Re: Introducing Slammer!

[Ivan X <ivanxqz@yahoo.com>]

On Jul 27, 4:43 am, Ivan Drucker<i...@ivanx.com> �wrote:

>>> the only instructions I had
>>> available were AND BIT BMI BVC CLI EOR JMPa LSR PHA PLP ROL RTI.
>>> (I used all but one of them.)

On Jul 30, 1:44�am, "Michael J. Mahon" <mjma...@aol.com> wrote:
>> I'm guessing that it was BVC that went unused...? �;-)

On 7/30/10 8:42 AM, in article
5491b850-7287-4c4f-81b3-dfed24773483@f42g2000yqn.googlegroups.com, "James
Littlejohn" <narguthru@yahoo.com> wrote:
 
> Just to prove that I was really listening to his presentation at
> Kfest, the correct answer is JMPa.

Very good -- you get an A! For Accumulator. Actually, I was mistaken -- I
didn't use PLP either.

Actually, given only two branch instructions, and no JSR or JMP, I used BVC
often for branch always; I ensure the V flag is cleared at the beginning of
Slammer by performing BIT on a location with 0 in it, and nothing else
Slammer does changes it. (I'm actually going to post the Slammer source and
an explanation of how I did the whole thing shortly.)

I actually use the V flag regularly even in my normal machine language
programming, because it is affected by so few instructions (ADC, SBC, CLV,
BIT, PLP, RTI). This means I can use it as a binary state indicator (among
other things); you can clear it with CLV, or set it with BIT $B1 (on a II+
or later). A four-way state in bits 6 and 7 of a byte makes for easy
branching without CMP, as you just BIT it, and then use BPL/BMI and BVC/BVS
accordingly. And CLV-BVC is a branch always for 6502 that doesn't disturb
the other flags, which is especially useful when writing relocatable code
where you can't easily use JSR and JMP.


--- news://freenews.netfront.net/ - complaints: news@netfront.net ---