From: james.harris.1@nospicedham.gmail.com   
      
   On 19/07/2017 21:25, Terje Mathisen wrote:   
   > James Harris wrote:   
   >> One more "wheel", though, just out of interest. I see that of the   
   >> opcodes available there are, in hex, these "ranges":   
   >>   
   >> the 3x range which gives some very useful xor and cmp the 7x range   
   >> which provides short jumps _if_ offsets fit   
   >   
   > Exactly right. :-)   
   >>   
   >> but there are not many useful instructions available. Has it already   
   >> been looked at to xor a run of bytes in order to make some other   
   >> opcode ranges available (and maybe make jCC offsets more amenable)?   
   >   
   > We use XOR in the SMC part to change known legal ascii values into the   
   > desired opcode bytes.   
   >   
   >> I've not worked out any particular xor value. Just asking if it's   
   >> already been considered.   
   >   
   > In my code I added one more (strong!) limitation: I wanted to write a   
   > primary decoder that would work with as little self-modification as   
   > possible, so when I figured out a way that only needed to fixup the   
   > absolute minimum (a single short (two-byte)) backwards branch   
   > instruction) I called it a day.   
      
   So the only thing you "self-modify" is a branch offset? That's   
   /definitely/ cool. I always hated self-modifying code. But branch   
   offsets have to be legit as it's very similar to doing the work of a   
   linker.   
      
   Interesting you mention imposing an additional limitation on yourself.   
   If I had a go at it I'd be itching to add some of my own. Things like   
      
   * A subsequent run from the start must also work.   
      
   * Need to cope with CR and LF at any point (after the first N bytes, for   
   some reasonable N).   
      
   * Only 8086 instructions.   
      
   * Not rely on [SP] being zero - in case of non-compliant environment. (I   
   haven't found the .COM spec but from what people have said I guess that   
   that 16-bit word is supposed to be zero ... perhaps so that a near   
   return gets to CS:0 which is termination code.)   
      
   Maybe others have done some of that. In any case, I am not volunteering   
   ... the challenge already looks more than hard enough. Now I know what   
   it was about I wish I'd followed from the beginning.   
      
      
   --   
   James Harris   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)