Startpoint: Small C v2.2 rev 117.

15/Apr/98 [All files]
 Started to port Small C to NT.  First step was to make it compileable by Visual C++
 which required come changes in Small C's own <stdio.h> as well as calls to
 'poll ()', 'iscons ()', 'getarg ()'.
 Changed trailer etc in CC4.C to give code for 386 and flat memory model.
 Made makefile and a new directory structure to separate source from make and
 executable.
 (And yes, VisualC 5.0 compiles it to an executable, but the resulting assembler 
 is not valid code for the flat memory model)

16/Apr/98
  Added CC5.C to print out debug info for the pcode in asm-listing.  Fix in
  CC2.C to also insert source-code in asm-listing.  Purified and fixed some
  uninitilized reads.  Still many assumptions that 'int' is 16bit.
  Must(?) make a 16bit version also, so I can compare.

17/Apr/98
  Fixed a bug regarding uninitialized variables that gave illegal code, added 
  function null_is in CC3.C.

19/Apr/98
  Revised index-operations in asm to 32 bits, added <o> in outcode() and added
  function offset() in CC4.C.  Can now assemble and link a 32bit application although
  it crashes when run.  Cannot have it all ...

20/Apr/98
  Changed primary and secondary register to work on 32 bits.  Pointer is also
  32 bit.  Leaved support for a 16bit short.  Now generates code that works
  if the C-code is 'nice'.  As I am now using Microsofts asm, linker and library
  I cannot call library-functions such as 'printf' since Smallc 'reverses' the 
  order of arguments in function calls.

22/Apr/98
  Changed calls to abort() into exit(). Decided not to reverse parameter order,
  instead will try to make an own CLIB.  Still problem when calling MSVC lib's
  like fprintf calls fputc.  Added complexity that stdout now is _iob[1].

23/Apr/98
  Changed fetch and store to work on 32bit if INT32 is #defined.
  No sign extension on word fetches though and also more serious
  bugs ...

27/Apr/98
  So bug-free that a printf should be handy to debug the rest. // (C++) comments
  added (always enabled, no switch)

28/Apr/98
  Beginning with libraries, eliminating implicit 'int = 16 bits' assumptions
  in the frontend.

...
05/May/98
  Numerous bugs in compiler and libraries...  Problem with INT 21H in libraries.

06/May/98
  Gave up trying to use MSDOS-personality under NT (as well as POSIX).  Have done
  some dirty coding to use MS's own clib instead.  Got fputc to work.  Next
  problem is to get correct alignment on data.

08/May/98
  Added -Ddefine to command-line args.  Problem with fetching parameters in
  fprintf, 'stacklepra'.

10/May/98
  puts and fputs working

25/May/98
  Lots of things, but now it can compile itself.

10/Nov/98
  Nobody seems to be using/having Microsoft MASM assembler, so let us give
  NASM a try.  NASM v0.98 should have been out by now but v0.97 will do.

16/Nov/98
  NASM-port can now compile itself, changed version to v3.1.
