;
_
Basil Johnson           UUCP: micor!beejay        |Oh, Fate! ...
Nepean, Ontario         InterNet:                 |Such a stupid thing:
CANADA  K2B 8E9         beejay@micor.ocunix.on.ca |Command a bird to fly
(613) 820-0804                                    |Then clip its wing!
__