From: muta...@gmail.com   
      
   On Monday, October 24, 2022 at 5:05:34 AM UTC+8, Joe Monk wrote:   
   > > The existing code base, some of which includes code I wrote,   
   > > mainly C code, does not arbitrarily hardcoded addresses like   
   > > 0x5555:0x0005 and then attempt to compare them to other   
   > > arbitrarily hardcoded addresses.   
   > >   
   > And thats why you dont understand the issue.   
      
   As far as I can tell, it is you who doesn't understand.   
      
   Please spend some time rechecking your assumptions.   
      
   Perhaps someone else can chime in. Someone else already   
   has chimed in and said that well-behaved application programs   
   should work even on the 80286 and have access to 16 MiB   
   of memory.   
      
   > The "shifts" are done by HARDWARE, not by your code. IOW, its not up to you!   
      
   Sure. And we're discussing THEORETICAL HARDWARE,   
   produced by a company such as NEC, that does 5 bit   
   segment shifts.   
      
   > By way of example:   
   >   
   > 2B0:013 and 2C0:003 point to the same physical address in memory.   
   >   
   > Thats why segments are LOGICAL, not physical.   
      
   And on the theoretical NEC hardware I described above, it   
   will instead be:   
      
   2B0:013 and 1E0:003 that point to the same physical address in memory.   
      
   And the existing code base doesn't care one iota because it doesn't   
   hardcode EITHER of those addresses, nor does it even attempt to   
   compare such disparate addresses (it would fail if it did, because   
   C code doesn't normally normalize pointers before comparison).   
      
   BFN. Paul.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)