<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>
Date   : Sun, 28 Dec 2008 16:43:28 +0000
From   : adsb@... (Andrew Benham)
Subject: Using stack page for temp vars

John Kortink wrote:

> What kind of approach to programming is this anyway,
> trying to make some assumption stick when you know
> you don't have to make any if you put in a little
> more effort ? Pretty much a waste of time, if you
> ask me.

Seconded.  If you're going to program making assumptions,
then you need to verify that those assumptions are correct
at run-time.  If the assumptions aren't correct, then you
can't continue.

So you could:

1/. Check that the tube isn't active

2/. Do an Osbyte 252 to get the current language ROM number,
    and an Osbyte 187 to get the BASIC ROM number.

3/. See if these ROM numbers are the same: iff they are, then
    you know that &70-&8F are available for the user.

4/. But the user may already be using some of these locations,
    so you'd have to push them on to the stack

5/. And hope that the user isn't using any of the locations for
    an event vector or similar

6/. And hope nobody else is doing the same as you
....

By the time you've done a vigorous run-time check, you might as
well have coded it properly in the first place.


I'm not on the moral high ground here.  I discovered that the Beebug
Master ROM (which I used to maintain) used the zero page locations
which Terminal on the Master used for (as I recall) vectors for
XON/XOFF processing :-(


-- 
Andrew Benham         adsb@...       
Southgate, London N14, United Kingdom

The gates in my computer are AND OR and NOT, not "Bill"



<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>