I was perusing the various news items, and found a nice tutorial on CMP, CPX, and CPY. I also found a nasty bug. The code as presented, in the very last tutorial, is as follows:



Oh, it'll make good on the tutorial's promise of moving up to 256 bytes of data. However, it'll move it from zero page on the 6502, and part of direct page and into the stack on the 65816. This code, therefore, is not portable, and is quite likely not what the author intended.

I think the intention of the code, as I understand it, is best expressed with the following code:



This way, we set up a pointer to the data to be moved at location $20, so that if the user really does want to move from zero page and part of the stack space, then he can do so. Besides, it gives the Y register some of the limelight that X always seems to get.

On the other hand, it's also 1:18 AM for me here, so maybe I'm just getting too sleepy...

I haven't seen the tutorial you mentioned, but that code could be improved. This is smaller and faster, but it is limited to addressing 128 bytes.

I actually very rarely use that style of indexing. I usually load the 2's compliment into the index register and increment, like this:



Your method works, of course. I've seen it used many times, but most of my applications for this type of thing is in moving overlapping regions of memory, where X *must* increase, not decrease, to prevent overwriting uncopied memory. That's where I first came up with the 2's compliment technique above, and it's been more or less a habit ever since.

Hi Guys,

Thanks for your input (and code). I will update the CMP tutorial during the next major update of 6502.org.

Best Regards,
Mike

_________________
- Mike Naberezny (mike@naberezny.com) http://6502.org