Found a minor bug in Merlin-8 v2.58

[Sheldon Simms <wssimms@gmail.com>]

I ran across this last night. It doesn't really matter, because the 
assembler bug is induced by a bug in the source program (failure to use 
immediate addressing).

Look at the symbol table addresses generated in comparison with the 
true addresses of the labels:


8000: A4 03	  1             ldy    endmsg-msg  ;should be immediate
8002: 60        2    loop	  rts
8003: C8 E8 A1  3    msg      asc    "Hi!"
               4    endmsg

--End assembly, 6 bytes, Errors: 0

Symbol table - alphabetical order:

  endmsg  =$8007   ?  loop    =$8003      msg     =$8004