Jump to content

Compiler/virtual machine interpreter

From Rosetta Code
Task
Compiler/virtual machine interpreter
You are encouraged to solve this task according to the task description, using any language you may know.
Virtual Machine Interpreter

A virtual machine implements a computer in software.

Write a virtual machine interpreter. This interpreter should be able to run virtual assembly language programs created via the task. This is a byte-coded, 32-bit word stack based virtual machine.

The program should read input from a file and/or stdin, and write output to a file and/or stdout.

Input format

Given the following program:

count = 1;
while (count < 10) {
    print("count is: ", count, "\n");
    count = count + 1;
}

The output from the Code generator is a virtual assembly code program:

Output from gen, input to VM
Datasize: 1 Strings: 2
"count is: "
"\n"
    0 push  1
    5 store [0]
   10 fetch [0]
   15 push  10
   20 lt
   21 jz     (43) 65
   26 push  0
   31 prts
   32 fetch [0]
   37 prti
   38 push  1
   43 prts
   44 fetch [0]
   49 push  1
   54 add
   55 store [0]
   60 jmp    (-51) 10
   65 halt

The first line of the input specifies the datasize required and the number of constant strings, in the order that they are reference via the code.

The data can be stored in a separate array, or the data can be stored at the beginning of the stack. Data is addressed starting at 0. If there are 3 variables, the 3rd one if referenced at address 2.

If there are one or more constant strings, they come next. The code refers to these strings by their index. The index starts at 0. So if there are 3 strings, and the code wants to reference the 3rd string, 2 will be used.

Next comes the actual virtual assembly code. The first number is the code address of that instruction. After that is the instruction mnemonic, followed by optional operands, depending on the instruction.

Registers

sp:

   the stack pointer - points to the next top of stack.  The stack is a 32-bit integer
   array.

pc:

   the program counter - points to the current instruction to be performed.  The code is an
   array of bytes.

Data:

   data
   string pool
Instructions

Each instruction is one byte. The following instructions also have a 32-bit integer operand:

fetch [index]

where index is an index into the data array.

store [index]

where index is an index into the data array.

push n

where value is a 32-bit integer that will be pushed onto the stack.

jmp (n) addr

where (n) is a 32-bit integer specifying the distance between the current location and the desired location. addr is an unsigned value of the actual code address.

jz (n) addr

where (n) is a 32-bit integer specifying the distance between the current location and the desired location. addr is an unsigned value of the actual code address.

The following instructions do not have an operand. They perform their operation directly against the stack:

For the following instructions, the operation is performed against the top two entries in the stack:

add
sub
mul
div
mod
lt
gt
le
ge
eq
ne
and
or

For the following instructions, the operation is performed against the top entry in the stack:

neg
not

Print the word at stack top as a character.

prtc

Print the word at stack top as an integer.

prti

Stack top points to an index into the string pool. Print that entry.

prts

Unconditional stop.

halt
A simple example virtual machine
def run_vm(data_size)
    int stack[data_size + 1000]
    set stack[0..data_size - 1] to 0
    int pc = 0
    while True:
        op = code[pc]
        pc += 1

        if op == FETCH:
            stack.append(stack[bytes_to_int(code[pc:pc+word_size])[0]]);
            pc += word_size
        elif op == STORE:
            stack[bytes_to_int(code[pc:pc+word_size])[0]] = stack.pop();
            pc += word_size
        elif op == PUSH:
            stack.append(bytes_to_int(code[pc:pc+word_size])[0]);
            pc += word_size
        elif op == ADD:   stack[-2] += stack[-1]; stack.pop()
        elif op == SUB:   stack[-2] -= stack[-1]; stack.pop()
        elif op == MUL:   stack[-2] *= stack[-1]; stack.pop()
        elif op == DIV:   stack[-2] /= stack[-1]; stack.pop()
        elif op == MOD:   stack[-2] %= stack[-1]; stack.pop()
        elif op == LT:    stack[-2] = stack[-2] <  stack[-1]; stack.pop()
        elif op == GT:    stack[-2] = stack[-2] >  stack[-1]; stack.pop()
        elif op == LE:    stack[-2] = stack[-2] <= stack[-1]; stack.pop()
        elif op == GE:    stack[-2] = stack[-2] >= stack[-1]; stack.pop()
        elif op == EQ:    stack[-2] = stack[-2] == stack[-1]; stack.pop()
        elif op == NE:    stack[-2] = stack[-2] != stack[-1]; stack.pop()
        elif op == AND:   stack[-2] = stack[-2] and stack[-1]; stack.pop()
        elif op == OR:    stack[-2] = stack[-2] or  stack[-1]; stack.pop()
        elif op == NEG:   stack[-1] = -stack[-1]
        elif op == NOT:   stack[-1] = not stack[-1]
        elif op == JMP:   pc += bytes_to_int(code[pc:pc+word_size])[0]
        elif op == JZ:    if stack.pop() then pc += word_size else pc += bytes_to_int(code[pc:pc+word_size])[0]
        elif op == PRTC:  print stack[-1] as a character; stack.pop()
        elif op == PRTS:  print the constant string referred to by stack[-1]; stack.pop()
        elif op == PRTI:  print stack[-1] as an integer; stack.pop()
        elif op == HALT:  break
Additional examples

Your solution should pass all the test cases above and the additional tests found Here.

Reference

The C and Python versions can be considered reference implementations.

Related Tasks

Ada

Works with: GNAT version Community 2021


This program outputs only the standard output, because I did not feel like implementing stream output to a named file. (Text I/O would have appended a newline or some such page-ender to the output.) One does not really need more than standard output for this task.

This Ada program is one of the faster implementations I have written, but you have to turn off runtime checks to get that speed.


--
-- The Rosetta Code Virtual Machine, in Ada.
--
-- It is assumed the platform on which this program is run
-- has two's-complement integers. (Otherwise one could modify
-- the vmint_to_vmsigned and vmsigned_to_vmint functions. But
-- the chances your binary integers are not two's-complement
-- seem pretty low.)
--

with Ada.Characters.Handling; use Ada.Characters.Handling;
with Ada.Command_Line;        use Ada.Command_Line;

with Ada.Strings.Unbounded;         use Ada.Strings.Unbounded;
with Ada.Strings.Unbounded.Text_IO; use Ada.Strings.Unbounded.Text_IO;

with Ada.Text_IO;              use Ada.Text_IO;
with Ada.Text_IO.Text_Streams; use Ada.Text_IO.Text_Streams;

with Ada.Unchecked_Conversion;

procedure VM
is
  bad_vm            : exception;
  vm_limit_exceeded : exception;
  vm_runtime_error  : exception;

  status           : Exit_Status;
  input_file_name  : Unbounded_String;
  output_file_name : Unbounded_String;
  input_file       : File_Type;
  output_file      : File_Type;

-- Some limits of this implementation. You can adjust these to taste.
  strings_size : constant := 2_048;
  stack_size   : constant := 2_048;
  data_size    : constant := 2_048;
  code_size    : constant := 32_768;

  type byte is mod 16#100#;
  type vmint is mod 16#1_0000_0000#;
  subtype vmsigned is Integer range -2_147_483_648 .. 2_147_483_647;

  op_halt  : constant byte := 0;
  op_add   : constant byte := 1;
  op_sub   : constant byte := 2;
  op_mul   : constant byte := 3;
  op_div   : constant byte := 4;
  op_mod   : constant byte := 5;
  op_lt    : constant byte := 6;
  op_gt    : constant byte := 7;
  op_le    : constant byte := 8;
  op_ge    : constant byte := 9;
  op_eq    : constant byte := 10;
  op_ne    : constant byte := 11;
  op_and   : constant byte := 12;
  op_or    : constant byte := 13;
  op_neg   : constant byte := 14;
  op_not   : constant byte := 15;
  op_prtc  : constant byte := 16;
  op_prti  : constant byte := 17;
  op_prts  : constant byte := 18;
  op_fetch : constant byte := 19;
  op_store : constant byte := 20;
  op_push  : constant byte := 21;
  op_jmp   : constant byte := 22;
  op_jz    : constant byte := 23;

  strings : array (0 .. strings_size - 1) of Unbounded_String;
  stack   : array (0 .. stack_size - 1) of vmint;
  data    : array (0 .. data_size - 1) of vmint;
  code    : array (0 .. code_size) of byte;
  sp      : vmint;
  pc      : vmint;

  output_stream : Stream_Access;

  function vmsigned_to_vmint is new Ada.Unchecked_Conversion
   (Source => vmsigned, Target => vmint);

  function vmint_to_vmsigned is new Ada.Unchecked_Conversion
   (Source => vmint, Target => vmsigned);

  function twos_complement
   (x : in vmint)
    return vmint
  is
  begin
    return (not x) + 1;
  end twos_complement;

  function vmint_to_digits
   (x : in vmint)
    return Unbounded_String
  is
    s : Unbounded_String;
    z : vmint;
  begin
    if x = 0 then
      s := To_Unbounded_String ("0");
    else
      s := To_Unbounded_String ("");
      z := x;
      while z /= 0 loop
        s := Character'Val ((z rem 10) + Character'Pos ('0')) & s;
        z := z / 10;
      end loop;
    end if;
    return s;
  end vmint_to_digits;

  function digits_to_vmint
   (s : in String)
    return vmint
  is
    zero     : constant Character := '0';
    zero_pos : constant Integer   := Character'Pos (zero);
    retval   : vmint;
  begin
    if s'Length < 1 then
      raise bad_vm with "expected a numeric literal";
    end if;
    retval := 0;
    for i in s'Range loop
      if Is_Decimal_Digit (s (i)) then
        retval :=
         (10 * retval) + vmint (Character'Pos (s (i)) - zero_pos);
      else
        raise bad_vm with "expected a decimal digit";
      end if;
    end loop;
    return retval;
  end digits_to_vmint;

  function string_to_vmint
   (s : in String)
    return vmint
  is
    retval : vmint;
  begin
    if s'Length < 1 then
      raise bad_vm with "expected a numeric literal";
    end if;
    if s (s'First) = '-' then
      if s'Length < 2 then
        raise bad_vm with "expected a numeric literal";
      end if;
      retval :=
       twos_complement (digits_to_vmint (s (s'First + 1 .. s'Last)));
    else
      retval := digits_to_vmint (s);
    end if;
    return retval;
  end string_to_vmint;

  procedure parse_header
   (s             : in     String;
    data_count    :    out vmint;
    strings_count :    out vmint)
  is
    i : Positive;
    j : Positive;
  begin
    i := s'First;
    while i <= s'Last and then not Is_Decimal_Digit (s (i)) loop
      i := i + 1;
    end loop;

    j := i;
    while j <= s'Last and then Is_Decimal_Digit (s (j)) loop
      j := j + 1;
    end loop;

    data_count := digits_to_vmint (s (i .. j - 1));

    i := j;
    while i <= s'Last and then not Is_Decimal_Digit (s (i)) loop
      i := i + 1;
    end loop;

    j := i;
    while j <= s'Last and then Is_Decimal_Digit (s (j)) loop
      j := j + 1;
    end loop;

    strings_count := digits_to_vmint (s (i .. j - 1));
  end parse_header;

  function parse_string_literal
   (s : in String)
    return Unbounded_String
  is
    t : Unbounded_String;
    i : Positive;

    --
    -- A little trick to get around mistaken highlighting on the
    -- Rosetta Code site.
    --
    quote_string : constant String    := """";
    quote        : constant Character := quote_string (1);

  begin
    t := To_Unbounded_String ("");

    i := s'First;
    while i <= s'Last and then s (i) /= quote loop
      i := i + 1;
    end loop;

    if s'Last < i or else s (i) /= quote then
      raise bad_vm with "expected a '""'";
    end if;

    i := i + 1;
    while i <= s'Last and then s (i) /= quote loop
      if s (i) /= '\' then
        Append (t, s (i));
        i := i + 1;
      elsif s'Last < i + 1 then
        raise bad_vm with "truncated string literal";
      elsif s (i + 1) = 'n' then
        Append (t, Character'Val (10));
        i := i + 2;
      elsif s (i + 1) = '\' then
        Append (t, '\');
        i := i + 2;
      else
        raise bad_vm with "unsupported escape sequence";
      end if;
    end loop;

    return t;
  end parse_string_literal;

  function name_to_opcode
   (s : in String)
    return byte
  is
    retval : byte;
  begin
    if s = "halt" then
      retval := op_halt;
    elsif s = "add" then
      retval := op_add;
    elsif s = "sub" then
      retval := op_sub;
    elsif s = "mul" then
      retval := op_mul;
    elsif s = "div" then
      retval := op_div;
    elsif s = "mod" then
      retval := op_mod;
    elsif s = "lt" then
      retval := op_lt;
    elsif s = "gt" then
      retval := op_gt;
    elsif s = "le" then
      retval := op_le;
    elsif s = "ge" then
      retval := op_ge;
    elsif s = "eq" then
      retval := op_eq;
    elsif s = "ne" then
      retval := op_ne;
    elsif s = "and" then
      retval := op_and;
    elsif s = "or" then
      retval := op_or;
    elsif s = "neg" then
      retval := op_neg;
    elsif s = "not" then
      retval := op_not;
    elsif s = "prtc" then
      retval := op_prtc;
    elsif s = "prti" then
      retval := op_prti;
    elsif s = "prts" then
      retval := op_prts;
    elsif s = "fetch" then
      retval := op_fetch;
    elsif s = "store" then
      retval := op_store;
    elsif s = "push" then
      retval := op_push;
    elsif s = "jmp" then
      retval := op_jmp;
    elsif s = "jz" then
      retval := op_jz;
    else
      raise bad_vm with ("unexpected opcode name");
    end if;
    return retval;
  end name_to_opcode;

  procedure parse_instruction
   (s       : in     String;
    address :    out vmint;
    opcode  :    out byte;
    arg     :    out vmint)
  is
    i : Positive;
    j : Positive;
  begin
    i := s'First;
    while i <= s'Last and then not Is_Decimal_Digit (s (i)) loop
      i := i + 1;
    end loop;

    j := i;
    while j <= s'Last and then Is_Decimal_Digit (s (j)) loop
      j := j + 1;
    end loop;

    address := digits_to_vmint (s (i .. j - 1));

    i := j;
    while i <= s'Last and then not Is_Letter (s (i)) loop
      i := i + 1;
    end loop;

    j := i;
    while j <= s'Last and then Is_Letter (s (j)) loop
      j := j + 1;
    end loop;

    opcode := name_to_opcode (s (i .. j - 1));

    i := j;
    while i <= s'Last and then Is_Space (s (i)) loop
      i := i + 1;
    end loop;

    if s'Last < i then
      arg := 0;
    else
      if not Is_Decimal_Digit (s (i)) and then s (i) /= '-' then
        i := i + 1;
      end if;
      j := i;
      while j <= s'Last
       and then (Is_Decimal_Digit (s (j)) or else s (j) = '-')
      loop
        j := j + 1;
      end loop;
      arg := string_to_vmint (s (i .. j - 1));
    end if;
  end parse_instruction;

  procedure read_and_parse_header
   (data_count    : out vmint;
    strings_count : out vmint)
  is
    line : Unbounded_String;
  begin
    Get_Line (Current_Input, line);
    parse_header (To_String (line), data_count, strings_count);
  end read_and_parse_header;

  procedure read_parse_and_store_strings
   (strings_count : in vmint)
  is
    line : Unbounded_String;
  begin
    if strings_count /= 0 then
      if strings_size < strings_count then
        raise vm_limit_exceeded with "strings limit exceeded";
      end if;
      for i in 0 .. strings_count - 1 loop
        Get_Line (Current_Input, line);
        strings (Integer (i)) :=
         parse_string_literal (To_String (line));
      end loop;
    end if;
  end read_parse_and_store_strings;

  function opcode_takes_arg
   (opcode : in byte)
    return Boolean
  is
    retval : Boolean;
  begin
    if opcode = op_fetch then
      retval := True;
    elsif opcode = op_store then
      retval := True;
    elsif opcode = op_push then
      retval := True;
    elsif opcode = op_jmp then
      retval := True;
    elsif opcode = op_jz then
      retval := True;
    else
      retval := False;
    end if;
    return retval;
  end opcode_takes_arg;

  procedure read_parse_and_store_instructions
  is
    line    : Unbounded_String;
    address : vmint;
    opcode  : byte;
    arg     : vmint;
    j       : Positive;
  begin
    while not End_Of_File (Current_Input) loop
      Get_Line (Current_Input, line);

      j := 1;
      while j <= Length (line) and then Is_Space (Element (line, j))
      loop
        j := j + 1;
      end loop;

      if j <= Length (line) then
        parse_instruction (To_String (line), address, opcode, arg);
        if opcode_takes_arg (opcode) then
          if code_size - 4 <= address then
            raise vm_limit_exceeded with "code space limit exceeded";
          end if;
          code (Integer (address)) := opcode;
          --
          -- Little-endian storage.
          --
          code (Integer (address) + 1) := byte (arg and 16#FF#);
          code (Integer (address) + 2) :=
           byte ((arg / 16#100#) and 16#FF#);
          code (Integer (address) + 3) :=
           byte ((arg / 16#1_0000#) and 16#FF#);
          code (Integer (address) + 4) :=
           byte ((arg / 16#100_0000#) and 16#FF#);
        else
          if code_size <= address then
            raise vm_limit_exceeded with "code space limit exceeded";
          end if;
          code (Integer (address)) := opcode;
        end if;
      end if;
    end loop;
  end read_parse_and_store_instructions;

  procedure read_parse_and_store_program
  is
    data_count    : vmint;
    strings_count : vmint;
  begin
    read_and_parse_header (data_count, strings_count);
    read_parse_and_store_strings (strings_count);
    read_parse_and_store_instructions;
  end read_parse_and_store_program;

  procedure pop_value
   (x : out vmint)
  is
  begin
    if sp = 0 then
      raise vm_runtime_error with "stack underflow";
    end if;
    sp := sp - 1;
    x  := stack (Integer (sp));
  end pop_value;

  procedure push_value
   (x : in vmint)
  is
  begin
    if stack_size <= sp then
      raise vm_runtime_error with "stack overflow";
    end if;
    stack (Integer (sp)) := x;
    sp                   := sp + 1;
  end push_value;

  procedure get_value
   (x : out vmint)
  is
  begin
    if sp = 0 then
      raise vm_runtime_error with "stack underflow";
    end if;
    x := stack (Integer (sp) - 1);
  end get_value;

  procedure put_value
   (x : in vmint)
  is
  begin
    if sp = 0 then
      raise vm_runtime_error with "stack underflow";
    end if;
    stack (Integer (sp) - 1) := x;
  end put_value;

  procedure fetch_value
   (i : in     vmint;
    x :    out vmint)
  is
  begin
    if data_size <= i then
      raise vm_runtime_error with "data boundary exceeded";
    end if;
    x := data (Integer (i));
  end fetch_value;

  procedure store_value
   (i : in vmint;
    x : in vmint)
  is
  begin
    if data_size <= i then
      raise vm_runtime_error with "data boundary exceeded";
    end if;
    data (Integer (i)) := x;
  end store_value;

  procedure immediate_value
   (x : out vmint)
  is
    b0, b1, b2, b3 : vmint;
  begin
    if code_size - 4 <= pc then
      raise vm_runtime_error with "code boundary exceeded";
    end if;
    --
    -- Little-endian order.
    --
    b0 := vmint (code (Integer (pc)));
    b1 := vmint (code (Integer (pc) + 1));
    b2 := vmint (code (Integer (pc) + 2));
    b3 := vmint (code (Integer (pc) + 3));
    x  :=
     b0 + (16#100# * b1) + (16#1_0000# * b2) + (16#100_0000# * b3);
  end immediate_value;

  procedure machine_add
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    put_value (x + y);
  end machine_add;

  procedure machine_sub
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    put_value (x - y);
  end machine_sub;

  procedure machine_mul
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    put_value
     (vmsigned_to_vmint
       (vmint_to_vmsigned (x) * vmint_to_vmsigned (y)));
  end machine_mul;

  procedure machine_div
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    put_value
     (vmsigned_to_vmint
       (vmint_to_vmsigned (x) / vmint_to_vmsigned (y)));
  end machine_div;

  procedure machine_mod
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    put_value
     (vmsigned_to_vmint
       (vmint_to_vmsigned (x) rem vmint_to_vmsigned (y)));
  end machine_mod;

  procedure machine_lt
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    if vmint_to_vmsigned (x) < vmint_to_vmsigned (y) then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_lt;

  procedure machine_gt
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    if vmint_to_vmsigned (x) > vmint_to_vmsigned (y) then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_gt;

  procedure machine_le
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    if vmint_to_vmsigned (x) <= vmint_to_vmsigned (y) then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_le;

  procedure machine_ge
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    if vmint_to_vmsigned (x) >= vmint_to_vmsigned (y) then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_ge;

  procedure machine_eq
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    if x = y then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_eq;

  procedure machine_ne
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    if x /= y then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_ne;

  procedure machine_and
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    if x /= 0 and y /= 0 then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_and;

  procedure machine_or
  is
    x, y : vmint;
  begin
    pop_value (y);
    get_value (x);
    if x /= 0 or y /= 0 then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_or;

  procedure machine_neg
  is
    x : vmint;
  begin
    get_value (x);
    put_value (twos_complement (x));
  end machine_neg;

  procedure machine_not
  is
    x : vmint;
  begin
    get_value (x);
    if x = 0 then
      put_value (1);
    else
      put_value (0);
    end if;
  end machine_not;

  procedure machine_prtc
  is
    x : vmint;
  begin
    pop_value (x);
    Character'Write (output_stream, Character'Val (x));
  end machine_prtc;

  procedure machine_prti
  is
    x : vmint;
  begin
    pop_value (x);
    if 16#7FFF_FFFF# < x then
      Character'Write (output_stream, '-');
      String'Write
       (output_stream,
        To_String (vmint_to_digits (twos_complement (x))));
    else
      String'Write (output_stream, To_String (vmint_to_digits (x)));
    end if;
  end machine_prti;

  procedure machine_prts
  is
    k : vmint;
  begin
    pop_value (k);
    if strings_size <= k then
      raise vm_runtime_error with "strings boundary exceeded";
    end if;
    String'Write (output_stream, To_String (strings (Integer (k))));
  end machine_prts;

  procedure machine_fetch
  is
    k : vmint;
    x : vmint;
  begin
    immediate_value (k);
    fetch_value (k, x);
    push_value (x);
    pc := pc + 4;
  end machine_fetch;

  procedure machine_store
  is
    k : vmint;
    x : vmint;
  begin
    immediate_value (k);
    pop_value (x);
    store_value (k, x);
    pc := pc + 4;
  end machine_store;

  procedure machine_push
  is
    x : vmint;
  begin
    immediate_value (x);
    push_value (x);
    pc := pc + 4;
  end machine_push;

  procedure machine_jmp
  is
    offset : vmint;
  begin
    immediate_value (offset);
    pc := pc + offset;
  end machine_jmp;

  procedure machine_jz
  is
    x      : vmint;
    offset : vmint;
  begin
    pop_value (x);
    if x = 0 then
      immediate_value (offset);
      pc := pc + offset;
    else
      pc := pc + 4;
    end if;
  end machine_jz;

  procedure machine_step
   (halt : out Boolean)
  is
    opcode             : byte;
    op_div_4, op_rem_4 : byte;
  begin
    if code_size <= pc then
      raise vm_runtime_error with "code boundary exceeded";
    end if;
    opcode   := code (Integer (pc));
    pc       := pc + 1;
    halt     := False;
    op_div_4 := opcode / 4;
    op_rem_4 := opcode rem 4;
    if op_div_4 = 0 then
      if op_rem_4 = 0 then
        halt := True;
      elsif op_rem_4 = 1 then
        machine_add;
      elsif op_rem_4 = 2 then
        machine_sub;
      else
        machine_mul;
      end if;
    elsif op_div_4 = 1 then
      if op_rem_4 = 0 then
        machine_div;
      elsif op_rem_4 = 1 then
        machine_mod;
      elsif op_rem_4 = 2 then
        machine_lt;
      else
        machine_gt;
      end if;
    elsif op_div_4 = 2 then
      if op_rem_4 = 0 then
        machine_le;
      elsif op_rem_4 = 1 then
        machine_ge;
      elsif op_rem_4 = 2 then
        machine_eq;
      else
        machine_ne;
      end if;
    elsif op_div_4 = 3 then
      if op_rem_4 = 0 then
        machine_and;
      elsif op_rem_4 = 1 then
        machine_or;
      elsif op_rem_4 = 2 then
        machine_neg;
      else
        machine_not;
      end if;
    elsif op_div_4 = 4 then
      if op_rem_4 = 0 then
        machine_prtc;
      elsif op_rem_4 = 1 then
        machine_prti;
      elsif op_rem_4 = 2 then
        machine_prts;
      else
        machine_fetch;
      end if;
    elsif op_div_4 = 5 then
      if op_rem_4 = 0 then
        machine_store;
      elsif op_rem_4 = 1 then
        machine_push;
      elsif op_rem_4 = 2 then
        machine_jmp;
      else
        machine_jz;
      end if;
    else
      -- Treat anything unrecognized as equivalent to a halt.
      halt := True;
    end if;
  end machine_step;

  procedure machine_continue
  is
    halt : Boolean;
  begin
    halt := False;
    while not halt loop
      machine_step (halt);
    end loop;
  end machine_continue;

  procedure machine_run
  is
  begin
    sp := 0;
    pc := 0;
    for i in data'Range loop
      data (i) := 0;
    end loop;
    machine_continue;
  end machine_run;

begin
  status := 0;

  input_file_name := To_Unbounded_String ("-");

  if Argument_Count = 0 then
    null;
  elsif Argument_Count = 1 then
    input_file_name := To_Unbounded_String (Argument (1));
  else
    Put ("Usage: ");
    Put (Command_Name);
    Put_Line (" [INPUTFILE]");
    Put ("If either INPUTFILE is missing or ""-"",");
    Put_Line (" standard input is used.");
    Put_Line ("Output is always to standard output.");
    status := 1;
  end if;

  if status = 0 then
    if input_file_name /= "-" then
      Open (input_file, In_File, To_String (input_file_name));
      Set_Input (input_file);
    end if;

    output_stream := Stream (Current_Output);
    read_parse_and_store_program;
    machine_run;

    if input_file_name /= "-" then
      Set_Input (Standard_Input);
      Close (input_file);
    end if;
  end if;

  Set_Exit_Status (status);
end VM;


Output:
$ gnatmake -q -gnatp -O3 -march=native vm.adb && ./vm compiler-tests/count.vm
count is: 1
count is: 2
count is: 3
count is: 4
count is: 5
count is: 6
count is: 7
count is: 8
count is: 9

Aime

integer n, pc, sp;
file f;
text s;
index code, Data;
list l, stack, strings;

f.affix(argv(1));

f.list(l, 0);

n = atoi(l[-1]);
while (n) {
    f.lead(s);
    strings.append(erase(s, -1, 0));
    n -= 1;
}

while (f.list(l, 0) ^ -1) {
    code.put(atoi(lf_x_text(l)), l);
}

pc = sp = 0;
while (1) {
    l = code[pc];
    s = l[0];
    if (s == "jz") {
        if (lb_pick(stack)) {
            isk_greater(code, pc, pc);
        } else {
            pc = atoi(l[-1]);
        }
    } elif (s == "jmp") {
        pc = atoi(l[-1]);
    } else {
        if (s == "push") {
            lb_push(stack, atoi(l[1]));
        } elif (s == "fetch") {
            lb_push(stack, Data[atoi(erase(l[1], -1, 0))]);
        } elif (s == "neg") {
            stack[-1] = -stack[-1];
        } elif (s == "not") {
            stack[-1] = !stack[-1];
        } elif (s == "halt") {
            break;
        } else {
            n = lb_pick(stack);
            if (s == "store") {
                Data[atoi(erase(l[1], -1, 0))] = n;
            } elif (s == "add") {
                stack[-1] = stack[-1] + n;
            } elif (s == "sub") {
                stack[-1] = stack[-1] - n;
            } elif (s == "mul") {
                stack[-1] = stack[-1] * n;
            } elif (s == "div") {
                stack[-1] = stack[-1] / n;
            } elif (s == "mod") {
                stack[-1] = stack[-1] % n;
            } elif (s == "lt") {
                stack[-1] = stack[-1] < n;
            } elif (s == "gt") {
                stack[-1] = stack[-1] > n;
            } elif (s == "le") {
                stack[-1] = stack[-1] <= n;
            } elif (s == "ge") {
                stack[-1] = stack[-1] >= n;
            } elif (s == "eq") {
                stack[-1] = stack[-1] == n;
            } elif (s == "ne") {
                stack[-1] = stack[-1] != n;
            } elif (s == "and") {
                stack[-1] = stack[-1] && n;
            } elif (s == "or") {
                stack[-1] = stack[-1] || n;
            } elif (s == "prtc") {
                o_byte(n);
            } elif (s == "prti") {
                o_(n);
            } elif (s == "prts") {
                o_(strings[n]);
            } else {
            }
        }

        isk_greater(code, pc, pc);
    }
}

ALGOL W

begin % virtual machine interpreter %
    % string literals %
    string(256) array stringValue  ( 0 :: 256 );
    integer     array stringLength ( 0 :: 256 );
    integer     MAX_STRINGS;
    % op codes %
    integer     oFetch, oStore, oPush
          ,     oAdd,   oSub,   oMul, oDiv, oMod, oLt, oGt,   oLe,   oGe,   oEq,   oNe
          ,     oAnd,   oOr,    oNeg, oNot, oJmp, oJz, oPrtc, oPrts, oPrti, oHalt
          ;
    string(6)   array opName       ( 1 :: 24 );
    integer     OP_MAX;
    % code %
    string(1)   array byteCode     ( 0 :: 4096 );
    integer     nextLocation, MAX_LOCATION;
    % data %
    integer     array data         ( 0 :: 4096 );
    integer     dataSize, MAX_DATA, MAX_STACK;
    % tracing %
    logical     trace;

    % reports an error and stops %
    procedure rtError( string(80) value message ); begin
        integer errorPos;
        write( s_w := 0, "**** Runtime error: " );
        errorPos := 0;
        while errorPos < 80 and message( errorPos // 1 ) not = "." do begin
            writeon( s_w := 0, message( errorPos // 1 ) );
            errorPos := errorPos + 1
        end while_not_at_end_of_message ;
        writeon( s_w := 0, "." );
        assert( false )
    end genError ;

    oFetch :=  1; opName( oFetch ) := "fetch"; oStore :=  2; opName( oStore ) := "store"; oPush :=  3; opName( oPush ) := "push";
    oAdd   :=  4; opName( oAdd   ) := "add";   oSub   :=  5; opName( oSub   ) := "sub";   oMul  :=  6; opName( oMul  ) := "mul";
    oDiv   :=  7; opName( oDiv   ) := "div";   oMod   :=  8; opName( oMod   ) := "mod";   oLt   :=  9; opName( oLt   ) := "lt";
    oGt    := 10; opName( oGt    ) := "gt";    oLe    := 11; opName( oLe    ) := "le";    oGe   := 12; opName( oGe   ) := "ge";
    oEq    := 13; opName( oEq    ) := "eq";    oNe    := 14; opName( oNe    ) := "ne";    oAnd  := 15; opName( oAnd  ) := "and";
    oOr    := 16; opName( oOr    ) := "or";    oNeg   := 17; opName( oNeg   ) := "neg";   oNot  := 18; opName( oNot  ) := "not";
    oJmp   := 19; opName( oJmp   ) := "jmp";   oJz    := 20; opName( oJz    ) := "jz";    oPrtc := 21; opName( oPrtc ) := "prtc";
    oPrts  := 22; opName( oPrts  ) := "prts";  oPrti  := 23; opName( oPrti  ) := "prti";  oHalt := 24; opName( oHalt ) := "halt";
    OP_MAX := oHalt;

    trace        := false;
    MAX_STACK    := 256;
    MAX_LOCATION := 4096;
    for pc := 0 until MAX_LOCATION do byteCode( pc ) := code( 0 );
    MAX_DATA := 4096;
    for dPos := 0 until MAX_DATA do data( dPos ) := 0;
    MAX_STRINGS := 256;
    for sPos := 0 until MAX_STRINGS do begin
        stringValue(  sPos ) := " ";
        stringLength( sPos ) := 0
    end for_sPos ;

    % load thge output from syntaxc analyser %
    begin % readCode %

        % skips spaces on the source line %
        procedure skipSpaces ; begin
            while line( lPos // 1 ) = " " do lPos := lPos + 1
        end skipSpaces ;

        % parses a string from line and stores it in the string literals table %
        procedure readString ( integer value stringNumber ) ; begin
            string(256) str;
            integer     sLen;
            str  := " ";
            sLen := 0;
            lPos := lPos + 1; % skip the opening double-quote %
            while lPos <= 255 and line( lPos // 1 ) not = """" do begin
                str( sLen // 1 ) := line( lPos // 1 );
                sLen := sLen + 1;
                lPos := lPos + 1
            end while_more_string ;
            if lPos > 255 then rtError( "Unterminated String." );
            % store the string %
            stringValue(  stringNumber ) := str;
            stringLength( stringNumber ) := sLen
        end readString ;

        % gets an integer from the line - checks for valid digits %
        integer procedure readInteger ; begin
            integer n;
            skipSpaces;
            n := 0;
            while line( lPos // 1 ) >= "0" and line( lPos // 1 ) <= "9" do begin
                n    := ( n * 10 ) + ( decode( line( lPos // 1 ) ) - decode( "0" ) );
                lPos := lPos + 1
            end while_not_end_of_integer ;
            n
        end readInteger ;

        % reads the next line from standard input %
        procedure readALine ; begin
            lPos := 0;
            readcard( line );
            if trace then write( s_w := 0, ">> ", line( 0 // 32 ) )
        end readALine ;

        % loads an instruction from the current source line %
        procedure loadCodeFromLine ; begin
            integer pc, opCode, operand, oPos;
            string(256) op;
            logical haveOperand;
            % get the code location %
            pc := readInteger;
            if pc > MAX_LOCATION then rtError( "Code too large." );
            % get the opCode %
            skipSpaces;
            oPos := 0;
            op := " ";
            while lPos <= 255 and line( lPos // 1 ) not = " " do begin
                op( oPos // 1 ) := line( lPos // 1 );
                oPos := oPos + 1;
                lPos := lPos + 1
            end while_more_opName ;
            % lookup the op code %
            opCode := 0;
            oPos   := 1;
            while oPos <= OP_MAX and opCode = 0 do begin
                if opName( oPos ) = op then opCode := oPos
                                       else oPos   := oPos + 1
            end while_op_not_found ;
            if opCode = 0 then rtError( "Unknown op code." );
            % get the operand if there is one %
            operand     := 0;
            haveOperand := false;
            if      opCode = oFetch or opCode = oStore then begin
                % fetch or store - operand is enclosed in square brackets %
                skipSpaces;
                if line( lPos // 1 ) not = "[" then rtError( """["" expected after fetch/store." );
                lPos        := lPos + 1;
                operand     := readInteger;
                if operand > dataSize then rtError( "fetch/store address out of range." );
                haveOperand := true
                end
            else if opCode = oPush then begin
                % push integer literal instruction %
                operand     := readInteger;
                haveOperand := true
                end
            else if opCode = oJmp or opCode = oJz then begin
                % jump - the operand is the relative address enclosed in parenthesis %
                % followed by the absolute address - we use the absolute address so  %
                % the opewrand will be >= 0 %
                skipSpaces;
                if line( lPos // 1 ) not = "(" then rtError( """("" expected after jmp/jz." );
                lPos        := lPos + 1;
                if line( lPos // 1 ) = "-" then % negative relative address % lPos := lPos + 1;
                operand     := readInteger;
                if line( lPos // 1 ) not = ")" then rtError( """)"" expected after jmp/jz." );
                lPos        := lPos + 1;
                operand     := readInteger;
                haveOperand := true
            end if_various_opcodes ;
            % store the code %
            byteCode( pc ) := code( opCode );
            if haveOperand then begin
                % have an operand for the op code %
                if ( pc + 4 ) > MAX_LOCATION then rtError( "Code too large." );
                for oPos := 1 until 4 do begin
                    pc := pc + 1;
                    byteCode( pc ) := code( operand rem 256 );
                    operand := operand div 256;
                end for_oPos
            end if_have_operand ;
        end loadCodeFromLine ;

        string(256) line;
        string(16)  name;
        integer     lPos, tPos, stringCount;

        % allow us to detect EOF %
        ENDFILE := EXCEPTION( false, 1, 0, false, "EOF" );

        % first line should be "Datasize: d Strings: s" where d = number variables %
        % and s = number of strings                                                %
        readALine;
        if line = "trace" then begin
            % extension - run in trace mode %
            trace := true;
            readALine
        end if_line_eq_trace ;
        if XCPNOTED(ENDFILE) then rtError( "Empty program file." );
        if line( 0 // 10 ) not = "Datasize: " then rtError( "Header line missing." );
        lPos := 10;
        dataSize := readInteger;
        if dataSize > MAX_DATA then rtError( "Datasize too large." );
        skipSpaces;
        if line( lPos // 9 ) not = "Strings: " then rtError( """Strings: "" missing on header line." );
        lPos := lPos + 9;
        stringCount := readInteger;
        if stringCount > MAX_STRINGS then rtError( "Too many strings." );
        % read the string table %
        for stringNumber := 0 until stringCount - 1 do begin
            string(256) str;
            integer     sLen, sPos;
            readALine;
            if XCPNOTED(ENDFILE) then rtError( "End-of-file in string table." );
            if line( lPos // 1 ) not = """" then rtError( "String literal expected." );
            str  := " ";
            sLen := 0;
            lPos := lPos + 1; % skip the opening double-quote %
            while lPos <= 255 and line( lPos // 1 ) not = """" do begin
                str( sLen // 1 ) := line( lPos // 1 );
                sLen := sLen + 1;
                lPos := lPos + 1
            end while_more_string ;
            if lPos > 255 then rtError( "Unterminated String." );
            % store the string %
            stringValue(  stringNumber ) := str;
            stringLength( stringNumber ) := sLen
        end for_sPos ;
        % read the code %
        readALine;
        while not XCPNOTED(ENDFILE) do begin
            if line not = " " then loadCodeFromLine;
            readALine
        end while_not_eof
    end;
    % run the program %
    begin
        integer pc, opCode, operand, sp;
        integer array st ( 0 :: MAX_STACK );
        logical halted;
        % prints a string from the string pool, escape sequences are interpreted %
        procedure writeOnString( integer value stringNumber ) ;
        begin
            integer     cPos, sLen;
            string(256) text;
            if stringNumber < 0 or stringNumber > MAX_STRINGS then rtError( "Invalid string number." );
            cPos := 0;
            sLen := stringLength( stringNumber );
            text := stringValue(  stringNumber );
            while cPos < stringLength( stringNumber ) do begin
                string(1) ch;
                ch := text( cPos // 1 );
                if ch not = "\" then writeon( s_w := 0, ch )
                else begin
                    % escaped character %
                    cPos := cPos + 1;
                    if cPos > sLen then rtError( "String terminates with ""\""." );
                    ch := text( cPos // 1 );
                    if ch = "n" then % newline % write()
                                else writeon( s_w := 0, ch )
                end;
                cPos := cPos + 1
            end while_not_end_of_string
        end writeOnString ;

        pc     := 0;
        sp     := -1;
        halted := false;
        while not halted do begin;
            % get the next op code and operand %
            opCode  := decode( byteCode( pc ) );
            pc      := pc + 1;
            operand := 0;
            if opCode = oFetch or opCode = oStore or opCode = oPush or opCode = oJmp or opCode = oJz then begin
                % this opCode has an operand %
                pc := pc + 4;
                for bPos := 1 until 4 do begin
                    operand := ( operand * 256 ) + decode( byteCode( pc - bPos ) );
                end for_bPos
            end if_opCode_with_an_operand ;
            if trace then begin
                write( i_w:= 1, s_w := 0, pc, " op(", opCode, "): ", opName( opCode ), " ", operand );
                write()
            end if_trace ;
            % interpret the instruction %
            if      opCode = oFetch then begin sp := sp + 1; st( sp ) := data( operand ) end
            else if opCode = oStore then begin data( operand ) := st( sp ); sp := sp - 1 end
            else if opCode = oPush  then begin sp := sp + 1; st( sp ) := operand         end
            else if opCode = oHalt  then halted := true
            else if opCode = oJmp   then pc     := operand
            else if oPCode = oJz    then begin
                if st( sp ) = 0 then pc := operand;
                sp := sp - 1
                end
            else if opCode = oPrtc  then begin writeon( i_w := 1, s_w := 0, code( st( sp ) ) ); sp := sp - 1 end
            else if opCode = oPrti  then begin writeon( i_w := 1, s_w := 0,       st( sp )   ); sp := sp - 1 end
            else if opCode = oPrts  then begin writeonString(                     st( sp )   ); sp := sp - 1 end
            else if opCode = oNeg   then st( sp ) := - st( sp )
            else if opCode = oNot   then st( sp ) := ( if st( sp ) = 0 then 1 else 0 )
            else begin
                operand := st( sp );
                sp      := sp - 1;
                if      opCode = oAdd   then st( sp ) :=    st( sp )    +  operand
                else if opCode = oSub   then st( sp ) :=    st( sp )    -  operand
                else if opCode = oMul   then st( sp ) :=    st( sp )    *  operand
                else if opCode = oDiv   then st( sp ) :=    st( sp )  div  operand
                else if opCode = oMod   then st( sp ) :=    st( sp )  rem  operand
                else if opCode = oLt    then st( sp ) := if st( sp )    <  operand then 1 else 0
                else if opCode = oGt    then st( sp ) := if st( sp )    >  operand then 1 else 0
                else if opCode = oLe    then st( sp ) := if st( sp )    <= operand then 1 else 0
                else if opCode = oGe    then st( sp ) := if st( sp )    >= operand then 1 else 0
                else if opCode = oEq    then st( sp ) := if st( sp )     = operand then 1 else 0
                else if opCode = oNe    then st( sp ) := if st( sp ) not = operand then 1 else 0
                else if opCode = oAnd   then st( sp ) := if st( sp ) not = 0 and operand not = 0 then 1 else 0
                else if opCode = oOr    then st( sp ) := if st( sp ) not = 0 or  operand not = 0 then 1 else 0
                else                         rtError( "Unknown opCode." )
            end if_various_opCodes
        end while_not_halted
    end
end.

ATS

Interpreter

Works with: ATS version Postiats 0.4.1

Compile with ‘patscc -O3 -DATS_MEMALLOC_LIBC -o vm vm-postiats.dats -latslib’

With the C optimizer turned on, like this, the program should run pretty fast, despite being relatively safe against going out of bounds, etc. Try it on the ASCII Mandelbrot example.

(Without the C optimizer, ATS code can run much, much more slowly. It is worth comparing the Mandelbrot example with and without the optimizer.)

(*
  Usage: vm [INPUTFILE [OUTPUTFILE]]
     If INPUTFILE or OUTPUTFILE is "-" or missing, then standard input
     or standard output is used, respectively.

  The Rosetta Code virtual machine task in ATS2 (also known as
  Postiats).
  
  Some implementation notes:

    * Values are stored as uint32, and it is checked that uint32
      really is 32 bits, two’s-complement. Addition and subtraction
      are allowed to roll around, and so can be done without casting
      to int32. (The C standard specifies that unsigned integer values
      will roll around, rather than signal an overflow.)

    * Where it matters, the uint32 are stored in little-endian
      order. I have *not* optimized the code for x86/AMD64 (which are
      little-endian and also can address unaligned data).

    * Here I am often writing out code instead of using some library
      function. Partly this is to improve code safety (proof at
      compile-time that buffers are not overrun, proof of loop
      termination, etc.). Partly this is because I do not feel like
      using the C library (or ATS interfaces to it) all that much.

    * I am using linear types and so forth, because I think it
      interesting to do so. It is unnecessary to use a garbage
      collector, because there (hopefully) are no memory leaks. (Not
      that we couldn’t simply let memory leak, for this little program
      with no REPL.)

*)

#define ATS_EXTERN_PREFIX "rosettacode_vm_"
#define ATS_DYNLOADFLAG 0       (* No initialization is needed. *)

#include "share/atspre_define.hats"
#include "share/atspre_staload.hats"

staload UN = "prelude/SATS/unsafe.sats"

#define NIL list_vt_nil ()
#define :: list_vt_cons

(* The stack has a fixed size but is very large. (Alternatively, one
could make the stack double in size whenever it overflows. Design
options such as using a linked list for the stack come with a
performance penalty.) *)
#define VMSTACK_SIZE 65536
macdef vmstack_size = (i2sz VMSTACK_SIZE)

(* In this program, exceptions are not meant to be caught, unless
   the catcher terminates the program. Linear types and
   general exception-catching do not go together well. *)
exception bad_vm of string
exception vm_runtime_error of string

(********************************************************************)
(*                                                                  *)
(* Some string functions that are safe against buffer overruns.     *)
(*                                                                  *)

fn
skip_whitespace {n, i : int | 0 <= i; i <= n}
                (s    : string n,
                 n    : size_t n,
                 i    : size_t i) :
    [j : int | i <= j; j <= n]
    size_t j =
  let
    fun
    loop {k : int | i <= k; k <= n} .<n - k>.
         (k : size_t k) :
        [j : int | i <= j; j <= n]
        size_t j =
      if k = n then
        k
      else if isspace (s[k]) then
        loop (succ k)
      else
        k
  in
    loop (i)
  end

fn
skip_non_whitespace {n, i : int | 0 <= i; i <= n}
                    (s    : string n,
                     n    : size_t n,
                     i    : size_t i) :
    [j : int | i <= j; j <= n]
    size_t j =
  let
    fun
    loop {k : int | i <= k; k <= n} .<n - k>.
         (k : size_t k) :
        [j : int | i <= j; j <= n]
        size_t j =
      if k = n then
        k
      else if isspace (s[k]) then
        k
      else
        loop (succ k)
  in
    loop (i)
  end

fn
substr_equal {n, i, j : int | 0 <= i; i <= j; j <= n}
             {m       : int | 0 <= m}
             (s       : string n,
              i       : size_t i,
              j       : size_t j,
              t       : string m) : bool =
  (* Is s[i .. j-1] equal to t? *)
  let
    val m = string_length t
  in
    if m <> j - i then
      false
    else
      let
        fun
        loop {k : int | 0 <= k; k <= m} .<m - k>.
             (k : size_t k) : bool =
          if k = m then
            true
          else if s[i + k] <> t[k] then
            false
          else
            loop (succ k)
      in
        loop (i2sz 0)
      end
  end

(********************************************************************)
(*                                                                  *)
(* vmint = 32-bit two’s-complement numbers.                         *)
(*                                                                  *)

stadef vmint_kind = uint32_kind
typedef vmint = uint32

extern castfn i2vm    : int -<> vmint
extern castfn u2vm    : uint -<> vmint
extern castfn byte2vm : byte -<> vmint

extern castfn vm2i    : vmint -<> int
extern castfn vm2sz   : vmint -<> size_t
extern castfn vm2byte : vmint -<> byte

%{^

/*
 * The ATS prelude might not have C implementations of all the
 * operations we would like to have, so here are some.
 */

typedef uint32_t vmint_t;

ATSinline() vmint_t
rosettacode_vm_g0uint_add_vmint (vmint_t x, vmint_t y)
{
  return (x + y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_sub_vmint (vmint_t x, vmint_t y)
{
  return (x - y);
}

ATSinline() int
rosettacode_vm_g0uint_eq_vmint (vmint_t x, vmint_t y)
{
  return (x == y);
}

ATSinline() int
rosettacode_vm_g0uint_neq_vmint (vmint_t x, vmint_t y)
{
  return (x != y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_equality_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) (x == y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_inequality_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) (x != y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_signed_lt_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) ((int32_t) x < (int32_t) y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_signed_gt_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) ((int32_t) x > (int32_t) y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_signed_lte_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) ((int32_t) x <= (int32_t) y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_signed_gte_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) ((int32_t) x >= (int32_t) y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_signed_mul_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) ((int32_t) x * (int32_t) y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_signed_div_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) ((int32_t) x / (int32_t) y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_signed_mod_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) ((int32_t) x % (int32_t) y);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_logical_not_vmint (vmint_t x)
{
  return (vmint_t) (!x);
}

ATSinline() vmint_t
rosettacode_vm_g0uint_logical_and_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) ((!!x) * (!!y));
}

ATSinline() vmint_t
rosettacode_vm_g0uint_logical_or_vmint (vmint_t x, vmint_t y)
{
  return (vmint_t) (1 - ((!x) * (!y)));
}

%}

extern fn g0uint_add_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn g0uint_sub_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn g0uint_eq_vmint (x : vmint, y : vmint) :<> bool = "mac#%"
extern fn g0uint_neq_vmint (x : vmint, y : vmint) :<> bool = "mac#%"

implement g0uint_add<vmint_kind> (x, y) = g0uint_add_vmint (x, y)
implement g0uint_sub<vmint_kind> (x, y) = g0uint_sub_vmint (x, y)
implement g0uint_eq<vmint_kind> (x, y) = g0uint_eq_vmint (x, y)
implement g0uint_neq<vmint_kind> (x, y) = g0uint_neq_vmint (x, y)

extern fn
g0uint_signed_mul_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_signed_div_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_signed_mod_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_equality_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_inequality_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_signed_lt_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_signed_gt_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_signed_lte_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_signed_gte_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_logical_not_vmint (x : vmint) :<> vmint = "mac#%"
extern fn
g0uint_logical_and_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"
extern fn
g0uint_logical_or_vmint (x : vmint, y : vmint) :<> vmint = "mac#%"

overload signed_mul with g0uint_signed_mul_vmint
overload signed_div with g0uint_signed_div_vmint
overload signed_mod with g0uint_signed_mod_vmint
overload equality with g0uint_equality_vmint
overload inequality with g0uint_inequality_vmint
overload signed_lt with g0uint_signed_lt_vmint
overload signed_gt with g0uint_signed_gt_vmint
overload signed_lte with g0uint_signed_lte_vmint
overload signed_gte with g0uint_signed_gte_vmint
overload logical_not with g0uint_logical_not_vmint
overload logical_and with g0uint_logical_and_vmint
overload logical_or with g0uint_logical_or_vmint

fn {}
twos_complement (x : vmint) :<>
    vmint =
  (~x) + i2vm 1

fn
ensure_that_vmint_is_suitable () : void =
  {
    val _ = assertloc (u2vm (0xFFFFFFFFU) + u2vm 1U = u2vm 0U)
    val _ = assertloc (u2vm 0U - u2vm 1U = u2vm (0xFFFFFFFFU))
    val _ = assertloc (i2vm (~1234) = twos_complement (i2vm 1234))
  }

fn
parse_digits {n, i, j : int | 0 <= i; i <= j; j <= n}
             (s       : string n,
              i       : size_t i,
              j       : size_t j) :
    vmint =
  let
    val bad_integer = "Bad integer."
    fun
    loop {k : int | i <= k; k <= j} .<j - k>.
         (k : size_t k,
          x : vmint) : vmint =
      if k = j then
        x
      else if ~isdigit (s[k]) then
        $raise bad_vm (bad_integer)
      else
        (* The result is allowed to overflow freely. *)
        loop (succ k, (i2vm 10 * x) + i2vm (char2i s[k] - char2i '0'))
  in
    if j = i then
      $raise bad_vm (bad_integer)
    else
      loop (i, i2vm 0)
  end

fn
parse_integer {n, i, j : int | 0 <= i; i <= j; j <= n}
              (s       : string n,
               i       : size_t i,
               j       : size_t j) :
    vmint =
  let
    val bad_integer = "Bad integer."
  in
    if j = i then
      $raise bad_vm (bad_integer)
    else if j = succ i && ~isdigit (s[i]) then
      $raise bad_vm (bad_integer)
    else if s[i] <> '-' then
      parse_digits (s, i, j)
    else if succ i = j then
      $raise bad_vm (bad_integer)
    else
      twos_complement (parse_digits (s, succ i, j))
  end

(********************************************************************)
(*                                                                  *)
(* A linear array type for elements of vmint, byte, etc.            *)
(*                                                                  *)

vtypedef vmarray_vt (t : t@ype+, n : int, p : addr) =
  @{
    pf = @[t][n] @ p,
    pfgc = mfree_gc_v p |
    n = size_t n,
    p = ptr p
  }
vtypedef vmarray_vt (t : t@ype+, n : int) =
  [p : addr] vmarray_vt (t, n, p)

fn {t : t@ype}
vmarray_vt_alloc {n    : int}
                 (n    : size_t n,
                  fill : t) :
    [p : addr | null < p]
    vmarray_vt (t, n, p) =
  let
    val @(pf, pfgc | p) = array_ptr_alloc<t> (n)
    val _ = array_initize_elt (!p, n, fill)
  in
    @{
      pf = pf,
      pfgc = pfgc |
      n = n,
      p = p
    }
  end

fn {t : t@ype}
vmarray_vt_free {n   : int}
                {p   : addr}
                (arr : vmarray_vt (t, n, p)) :
    void =
  let
    val @{
          pf = pf,
          pfgc = pfgc |
          n = n,
          p = p
        } = arr
  in
    array_ptr_free (pf, pfgc | p)
  end

fn {t : t@ype}
vmarray_vt_fill {n    : int}
                {p    : addr}
                (arr  : !vmarray_vt (t, n, p),
                 fill : t) :
    void =
  array_initize_elt (!(arr.p), (arr.n), fill)

fn {t  : t@ype}
   {tk : tkind}
vmarray_vt_get_at_g1int {n, i : int | 0 <= i; i < n}
                        (arr  : !vmarray_vt (t, n),
                         i    : g1int (tk, i)) :
    t =
  array_get_at (!(arr.p), i)

fn {t  : t@ype}
   {tk : tkind}
vmarray_vt_get_at_g1uint {n, i : int | 0 <= i; i < n}
                         (arr  : !vmarray_vt (t, n),
                          i    : g1uint (tk, i)) :
    t =
  array_get_at (!(arr.p), i)

overload [] with vmarray_vt_get_at_g1int
overload [] with vmarray_vt_get_at_g1uint

fn {t  : t@ype}
   {tk : tkind}
vmarray_vt_set_at_g1int {n, i : int | 0 <= i; i < n}
                        (arr  : !vmarray_vt (t, n),
                         i    : g1int (tk, i),
                         x    : t) :
    void =
  array_set_at (!(arr.p), i, x)

fn {t  : t@ype}
   {tk : tkind}
vmarray_vt_set_at_g1uint {n, i : int | 0 <= i; i < n}
                         (arr  : !vmarray_vt (t, n),
                          i    : g1uint (tk, i),
                          x    : t) :
    void =
  array_set_at (!(arr.p), i, x)

overload [] with vmarray_vt_set_at_g1int
overload [] with vmarray_vt_set_at_g1uint

fn {t : t@ype}
vmarray_vt_length {n   : int}
                  (arr : !vmarray_vt (t, n)) :<>
    size_t n =
  arr.n

(********************************************************************)
(*                                                                  *)
(* Storage for the strings section.                                 *)
(*                                                                  *)

vtypedef vmstring_vt (n : int, p : addr) =
  @{
    (* A vmstring_vt is NUL-terminated, and thus there is [n + 1]
       instead of [n] in the following declaration. *)
    pf = @[char][n + 1] @ p,
    pfgc = mfree_gc_v p |
    length = size_t n,
    p = ptr p
  }
vtypedef vmstring_vt (n : int) = [p : addr] vmstring_vt (n, p)
vtypedef vmstring_vt = [n : int | 0 <= n] vmstring_vt (n)

vtypedef vmstrings_section_vt (n : int, p : addr) =
  @{
    pf = @[vmstring_vt][n] @ p,
    pfgc = mfree_gc_v p |
    n = size_t n,
    p = ptr p
  }
vtypedef vmstrings_section_vt (n : int) =
  [p : addr] vmstrings_section_vt (n, p)

fn {t : t@ype}
vmstrings_section_vt_length {n   : int}
                            (arr : !vmstrings_section_vt (n)) :<>
    size_t n =
  arr.n

fn
vmstring_vt_free {n : int}
                 {p : addr}
                 (s : vmstring_vt (n, p)) :
    void =
  array_ptr_free (s.pf, s.pfgc | s.p)

fn
vmstrings_section_vt_free {n       : int}
                          {p       : addr}
                          (strings : vmstrings_section_vt (n, p)) :
    void =
  {
    fun
    free_the_strings {n  : int | 0 <= n}
                     {p  : addr} .<n>.
                     (pf : !(@[vmstring_vt][n] @ p) >>
                              @[vmstring_vt?][n] @ p |
                      n  : size_t n,
                      p  : ptr p) : void =
      if n = 0 then
        {
          prval _ = pf :=
            array_v_unnil_nil {vmstring_vt, vmstring_vt?} pf
        }
      else
        {
          prval @(pf_element, pf_rest) = array_v_uncons pf
          val _ = vmstring_vt_free (!p)
          val p_next = ptr_succ<vmstring_vt> (p)
          val _ = free_the_strings (pf_rest | pred n, p_next)
          prval _ = pf := array_v_cons (pf_element, pf_rest)
        }

    val @{
          pf = pf,
          pfgc = pfgc |
          n = n,
          p = p
        } = strings
    prval _ = lemma_g1uint_param n
    val _ = free_the_strings (pf | n, p)
    val _ = array_ptr_free (pf, pfgc | p)
  }

fn
quoted_string_length {n : int | 0 <= n}
                     (s : string n,
                      n : size_t n) :
    [m : int | 0 <= m; m <= n - 2]
    size_t m =
  let
    val bad_quoted_string = "Bad quoted string."

    fun
    loop {i : int | 1 <= i; i <= n - 1}
         {j : int | 0 <= j; j <= i - 1} .<n - i>.
         (i : size_t i,
          j : size_t j) :
        [k : int | 0 <= k; k <= n - 2]
        size_t k =
      if i = pred n then
        j
      else if s[i] <> '\\' then
        loop (succ i, succ j)
      else if succ i = pred n then
        $raise bad_vm (bad_quoted_string)
      else if s[succ i] = 'n' || s[succ i] = '\\' then
        loop (succ (succ i), succ j)
      else
        $raise bad_vm (bad_quoted_string)
  in
    if n < i2sz 2 then
      $raise bad_vm (bad_quoted_string)
    else if s[0] <> '"' then
      $raise bad_vm (bad_quoted_string)
    else if s[pred n] <> '"' then
      $raise bad_vm (bad_quoted_string)
    else    
      loop (i2sz 1, i2sz 0)
  end

fn
dequote_string {m, n : int | 0 <= m; m <= n - 2}
               (s : string n,
                n : size_t n,
                t : !vmstring_vt m) :
    void =
  let
    fun
    loop {i : int | 1 <= i; i <= n - 1}
         {j : int | 0 <= j; j <= i - 1} .<n - i>.
         (t : !vmstring_vt m,
          i : size_t i,
          j : size_t j) : void =
      let
        macdef t_str = !(t.p)
      in
        if i = pred n then
          ()
        else if (t.length) < j then
          assertloc (false)
        else if s[i] <> '\\' then
          begin
            t_str[j] := s[i];
            loop (t, succ i, succ j)
          end
        else if succ i = pred n then
          assertloc (false)
        else if s[succ i] = 'n' then
          begin
            t_str[j] := '\n';
            loop (t, succ (succ i), succ j)
          end
        else
          begin
            t_str[j] := s[succ i];
            loop (t, succ (succ i), succ j)
          end
      end
  in
    loop (t, i2sz 1, i2sz 0)
  end        

fn
read_vmstrings {strings_size : int}
               {strings_addr : addr}
               (pf_strings   :
                    !(@[vmstring_vt?][strings_size] @ strings_addr) >>
                        @[vmstring_vt][strings_size] @ strings_addr |
                f            : FILEref,
                strings_size : size_t strings_size,
                strings      : ptr strings_addr) :
    void =
  let
    prval _ = lemma_g1uint_param strings_size

    fun
    loop {k   : int | 0 <= k; k <= strings_size} .<strings_size - k>.
         (lst : list_vt (vmstring_vt, k),
          k   : size_t k) :
        list_vt (vmstring_vt, strings_size) =
      if k = strings_size then
        list_vt_reverse (lst)
      else
        let
          val bad_quoted_string = "Bad quoted string."
          val line = fileref_get_line_string (f)
          val s = $UN.strptr2string (line)
          val n = string_length s
              val str_length = quoted_string_length (s, n)
          val (pf, pfgc | p) =
            array_ptr_alloc<char> (succ str_length)
          val _ = array_initize_elt (!p, succ str_length, '\0')
          val vmstring =
            @{
              pf = pf,
              pfgc = pfgc |
              length = str_length,
              p = p
            }
        in
          dequote_string (s, n, vmstring);
          free line;
          loop (vmstring :: lst, succ k)
        end

    val lst = loop (NIL, i2sz 0)
  in
    array_initize_list_vt<vmstring_vt>
      (!strings, sz2i strings_size, lst)
  end

fn
vmstrings_section_vt_read {strings_size : int}
                          (f            : FILEref,
                           strings_size : size_t strings_size) :
    [p : addr]
    vmstrings_section_vt (strings_size, p) =
  let
    val @(pf, pfgc | p) = array_ptr_alloc<vmstring_vt> strings_size
    val _ = read_vmstrings (pf | f, strings_size, p)
  in
    @{
      pf = pf,
      pfgc = pfgc |
      n = strings_size,
      p = p
    }
  end

fn
vmstring_fprint {n, i    : int | i < n}
                (f       : FILEref,
                 strings : !vmstrings_section_vt n,
                 i       : size_t i) :
    void =
  {

    (*
     * The following code does some ‘unsafe’ tricks. For instance, it
     * is assumed each stored string is NUL-terminated.
     *)

    fn
    print_it (str : !vmstring_vt) : void =
      fileref_puts (f, $UN.cast{string} (str.p))

    prval _ = lemma_g1uint_param i
    val p_element = array_getref_at (!(strings.p), i)
    val @(pf_element | p_element) =
      $UN.castvwtp0
        {[n : int; p : addr] @(vmstring_vt @ p | ptr p)}
        (p_element)
    val _ = print_it (!p_element)
    prval _ = $UN.castview0{void} pf_element
  }

(********************************************************************)
(*                                                                  *)
(* vm_vt: the dataviewtype for a virtual machine.                   *)
(*                                                                  *)

datavtype instruction_vt =
| instruction_vt_1 of (byte)
| instruction_vt_5 of (byte, byte, byte, byte, byte)

#define OPCODE_COUNT 24

#define OP_HALT    0x0000  // 00000
#define OP_ADD     0x0001  // 00001
#define OP_SUB     0x0002  // 00010
#define OP_MUL     0x0003  // 00011
#define OP_DIV     0x0004  // 00100
#define OP_MOD     0x0005  // 00101
#define OP_LT      0x0006  // 00110
#define OP_GT      0x0007  // 00111
#define OP_LE      0x0008  // 01000
#define OP_GE      0x0009  // 01001
#define OP_EQ      0x000A  // 01010
#define OP_NE      0x000B  // 01011
#define OP_AND     0x000C  // 01100
#define OP_OR      0x000D  // 01101
#define OP_NEG     0x000E  // 01110
#define OP_NOT     0x000F  // 01111
#define OP_PRTC    0x0010  // 10000
#define OP_PRTI    0x0011  // 10001
#define OP_PRTS    0x0012  // 10010
#define OP_FETCH   0x0013  // 10011
#define OP_STORE   0x0014  // 10100
#define OP_PUSH    0x0015  // 10101
#define OP_JMP     0x0016  // 10110
#define OP_JZ      0x0017  // 10111

#define REGISTER_PC 0
#define REGISTER_SP 1
#define MAX_REGISTER REGISTER_SP

vtypedef vm_vt (strings_size : int,
                strings_addr : addr,
                code_size    : int,
                code_addr    : addr,
                data_size    : int,
                data_addr    : addr,
                stack_size   : int,
                stack_addr   : addr) =
  @{
    strings = vmstrings_section_vt (strings_size, strings_addr),
    code = vmarray_vt (byte, code_size, code_addr),
    data = vmarray_vt (vmint, data_size, data_addr),
    stack = vmarray_vt (vmint, stack_size, stack_addr),
    registers = vmarray_vt (vmint, MAX_REGISTER + 1)
  }

vtypedef vm_vt (strings_size : int,
                code_size    : int,
                data_size    : int,
                stack_size   : int) =
  [strings_addr : addr]
  [code_addr    : addr]
  [data_addr    : addr]
  [stack_addr   : addr]
  vm_vt (strings_size, strings_addr,
         code_size, code_addr,
         data_size, data_addr,
         stack_size, stack_addr)

vtypedef vm_vt =
  [strings_size : int]
  [code_size    : int]
  [data_size    : int]
  [stack_size   : int]
  vm_vt (strings_size, code_size, data_size, stack_size)

fn
vm_vt_free (vm : vm_vt) :
    void =
  let
    val @{
          strings = strings,
          code = code,
          data = data,
          stack = stack,
          registers = registers
        } = vm
  in
    vmstrings_section_vt_free strings;
    vmarray_vt_free<byte> code;
    vmarray_vt_free<vmint> data;
    vmarray_vt_free<vmint> stack;
    vmarray_vt_free<vmint> registers
  end

fn
opcode_name_to_byte {n, i, j : int | 0 <= i; i <= j; j <= n}
                    (arr : &(@[String0][OPCODE_COUNT]),
                     str : string n,
                     i   : size_t i,
                     j   : size_t j) :
    byte =
  let
    fun
    loop {k   : int | 0 <= k; k <= OPCODE_COUNT} .<OPCODE_COUNT - k>.
         (arr : &(@[String0][OPCODE_COUNT]),
          k   : int k) : byte =
      if k = OPCODE_COUNT then
        $raise bad_vm ("Unrecognized opcode name.")
      else if substr_equal (str, i, j, arr[k]) then
        i2byte k
      else
        loop (arr, succ k)
  in
    loop (arr, 0)
  end

fn {}
vmint_byte0 (i : vmint) :<>
    byte =
  vm2byte (i land (u2vm 0xFFU))

fn {}
vmint_byte1 (i : vmint) :<>
    byte =
  vm2byte ((i >> 8) land (u2vm 0xFFU))

fn {}
vmint_byte2 (i : vmint) :<>
    byte =
  vm2byte ((i >> 16) land (u2vm 0xFFU))

fn {}
vmint_byte3 (i : vmint) :<>
    byte =
  vm2byte (i >> 24)

fn
parse_instruction {n    : int | 0 <= n}
                  (arr  : &(@[String0][OPCODE_COUNT]),
                   line : string n) :
    instruction_vt =
  let
    val bad_instruction = "Bad VM instruction."
    val n = string_length (line)
    val i = skip_whitespace (line, n, i2sz 0)

    (* Skip the address field*)
    val i = skip_non_whitespace (line, n, i)

    val i = skip_whitespace (line, n, i)
    val j = skip_non_whitespace (line, n, i)
    val opcode = opcode_name_to_byte (arr, line, i, j)

    val start_of_argument = j

    fn
    finish_push () :
        instruction_vt =
      let
        val i1 = skip_whitespace (line, n, start_of_argument)
        val j1 = skip_non_whitespace (line, n, i1)
        val arg = parse_integer (line, i1, j1)
      in
        (* Little-endian storage. *)
        instruction_vt_5 (opcode, vmint_byte0 arg, vmint_byte1 arg,
                          vmint_byte2 arg, vmint_byte3 arg)
      end

    fn
    finish_fetch_or_store () :
        instruction_vt =
      let
        val i1 = skip_whitespace (line, n, start_of_argument)
        val j1 = skip_non_whitespace (line, n, i1)
      in
        if j1 - i1 < i2sz 3 then
          $raise bad_vm (bad_instruction)
        else if line[i1] <> '\[' || line[pred j1] <> ']' then
          $raise bad_vm (bad_instruction)
        else
          let
            val arg = parse_integer (line, succ i1, pred j1)
          in
            (* Little-endian storage. *)
            instruction_vt_5 (opcode, vmint_byte0 arg, vmint_byte1 arg,
                              vmint_byte2 arg, vmint_byte3 arg)
          end
      end

    fn
    finish_jmp_or_jz () :
        instruction_vt =
      let
        val i1 = skip_whitespace (line, n, start_of_argument)
        val j1 = skip_non_whitespace (line, n, i1)
      in
        if j1 - i1 < i2sz 3 then
          $raise bad_vm (bad_instruction)
        else if line[i1] <> '\(' || line[pred j1] <> ')' then
          $raise bad_vm (bad_instruction)
        else
          let
            val arg = parse_integer (line, succ i1, pred j1)
          in
            (* Little-endian storage. *)
            instruction_vt_5 (opcode, vmint_byte0 arg, vmint_byte1 arg,
                              vmint_byte2 arg, vmint_byte3 arg)
          end
      end
  in
    case+ byte2int0 opcode of
    | OP_PUSH => finish_push ()
    | OP_FETCH => finish_fetch_or_store ()
    | OP_STORE => finish_fetch_or_store ()
    | OP_JMP => finish_jmp_or_jz ()
    | OP_JZ => finish_jmp_or_jz ()
    | _ => instruction_vt_1 (opcode)
  end

fn
read_instructions (f   : FILEref,
                   arr : &(@[String0][OPCODE_COUNT])) :
    (List_vt (instruction_vt), Size_t) =
  (* Read the instructions from the input, producing a list of
     instruction_vt objects, and also calculating the total
     number of bytes in the instructions. *)
  let
    fun
    loop (arr          : &(@[String0][OPCODE_COUNT]),
          lst          : List_vt (instruction_vt),
          bytes_needed : Size_t) :
        @(List_vt (instruction_vt), Size_t) =
      if fileref_is_eof f then
        @(list_vt_reverse lst, bytes_needed)
      else
        let
          val line = fileref_get_line_string (f)
        in
          if fileref_is_eof f then
            begin
              free line;
              @(list_vt_reverse lst, bytes_needed)
            end
          else
            let
              val instruction =
                parse_instruction (arr, $UN.strptr2string line)
              val _ = free line
              prval _ = lemma_list_vt_param lst
            in
              case+ instruction of
              | instruction_vt_1 _ =>
                loop (arr, instruction :: lst, bytes_needed + i2sz 1)
              | instruction_vt_5 _ =>
                loop (arr, instruction :: lst, bytes_needed + i2sz 5)
            end
        end
  in
    loop (arr, NIL, i2sz 0)
  end

fn
list_of_instructions_to_code {bytes_needed : int}
                             (lst          : List_vt (instruction_vt),
                              bytes_needed : size_t bytes_needed) :
    [bytes_needed : int]
    vmarray_vt (byte, bytes_needed) =
  (* This routine consumes and destroys lst. *)
  let
    fun
    loop {n    : int | 0 <= n} .<n>.
         (code : &vmarray_vt (byte, bytes_needed),
          lst  : list_vt (instruction_vt, n),
          i    : Size_t) : void =
      case+ lst of
      | ~ NIL => ()
      | ~ head :: tail =>
        begin
          case head of
          | ~ instruction_vt_1 (byte1) =>
            let
              val _ = assertloc (i < bytes_needed)
            in
              code[i] := byte1;
              loop (code, tail, i + i2sz 1)
            end
          | ~ instruction_vt_5 (byte1, byte2, byte3, byte4, byte5) =>
            let
              val _ = assertloc (i + i2sz 4 < bytes_needed)
            in
              code[i] := byte1;
              code[i + i2sz 1] := byte2;
              code[i + i2sz 2] := byte3;
              code[i + i2sz 3] := byte4;
              code[i + i2sz 4] := byte5;
              loop (code, tail, i + i2sz 5)
            end
        end

    var code = vmarray_vt_alloc<byte> (bytes_needed, i2byte OP_HALT)

    prval _ = lemma_list_vt_param lst
    prval _ = lemma_g1uint_param bytes_needed
    val _ = loop (code, lst, i2sz 0)
  in
    code
  end

fn
read_and_parse_code (f   : FILEref,
                     arr : &(@[String0][OPCODE_COUNT])) :
    [bytes_needed : int]
    vmarray_vt (byte, bytes_needed) =
  let
    val @(instructions, bytes_needed) = read_instructions (f, arr)
  in
    list_of_instructions_to_code (instructions, bytes_needed)
  end

fn
parse_header_line {n    : int | 0 <= n}
                  (line : string n) :
    @(vmint, vmint) =
  let
    val bad_vm_header_line = "Bad VM header line."
    val n = string_length (line)
    val i = skip_whitespace (line, n, i2sz 0)
    val j = skip_non_whitespace (line, n, i)
    val _ = if ~substr_equal (line, i, j, "Datasize:") then
              $raise bad_vm (bad_vm_header_line)
    val i = skip_whitespace (line, n, j)
    val j = skip_non_whitespace (line, n, i)
    val data_size = parse_integer (line, i, j)
    val i = skip_whitespace (line, n, j)
    val j = skip_non_whitespace (line, n, i)
    val _ = if ~substr_equal (line, i, j, "Strings:") then
              $raise bad_vm (bad_vm_header_line)
    val i = skip_whitespace (line, n, j)
    val j = skip_non_whitespace (line, n, i)
    val strings_size = parse_integer (line, i, j)
  in
    @(data_size, strings_size)
  end

fn
read_vm (f                : FILEref,
         opcode_names_arr : &(@[String0][OPCODE_COUNT])) :
    vm_vt =
  let
    val line = fileref_get_line_string (f)

    val @(data_size, strings_size) =
      parse_header_line ($UN.strptr2string line)

    val _ = free line

    val [data_size : int] data_size =
      g1ofg0 (vm2sz data_size)
    val [strings_size : int] strings_size =
      g1ofg0 (vm2sz strings_size)

    prval _ = lemma_g1uint_param data_size
    prval _ = lemma_g1uint_param strings_size

    prval _ = prop_verify {0 <= data_size} ()
    prval _ = prop_verify {0 <= strings_size} ()

    val strings = vmstrings_section_vt_read (f, strings_size)
    val code = read_and_parse_code (f, opcode_names_arr)
    val data = vmarray_vt_alloc<vmint> (data_size, i2vm 0)
    val stack = vmarray_vt_alloc<vmint> (vmstack_size, i2vm 0)
    val registers = vmarray_vt_alloc<vmint> (i2sz (MAX_REGISTER + 1),
                                             i2vm 0)
  in
    @{
      strings = strings,
      code = code,
      data = data,
      stack = stack,
      registers = registers
    }
  end

fn {}
pop (vm : &vm_vt) :
    vmint =
  let
    macdef registers = vm.registers
    macdef stack = vm.stack
    val sp_before = registers[REGISTER_SP]
  in
    if sp_before = i2vm 0 then
      $raise vm_runtime_error ("Stack underflow.")
    else
      let
        val sp_after = sp_before - i2vm 1
        val _ = registers[REGISTER_SP] := sp_after
        val i = g1ofg0 (vm2sz sp_after)

        (* What follows is a runtime assertion that the upper stack
           boundary is not gone past, even though it certainly will
           not. This is necessary (assuming one does not use something
           such as $UN.prop_assert) because the stack pointer is a
           vmint, whose bounds cannot be proven at compile time.

           If you comment out the assertloc, the program will not pass
           typechecking.

           Compilers for many other languages will just insert such
           checks willy-nilly, leading programmers to turn off such
           instrumentation in the very code they provide to users.

           One might be tempted to use Size_t instead for the stack
           pointer, but what if the instruction set were later
           augmented with ways to read from or write into the stack
           pointer? *)
          val _ = assertloc (i < vmarray_vt_length stack)
      in
        stack[i]
      end
  end

fn {}
push (vm : &vm_vt,
      x  : vmint) :
    void =
  let
    macdef registers = vm.registers
    macdef stack = vm.stack
    val sp_before = registers[REGISTER_SP]
    val i = g1ofg0 (vm2sz sp_before)
  in
    if vmarray_vt_length stack <= i then
      $raise vm_runtime_error ("Stack overflow.")
    else
      let
        val sp_after = sp_before + i2vm 1
      in
        registers[REGISTER_SP] := sp_after;
        stack[i] := x
      end
  end

fn {}
fetch_data (vm    : &vm_vt,
            index : vmint) :
    vmint =
  let
    macdef data = vm.data
    val i = g1ofg0 (vm2sz index)
  in
    if vmarray_vt_length data <= i then
      $raise vm_runtime_error ("Fetch from outside the data section.")
    else
      data[i]
  end

fn {}
store_data (vm    : &vm_vt,
            index : vmint,
            x     : vmint) :
    void =
  let
    macdef data = vm.data
    val i = g1ofg0 (vm2sz index)
  in
    if vmarray_vt_length data <= i then
      $raise vm_runtime_error ("Store to outside the data section.")
    else
      data[i] := x
  end

fn {}
get_argument (vm : &vm_vt) :
    vmint =
  let
    macdef code = vm.code
    macdef registers = vm.registers
    val pc = registers[REGISTER_PC]
    val i = g1ofg0 (vm2sz pc)
  in
    if vmarray_vt_length code <= i + i2sz 4 then
      $raise (vm_runtime_error
                ("The program counter is out of bounds."))
    else
      let
        (* The data is stored little-endian. *)
        val byte0 = byte2vm code[i]
        val byte1 = byte2vm code[i + i2sz 1]
        val byte2 = byte2vm code[i + i2sz 2]
        val byte3 = byte2vm code[i + i2sz 3]
      in
        (byte0) lor (byte1 << 8) lor (byte2 << 16) lor (byte3 << 24)
      end      
  end

fn {}
skip_argument (vm : &vm_vt) :
    void =
  let
    macdef registers = vm.registers
    val pc = registers[REGISTER_PC]
  in
    registers[REGISTER_PC] := pc + i2vm 4
  end

extern fun {}
unary_operation$inner : vmint -<> vmint
fn {}
unary_operation (vm : &vm_vt) :
    void =
  let
    macdef registers = vm.registers
    macdef stack = vm.stack
    val sp = registers[REGISTER_SP]
    val i = g1ofg0 (vm2sz (sp))
    prval _ = lemma_g1uint_param i
  in
    if i = i2sz 0 then
      $raise vm_runtime_error ("Stack underflow.")
    else
      let
        val _ = assertloc (i < vmarray_vt_length stack)

        (* The actual unary operation is inserted here during
           template expansion. *)
        val result = unary_operation$inner<> (stack[i - 1])
      in
        stack[i - 1] := result
      end
  end

extern fun {}
binary_operation$inner : (vmint, vmint) -<> vmint
fn {}
binary_operation (vm : &vm_vt) :
    void =
  let
    macdef registers = vm.registers
    macdef stack = vm.stack
    val sp_before = registers[REGISTER_SP]
    val i = g1ofg0 (vm2sz (sp_before))
    prval _ = lemma_g1uint_param i
  in
    if i <= i2sz 1 then
      $raise vm_runtime_error ("Stack underflow.")
    else
      let
        val _ = registers[REGISTER_SP] := sp_before - i2vm 1
        val _ = assertloc (i < vmarray_vt_length stack)

        (* The actual binary operation is inserted here during
           template expansion. *)
        val result =
          binary_operation$inner<> (stack[i - 2], stack[i - 1])
      in
        stack[i - 2] := result
      end
  end

fn {}
uop_neg (vm : &vm_vt) :
    void =
  let
    implement {}
    unary_operation$inner (x) =
      twos_complement x
  in
    unary_operation (vm)
  end

fn {}
uop_not (vm : &vm_vt) :
    void =
  let
    implement {}
    unary_operation$inner (x) =
      logical_not x
  in
    unary_operation (vm)
  end

fn {}
binop_add (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x + y
  in
    binary_operation (vm)
  end

fn {}
binop_sub (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x - y
  in
    binary_operation (vm)
  end

fn {}
binop_mul (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \signed_mul y
  in
    binary_operation (vm)
  end

fn {}
binop_div (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \signed_div y
  in
    binary_operation (vm)
  end

fn {}
binop_mod (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \signed_mod y
  in
    binary_operation (vm)
  end

fn {}
binop_eq (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \equality y
  in
    binary_operation (vm)
  end

fn {}
binop_ne (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \inequality y
  in
    binary_operation (vm)
  end

fn {}
binop_lt (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \signed_lt y
  in
    binary_operation (vm)
  end

fn {}
binop_gt (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \signed_gt y
  in
    binary_operation (vm)
  end

fn {}
binop_le (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \signed_lte y
  in
    binary_operation (vm)
  end

fn {}
binop_ge (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \signed_gte y
  in
    binary_operation (vm)
  end

fn {}
binop_and (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \logical_and y
  in
    binary_operation (vm)
  end

fn {}
binop_or (vm : &vm_vt) :
    void =
  let
    implement {}
    binary_operation$inner (x, y) =
      x \logical_or y
  in
    binary_operation (vm)
  end

fn {}
do_push (vm : &vm_vt) :
    void =
  let
    val arg = get_argument (vm)
  in
    push (vm, arg);
    skip_argument (vm)
  end

fn {}
do_fetch (vm : &vm_vt) :
    void =
  let
    val i = get_argument (vm)
    val x = fetch_data (vm, i)
  in
    push (vm, x);
    skip_argument (vm)
  end

fn {}
do_store (vm : &vm_vt) :
    void =
  let
    val i = get_argument (vm)
    val x = pop (vm)
  in
    store_data (vm, i, x);
    skip_argument (vm)
  end

fn {}
do_jmp (vm : &vm_vt) :
    void =
  let
    macdef registers = vm.registers
    val arg = get_argument (vm)
    val pc = registers[REGISTER_PC]
  in
    registers[REGISTER_PC] := pc + arg
  end

fn {}
do_jz (vm : &vm_vt) :
    void =
  let
    val x = pop (vm)
  in
    if x = i2vm 0 then
      do_jmp (vm)
    else
      skip_argument (vm)
  end

fn {}
do_prtc (f_output : FILEref,
         vm       : &vm_vt) :
    void =
  let
    val x = pop (vm)
  in
    fileref_putc (f_output, vm2i x)
  end

fn {}
do_prti (f_output : FILEref,
         vm       : &vm_vt) :
    void =
  let
    val x = pop (vm)
  in
    fprint! (f_output, vm2i x)
  end

fn {}
do_prts (f_output : FILEref,
         vm       : &vm_vt) :
    void =
  let
    val i = g1ofg0 (vm2sz (pop (vm)))
  in
    if vmstrings_section_vt_length (vm.strings) <= i then
      $raise vm_runtime_error ("String index out of bounds.")
    else
      vmstring_fprint (f_output, vm.strings, i)
  end

fn
vm_step (f_output     : FILEref,
         vm           : &vm_vt,
         machine_halt : &bool,
         bad_opcode   : &bool) :
    void =
  let
    macdef code = vm.code
    macdef registers = vm.registers

    val pc = registers[REGISTER_PC]

    val i = g1ofg0 (vm2sz (pc))
    prval _ = lemma_g1uint_param i
  in
    if vmarray_vt_length (code) <= i then
      $raise (vm_runtime_error
                ("The program counter is out of bounds."))
    else
      let
        val _ = registers[REGISTER_PC] := pc + i2vm 1

        val opcode = code[i]
        val u_opcode = byte2uint0 opcode
      in
        (* Dispatch by bifurcation on the bit pattern of the
           opcode. This method is logarithmic in the number
           of opcode values. *)
        machine_halt := false;
        bad_opcode := false;
        if (u_opcode land (~(0x1FU))) = 0U then
          begin
            if (u_opcode land 0x10U) = 0U then
              begin
                if (u_opcode land 0x08U) = 0U then
                  begin
                    if (u_opcode land 0x04U) = 0U then
                      begin
                        if (u_opcode land 0x02U) = 0U then
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              (* OP_HALT *)
                              machine_halt := true
                            else
                              binop_add (vm)
                          end
                        else
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              binop_sub (vm)
                            else
                              binop_mul (vm)
                          end
                      end
                    else
                      begin
                        if (u_opcode land 0x02U) = 0U then
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              binop_div (vm)
                            else
                              binop_mod (vm)
                          end
                        else
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              binop_lt (vm)
                            else
                              binop_gt (vm)
                          end
                      end
                  end
                else
                  begin
                    if (u_opcode land 0x04U) = 0U then
                      begin
                        if (u_opcode land 0x02U) = 0U then
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              binop_le (vm)
                            else
                              binop_ge (vm)
                          end
                        else
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              binop_eq (vm)
                            else
                              binop_ne (vm)
                          end
                      end
                    else
                      begin
                        if (u_opcode land 0x02U) = 0U then
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              binop_and (vm)
                            else
                              binop_or (vm)
                          end
                        else
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              uop_neg (vm)
                            else
                              uop_not (vm)
                          end
                      end
                  end
              end
            else
              begin
                if (u_opcode land 0x08U) = 0U then
                  begin
                    if (u_opcode land 0x04U) = 0U then
                      begin
                        if (u_opcode land 0x02U) = 0U then
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              do_prtc (f_output, vm)
                            else
                              do_prti (f_output, vm)
                          end
                        else
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              do_prts (f_output, vm)
                            else
                              do_fetch (vm)
                          end
                      end
                    else
                      begin
                        if (u_opcode land 0x02U) = 0U then
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              do_store (vm)
                            else
                              do_push (vm)
                          end
                        else
                          begin
                            if (u_opcode land 0x01U) = 0U then
                              do_jmp (vm)
                            else
                              do_jz (vm)
                          end
                      end
                  end
                else
                  bad_opcode := true
              end
          end
        else
          bad_opcode := true
      end
  end

fn
vm_continue (f_output : FILEref,
             vm       : &vm_vt) :
    void =
  let
    fun
    loop (vm           : &vm_vt,
          machine_halt : &bool,
          bad_opcode   : &bool) : void =
      if ~machine_halt && ~bad_opcode then
        begin
          vm_step (f_output, vm, machine_halt, bad_opcode);
          loop (vm, machine_halt, bad_opcode)
        end

    var machine_halt : bool = false
    var bad_opcode : bool = false
  in
    loop (vm, machine_halt, bad_opcode);
    if bad_opcode then
      $raise vm_runtime_error ("Unrecognized opcode at runtime.")
  end

fn
vm_initialize (vm : &vm_vt) :
    void =
  let
    macdef data = vm.data
    macdef registers = vm.registers
  in
    vmarray_vt_fill (data, i2vm 0);
    registers[REGISTER_PC] := i2vm 0;
    registers[REGISTER_SP] := i2vm 0
  end


fn
vm_run (f_output : FILEref,
        vm       : &vm_vt) :
    void =
  begin
    vm_initialize (vm);
    vm_continue (f_output, vm)
  end

(********************************************************************)

implement
main0 (argc, argv) =
  {
    val inpfname =
      if 2 <= argc then
        $UN.cast{string} argv[1]
      else
        "-"
    val outfname =
      if 3 <= argc then
        $UN.cast{string} argv[2]
      else
        "-"

    val inpf =
      if (inpfname : string) = "-" then
        stdin_ref
      else
        fileref_open_exn (inpfname, file_mode_r)

    val outf =
      if (outfname : string) = "-" then
        stdout_ref
      else
        fileref_open_exn (outfname, file_mode_w)

    (* The following order must match that established by
       OP_HALT, OP_ADD, OP_SUB, etc. *)
    var opcode_order =
      @[String0][OPCODE_COUNT] ("halt",  // 00000   bit pattern
                                "add",   // 00001
                                "sub",   // 00010
                                "mul",   // 00011
                                "div",   // 00100
                                "mod",   // 00101
                                "lt",    // 00110
                                "gt",    // 00111
                                "le",    // 01000
                                "ge",    // 01001
                                "eq",    // 01010
                                "ne",    // 01011
                                "and",   // 01100
                                "or",    // 01101
                                "neg",   // 01110
                                "not",   // 01111
                                "prtc",  // 10000
                                "prti",  // 10001
                                "prts",  // 10010
                                "fetch", // 10011
                                "store", // 10100
                                "push",  // 10101
                                "jmp",   // 10110
                                "jz")    // 10111

    val _ = ensure_that_vmint_is_suitable ()
    var vm = read_vm (inpf, opcode_order)
    val _ = vm_run (outf, vm)
    val _ = vm_vt_free vm
  }

(********************************************************************)
Output:

$ patscc -O3 -DATS_MEMALLOC_LIBC -o vm vm-postiats.dats -latslib && ./lex < compiler-tests/count.t | ./parse | ./gen | ./vm

count is: 1
count is: 2
count is: 3
count is: 4
count is: 5
count is: 6
count is: 7
count is: 8
count is: 9

Compiler

It seemed interesting to write translators from virtual machine code to other languages. Find at https://pastebin.com/pntTVTN3 a translator from Rosetta Code VM assembly language to ATS. The ATS program can be compiled to native code, which should run pretty fast if you use the C optimizer.

An ongoing project, to extend the translator to output languages other than ATS, is at https://sourceforge.net/p/chemoelectric/rosettacode-contributions/ci/default/tree/vmc.dats

AWK

Tested with gawk 4.1.1 and mawk 1.3.4.

function error(msg) {
  printf("%s\n", msg)
  exit(1)
}

function bytes_to_int(bstr,          i, sum) {
  sum = 0
  for (i=word_size-1; i>=0; i--) {
    sum *= 256
    sum += code[bstr+i]
  }
  return sum
}

function emit_byte(x) {
  code[next_free_code_index++] = x
}

function emit_word(x,       i) {
  for (i=0; i<word_size; i++) {
    emit_byte(int(x)%256);
    x = int(x/256)
  }
}

function run_vm(data_size) {
  sp = data_size + 1
  pc = 0
  while (1) {
    op = code[pc++]
    if (op == FETCH) {
      stack[sp++] = stack[bytes_to_int(pc)]
      pc += word_size
    } else if (op == STORE) {
      stack[bytes_to_int(pc)] = stack[--sp]
      pc += word_size
    } else if (op == PUSH) {
      stack[sp++] = bytes_to_int(pc)
      pc += word_size
    } else if (op == ADD ) { stack[sp-2] += stack[sp-1]; sp--
    } else if (op == SUB ) { stack[sp-2] -= stack[sp-1]; sp--
    } else if (op == MUL ) { stack[sp-2] *= stack[sp-1]; sp--
    } else if (op == DIV ) { stack[sp-2]  = int(stack[sp-2] / stack[sp-1]); sp--
    } else if (op == MOD ) { stack[sp-2] %= stack[sp-1]; sp--
    } else if (op == LT  ) { stack[sp-2] = stack[sp-2] <  stack[sp-1]; sp--
    } else if (op == GT  ) { stack[sp-2] = stack[sp-2] >  stack[sp-1]; sp--
    } else if (op == LE  ) { stack[sp-2] = stack[sp-2] <= stack[sp-1]; sp--
    } else if (op == GE  ) { stack[sp-2] = stack[sp-2] >= stack[sp-1]; sp--
    } else if (op == EQ  ) { stack[sp-2] = stack[sp-2] == stack[sp-1]; sp--
    } else if (op == NE  ) { stack[sp-2] = stack[sp-2] != stack[sp-1]; sp--
    } else if (op == AND ) { stack[sp-2] = stack[sp-2] && stack[sp-1]; sp--
    } else if (op == OR  ) { stack[sp-2] = stack[sp-2] || stack[sp-1]; sp--
    } else if (op == NEG ) { stack[sp-1] = - stack[sp-1]
    } else if (op == NOT ) { stack[sp-1] = ! stack[sp-1]
    } else if (op == JMP ) { pc += bytes_to_int(pc)
    } else if (op == JZ  ) { if (stack[--sp]) { pc += word_size } else { pc += bytes_to_int(pc) }
    } else if (op == PRTC) { printf("%c", stack[--sp])
    } else if (op == PRTS) { printf("%s", string_pool[stack[--sp]])
    } else if (op == PRTI) { printf("%d", stack[--sp])
    } else if (op == HALT) { break
    }
  } # while
}

function str_trans(srce,           dest, i) {
  dest = ""
  for (i=1; i <= length(srce); ) {
    if (substr(srce, i, 1) == "\\" && i < length(srce)) {
      if (substr(srce, i+1, 1) == "n") {
        dest = dest "\n"
        i += 2
      } else if (substr(srce, i+1, 1) == "\\") {
        dest = dest "\\"
        i += 2
      }
    } else {
      dest = dest substr(srce, i, 1)
      i += 1
    }
  }
  return dest
}

function load_code(            n, i) {
  getline line
  if (line ==  "")
    error("empty line")
  n=split(line, line_list)
  data_size = line_list[2]
  n_strings = line_list[4]
  for (i=0; i<n_strings; i++) {
    getline line
    gsub(/\n/, "", line)
    gsub(/"/ , "", line)
    string_pool[i] = str_trans(line)
  }
  while (getline) {
    offset = int($1)
    instr  = $2
    opcode = code_map[instr]
    if (opcode == "")
      error("Unknown instruction " instr " at " offset)
    emit_byte(opcode)
    if (opcode == JMP || opcode == JZ) {
      p = int($4)
      emit_word(p - (offset + 1))
    } else if (opcode == PUSH) {
      value = int($3)
      emit_word(value)
    } else if (opcode == FETCH || opcode == STORE) {
      gsub(/\[/, "", $3)
      gsub(/\]/, "", $3)
      value = int($3)
      emit_word(value)
    }
  }
  return data_size
}

BEGIN {
  code_map["fetch"] = FETCH =  1
  code_map["store"] = STORE =  2
  code_map["push" ] = PUSH  =  3
  code_map["add"  ] = ADD   =  4
  code_map["sub"  ] = SUB   =  5
  code_map["mul"  ] = MUL   =  6
  code_map["div"  ] = DIV   =  7
  code_map["mod"  ] = MOD   =  8
  code_map["lt"   ] = LT    =  9
  code_map["gt"   ] = GT    = 10
  code_map["le"   ] = LE    = 11
  code_map["ge"   ] = GE    = 12
  code_map["eq"   ] = EQ    = 13
  code_map["ne"   ] = NE    = 14
  code_map["and"  ] = AND   = 15
  code_map["or"   ] = OR    = 16
  code_map["neg"  ] = NEG   = 17
  code_map["not"  ] = NOT   = 18
  code_map["jmp"  ] = JMP   = 19
  code_map["jz"   ] = JZ    = 20
  code_map["prtc" ] = PRTC  = 21
  code_map["prts" ] = PRTS  = 22
  code_map["prti" ] = PRTI  = 23
  code_map["halt" ] = HALT  = 24

  next_free_node_index = 1
  next_free_code_index = 0
  word_size   = 4
  input_file = "-"
  if (ARGC > 1)
    input_file = ARGV[1]
  data_size = load_code()
  run_vm(data_size)
}
Output  —  count:

count is: 1
count is: 2
count is: 3
count is: 4
count is: 5
count is: 6
count is: 7
count is: 8
count is: 9

C

Tested with gcc 4.81 and later, compiles warning free with -Wall -Wextra

#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <string.h>
#include <stdint.h>
#include <ctype.h>

#define NELEMS(arr) (sizeof(arr) / sizeof(arr[0]))

#define da_dim(name, type)  type *name = NULL;          \
                            int _qy_ ## name ## _p = 0;  \
                            int _qy_ ## name ## _max = 0

#define da_redim(name)      do {if (_qy_ ## name ## _p >= _qy_ ## name ## _max) \
                                name = realloc(name, (_qy_ ## name ## _max += 32) * sizeof(name[0]));} while (0)

#define da_rewind(name)     _qy_ ## name ## _p = 0

#define da_append(name, x)  do {da_redim(name); name[_qy_ ## name ## _p++] = x;} while (0)

typedef unsigned char uchar;
typedef uchar code;

typedef enum { FETCH, STORE, PUSH, ADD, SUB, MUL, DIV, MOD, LT, GT, LE, GE, EQ, NE, AND,
    OR, NEG, NOT, JMP, JZ, PRTC, PRTS, PRTI, HALT
} Code_t;

typedef struct Code_map {
    char    *text;
    Code_t   op;
} Code_map;

Code_map code_map[] = {
    {"fetch",  FETCH},
    {"store",  STORE},
    {"push",   PUSH },
    {"add",    ADD  },
    {"sub",    SUB  },
    {"mul",    MUL  },
    {"div",    DIV  },
    {"mod",    MOD  },
    {"lt",     LT   },
    {"gt",     GT   },
    {"le",     LE   },
    {"ge",     GE   },
    {"eq",     EQ   },
    {"ne",     NE   },
    {"and",    AND  },
    {"or",     OR   },
    {"neg",    NEG  },
    {"not",    NOT  },
    {"jmp",    JMP  },
    {"jz",     JZ   },
    {"prtc",   PRTC },
    {"prts",   PRTS },
    {"prti",   PRTI },
    {"halt",   HALT },
};

FILE *source_fp;
da_dim(object, code);

void error(const char *fmt, ... ) {
    va_list ap;
    char buf[1000];

    va_start(ap, fmt);
    vsprintf(buf, fmt, ap);
    va_end(ap);
    printf("error: %s\n", buf);
    exit(1);
}

/*** Virtual Machine interpreter ***/
void run_vm(const code obj[], int32_t data[], int g_size, char **string_pool) {
    int32_t *sp = &data[g_size + 1];
    const code *pc = obj;

    again:
    switch (*pc++) {
        case FETCH: *sp++ = data[*(int32_t *)pc];  pc += sizeof(int32_t); goto again;
        case STORE: data[*(int32_t *)pc] = *--sp;  pc += sizeof(int32_t); goto again;
        case PUSH:  *sp++ = *(int32_t *)pc;        pc += sizeof(int32_t); goto again;
        case ADD:   sp[-2] += sp[-1]; --sp;                             goto again;
        case SUB:   sp[-2] -= sp[-1]; --sp;                             goto again;
        case MUL:   sp[-2] *= sp[-1]; --sp;                             goto again;
        case DIV:   sp[-2] /= sp[-1]; --sp;                             goto again;
        case MOD:   sp[-2] %= sp[-1]; --sp;                             goto again;
        case LT:    sp[-2] = sp[-2] <  sp[-1]; --sp;                    goto again;
        case GT:    sp[-2] = sp[-2] >  sp[-1]; --sp;                    goto again;
        case LE:    sp[-2] = sp[-2] <= sp[-1]; --sp;                    goto again;
        case GE:    sp[-2] = sp[-2] >= sp[-1]; --sp;                    goto again;
        case EQ:    sp[-2] = sp[-2] == sp[-1]; --sp;                    goto again;
        case NE:    sp[-2] = sp[-2] != sp[-1]; --sp;                    goto again;
        case AND:   sp[-2] = sp[-2] && sp[-1]; --sp;                    goto again;
        case OR:    sp[-2] = sp[-2] || sp[-1]; --sp;                    goto again;
        case NEG:   sp[-1] = -sp[-1];                                   goto again;
        case NOT:   sp[-1] = !sp[-1];                                   goto again;
        case JMP:   pc += *(int32_t *)pc;                               goto again;
        case JZ:    pc += (*--sp == 0) ? *(int32_t *)pc : (int32_t)sizeof(int32_t); goto again;
        case PRTC:  printf("%c", sp[-1]); --sp;                         goto again;
        case PRTS:  printf("%s", string_pool[sp[-1]]); --sp;            goto again;
        case PRTI:  printf("%d", sp[-1]); --sp;                         goto again;
        case HALT:                                                      break;
        default:    error("Unknown opcode %d\n", *(pc - 1));
    }
}

char *read_line(int *len) {
    static char *text = NULL;
    static int textmax = 0;

    for (*len = 0; ; (*len)++) {
        int ch = fgetc(source_fp);
        if (ch == EOF || ch == '\n') {
            if (*len == 0)
                return NULL;
            break;
        }
        if (*len + 1 >= textmax) {
            textmax = (textmax == 0 ? 128 : textmax * 2);
            text = realloc(text, textmax);
        }
        text[*len] = ch;
    }
    text[*len] = '\0';
    return text;
}

char *rtrim(char *text, int *len) {         // remove trailing spaces
    for (; *len > 0 && isspace(text[*len - 1]); --(*len))
        ;

    text[*len] = '\0';
    return text;
}

char *translate(char *st) {
    char *p, *q;
    if (st[0] == '"')                       // skip leading " if there
        ++st;
    p = q = st;

    while ((*p++ = *q++) != '\0') {
        if (q[-1] == '\\') {
            if (q[0] == 'n') {
                p[-1] = '\n';
                ++q;
            } else if (q[0] == '\\') {
                ++q;
            }
        }
        if (q[0] == '"' && q[1] == '\0')    // skip trialing " if there
            ++q;
    }

    return st;
}

/* convert an opcode string into its byte value */
int findit(const char text[], int offset) {
    for (size_t i = 0; i < sizeof(code_map) / sizeof(code_map[0]); i++) {
        if (strcmp(code_map[i].text, text) == 0)
            return code_map[i].op;
    }
    error("Unknown instruction %s at %d\n", text, offset);
    return -1;
}

void emit_byte(int c) {
    da_append(object, (uchar)c);
}

void emit_int(int32_t n) {
    union {
        int32_t n;
        unsigned char c[sizeof(int32_t)];
    } x;

    x.n = n;

    for (size_t i = 0; i < sizeof(x.n); ++i) {
        emit_byte(x.c[i]);
    }
}

/*
Datasize: 5 Strings: 3
" is prime\n"
"Total primes found: "
"\n"
 154 jmp    (-73) 82
 164 jz     (32) 197
 175 push  0
 159 fetch [4]
 149 store [3]
 */

/* Load code into global array object, return the string pool and data size */
char **load_code(int *ds) {
    int line_len, n_strings;
    char **string_pool;
    char *text = read_line(&line_len);
    text = rtrim(text, &line_len);

    strtok(text, " ");                      // skip "Datasize:"
    *ds = atoi(strtok(NULL, " "));          // get actual data_size
    strtok(NULL, " ");                      // skip "Strings:"
    n_strings = atoi(strtok(NULL, " "));    // get number of strings

    string_pool = malloc(n_strings * sizeof(char *));
    for (int i = 0; i < n_strings; ++i) {
        text = read_line(&line_len);
        text = rtrim(text, &line_len);
        text = translate(text);
        string_pool[i] = strdup(text);
    }

    for (;;) {
        int len;

        text = read_line(&line_len);
        if (text == NULL)
            break;
        text = rtrim(text, &line_len);

        int offset = atoi(strtok(text, " "));   // get the offset
        char *instr = strtok(NULL, " ");    // get the instruction
        int opcode = findit(instr, offset);
        emit_byte(opcode);
        char *operand = strtok(NULL, " ");

        switch (opcode) {
            case JMP: case JZ:
                operand++;                  // skip the '('
                len = strlen(operand);
                operand[len - 1] = '\0';    // remove the ')'
                emit_int(atoi(operand));
                break;
            case PUSH:
                emit_int(atoi(operand));
                break;
            case FETCH: case STORE:
                operand++;                  // skip the '['
                len = strlen(operand);
                operand[len - 1] = '\0';    // remove the ']'
                emit_int(atoi(operand));
                break;
        }
    }
    return string_pool;
}

void init_io(FILE **fp, FILE *std, const char mode[], const char fn[]) {
    if (fn[0] == '\0')
        *fp = std;
    else if ((*fp = fopen(fn, mode)) == NULL)
        error(0, 0, "Can't open %s\n", fn);
}

int main(int argc, char *argv[]) {
    init_io(&source_fp, stdin,  "r",  argc > 1 ? argv[1] : "");
    int data_size;
    char **string_pool = load_code(&data_size);
    int data[1000 + data_size];
    run_vm(object, data, data_size, string_pool);
}

C++

This example passes all tests, although for brevity of output only one test result is shown.

#include <cstdint>
#include <fstream>
#include <iostream>
#include <sstream>
#include <string>
#include <unordered_map>
#include <vector>

std::vector<std::string> split_string(const std::string& text, const char& delimiter) {
	std::vector<std::string> lines;
	std::istringstream stream(text);
	std::string line;
	while ( std::getline(stream, line, delimiter) ) {
		if ( ! line.empty() ) {
			lines.emplace_back(line);
		}
	}
    return lines;
}

std::string parseString(const std::string& text) {
	std::string result = "";
	uint32_t i = 0;
	while ( i < text.length() ) {
		if ( text[i] == '\\' && i + 1 < text.length() ) {
			if ( text[i + 1] == 'n' ) {
				result += "\n";
				i++;
			} else if ( text[i + 1] == '\\') {
				result += "\\";
				i++;
			}
		} else {
			result += text[i];
		}
		i++;
	}

	return result;
}

void add_to_codes(const uint32_t& number, std::vector<uint8_t>& codes) {
	for ( uint32_t i = 0; i < 32; i += 8 ) {
		codes.emplace_back((number >> i) & 0xff);
	}
}

uint32_t operand(const uint32_t& index, const std::vector<uint8_t>& codes) {
	uint32_t result = 0;
	for ( uint32_t i = index + 3; i >= index; --i ) {
		 result = ( result << 8 ) + codes[i];
	}

	return result;
}

struct VirtualMachineInfo {
	uint32_t data_size;
	std::vector<std::string> vm_strings;
	std::vector<uint8_t> codes;
};

enum class Op_code {
	HALT, ADD, SUB, MUL, DIV, MOD, LT, GT, LE, GE, EQ, NE, AND, OR, NEG, NOT,
	PRTC, PRTI, PRTS, FETCH, STORE, PUSH, JMP, JZ
};

std::unordered_map<std::string, Op_code> string_to_enum = {
	{ "halt",  Op_code::HALT  }, { "add",   Op_code::ADD   }, { "sub",   Op_code::SUB   },
	{ "mul",   Op_code::MUL   }, { "div",   Op_code::DIV   }, { "mod",   Op_code::MOD   },
	{ "lt",    Op_code::LT    }, { "gt",    Op_code::GT    }, { "le",    Op_code::LE    },
	{ "ge",    Op_code::GE    }, { "eq",    Op_code::EQ    }, { "ne",    Op_code::NE    },
	{ "and",   Op_code::AND   }, { "or",    Op_code::OR    }, { "neg",   Op_code::NEG   },
	{ "not",   Op_code::NOT   }, { "prtc",  Op_code::PRTC  }, { "prti",  Op_code::PRTI  },
	{ "prts",  Op_code::PRTS  }, { "fetch", Op_code::FETCH }, { "store", Op_code::STORE },
	{ "push",  Op_code::PUSH  }, { "jmp",   Op_code::JMP   }, { "jz",    Op_code::JZ    }
};

VirtualMachineInfo load_code(const std::string& file_path) {
	std::ifstream stream(file_path);
	std::vector<std::string> lines;
	std::string line;

	while ( std::getline(stream, line) ) {
	    lines.emplace_back(line);
	}

	line = lines.front();
	if ( line.substr(0, 3) == "lex" ) {
		lines.erase(lines.begin());
		line = lines.front();
	}

	std::vector<std::string> sections = split_string(line, ' ');
	const uint32_t data_size = std::stoi(sections[1]);
	const uint32_t string_count = std::stoi(sections[3]);

	std::vector<std::string> vm_strings = { };
	for ( uint32_t i = 1; i <= string_count; ++i ) {
		std::string content = lines[i].substr(1, lines[i].length() - 2);
		vm_strings.emplace_back(parseString(content));
	}

	uint32_t offset = 0;
	std::vector<uint8_t> codes = { };
	for ( uint32_t i = string_count + 1; i < lines.size(); ++i ) {
		sections = split_string(lines[i], ' ');
		offset = std::stoi(sections[0]);
		Op_code op_code = string_to_enum[sections[1]];
		codes.emplace_back(static_cast<uint8_t>(op_code));

		switch ( op_code ) {
			case Op_code::FETCH :
			case Op_code::STORE :
			    add_to_codes(std::stoi(sections[2].substr(1, sections[2].length() - 2)), codes); break;
			case Op_code::PUSH  : add_to_codes(std::stoi(sections[2]), codes); break;
			case Op_code::JMP   :
			case Op_code::JZ    : add_to_codes(std::stoi(sections[3]) - offset - 1, codes); break;
			default : break;
		}
	}

	return VirtualMachineInfo(data_size, vm_strings, codes);
}

void runVirtualMachine(const uint32_t& data_size, const std::vector<std::string>& vm_strings,
					   const std::vector<uint8_t>& codes) {
	const uint32_t word_size = 4;
	std::vector<int32_t> stack(data_size, 0);
	uint32_t index = 0;
	Op_code op_code;

	while ( op_code != Op_code::HALT ) {
		op_code = static_cast<Op_code>(codes[index]);
		index++;

		switch ( op_code ) {
			case Op_code::HALT  : break;
			case Op_code::ADD   : stack[stack.size() - 2] += stack.back(); stack.pop_back(); break;
			case Op_code::SUB   : stack[stack.size() - 2] -= stack.back(); stack.pop_back(); break;
			case Op_code::MUL   : stack[stack.size() - 2] *= stack.back(); stack.pop_back(); break;
			case Op_code::DIV   : stack[stack.size() - 2] /= stack.back(); stack.pop_back(); break;
			case Op_code::MOD   : stack[stack.size() - 2] %= stack.back(); stack.pop_back(); break;
			case Op_code::LT    :
				{ stack[stack.size() - 2] = ( stack[stack.size() - 2] < stack.back() ) ? 1 : 0;
				  stack.pop_back(); break;
				}
			case Op_code::GT    :
				{ stack[stack.size() - 2] = ( stack[stack.size() - 2] > stack.back() ) ? 1 : 0;
				  stack.pop_back(); break;
				}
			case Op_code::LE    :
			    { stack[stack.size() - 2] = ( stack[stack.size() - 2] <= stack.back() ) ? 1 : 0;
			      stack.pop_back(); break;
			    }
			case Op_code::GE    :
				{ stack[stack.size() - 2] = ( stack[stack.size() - 2] >= stack.back() ) ? 1 : 0;
				  stack.pop_back(); break;
				}
			case Op_code::EQ    :
				{ stack[stack.size() - 2] = ( stack[stack.size() - 2] == stack.back() ) ? 1 : 0;
				  stack.pop_back(); break;
				}
			case Op_code::NE    :
				{ stack[stack.size() - 2] = ( stack[stack.size() - 2] != stack.back() ) ? 1 : 0;
			      stack.pop_back(); break;
				}
			case Op_code::AND   :
				{ uint32_t value = ( stack[stack.size() - 2] != 0 && stack.back() != 0 ) ? 1 : 0;
				  stack[stack.size() - 2] = value; stack.pop_back(); break;
				}
			case Op_code::OR    :
				{ uint32_t value = ( stack[stack.size() - 2] != 0 || stack.back() != 0 ) ? 1 : 0;
				  stack[stack.size() - 2] = value; stack.pop_back(); break;
				}
			case Op_code::NEG   : stack.back() = -stack.back(); break;
			case Op_code::NOT   : stack.back() = ( stack.back() == 0 ) ? 1 : 0; break;
			case Op_code::PRTC  : std::cout << static_cast<char>(stack.back()); stack.pop_back(); break;
			case Op_code::PRTI  : std::cout << stack.back(); stack.pop_back(); break;
			case Op_code::PRTS  : std::cout << vm_strings[stack.back()]; stack.pop_back(); break;
			case Op_code::FETCH : { stack.emplace_back(stack[operand(index, codes)]);
									index += word_size; break;
								  }
			case Op_code::STORE : { stack[operand(index, codes)] = stack.back(); index += word_size;
									stack.pop_back(); break;
								  }
			case Op_code::PUSH  : stack.emplace_back(operand(index, codes)); index += word_size; break;
			case Op_code::JMP   : index += operand(index, codes); break;
			case Op_code::JZ    : { index += ( stack.back() == 0 ) ? operand(index, codes) : word_size;
									stack.pop_back(); break;
								  }
		}
	}
}

int main() {
	VirtualMachineInfo info = load_code("Compiler Test Cases/AsciiMandlebrot.txt");
	runVirtualMachine(info.data_size, info.vm_strings, info.codes);
}
Output:
1111111111111111111111122222222222222222222222222222222222222222222222222222222222222222222222222211111
1111111111111111111122222222222222222222222222222222222222222222222222222222222222222222222222222222211
1111111111111111112222222222222222222222222222222222222222222222222222222222222222222222222222222222222
1111111111111111222222222222222222233333333333333333333333222222222222222222222222222222222222222222222
1111111111111112222222222222333333333333333333333333333333333333222222222222222222222222222222222222222
1111111111111222222222233333333333333333333333344444456655544443333332222222222222222222222222222222222
1111111111112222222233333333333333333333333444444445567@@6665444444333333222222222222222222222222222222
11111111111222222333333333333333333333334444444445555679@@@@7654444443333333222222222222222222222222222
1111111112222223333333333333333333333444444444455556789@@@@98755544444433333332222222222222222222222222
1111111122223333333333333333333333344444444445556668@@@    @@@76555544444333333322222222222222222222222
1111111222233333333333333333333344444444455566667778@@      @987666555544433333333222222222222222222222
111111122333333333333333333333444444455556@@@@@99@@@@@@    @@@@@@877779@5443333333322222222222222222222
1111112233333333333333333334444455555556679@   @@@               @@@@@@ 8544333333333222222222222222222
1111122333333333333333334445555555556666789@@@                        @86554433333333322222222222222222
1111123333333333333444456666555556666778@@ @                         @@87655443333333332222222222222222
111123333333344444455568@887789@8777788@@@                            @@@@65444333333332222222222222222
111133334444444455555668@@@@@@@@@@@@99@@@                              @@765444333333333222222222222222
111133444444445555556778@@@         @@@@                                @855444333333333222222222222222
11124444444455555668@99@@             @                                 @655444433333333322222222222222
11134555556666677789@@                                                @86655444433333333322222222222222
111                                                                 @@876555444433333333322222222222222
11134555556666677789@@                                                @86655444433333333322222222222222
11124444444455555668@99@@             @                                 @655444433333333322222222222222
111133444444445555556778@@@         @@@@                                @855444333333333222222222222222
111133334444444455555668@@@@@@@@@@@@99@@@                              @@765444333333333222222222222222
111123333333344444455568@887789@8777788@@@                            @@@@65444333333332222222222222222
1111123333333333333444456666555556666778@@ @                         @@87655443333333332222222222222222
1111122333333333333333334445555555556666789@@@                        @86554433333333322222222222222222
1111112233333333333333333334444455555556679@   @@@               @@@@@@ 8544333333333222222222222222222
111111122333333333333333333333444444455556@@@@@99@@@@@@    @@@@@@877779@5443333333322222222222222222222
1111111222233333333333333333333344444444455566667778@@      @987666555544433333333222222222222222222222
1111111122223333333333333333333333344444444445556668@@@    @@@76555544444333333322222222222222222222222
1111111112222223333333333333333333333444444444455556789@@@@98755544444433333332222222222222222222222222
11111111111222222333333333333333333333334444444445555679@@@@7654444443333333222222222222222222222222222
1111111111112222222233333333333333333333333444444445567@@6665444444333333222222222222222222222222222222
1111111111111222222222233333333333333333333333344444456655544443333332222222222222222222222222222222222
1111111111111112222222222222333333333333333333333333333333333333222222222222222222222222222222222222222
1111111111111111222222222222222222233333333333333333333333222222222222222222222222222222222222222222222
1111111111111111112222222222222222222222222222222222222222222222222222222222222222222222222222222222222
1111111111111111111122222222222222222222222222222222222222222222222222222222222222222222222222222222211

COBOL

Code by Steve Williams (with changes to work around code highlighting issues). Tested with GnuCOBOL 2.2.

        >>SOURCE FORMAT IS FREE
identification division.
*> this code is dedicated to the public domain
*> (GnuCOBOL) 2.3-dev.0
program-id. vminterpreter.
environment division.
configuration section.
repository.  function all intrinsic.
input-output section.
file-control.
    select input-file assign using input-name
        status is input-status
        organization is line sequential.
data division.

file section.
fd  input-file.
01  input-record pic x(64).

working-storage section.
01  program-name pic x(32).
01  input-name pic x(32).
01  input-status pic xx.

01  error-record pic x(64) value spaces global.

01  v-max pic 99.
01  parameters.
    03  offset pic 999.
    03  opcode pic x(8).
    03  parm0 pic x(16).
    03  parm1 pic x(16).
    03  parm2 pic x(16).

01  opcodes.
    03  opFETCH pic x value x'00'.
    03  opSTORE pic x value x'01'.
    03  opPUSH  pic x value x'02'.
    03  opADD   pic x value x'03'.
    03  opSUB   pic x value x'04'.
    03  opMUL   pic x value x'05'.
    03  opDIV   pic x value x'06'.
    03  opMOD   pic x value x'07'.
    03  opLT    pic x value x'08'.
    03  opGT    pic x value x'09'.
    03  opLE    pic x value x'0A'.
    03  opGE    pic x value x'0B'.
    03  opEQ    pic x value x'0C'.
    03  opNE    pic x value x'0D'.
    03  opAND   pic x value x'0E'.
    03  opOR    pic x value x'0F'.
    03  opNEG   pic x value x'10'.
    03  opNOT   pic x value x'11'.
    03  opJMP   pic x value x'13'.
    03  opJZ    pic x value x'14'.
    03  opPRTC  pic x value x'15'.
    03  opPRTS  pic x value x'16'.
    03  opPRTI  pic x value x'17'.
    03  opHALT  pic x value x'18'.

01  filler.
    03  s pic 99.
    03  s-max pic 99 value 0.
    03  s-lim pic 99 value 16.
    03  filler occurs 16.
        05  string-length pic 99.
        05  string-entry pic x(48).

01  filler.
    03  v pic 99.
    03  v-lim pic 99 value 16.
    03  variables occurs 16 usage binary-int.

01  generated-code global.
    03  c  pic 999 value 1.
    03  pc pic 999.
    03  c-lim pic 999 value 512.
    03  kode pic x(512).

01  filler.
    03  stack1 pic 999 value 2.
    03  stack2 pic 999 value 1.
    03  stack-lim pic 999 value 998.
    03  stack occurs 998 usage binary-int.

01  display-definitions global.
    03  ascii-character.
        05  numeric-value usage binary-char.
    03  display-integer pic -(9)9.
    03  word-x.
        05  word usage binary-int.
    03  word-length pic 9.
    03  string1 pic 99.
    03  length1 pic 99.
    03  count1 pic 99.
    03  display-pending pic x.

procedure division.
start-vminterpreter.
    display 1 upon command-line *> get arg(1)
    accept program-name from argument-value
    move length(word) to word-length
    perform load-code
    perform run-code
    stop run
    .
run-code.
    move 1 to pc 
    perform until pc >= c
        evaluate kode(pc:1) 
        when opFETCH
            perform push-stack
            move kode(pc + 1:word-length) to word-x
            add 1 to word *> convert offset to subscript
            move variables(word) to stack(stack1)
            add word-length to pc
        when opPUSH
            perform push-stack
            move kode(pc + 1:word-length) to word-x
            move word to stack(stack1)
            add word-length to pc
        when opNEG
            compute stack(stack1) = -stack(stack1)
        when opNOT
            if stack(stack1) = 0
                move 1 to stack(stack1)
            else
                move 0 to stack(stack1)
            end-if
        when opJMP
            move kode(pc + 1:word-length) to word-x
            move word to pc
        when opHALT
            if display-pending = 'Y'
                display space
            end-if
            exit perform
        when opJZ
            if stack(stack1) = 0
                move kode(pc + 1:word-length) to word-x
                move word to pc
            else
                add word-length to pc
            end-if
            perform pop-stack
        when opSTORE
            move kode(pc + 1:word-length) to word-x
            add 1 to word *> convert offset to subscript
            move stack(stack1) to variables(word)
            add word-length to pc
            perform pop-stack
        when opADD
            add stack(stack1) to stack(stack2)
            perform pop-stack
        when opSUB
            subtract stack(stack1) from stack(stack2)
            perform pop-stack
        when opMUL
            multiply stack(stack1) by stack(stack2)
                *>rounded mode nearest-toward-zero *> doesn't match python 
            perform pop-stack
        when opDIV
            divide stack(stack1) into stack(stack2)
                *>rounded mode nearest-toward-zero *> doesn't match python 
            perform pop-stack
        when opMOD
            move mod(stack(stack2),stack(stack1)) to stack(stack2)
            perform pop-stack
        when opLT
            if stack(stack2) <  stack(stack1)
                move 1 to stack(stack2)
            else
                move 0 to stack(stack2)
            end-if
            perform pop-stack
        when opGT
            if stack(stack2) >  stack(stack1)
                move 1 to stack(stack2)
            else
                move 0 to stack(stack2)
            end-if
            perform pop-stack
        when opLE
            if stack(stack2) <= stack(stack1)
                move 1 to stack(stack2)
            else
                move 0 to stack(stack2)
            end-if
            perform pop-stack
        when opGE
            if stack(stack2) >= stack(stack1)
                move 1 to stack(stack2)
            else
                move 0 to stack(stack2)
            end-if
            perform pop-stack
        when opEQ
            if stack(stack2) = stack(stack1)
                move 1 to stack(stack2)
            else
                move 0 to stack(stack2)
            end-if
            perform pop-stack
        when opNE
            if stack(stack2) <> stack(stack1)
                move 1 to stack(stack2)
            else
                move 0 to stack(stack2)
            end-if
            perform pop-stack
        when opAND
            call "CBL_AND" using stack(stack1) stack(stack2) by value word-length
            perform pop-stack
        when opOR
            call "CBL_OR" using stack(stack1) stack(stack2) by value word-length
            perform pop-stack
        when opPRTC
            move stack(stack1) to numeric-value
            if numeric-value = 10
                display space
                move 'N' to display-pending
            else
                display ascii-character with no advancing
                move 'Y' to display-pending
            end-if
            perform pop-stack
        when opPRTS
            add 1 to word *> convert offset to subscript
            move 1 to string1
            move string-length(word) to length1
            perform until string1 > string-length(word)
                move 0 to count1
                inspect string-entry(word)(string1:length1)
                    tallying count1 for characters before initial '\'   *> ' workaround code highlighter problem
                evaluate true
                when string-entry(word)(string1 + count1 + 1:1) = 'n' *> \n
                    display string-entry(word)(string1:count1)
                    move 'N' to display-pending
                    compute string1 = string1 + 2 + count1
                    compute length1 = length1 - 2 - count1
                when string-entry(word)(string1 + count1 + 1:1) = '\' *> ' \\
                    display string-entry(word)(string1:count1 + 1) with no advancing
                    move 'Y' to display-pending
                    compute string1 = string1 + 2 + count1
                    compute length1 = length1 - 2 - count1
                when other
                    display string-entry(word)(string1:count1) with no advancing
                    move 'Y' to display-pending
                    add count1 to string1
                    subtract count1 from length1
                end-evaluate
            end-perform 
            perform pop-stack
        when opPRTI
            move stack(stack1) to display-integer
            display trim(display-integer) with no advancing
            move 'Y' to display-pending
            perform pop-stack
        end-evaluate
        add 1 to pc
    end-perform
    .
push-stack.
    if stack1 >= stack-lim
        string 'in vminterpreter at ' pc ' stack overflow at ' stack-lim into error-record
        perform report-error
    end-if
    add 1 to stack1 stack2
    >>d display ' push at ' pc space stack1 space stack2
    .
pop-stack.
    if stack1 < 2
        string 'in vminterpreter at ' pc ' stack underflow' into error-record
        perform report-error
    end-if
    >>d display ' pop at ' pc space stack1 space stack2
    subtract 1 from stack1 stack2
    .
load-code.
    perform read-input
    if input-status <> '00'
        string 'in vminterpreter no input data' into error-record
        perform report-error
    end-if

    unstring input-record delimited by all spaces into parm1 v-max parm2 s-max
    if v-max > v-lim
        string 'in vminterpreter datasize exceeds ' v-lim into error-record
        perform report-error
    end-if
    if s-max > s-lim
        string 'in vminterpreter number of strings exceeds ' s-lim into error-record
        perform report-error
    end-if

    perform read-input
    perform varying s from 1 by 1 until s > s-max
    or input-status <> '00'
        compute string-length(s) string-length(word) = length(trim(input-record)) - 2
        move input-record(2:string-length(word)) to string-entry(s)
        perform read-input
    end-perform
    if s <= s-max
        string 'in vminterpreter not all strings found' into error-record
        perform report-error
    end-if

    perform until input-status <> '00'
        initialize parameters
        unstring input-record delimited by all spaces into
            parm0 offset opcode parm1 parm2
        evaluate opcode
        when 'fetch'
            call 'emitbyte' using opFETCH
            call 'emitword' using parm1
        when 'store'
            call 'emitbyte' using opSTORE
            call 'emitword' using parm1
        when 'push'
            call 'emitbyte' using opPUSH
            call 'emitword' using parm1
        when 'add' call 'emitbyte' using opADD
        when 'sub' call 'emitbyte' using opSUB
        when 'mul' call 'emitbyte' using opMUL
        when 'div' call 'emitbyte' using opDIV
        when 'mod' call 'emitbyte' using opMOD
        when 'lt'  call 'emitbyte' using opLT
        when 'gt'  call 'emitbyte' using opGT
        when 'le'  call 'emitbyte' using opLE
        when 'ge'  call 'emitbyte' using opGE
        when 'eq'  call 'emitbyte' using opEQ
        when 'ne'  call 'emitbyte' using opNE
        when 'and' call 'emitbyte' using opAND
        when 'or'  call 'emitbyte' using opOR
        when 'not' call 'emitbyte' using opNOT
        when 'neg' call 'emitbyte' using opNEG
        when 'jmp'
             call 'emitbyte' using opJMP
             call 'emitword' using parm2
        when 'jz'
             call 'emitbyte' using opJZ
             call 'emitword' using parm2
        when 'prtc' call 'emitbyte' using opPRTC
        when 'prts' call 'emitbyte' using opPRTS
        when 'prti' call 'emitbyte' using opPRTI
        when 'halt' call 'emitbyte' using opHALT
        when other
            string 'in vminterpreter unknown opcode ' trim(opcode) ' at ' offset into error-record
            perform report-error
        end-evaluate
        perform read-input
    end-perform
    .
read-input.
    if program-name = spaces
        move '00' to input-status
        accept input-record on exception move '10' to input-status end-accept
        exit paragraph
    end-if
    if input-name = spaces
        string program-name delimited by space '.gen' into input-name
        open input input-file
        if input-status <> '00'
            string 'in vminterpreter ' trim(input-name) ' file open status ' input-status
                into error-record
            perform report-error
        end-if
    end-if
    read input-file into input-record
    evaluate input-status
    when '00'
        continue
    when '10'
        close input-file
    when other
        string 'in vminterpreter unexpected input-status: ' input-status into error-record
        perform report-error
    end-evaluate
    .
report-error.
    display error-record upon syserr
    stop run with error status -1
    .
identification division.
program-id. emitbyte.
data division.
linkage section.
01  opcode pic x.
procedure division using opcode.
start-emitbyte.
    if c >= c-lim
        string 'in vminterpreter emitbyte c exceeds ' c-lim into error-record
        call 'reporterror'
    end-if
    move opcode to kode(c:1)
    add 1 to c
    .
end program emitbyte.
   
identification division.
program-id. emitword.
data division.
working-storage section.
01  word-temp pic x(8).
linkage section.
01  word-value any length.
procedure division using word-value.
start-emitword.
    if c + word-length >= c-lim
        string 'in vminterpreter emitword c exceeds ' c-lim into error-record
        call 'reporterror'
    end-if
    move word-value to word-temp
    inspect word-temp converting '[' to ' '
    inspect word-temp converting ']' to ' '
    move numval(trim(word-temp)) to word
    move word-x to kode(c:word-length)
    add word-length to c
    .
end program emitword.

end program vminterpreter.
Output  —  Count:
prompt$ ./lexer <testcases/Count | ./parser | ./generator | ./vminterpreter 
count is: 1
count is: 2
count is: 3
count is: 4
count is: 5
count is: 6
count is: 7
count is: 8
count is: 9

Common Lisp

Works with: roswell version 21.10.14.111
Works with: SBCL version 2.2.3
Library: cl-ppcre
Library: trivia


I ran it with SBCL, CCL, and ECL. SBCL gave by far the best performance on mandel.vm, although I do not know all the optimization tricks one can employ.


#!/bin/sh
#|-*- mode:lisp -*-|#
#|
exec ros -Q -- $0 "$@"
|#
(progn ;;init forms
  (ros:ensure-asdf)
  #+quicklisp(ql:quickload '() :silent t)
  )

(defpackage :ros.script.vm.3858678051
  (:use :cl))
(in-package :ros.script.vm.3858678051)

;;;
;;; The Rosetta Code Virtual Machine, in Common Lisp.
;;;
;;; Notes:
;;;
;;;   * I have tried not to use foreign types or similar means of
;;;     optimization.
;;;
;;;   * Integers are stored in the VM's executable memory in
;;;     big-endian order. Not because I prefer it, but because I do
;;;     not want to get myself into a little-endian rut.
;;;

(require "cl-ppcre")
(require "trivia")

;;; Yes, I could compute how much memory is needed, or I could assume
;;; that the instructions are in address order. However, for *this*
;;; implementation I am going to use a large fixed-size memory and use
;;; the address fields of instructions to place the instructions.
(defconstant executable-memory-size 65536
  "The size of memory for executable code, in 8-bit words.")

;;; Similarly, I am going to have fixed size data and stack memory.
(defconstant data-memory-size 2048
  "The size of memory for stored data, in 32-bit words.")
(defconstant stack-memory-size 2048
  "The size of memory for the stack, in 32-bit words.")

;;; And so I am going to have specialized types for the different
;;; kinds of memory the platform contains. Also for its "word" and
;;; register types.
(deftype word ()
  '(unsigned-byte 32))
(deftype register ()
  '(simple-array word (1)))
(deftype executable-memory ()
  `(simple-array (unsigned-byte 8) ,(list executable-memory-size)))
(deftype data-memory ()
  `(simple-array word ,(list data-memory-size)))
(deftype stack-memory ()
  `(simple-array word ,(list stack-memory-size)))

(defconstant re-blank-line
  (ppcre:create-scanner "^\\s*$"))

(defconstant re-parse-instr-1
  (ppcre:create-scanner "^\\s*(\\d+)\\s*(.*\\S)"))

(defconstant re-parse-instr-2
  (ppcre:create-scanner "(?i)^(\\S+)\\s*(.*)"))

(defconstant re-parse-instr-3
  (ppcre:create-scanner "^[[(]?([0-9-]+)"))

(defconstant opcode-names
  #("halt"
    "add"
    "sub"
    "mul"
    "div"
    "mod"
    "lt"
    "gt"
    "le"
    "ge"
    "eq"
    "ne"
    "and"
    "or"
    "neg"
    "not"
    "prtc"
    "prti"
    "prts"
    "fetch"
    "store"
    "push"
    "jmp"
    "jz"))

(defun blank-line-p (s)
  (not (not (ppcre:scan re-blank-line s))))

(defun opcode-from-name (s)
  (position-if (lambda (name)
                 (string= s name))
               opcode-names))

(defun create-executable-memory ()
  (coerce (make-list executable-memory-size
                     :initial-element (opcode-from-name "halt"))
          'executable-memory))

(defun create-data-memory ()
  (coerce (make-list data-memory-size :initial-element 0)
          'data-memory))

(defun create-stack-memory ()
  (coerce (make-list stack-memory-size :initial-element 0)
          'stack-memory))

(defun create-register ()
  (coerce (make-list 1 :initial-element 0) 'register))

(defstruct machine
  (sp (create-register) :type register) ; Stack pointer.
  (ip (create-register) :type register) ; Instruction pointer (same
  ; thing as program counter).
  (code (create-executable-memory) :type executable-memory)
  (data (create-data-memory) :type data-memory)
  (stack (create-stack-memory) :type stack-memory)
  (strings nil)
  output *standard-output*)

(defun insert-instruction (memory instr)
  (declare (type executable-memory memory))
  (trivia:match instr
    ((list address opcode arg)
     (let ((instr-size (if arg 5 1)))
       (unless (<= (+ address instr-size) executable-memory-size)
         (warn "the VM's executable memory size is exceeded")
         (uiop:quit 1))
       (setf (elt memory address) opcode)
       (when arg
         ;; Big-endian order.
         (setf (elt memory (+ address 1)) (ldb (byte 8 24) arg))
         (setf (elt memory (+ address 2)) (ldb (byte 8 16) arg))
         (setf (elt memory (+ address 3)) (ldb (byte 8 8) arg))
         (setf (elt memory (+ address 4)) (ldb (byte 8 0) arg)))))))

(defun load-executable-memory (memory instr-lst)
  (declare (type executable-memory memory))
  (loop for instr in instr-lst
        do (insert-instruction memory instr)))

(defun parse-instruction (s)
  (if (blank-line-p s)
      nil
      (let* ((strings (nth-value 1 (ppcre:scan-to-strings
                                    re-parse-instr-1 s)))
             (address (parse-integer (elt strings 0)))
             (split (nth-value 1 (ppcre:scan-to-strings
                                  re-parse-instr-2 (elt strings 1))))
             (opcode-name (string-downcase (elt split 0)))
             (opcode (opcode-from-name opcode-name))
             (arguments (elt split 1))
             (has-arg (trivia:match opcode-name
                        ((or "fetch" "store" "push" "jmp" "jz") t)
                        (_ nil))))
        (if has-arg
            (let* ((argstr-lst
                     (nth-value 1 (ppcre:scan-to-strings
                                   re-parse-instr-3 arguments)))
                   (argstr (elt argstr-lst 0)))
              `(,address ,opcode ,(parse-integer argstr)))
            `(,address ,opcode ())))))

(defun read-instructions (inpf)
  (loop for line = (read-line inpf nil 'eoi)
        until (eq line 'eoi)
        for instr = (parse-instruction line)
        when instr collect instr))

(defun read-datasize-and-strings-count (inpf)
  (let ((line (read-line inpf)))
    (multiple-value-bind (_whole-match strings)
        ;; This is a permissive implementation.
        (ppcre:scan-to-strings
         "(?i)^\\s*Datasize\\s*:\\s*(\\d+)\\s*Strings\\s*:\\s*(\\d+)"
         line)
      (declare (ignore _whole-match))
      `(,(parse-integer (elt strings 0))
        ,(parse-integer (elt strings 1))))))

(defun parse-string-literal (s)
  ;; This is a permissive implementation, but only in that it skips
  ;; any leading space. It does not check carefully for outright
  ;; mistakes.
  (let* ((s (ppcre:regex-replace "^\\s*" s ""))
         (quote-mark (elt s 0))
         (i 1)
         (lst
           (loop until (char= (elt s i) quote-mark)
                 collect (let ((c (elt s i)))
                           (if (char= c #\\)
                               (let ((c0 (trivia:match (elt s (1+ i))
                                           (#\n #\newline)
                                           (c1 c1))))
                                 (setq i (+ i 2))
                                 c0)
                               (progn
                                 (setq i (1+ i))
                                 c))))))
    (coerce lst 'string)))

(defun read-string-literals (inpf strings-count)
  (loop for i from 1 to strings-count
        collect (parse-string-literal (read-line inpf))))

(defun open-inpf (inpf-filename)
  (if (string= inpf-filename "-")
      *standard-input*
      (open inpf-filename :direction :input)))

(defun open-outf (outf-filename)
  (if (string= outf-filename "-")
      *standard-output*
      (open outf-filename :direction :output
                          :if-exists :overwrite
                          :if-does-not-exist :create)))

(defun word-signbit-p (x)
  "True if and only if the sign bit is set."
  (declare (type word x))
  (/= 0 (logand x #x80000000)))

(defun word-add (x y)
  "Addition with overflow freely allowed."
  (declare (type word x))
  (declare (type word y))
  (coerce (logand (+ x y) #xFFFFFFFF) 'word))

(defun word-neg (x)
  "The two's complement."
  (declare (type word x))
  (word-add (logxor x #xFFFFFFFF) 1))

(defun word-sub (x y)
  "Subtraction with overflow freely allowed."
  (declare (type word x))
  (declare (type word y))
  (word-add x (word-neg y)))

(defun word-mul (x y)
  "Signed multiplication."
  (declare (type word x))
  (declare (type word y))
  (let ((x<0 (word-signbit-p x))
        (y<0 (word-signbit-p y)))
    (let ((abs-x (if x<0 (word-neg x) x))
          (abs-y (if y<0 (word-neg y) y)))
      (let* ((abs-xy (the word
                          (logand (* abs-x abs-y) #xFFFFFFFF))))
        (if x<0
            (if y<0 abs-xy (word-neg abs-xy))
            (if y<0 (word-neg abs-xy) abs-xy))))))

(defun word-div (x y)
  "The quotient after signed integer division with truncation towards
zero."
  (declare (type word x))
  (declare (type word y))
  (let ((x<0 (word-signbit-p x))
        (y<0 (word-signbit-p y)))
    (let ((abs-x (if x<0 (word-neg x) x))
          (abs-y (if y<0 (word-neg y) y)))
      (let* ((abs-x/y (the word
                           (logand (floor abs-x abs-y) #xFFFFFFFF))))
        (if x<0
            (if y<0 abs-x/y (word-neg abs-x/y))
            (if y<0 (word-neg abs-x/y) abs-x/y))))))

(defun word-mod (x y)
  "The remainder after signed integer division with truncation towards
zero."
  (declare (type word x))
  (declare (type word y))
  (let ((x<0 (word-signbit-p x))
        (y<0 (word-signbit-p y)))
    (let ((abs-x (if x<0 (word-neg x) x))
          (abs-y (if y<0 (word-neg y) y)))
      (let* ((abs-x%y (the word
                           (logand (nth-value 1 (floor abs-x abs-y))
                                   #xFFFFFFFF))))
        (if x<0 (word-neg abs-x%y) abs-x%y)))))

(defun b2i (b)
  (declare (type boolean b))
  (if b 1 0))

(defun word-lt (x y)
  "Signed comparison: is x less than y?"
  (declare (type word x))
  (declare (type word y))
  (let ((x<0 (word-signbit-p x))
        (y<0 (word-signbit-p y)))
    (b2i (if x<0
             (if y<0 (< x y) t)
             (if y<0 nil (< x y))))))

(defun word-le (x y)
  "Signed comparison: is x less than or equal to y?"
  (declare (type word x))
  (declare (type word y))
  (let ((x<0 (word-signbit-p x))
        (y<0 (word-signbit-p y)))
    (b2i (if x<0
             (if y<0 (<= x y) t)
             (if y<0 nil (<= x y))))))

(defun word-gt (x y)
  "Signed comparison: is x greater than y?"
  (declare (type word x))
  (declare (type word y))
  (let ((x<0 (word-signbit-p x))
        (y<0 (word-signbit-p y)))
    (b2i (if x<0
             (if y<0 (> x y) nil)
             (if y<0 t (> x y))))))

(defun word-ge (x y)
  "Signed comparison: is x greater than or equal to y?"
  (declare (type word x))
  (declare (type word y))
  (let ((x<0 (word-signbit-p x))
        (y<0 (word-signbit-p y)))
    (b2i (if x<0
             (if y<0 (>= x y) nil)
             (if y<0 t (>= x y))))))

(defun word-eq (x y)
  "Is x equal to y?"
  (declare (type word x))
  (declare (type word y))
  (b2i (= x y)))

(defun word-ne (x y)
  "Is x not equal to y?"
  (declare (type word x))
  (declare (type word y))
  (b2i (/= x y)))

(defun word-cmp (x)
  "The logical complement."
  (declare (type word x))
  (b2i (= x 0)))

(defun word-and (x y)
  "The logical conjunction."
  (declare (type word x))
  (declare (type word y))
  (b2i (and (/= x 0) (/= y 0))))

(defun word-or (x y)
  "The logical disjunction."
  (declare (type word x))
  (declare (type word y))
  (b2i (or (/= x 0) (/= y 0))))

(defun unop (stack sp operation)
  "Perform a unary operation on the stack."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (declare (type (function (word) word) operation))
  (let ((i (elt sp 0)))
    (unless (<= 1 i)
      (warn "stack underflow")
      (uiop:quit 1))
    (let ((x (elt stack (1- i))))
      (setf (elt stack (1- i)) (funcall operation x)))))

(defun binop (stack sp operation)
  "Perform a binary operation on the stack."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (declare (type (function (word word) word) operation))
  (let ((i (elt sp 0)))
    (unless (<= 2 i)
      (warn "stack underflow")
      (uiop:quit 1))
    (let ((x (elt stack (- i 2)))
          (y (elt stack (1- i))))
      (setf (elt stack (- i 2)) (funcall operation x y)))
    (setf (elt sp 0) (1- i))))

(defun jri (code ip)
  "Jump relative immediate."
  (declare (type executable-memory code))
  (declare (type register ip))
  ;; Big-endian order.
  (let ((j (elt ip 0)))
    (unless (<= (+ j 4) executable-memory-size)
      (warn "address past end of executable memory")
      (uiop:quit 1))
    (let* ((offset (elt code (+ j 3)))
           (offset (dpb (elt code (+ j 2)) (byte 8 8) offset))
           (offset (dpb (elt code (+ j 1)) (byte 8 16) offset))
           (offset (dpb (elt code j) (byte 8 24) offset)))
      (setf (elt ip 0) (word-add j offset)))))

(defun jriz (stack sp code ip)
  "Jump relative immediate, if zero."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (declare (type executable-memory code))
  (declare (type register ip))
  (let ((i (elt sp 0)))
    (unless (<= 1 i)
      (warn "stack underflow")
      (uiop:quit 1))
    (let ((x (elt stack (1- i))))
      (setf (elt sp 0) (1- i))
      (if (= x 0)
          (jri code ip)
          (setf (elt ip 0) (+ (elt ip 0) 4))))))

(defun get-immediate-value (code ip)
  (declare (type executable-memory code))
  (declare (type register ip))
  ;; Big-endian order.
  (let ((j (elt ip 0)))
    (unless (<= (+ j 4) executable-memory-size)
      (warn "address past end of executable memory")
      (uiop:quit 1))
    (let* ((x (elt code (+ j 3)))
           (x (dpb (elt code (+ j 2)) (byte 8 8) x))
           (x (dpb (elt code (+ j 1)) (byte 8 16) x))
           (x (dpb (elt code j) (byte 8 24) x)))
      (setf (elt ip 0) (+ j 4))
      x)))

(defun pushi (stack sp code ip)
  "Push-immediate a value from executable memory onto the stack."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (declare (type executable-memory code))
  (declare (type register ip))
  (let ((i (elt sp 0)))
    (unless (< i stack-memory-size)
      (warn "stack overflow")
      (uiop:quit 1))
    (setf (elt stack i) (get-immediate-value code ip))
    (setf (elt sp 0) (1+ i))))

(defun fetch (stack sp code ip data)
  "Fetch data to the stack, using the storage location given in
executable memory."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (declare (type executable-memory code))
  (declare (type register ip))
  (declare (type data-memory data))
  (let ((i (elt sp 0)))
    (unless (< i stack-memory-size)
      (warn "stack overflow")
      (uiop:quit 1))
    (let* ((k (get-immediate-value code ip))
           (x (elt data k)))
      (setf (elt stack i) x)
      (setf (elt sp 0) (1+ i)))))

(defun pop-one (stack sp)
  (let ((i (elt sp 0)))
    (unless (<= 1 i)
      (warn "stack underflow")
      (uiop:quit 1))
    (let* ((x (elt stack (1- i))))
      (setf (elt sp 0) (1- i))
      x)))

(defun store (stack sp code ip data)
  "Store data from the stack, using the storage location given in
executable memory."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (declare (type executable-memory code))
  (declare (type register ip))
  (declare (type data-memory data))
  (let ((i (elt sp 0)))
    (unless (<= 1 i)
      (warn "stack underflow")
      (uiop:quit 1))
    (let ((k (get-immediate-value code ip))
          (x (pop-one stack sp)))
      (setf (elt data k) x))))

(defun prti (stack sp outf)
  "Print the top value of the stack, as a signed decimal value."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (let* ((n (pop-one stack sp))
         (n<0 (word-signbit-p n)))
    (if n<0
        (format outf "-~D" (word-neg n))
        (format outf "~D" n))))

(defun prtc (stack sp outf)
  "Print the top value of the stack, as a character."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (let* ((c (pop-one stack sp)))
    (format outf "~C" (code-char c))))

(defun prts (stack sp strings outf)
  "Print the string specified by the top of the stack."
  (declare (type stack-memory stack))
  (declare (type register sp))
  (let* ((k (pop-one stack sp))
         (s (elt strings k)))
    (format outf "~A" s)))

(defmacro defun-machine-binop (op)
  (let ((machine-op (read-from-string
                     (concatenate 'string "machine-" (string op))))
        (word-op (read-from-string
                  (concatenate 'string "word-" (string op)))))
    `(defun ,machine-op (mach)
       (declare (type machine mach))
       (binop (machine-stack mach)
              (machine-sp mach)
              #',word-op))))

(defmacro defun-machine-unop (op)
  (let ((machine-op (read-from-string
                     (concatenate 'string "machine-" (string op))))
        (word-op (read-from-string
                  (concatenate 'string "word-" (string op)))))
    `(defun ,machine-op (mach)
       (declare (type machine mach))
       (unop (machine-stack mach)
             (machine-sp mach)
             #',word-op))))

(defun-machine-binop "add")
(defun-machine-binop "sub")
(defun-machine-binop "mul")
(defun-machine-binop "div")
(defun-machine-binop "mod")
(defun-machine-binop "lt")
(defun-machine-binop "gt")
(defun-machine-binop "le")
(defun-machine-binop "ge")
(defun-machine-binop "eq")
(defun-machine-binop "ne")
(defun-machine-binop "and")
(defun-machine-binop "or")

(defun-machine-unop "neg")
(defun machine-not (mach)
  (declare (type machine mach))
  (unop (machine-stack mach)
        (machine-sp mach)
        #'word-cmp))

(defun machine-prtc (mach)
  (declare (type machine mach))
  (prtc (machine-stack mach)
        (machine-sp mach)
        (machine-output mach)))

(defun machine-prti (mach)
  (declare (type machine mach))
  (prti (machine-stack mach)
        (machine-sp mach)
        (machine-output mach)))

(defun machine-prts (mach)
  (declare (type machine mach))
  (prts (machine-stack mach)
        (machine-sp mach)
        (machine-strings mach)
        (machine-output mach)))

(defun machine-fetch (mach)
  (declare (type machine mach))
  (fetch (machine-stack mach)
         (machine-sp mach)
         (machine-code mach)
         (machine-ip mach)
         (machine-data mach)))

(defun machine-store (mach)
  (declare (type machine mach))
  (store (machine-stack mach)
         (machine-sp mach)
         (machine-code mach)
         (machine-ip mach)
         (machine-data mach)))

(defun machine-push (mach)
  (declare (type machine mach))
  (pushi (machine-stack mach)
         (machine-sp mach)
         (machine-code mach)
         (machine-ip mach)))

(defun machine-jmp (mach)
  (declare (type machine mach))
  (jri (machine-code mach)
       (machine-ip mach)))

(defun machine-jz (mach)
  (declare (type machine mach))
  (jriz (machine-stack mach)
        (machine-sp mach)
        (machine-code mach)
        (machine-ip mach)))

(defun get-opcode (mach)
  (declare (type machine mach))
  (let ((code (machine-code mach))
        (ip (machine-ip mach)))
    (let ((j (elt ip 0)))
      (unless (< j executable-memory-size)
        (warn "address past end of executable memory")
        (uiop:quit 1))
      (let ((opcode (elt code j)))
        (setf (elt ip 0) (1+ j))
        opcode))))

(defun run-instruction (mach opcode)
  (declare (type machine mach))
  (declare (type fixnum opcode))
  (let ((op-mod-4 (logand opcode #x3))
        (op-div-4 (ash opcode -2)))
    (trivia:match op-div-4
      (0 (trivia:match op-mod-4
           (1 (machine-add mach))
           (2 (machine-sub mach))
           (3 (machine-mul mach))))
      (1 (trivia:match op-mod-4
           (0 (machine-div mach))
           (1 (machine-mod mach))
           (2 (machine-lt mach))
           (3 (machine-gt mach))))
      (2 (trivia:match op-mod-4
           (0 (machine-le mach))
           (1 (machine-ge mach))
           (2 (machine-eq mach))
           (3 (machine-ne mach))))
      (3 (trivia:match op-mod-4
           (0 (machine-and mach))
           (1 (machine-or mach))
           (2 (machine-neg mach))
           (3 (machine-not mach))))
      (4 (trivia:match op-mod-4
           (0 (machine-prtc mach))
           (1 (machine-prti mach))
           (2 (machine-prts mach))
           (3 (machine-fetch mach))))
      (5 (trivia:match op-mod-4
           (0 (machine-store mach))
           (1 (machine-push mach))
           (2 (machine-jmp mach))
           (3 (machine-jz mach)))))))

(defun run-vm (mach)
  (declare (type machine mach))
  (let ((opcode-for-halt (the fixnum (opcode-from-name "halt")))
        (opcode-for-add (the fixnum (opcode-from-name "add")))
        (opcode-for-jz (the fixnum (opcode-from-name "jz"))))
    (loop for opcode = (the fixnum (get-opcode mach))
          until (= opcode opcode-for-halt)
          do (progn (when (or (< opcode opcode-for-add)
                              (< opcode-for-jz opcode))
                      (warn "unsupported opcode")
                      (uiop:quit 1))
                    (run-instruction mach opcode)))))

(defun usage-error ()
  (princ "Usage: vm [INPUTFILE [OUTPUTFILE]]" *standard-output*)
  (terpri *standard-output*)
  (princ "If either INPUTFILE or OUTPUTFILE is \"-\", the respective"
         *standard-output*)
  (princ " standard I/O is used." *standard-output*)
  (terpri *standard-output*)
  (uiop:quit 1))

(defun get-filenames (argv)
  (trivia:match argv
    ((list) '("-" "-"))
    ((list inpf-filename) `(,inpf-filename "-"))
    ((list inpf-filename outf-filename) `(,inpf-filename
                                          ,outf-filename))
    (_ (usage-error))))

(defun main (&rest argv)
  (let* ((filenames (get-filenames argv))
         (inpf-filename (car filenames))
         (inpf (open-inpf inpf-filename))
         (outf-filename (cadr filenames))
         (outf (open-outf outf-filename))

         (sizes (read-datasize-and-strings-count inpf))
         (datasize (car sizes))
         (strings-count (cadr sizes))
         (strings (read-string-literals inpf strings-count))
         (instructions (read-instructions inpf))
         ;; We shall remain noncommittal about how strings are stored
         ;; on the hypothetical machine.
         (strings (coerce strings 'simple-vector))

         (mach (make-machine :strings strings
                             :output outf)))

    (unless (<= datasize data-memory-size)
      (warn "the VM's data memory size is exceeded")
      (uiop:quit 1))

    (load-executable-memory (machine-code mach) instructions)
    (run-vm mach)

    (unless (string= inpf-filename "-")
      (close inpf))
    (unless (string= outf-filename "-")
      (close outf))

    (uiop:quit 0)))

;;; vim: set ft=lisp lisp:


Output:
$ ./vm.ros compiler-tests/count.vm
count is: 1
count is: 2
count is: 3
count is: 4
count is: 5
count is: 6
count is: 7
count is: 8
count is: 9

D

Works with: gcc version 11.2.1
Works with: dmd version 2.096.1
Translation of: ATS


This program is fairly close to the ATS from which it was derived, although it differs greatly in certain details where the D code is significantly simpler, and was much easier to write.

If the D is optimized and compiled without bounds checks, the performance on the ASCII Mandelbrot seems comparable to that of the ATS. Differences, indeed, might rest mainly in the I/O library routines. It should be noted, though, that the ATS achieves bounds safety without runtime bounds checks; that is a major point in using it. Also, when debugging the D, I encountered a segfault due to assignment to a null class object; the ATS compiler would be much more likely to detect that kind of mistake.


//
// The Rosetta Code Virtual Machine in D.
//
// This code was migrated from an implementation in ATS. I have tried
// to keep it possible to compare the two languages easily, although
// in some cases the demonstration of "low level" techniques in ATS
// (such as avoiding memory leaks that might require garbage
// collection), or the use of linked lists as intermediate storage, or
// other such matters, seemed inappropriate to duplicate in D
// programming.
//
// (For example: in ATS, using a fully built linked list to initialize
// an array solves typechecking issues that simply do not exist in D's
// type system.)
//

import std.ascii;
import std.conv;
import std.stdint;
import std.stdio;
import std.string;
import std.typecons;

enum Op {
  HALT    = 0x0000,  // 00000
  ADD     = 0x0001,  // 00001
  SUB     = 0x0002,  // 00010
  MUL     = 0x0003,  // 00011
  DIV     = 0x0004,  // 00100
  MOD     = 0x0005,  // 00101
  LT      = 0x0006,  // 00110
  GT      = 0x0007,  // 00111
  LE      = 0x0008,  // 01000
  GE      = 0x0009,  // 01001
  EQ      = 0x000A,  // 01010
  NE      = 0x000B,  // 01011
  AND     = 0x000C,  // 01100
  OR      = 0x000D,  // 01101
  NEG     = 0x000E,  // 01110
  NOT     = 0x000F,  // 01111
  PRTC    = 0x0010,  // 10000
  PRTI    = 0x0011,  // 10001
  PRTS    = 0x0012,  // 10010
  FETCH   = 0x0013,  // 10011
  STORE   = 0x0014,  // 10100
  PUSH    = 0x0015,  // 10101
  JMP     = 0x0016,  // 10110
  JZ      = 0x0017   // 10111
}

const string[] opcodeOrder =
  ["halt",  // 00000   bit pattern
   "add",   // 00001
   "sub",   // 00010
   "mul",   // 00011
   "div",   // 00100
   "mod",   // 00101
   "lt",    // 00110
   "gt",    // 00111
   "le",    // 01000
   "ge",    // 01001
   "eq",    // 01010
   "ne",    // 01011
   "and",   // 01100
   "or",    // 01101
   "neg",   // 01110
   "not",   // 01111
   "prtc",  // 10000
   "prti",  // 10001
   "prts",  // 10010
   "fetch", // 10011
   "store", // 10100
   "push",  // 10101
   "jmp",   // 10110
   "jz"];   // 10111

enum Register {
  PC = 0,
  SP = 1,
  MAX = SP
}

alias vmint = uint32_t;

class VM {
  string[] strings;
  ubyte[] code;
  vmint[] data;
  vmint[] stack;
  vmint[Register.MAX + 1] registers;
}

class BadVMException : Exception
{
  this(string msg, string file = __FILE__, size_t line = __LINE__)
  {
    super(msg, file, line);
  }
}

class VMRuntimeException : Exception
{
  this(string msg, string file = __FILE__, size_t line = __LINE__)
  {
    super(msg, file, line);
  }
}

vmint
twosComplement (vmint x)
{
  // This computes the negative of x, if x is regarded as signed.
  pragma(inline);
  return (~x) + vmint(1U);
}

vmint
add (vmint x, vmint y)
{
  // This works whether x or y is regarded as unsigned or signed.
  pragma(inline);
  return x + y;
}

vmint
sub (vmint x, vmint y)
{
  // This works whether x or y is regarded as unsigned or signed.  
  pragma(inline);
  return x - y;
}

vmint
equality (vmint x, vmint y)
{
  pragma(inline);
  return vmint(x == y);
}

vmint
inequality (vmint x, vmint y)
{
  pragma(inline);
  return vmint(x != y);
}

vmint
signedLt (vmint x, vmint y)
{
  pragma(inline);
  return vmint(int32_t(x) < int32_t(y));
}

vmint
signedGt (vmint x, vmint y)
{
  pragma(inline);
  return vmint(int32_t(x) > int32_t(y));
}

vmint
signedLte (vmint x, vmint y)
{
  pragma(inline);
  return vmint(int32_t(x) <= int32_t(y));
}

vmint
signedGte (vmint x, vmint y)
{
  pragma(inline);
  return vmint(int32_t(x) >= int32_t(y));
}

vmint
signedMul (vmint x, vmint y)
{
  pragma(inline);
  return vmint(int32_t(x) * int32_t(y));
}

vmint
signedDiv (vmint x, vmint y)
{
  pragma(inline);
  return vmint(int32_t(x) / int32_t(y));
}

vmint
signedMod (vmint x, vmint y)
{
  pragma(inline);
  return vmint(int32_t(x) % int32_t(y));
}

vmint
logicalNot (vmint x)
{
  pragma(inline);
  return vmint(!x);
}

vmint
logicalAnd (vmint x, vmint y)
{
  pragma(inline);
  return vmint((!!x) * (!!y));
}

vmint
logicalOr (vmint x, vmint y)
{
  pragma(inline);
  return (vmint(1) - vmint((!x) * (!y)));
}

vmint
parseDigits (string s, size_t i, size_t j)
{
  const badInteger = "bad integer";

  if (j == i)
    throw new BadVMException (badInteger);
  auto x = vmint(0);
  for (size_t k = i; k < j; k += 1)
    if (!isDigit (s[k]))
      throw new BadVMException (badInteger);
    else
      // The result is allowed to overflow freely.
      x = (vmint(10) * x) + vmint(s[k] - '0');
  return x;
}

vmint
parseInteger (string s, size_t i, size_t j)
{
  const badInteger = "bad integer";

  vmint retval;
  if (j == i)
    throw new BadVMException (badInteger);
  else if (j == i + vmint(1) && !isDigit (s[i]))
    throw new BadVMException (badInteger);
  else if (s[i] != '-')
    retval = parseDigits (s, i, j);
  else if (j == i + vmint(1))
    throw new BadVMException (badInteger);
  else
    retval = twosComplement (parseDigits (s, i + vmint(1), j));
  return retval;
}

size_t
skipWhitespace (string s, size_t n, size_t i)
{
  while (i < n && isWhite (s[i]))
    i += 1;
  return i;
}

size_t
skipNonwhitespace (string s, size_t n, size_t i)
{
  while (i < n && !isWhite (s[i]))
    i += 1;
  return i;
}

bool
substrEqual (string s, size_t i, size_t j, string t)
{
  // Is s[i .. j-1] equal to t?

  auto retval = false;
  auto m = t.length;
  if (m == j - i)
    {
      auto k = size_t(0);
      while (k < m && s[i + k] == t[k])
        k += 1;
      retval = (k == m);
    }
  return retval;
}

string
dequoteString (string s, size_t n)
{
  const badQuotedString = "bad quoted string";

  string t = "";
  s = strip(s);
  if (s.length < 2 || s[0] != '"' || s[$ - 1] != '"')
    throw new BadVMException (badQuotedString);
  auto i = 1;
  while (i < s.length - 1)
    if (s[i] != '\\')
      {
        t ~= s[i];
        i += 1;
      }
    else if (i + 1 == s.length - 1)
      throw new BadVMException (badQuotedString);
    else if (s[i + 1] == 'n')
      {
        t ~= '\n';
        i += 2;
      }
    else if (s[i + 1] == '\\')
      {
        t ~= '\\';
        i += 2;
      }
    else
      throw new BadVMException (badQuotedString);
  return t;
}

string[]
readStrings (File f, size_t stringsSize)
{
  const badQuotedString = "Bad quoted string.";

  string[] strings;
  strings.length = stringsSize;
  for (size_t k = 0; k < stringsSize; k += 1)
    {
      auto line = f.readln();
      strings[k] = dequoteString (line, line.length);
    }
  return strings;
}

ubyte
opcodeNameTo_ubyte (string str, size_t i, size_t j)
{
  size_t k = 0;
  while (k < opcodeOrder.length &&
         !substrEqual (str, i, j, opcodeOrder[k]))
    k += 1;
  if (k == opcodeOrder.length)
    throw new BadVMException ("unrecognized opcode name");
  return to!ubyte(k);
}

ubyte
vmintByte0 (vmint i)
{
  return (i & 0xFF);
}

ubyte
vmintByte1 (vmint i)
{
  return ((i >> 8) & 0xFF);
}

ubyte
vmintByte2 (vmint i)
{
  return ((i >> 16) & 0xFF);
}

ubyte
vmintByte3 (vmint i)
{
  return (i >> 24);
}

ubyte[]
parseInstruction (string line)
{
  const bad_instruction = "bad VM instruction";

  const n = line.length;
  auto i = skipWhitespace (line, n, 0);

  // Skip the address field.
  i = skipNonwhitespace (line, n, i);

  i = skipWhitespace (line, n, i);
  auto j = skipNonwhitespace (line, n, i);
  auto opcode = opcodeNameTo_ubyte (line, i, j);

  auto startOfArgument = j;

  ubyte[] finishPush ()
  {
    const i1 = skipWhitespace (line, n, startOfArgument);
    const j1 = skipNonwhitespace (line, n, i1);
    const arg = parseInteger (line, i1, j1);
    // Little-endian storage.
    return [opcode, vmintByte0 (arg), vmintByte1 (arg),
            vmintByte2 (arg), vmintByte3 (arg)];
  }

  ubyte[] finishFetchOrStore ()
  {
    const i1 = skipWhitespace (line, n, startOfArgument);
    const j1 = skipNonwhitespace (line, n, i1);
    if (j1 - i1 < 3 || line[i1] != '[' || line[j1 - 1] != ']')
      throw new BadVMException (bad_instruction);
    const arg = parseInteger (line, i1 + 1, j1 - 1);
    // Little-endian storage.
    return [opcode, vmintByte0 (arg), vmintByte1 (arg),
            vmintByte2 (arg), vmintByte3 (arg)];
  }

  ubyte[] finishJmpOrJz ()
  {
    const i1 = skipWhitespace (line, n, startOfArgument);
    const j1 = skipNonwhitespace (line, n, i1);
    if (j1 - i1 < 3 || line[i1] != '(' || line[j1 - 1] != ')')
      throw new BadVMException (bad_instruction);
    const arg = parseInteger (line, i1 + 1, j1 - 1);
    // Little-endian storage.
    return [opcode, vmintByte0 (arg), vmintByte1 (arg),
            vmintByte2 (arg), vmintByte3 (arg)];
  }

  ubyte[] retval;
  switch (opcode)
    {
    case Op.PUSH:
      retval = finishPush ();
      break;
    case Op.FETCH:
    case Op.STORE:
      retval = finishFetchOrStore ();
      break;
    case Op.JMP:
    case Op.JZ:
      retval = finishJmpOrJz ();
      break;
    default:
      retval = [opcode];
      break;
    }

  return retval;
}

ubyte[]
readCode (File f)
{
  // Read the instructions from the input, producing an array of
  // array of instruction bytes.
  ubyte[] code = [];
  auto line = f.readln();
  while (line !is null)
    {
      code ~= parseInstruction (line);
      line = f.readln();
    }
  return code;
}

void
parseHeaderLine (string line, ref size_t dataSize,
                 ref size_t stringsSize)
{
  const bad_vm_header_line = "bad VM header line";

  const n = line.length;
  auto i = skipWhitespace (line, n, 0);
  auto j = skipNonwhitespace (line, n, i);
  if (!substrEqual (line, i, j, "Datasize:"))
    throw new BadVMException (bad_vm_header_line);
  i = skipWhitespace (line, n, j);
  j = skipNonwhitespace (line, n, i);
  dataSize = parseInteger (line, i, j);
  i = skipWhitespace (line, n, j);
  j = skipNonwhitespace (line, n, i);
  if (!substrEqual (line, i, j, "Strings:"))
    throw new BadVMException (bad_vm_header_line);
  i = skipWhitespace (line, n, j);
  j = skipNonwhitespace (line, n, i);
  stringsSize = parseInteger (line, i, j);
}

VM
readVM (File f)
{
  const line = f.readln();

  size_t dataSize;
  size_t stringsSize;
  parseHeaderLine (line, dataSize, stringsSize);

  VM vm = new VM();
  vm.strings = readStrings (f, stringsSize);
  vm.code = readCode (f);
  vm.data.length = dataSize;
  vm.stack.length = 65536; // A VERY big stack, MUCH bigger than is
                           // "reasonable" for this VM. The same size
                           // as in the ATS, however.
  vm.registers[Register.PC] = vmint(0);
  vm.registers[Register.SP] = vmint(0);

  return vm;
}

vmint
pop (VM vm)
{
  pragma(inline);
  const spBefore = vm.registers[Register.SP];
  if (spBefore == 0)
    throw new VMRuntimeException ("stack underflow");
  const spAfter = spBefore - vmint(1);
  vm.registers[Register.SP] = spAfter;
  return vm.stack[spAfter];
}

void
push (VM vm, vmint x)
{
  pragma(inline);
  const spBefore = vm.registers[Register.SP];
  if (vm.stack.length <= spBefore)
    throw new VMRuntimeException ("stack overflow");
  vm.stack[spBefore] = x;
  const spAfter = spBefore + vmint(1);
  vm.registers[Register.SP] = spAfter;
}

vmint
fetchData (VM vm, vmint index)
{
  pragma(inline);
  if (vm.data.length <= index)
    throw new VMRuntimeException
      ("fetch from outside the data section");
  return vm.data[index];
}

void
storeData (VM vm, vmint index, vmint x)
{
  pragma(inline);
  if (vm.data.length <= index)
    throw new VMRuntimeException
      ("store to outside the data section");
  vm.data[index] = x;
}

vmint
getArgument (VM vm)
{
  pragma(inline);
  auto pc = vm.registers[Register.PC];
  if (vm.code.length <= pc + vmint(4))
    throw new VMRuntimeException
      ("the program counter is out of bounds");
  // The data is stored little-endian.
  const byte0 = vmint (vm.code[pc]);
  const byte1 = vmint (vm.code[pc + vmint(1)]);
  const byte2 = vmint (vm.code[pc + vmint(2)]);
  const byte3 = vmint (vm.code[pc + vmint(3)]);
  return (byte0) | (byte1 << 8) | (byte2 << 16) | (byte3 << 24);
}

void
skipArgument (VM vm)
{
  pragma(inline);
  vm.registers[Register.PC] += vmint(4);
}

//
// The string mixins below are going to do for us *some* of what the
// ATS template system did for us. The two methods hardly resemble
// each other, but both can be used to construct function definitions
// at compile time.
//

template
UnaryOperation (alias name, alias func)
{
  const char[] UnaryOperation =
    "void " ~
    name ~ " (VM vm)
    {
      pragma(inline);
      const sp = vm.registers[Register.SP];
      if (sp == vmint(0))
        throw new VMRuntimeException (\"stack underflow\");
      const x = vm.stack[sp - vmint(1)];
      const z = " ~ func ~ " (x);
      vm.stack[sp - vmint(1)] = z;
    }";
}

template
BinaryOperation (alias name, alias func)
{
  const char[] BinaryOperation =
    "void " ~
    name ~ " (VM vm)
    {
      pragma(inline);
      const spBefore = vm.registers[Register.SP];
      if (spBefore <= vmint(1))
        throw new VMRuntimeException (\"stack underflow\");
      const spAfter = spBefore - vmint(1);
      vm.registers[Register.SP] = spAfter;
      const x = vm.stack[spAfter - vmint(1)];
      const y = vm.stack[spAfter];
      const z = " ~ func ~ "(x, y);
      vm.stack[spAfter - vmint(1)] = z;
    }";
}

mixin (UnaryOperation!("uopNeg", "twosComplement"));
mixin (UnaryOperation!("uopNot", "logicalNot"));

mixin (BinaryOperation!("binopAdd", "add"));
mixin (BinaryOperation!("binopSub", "sub"));
mixin (BinaryOperation!("binopMul", "signedMul"));
mixin (BinaryOperation!("binopDiv", "signedDiv"));
mixin (BinaryOperation!("binopMod", "signedMod"));
mixin (BinaryOperation!("binopEq", "equality"));
mixin (BinaryOperation!("binopNe", "inequality"));
mixin (BinaryOperation!("binopLt", "signedLt"));
mixin (BinaryOperation!("binopGt", "signedGt"));
mixin (BinaryOperation!("binopLe", "signedLte"));
mixin (BinaryOperation!("binopGe", "signedGte"));
mixin (BinaryOperation!("binopAnd", "logicalAnd"));
mixin (BinaryOperation!("binopOr", "logicalOr"));

void
doPush (VM vm)
{
  pragma(inline);
  const arg = getArgument (vm);
  push (vm, arg);
  skipArgument (vm);
}

void
doFetch (VM vm)
{
  pragma(inline);
  const i = getArgument (vm);
  const x = fetchData (vm, i);
  push (vm, x);
  skipArgument (vm);
}

void
doStore (VM vm)
{
  pragma(inline);
  const i = getArgument (vm);
  const x = pop (vm);
  storeData (vm, i, x);
  skipArgument (vm);
}

void
doJmp (VM vm)
{
  pragma(inline);
  const arg = getArgument (vm);
  vm.registers[Register.PC] += arg;
}

void
doJz (VM vm)
{
  pragma(inline);
  const x = pop (vm);
  if (x == vmint(0))
    doJmp (vm);
  else
    skipArgument (vm);
}

void
doPrtc (File fOut, VM vm)
{
  const x = pop (vm);
  fOut.write (to!char(x));
}

void
doPrti (File fOut, VM vm)
{
  const x = pop (vm);
  fOut.write (int32_t(x));
}

void
doPrts (File fOut, VM vm)
{
  const i = pop (vm);
  if (vm.strings.length <= i)
    throw new VMRuntimeException ("string index out of bounds");
  fOut.write (vm.strings[i]);
}

void
vmStep (File fOut, VM vm, ref bool machineHalt, ref bool badOpcode)
{
  const pc = vm.registers[Register.PC];
  if (vm.code.length <= pc)
    throw new VMRuntimeException
      ("the program counter is out of bounds");
  vm.registers[Register.PC] = pc + vmint(1);
  const opcode = vm.code[pc];
  const uOpcode = uint(opcode);

  // Dispatch by bifurcation on the bit pattern of the opcode. This
  // method is logarithmic in the number of opcode values.

  machineHalt = false;
  badOpcode = false;
  if ((uOpcode & (~0x1FU)) == 0U)
    {
      if ((uOpcode & 0x10U) == 0U)
        {
          if ((uOpcode & 0x08U) == 0U)
            {
              if ((uOpcode & 0x04U) == 0U)
                {
                  if ((uOpcode & 0x02U) == 0U)
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        machineHalt = true;
                      else
                        binopAdd (vm);
                    }
                  else
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        binopSub (vm);
                      else
                        binopMul (vm);
                    }
                }
              else
                {
                  if ((uOpcode & 0x02U) == 0U)
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        binopDiv (vm);
                      else
                        binopMod (vm);
                    }
                  else
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        binopLt (vm);
                      else
                        binopGt (vm);
                    }
                }
            }
          else
            {
              if ((uOpcode & 0x04U) == 0U)
                {
                  if ((uOpcode & 0x02U) == 0U)
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        binopLe (vm);
                      else
                        binopGe (vm);
                    }
                  else
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        binopEq (vm);
                      else
                        binopNe (vm);
                    }
                }
              else
                {
                  if ((uOpcode & 0x02U) == 0U)
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        binopAnd (vm);
                      else
                        binopOr (vm);
                    }
                  else
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        uopNeg (vm);
                      else
                        uopNot (vm);
                    }
                }
            }
        }
      else
        {
          if ((uOpcode & 0x08U) == 0U)
            {
              if ((uOpcode & 0x04U) == 0U)
                {
                  if ((uOpcode & 0x02U) == 0U)
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        doPrtc (fOut, vm);
                      else
                        doPrti (fOut, vm);
                    }
                  else
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        doPrts (fOut, vm);
                      else
                        doFetch (vm);
                    }
                }
              else
                {
                  if ((uOpcode & 0x02U) == 0U)
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        doStore (vm);
                      else
                        doPush (vm);
                    }
                  else
                    {
                      if ((uOpcode & 0x01U) == 0U)
                        doJmp (vm);
                      else
                        doJz (vm);
                    }
                }
            }
          else
            badOpcode = true;
        }
    }
  else
    badOpcode = true;
}

void
vmContinue (File fOut, VM vm)
{
  auto machineHalt = false;
  auto badOpcode = false;
  while (!machineHalt && !badOpcode)
    vmStep (fOut, vm, machineHalt, badOpcode);
  if (badOpcode)
    throw new VMRuntimeException ("unrecognized opcode at runtime");
}

void
vmInitialize (VM vm)
{
  foreach (ref x; vm.data)
    x = vmint(0);
  vm.registers[Register.PC] = vmint(0);
  vm.registers[Register.SP] = vmint(0);
}

void
vmRun (File fOut, VM vm)
{
  vmInitialize (vm);
  vmContinue (fOut, vm);
}

void
ensure_that_vmint_is_suitable ()
{
  // Try to guarantee that vmint is exactly 32 bits, and that it
  // allows overflow in either direction.
  assert (vmint(0xFFFFFFFFU) + vmint(1U) == vmint(0U));
  assert (vmint(0U) - vmint(1U) == vmint(0xFFFFFFFFU));
  assert (vmint(-1234) == twosComplement (vmint(1234)));
}

int
main (char[][] args)
{
  auto inpFilename = "-";
  auto outFilename = "-";
  if (2 <= args.length)
    inpFilename = to!string (args[1]);
  if (3 <= args.length)
    outFilename = to!string (args[2]);

  auto inpF = stdin;
  if (inpFilename != "-")
    inpF = File (inpFilename, "r");
  auto vm = readVM (inpF);
  if (inpFilename != "-")
    inpF.close();

  auto outF = stdout;
  if (outFilename != "-")
    outF = File (outFilename, "w");
  ensure_that_vmint_is_suitable ();
  vmRun (outF, vm);
  if (outFilename != "-")
    outF.close();

  return 0;
}


Output:
$ gdc -Wall -Wextra -fno-bounds-check -O3 -march=native -fno-stack-protector vm_in_D.d && ./a.out compiler-tests/count.vm
count is: 1
count is: 2
count is: 3
count is: 4
count is: 5
count is: 6
count is: 7
count is: 8
count is: 9


Forth

Tested with Gforth 0.7.3

CREATE BUF 0 ,              \ single-character look-ahead buffer
: PEEK   BUF @ 0= IF KEY BUF ! THEN BUF @ ;
: GETC   PEEK  0 BUF ! ;
: SPACE?   DUP BL = SWAP 9 14 WITHIN OR ;
: >SPACE   BEGIN PEEK SPACE? WHILE GETC DROP REPEAT ;
: DIGIT?   48 58 WITHIN ;
: >INT ( -- n)   >SPACE  0
   BEGIN  PEEK DIGIT?
   WHILE  GETC [CHAR] 0 -  SWAP 10 * +  REPEAT ;
CREATE A 0 ,
: C@A ( -- c)  A @ C@ ;
: C@A+ ( -- c)  C@A  1 CHARS A +! ;
: C!A+ ( c --)  A @ C!  1 CHARS A +! ;
: WORD ( -- c-addr)  >SPACE  PAD 1+ A !
   BEGIN PEEK SPACE? INVERT WHILE GETC C!A+ REPEAT
   >SPACE  PAD A @ OVER - 1- PAD C! ;
: >STRING ( -- c-addr)  >SPACE GETC DROP  PAD 1+ A !
   BEGIN PEEK [CHAR] " <> WHILE GETC C!A+ REPEAT
   GETC DROP  PAD A @ OVER - 1- PAD C! ;
: \INTERN ( c-addr -- c-addr)  HERE >R  A ! C@A+ DUP C,
   BEGIN DUP WHILE C@A+  
     DUP [CHAR] \ = IF DROP -1 R@ +!  C@A+ 
       [CHAR] n = IF 10 ELSE [CHAR] \ THEN
     THEN C,  1-
   REPEAT  DROP R> ;
: .   0 .R ;

CREATE DATA 0 ,
CREATE STRINGS 0 ,
: >DATA   HERE DATA !
   WORD DROP  >INT 4 * BEGIN DUP WHILE 0 C, 1- REPEAT DROP ;
: >STRINGS   HERE STRINGS !
   WORD DROP  >INT DUP >R CELLS  ALLOT
   0 BEGIN DUP R@ < WHILE 
     DUP CELLS >STRING \INTERN STRINGS @ ROT + !  1+
   REPEAT R> DROP DROP ;
: >HEADER   >DATA >STRINGS ;
: i32! ( n addr --)
   OVER           $FF AND OVER C! 1+
   OVER  8 RSHIFT $FF AND OVER C! 1+
   OVER 16 RSHIFT $FF AND OVER C! 1+
   SWAP 24 RSHIFT $FF AND SWAP C! ;
: i32@ ( addr -- n) >R  \ This is kinda slow... hmm
   R@     C@
   R@ 1 + C@  8 LSHIFT OR
   R@ 2 + C@ 16 LSHIFT OR
   R> 3 + C@ 24 LSHIFT OR
   DUP $7FFFFFFF AND SWAP $80000000 AND - ;  \ sign extend
: i32, ( n --)  HERE  4 ALLOT  i32! ;
: i32@+ ( -- n)  A @ i32@  A @ 4 + A ! ;
CREATE BYTECODE 0 ,
: @fetch   i32@+ 4 * DATA @ + i32@ ;
: @store   i32@+ 4 * DATA @ + i32! ;
: @jmp     i32@+ BYTECODE @ + A ! ;
: @jz      IF 4 A +! ELSE @jmp THEN ;
: @prts    CELLS STRINGS @ + @ COUNT TYPE ;
: @div     >R S>D R> SM/REM SWAP DROP ;
CREATE OPS
' @fetch , ' @store , ' i32@+ , ' @jmp ,   ' @jz ,
' EMIT ,   ' . ,      ' @prts , ' NEGATE , ' 0= ,
' + ,      ' - ,      ' * ,     ' @div ,   ' MOD ,
' < ,      ' > ,      ' <= ,    ' >= ,
' = ,      ' <> ,     ' AND ,   ' OR ,     ' BYE ,
CREATE #OPS 0 ,
: OP:   CREATE #OPS @ ,  1 #OPS +!  DOES> @ ;
OP: fetch  OP: store  OP: push  OP: jmp  OP: jz
OP: prtc   OP: prti   OP: prts  OP: neg  OP: not
OP: add    OP: sub    OP: mul   OP: div  OP: mod
OP: lt     OP: gt     OP: le    OP: ge
OP: eq     OP: ne     OP: and   OP: or   OP: halt
: >OP   WORD FIND
   0= IF ." Unrecognized opcode" ABORT THEN EXECUTE ;
: >i32   >INT i32, ;
: >[i32]  GETC DROP >i32 GETC DROP ;
: >OFFSET   WORD DROP ( drop relative offset) >i32 ;
CREATE >PARAM  ' >[i32] DUP , , ' >i32 , ' >OFFSET DUP , ,
: >BYTECODE   HERE >R
   BEGIN >INT DROP  >OP >R  R@ C,
     R@ 5 < IF R@ CELLS >PARAM + @ EXECUTE THEN
   R> halt = UNTIL  R> BYTECODE ! ;
: RUN   BYTECODE @ A !
   BEGIN C@A+ CELLS OPS + @ EXECUTE AGAIN ;
>HEADER >BYTECODE RUN

Fortran

Works with: gfortran version 11.2.1

Fortran 2008/2018 code with some limited use of the C preprocessor. If you are on a platform with case-sensitive filenames, and call the source file vm.F90, then gfortran will know to use the C preprocessor.

module compiler_type_kinds
  use, intrinsic :: iso_fortran_env, only: int32
  use, intrinsic :: iso_fortran_env, only: int64

  implicit none
  private

  ! Synonyms.
  integer, parameter, public :: size_kind = int64
  integer, parameter, public :: length_kind = size_kind
  integer, parameter, public :: nk = size_kind

  ! Synonyms for character capable of storing a Unicode code point.
  integer, parameter, public :: unicode_char_kind = selected_char_kind ('ISO_10646')
  integer, parameter, public :: ck = unicode_char_kind

  ! Synonyms for integers capable of storing a Unicode code point.
  integer, parameter, public :: unicode_ichar_kind = int32
  integer, parameter, public :: ick = unicode_ichar_kind

  ! Synonyms for integers in the virtual machine or the interpreter’s
  ! runtime. (The Rosetta Code task says integers in the virtual
  ! machine are 32-bit, but there is nothing in the task that prevents
  ! us using 64-bit integers in the compiler and interpreter.)
  integer, parameter, public :: runtime_int_kind = int64
  integer, parameter, public :: rik = runtime_int_kind
end module compiler_type_kinds

module helpers
  use, non_intrinsic :: compiler_type_kinds, only: nk, rik, ck

  implicit none
  private

  public :: new_storage_size
  public :: next_power_of_two

  public :: isspace
  public :: quoted_string

  public :: int32_to_vm_bytes
  public :: uint32_to_vm_bytes
  public :: int32_from_vm_bytes
  public :: uint32_from_vm_bytes

  public :: bool2int

  character(1, kind = ck), parameter, public :: horizontal_tab_char = char (9, kind = ck)
  character(1, kind = ck), parameter, public :: linefeed_char = char (10, kind = ck)
  character(1, kind = ck), parameter, public :: vertical_tab_char = char (11, kind = ck)
  character(1, kind = ck), parameter, public :: formfeed_char = char (12, kind = ck)
  character(1, kind = ck), parameter, public :: carriage_return_char = char (13, kind = ck)
  character(1, kind = ck), parameter, public :: space_char = ck_' '

  ! The following is correct for Unix and its relatives.
  character(1, kind = ck), parameter, public :: newline_char = linefeed_char

  character(1, kind = ck), parameter, public :: backslash_char = char (92, kind = ck)

contains

  elemental function new_storage_size (length_needed) result (size)
    integer(kind = nk), intent(in) :: length_needed
    integer(kind = nk) :: size

    ! Increase storage by orders of magnitude.

    if (2_nk**32 < length_needed) then
       size = huge (1_nk)
    else
       size = next_power_of_two (length_needed)
    end if
  end function new_storage_size

  elemental function next_power_of_two (x) result (y)
    integer(kind = nk), intent(in) :: x
    integer(kind = nk) :: y

    !
    ! It is assumed that no more than 64 bits are used.
    !
    ! The branch-free algorithm is that of
    ! https://archive.is/nKxAc#RoundUpPowerOf2
    !
    ! Fill in bits until one less than the desired power of two is
    ! reached, and then add one.
    !

    y = x - 1
    y = ior (y, ishft (y, -1))
    y = ior (y, ishft (y, -2))
    y = ior (y, ishft (y, -4))
    y = ior (y, ishft (y, -8))
    y = ior (y, ishft (y, -16))
    y = ior (y, ishft (y, -32))
    y = y + 1
  end function next_power_of_two

  elemental function isspace (ch) result (bool)
    character(1, kind = ck), intent(in) :: ch
    logical :: bool

    bool = (ch == horizontal_tab_char) .or.  &
         & (ch == linefeed_char) .or.        &
         & (ch == vertical_tab_char) .or.    &
         & (ch == formfeed_char) .or.        &
         & (ch == carriage_return_char) .or. &
         & (ch == space_char)
  end function isspace

  function quoted_string (str) result (qstr)
    character(*, kind = ck), intent(in) :: str
    character(:, kind = ck), allocatable :: qstr

    integer(kind = nk) :: n, i, j

    ! Compute n = the size of qstr.
    n = 2_nk
    do i = 1_nk, len (str, kind = nk)
       select case (str(i:i))
       case (newline_char, backslash_char)
          n = n + 2
       case default
          n = n + 1
       end select
    end do

    allocate (character(n, kind = ck) :: qstr)

    ! Quote the string.
    qstr(1:1) = ck_'"'
    j = 2_nk
    do i = 1_nk, len (str, kind = nk)
       select case (str(i:i))
       case (newline_char)
          qstr(j:j) = backslash_char
          qstr((j + 1):(j + 1)) = ck_'n'
          j = j + 2
       case (backslash_char)
          qstr(j:j) = backslash_char
          qstr((j + 1):(j + 1)) = backslash_char
          j = j + 2
       case default
          qstr(j:j) = str(i:i)
          j = j + 1
       end select
    end do
    if (j /= n) error stop      ! Check code correctness.
    qstr(n:n) = ck_'"'
  end function quoted_string

  subroutine int32_to_vm_bytes (n, bytes, i)
    integer(kind = rik), intent(in) :: n
    character(1), intent(inout) :: bytes(0:*)
    integer(kind = rik), intent(in) :: i

    !
    ! The virtual machine is presumed to be little-endian. Because I
    ! slightly prefer little-endian.
    !

    bytes(i) = achar (ibits (n, 0, 8))
    bytes(i + 1) = achar (ibits (n, 8, 8))
    bytes(i + 2) = achar (ibits (n, 16, 8))
    bytes(i + 3) = achar (ibits (n, 24, 8))
  end subroutine int32_to_vm_bytes

  subroutine uint32_to_vm_bytes (n, bytes, i)
    integer(kind = rik), intent(in) :: n
    character(1), intent(inout) :: bytes(0:*)
    integer(kind = rik), intent(in) :: i

    call int32_to_vm_bytes (n, bytes, i)
  end subroutine uint32_to_vm_bytes

  subroutine int32_from_vm_bytes (n, bytes, i)
    integer(kind = rik), intent(out) :: n
    character(1), intent(in) :: bytes(0:*)
    integer(kind = rik), intent(in) :: i

    !
    ! The virtual machine is presumed to be little-endian. Because I
    ! slightly prefer little-endian.
    !

    call uint32_from_vm_bytes (n, bytes, i)
    if (ibits (n, 31, 1) == 1) then
       ! Extend the sign bit.
       n = ior (n, not ((2_rik ** 32) - 1))
    end if
  end subroutine int32_from_vm_bytes

  subroutine uint32_from_vm_bytes (n, bytes, i)
    integer(kind = rik), intent(out) :: n
    character(1), intent(in) :: bytes(0:*)
    integer(kind = rik), intent(in) :: i

    !
    ! The virtual machine is presumed to be little-endian. Because I
    ! slightly prefer little-endian.
    !

    integer(kind = rik) :: n0, n1, n2, n3

    n0 = iachar (bytes(i), kind = rik)
    n1 = ishft (iachar (bytes(i + 1), kind = rik), 8)
    n2 = ishft (iachar (bytes(i + 2), kind = rik), 16)
    n3 = ishft (iachar (bytes(i + 3), kind = rik), 24)
    n = ior (n0, ior (n1, ior (n2, n3)))
  end subroutine uint32_from_vm_bytes

  elemental function bool2int (bool) result (int)
    logical, intent(in) :: bool
    integer(kind = rik) :: int

    if (bool) then
       int = 1_rik
    else
       int = 0_rik
    end if
  end function bool2int

end module helpers

module string_buffers
  use, intrinsic :: iso_fortran_env, only: error_unit
  use, intrinsic :: iso_fortran_env, only: int64
  use, non_intrinsic :: compiler_type_kinds, only: nk, ck, ick
  use, non_intrinsic :: helpers

  implicit none
  private

  public :: strbuf_t
  public :: skip_whitespace
  public :: skip_non_whitespace
  public :: skip_whitespace_backwards
  public :: at_end_of_line

  type :: strbuf_t
     integer(kind = nk), private :: len = 0
     !
     ! ‘chars’ is made public for efficient access to the individual
     ! characters.
     !
     character(1, kind = ck), allocatable, public :: chars(:)
   contains
     procedure, pass, private :: ensure_storage => strbuf_t_ensure_storage
     procedure, pass :: to_unicode_full_string => strbuf_t_to_unicode_full_string
     procedure, pass :: to_unicode_substring => strbuf_t_to_unicode_substring
     procedure, pass :: length => strbuf_t_length
     procedure, pass :: set => strbuf_t_set
     procedure, pass :: append => strbuf_t_append
     generic :: to_unicode => to_unicode_full_string
     generic :: to_unicode => to_unicode_substring
     generic :: assignment(=) => set
  end type strbuf_t

contains

  function strbuf_t_to_unicode_full_string (strbuf) result (s)
    class(strbuf_t), intent(in) :: strbuf
    character(:, kind = ck), allocatable :: s

    !
    ! This does not actually ensure that the string is valid Unicode;
    ! any 31-bit ‘character’ is supported.
    !

    integer(kind = nk) :: i

    allocate (character(len = strbuf%len, kind = ck) :: s)
    do i = 1, strbuf%len
       s(i:i) = strbuf%chars(i)
    end do
  end function strbuf_t_to_unicode_full_string

  function strbuf_t_to_unicode_substring (strbuf, i, j) result (s)
    !
    ! ‘Extreme’ values of i and j are allowed, as shortcuts for ‘from
    ! the beginning’, ‘up to the end’, or ‘empty substring’.
    !
    class(strbuf_t), intent(in) :: strbuf
    integer(kind = nk), intent(in) :: i, j
    character(:, kind = ck), allocatable :: s

    !
    ! This does not actually ensure that the string is valid Unicode;
    ! any 31-bit ‘character’ is supported.
    !

    integer(kind = nk) :: i1, j1
    integer(kind = nk) :: n
    integer(kind = nk) :: k

    i1 = max (1_nk, i)
    j1 = min (strbuf%len, j)
    n = max (0_nk, (j1 - i1) + 1_nk)

    allocate (character(n, kind = ck) :: s)
    do k = 1, n
       s(k:k) = strbuf%chars(i1 + (k - 1_nk))
    end do
  end function strbuf_t_to_unicode_substring

  elemental function strbuf_t_length (strbuf) result (n)
    class(strbuf_t), intent(in) :: strbuf
    integer(kind = nk) :: n

    n = strbuf%len
  end function strbuf_t_length

  subroutine strbuf_t_ensure_storage (strbuf, length_needed)
    class(strbuf_t), intent(inout) :: strbuf
    integer(kind = nk), intent(in) :: length_needed

    integer(kind = nk) :: len_needed
    integer(kind = nk) :: new_size
    type(strbuf_t) :: new_strbuf

    len_needed = max (length_needed, 1_nk)

    if (.not. allocated (strbuf%chars)) then
       ! Initialize a new strbuf%chars array.
       new_size = new_storage_size (len_needed)
       allocate (strbuf%chars(1:new_size))
    else if (ubound (strbuf%chars, 1) < len_needed) then
       ! Allocate a new strbuf%chars array, larger than the current
       ! one, but containing the same characters.
       new_size = new_storage_size (len_needed)
       allocate (new_strbuf%chars(1:new_size))
       new_strbuf%chars(1:strbuf%len) = strbuf%chars(1:strbuf%len)
       call move_alloc (new_strbuf%chars, strbuf%chars)
    end if
  end subroutine strbuf_t_ensure_storage

  subroutine strbuf_t_set (dst, src)
    class(strbuf_t), intent(inout) :: dst
    class(*), intent(in) :: src

    integer(kind = nk) :: n
    integer(kind = nk) :: i

    select type (src)
    type is (character(*, kind = ck))
       n = len (src, kind = nk)
       call dst%ensure_storage(n)
       do i = 1, n
          dst%chars(i) = src(i:i)
       end do
       dst%len = n
    type is (character(*))
       n = len (src, kind = nk)
       call dst%ensure_storage(n)
       do i = 1, n
          dst%chars(i) = src(i:i)
       end do
       dst%len = n
    class is (strbuf_t)
       n = src%len
       call dst%ensure_storage(n)
       dst%chars(1:n) = src%chars(1:n)
       dst%len = n
    class default
       error stop
    end select
  end subroutine strbuf_t_set

  subroutine strbuf_t_append (dst, src)
    class(strbuf_t), intent(inout) :: dst
    class(*), intent(in) :: src

    integer(kind = nk) :: n_dst, n_src, n
    integer(kind = nk) :: i

    select type (src)
    type is (character(*, kind = ck))
       n_dst = dst%len
       n_src = len (src, kind = nk)
       n = n_dst + n_src
       call dst%ensure_storage(n)
       do i = 1, n_src
          dst%chars(n_dst + i) = src(i:i)
       end do
       dst%len = n
    type is (character(*))
       n_dst = dst%len
       n_src = len (src, kind = nk)
       n = n_dst + n_src
       call dst%ensure_storage(n)
       do i = 1, n_src
          dst%chars(n_dst + i) = src(i:i)
       end do
       dst%len = n
    class is (strbuf_t)
       n_dst = dst%len
       n_src = src%len
       n = n_dst + n_src
       call dst%ensure_storage(n)
       dst%chars((n_dst + 1):n) = src%chars(1:n_src)
       dst%len = n
    class default
       error stop
    end select
  end subroutine strbuf_t_append

  function skip_whitespace (strbuf, i) result (j)
    class(strbuf_t), intent(in) :: strbuf
    integer(kind = nk), intent(in) :: i
    integer(kind = nk) :: j

    logical :: done

    j = i
    done = .false.
    do while (.not. done)
       if (at_end_of_line (strbuf, j)) then
          done = .true.
       else if (.not. isspace (strbuf%chars(j))) then
          done = .true.
       else
          j = j + 1
       end if
    end do
  end function skip_whitespace

  function skip_non_whitespace (strbuf, i) result (j)
    class(strbuf_t), intent(in) :: strbuf
    integer(kind = nk), intent(in) :: i
    integer(kind = nk) :: j

    logical :: done

    j = i
    done = .false.
    do while (.not. done)
       if (at_end_of_line (strbuf, j)) then
          done = .true.
       else if (isspace (strbuf%chars(j))) then
          done = .true.
       else
          j = j + 1
       end if
    end do
  end function skip_non_whitespace

  function skip_whitespace_backwards (strbuf, i) result (j)
    class(strbuf_t), intent(in) :: strbuf
    integer(kind = nk), intent(in) :: i
    integer(kind = nk) :: j

    logical :: done

    j = i
    done = .false.
    do while (.not. done)
       if (j == -1) then
          done = .true.
       else if (.not. isspace (strbuf%chars(j))) then
          done = .true.
       else
          j = j - 1
       end if
    end do
  end function skip_whitespace_backwards

  function at_end_of_line (strbuf, i) result (bool)
    class(strbuf_t), intent(in) :: strbuf
    integer(kind = nk), intent(in) :: i
    logical :: bool

    bool = (strbuf%length() < i)
  end function at_end_of_line

end module string_buffers

module reading_one_line_from_a_stream
  use, intrinsic :: iso_fortran_env, only: input_unit
  use, intrinsic :: iso_fortran_env, only: error_unit
  use, non_intrinsic :: compiler_type_kinds, only: nk, ck, ick
  use, non_intrinsic :: string_buffers

  implicit none
  private

  ! get_line_from_stream: read an entire input line from a stream into
  ! a strbuf_t.
  public :: get_line_from_stream

  character(1, kind = ck), parameter :: linefeed_char = char (10, kind = ck)

  ! The following is correct for Unix and its relatives.
  character(1, kind = ck), parameter :: newline_char = linefeed_char

contains

  subroutine get_line_from_stream (unit_no, eof, no_newline, strbuf)
    integer, intent(in) :: unit_no
    logical, intent(out) :: eof ! End of file?
    logical, intent(out) :: no_newline ! There is a line but it has no
                                       ! newline? (Thus eof also must
                                       ! be .true.)
    class(strbuf_t), intent(inout) :: strbuf

    character(1, kind = ck) :: ch

    strbuf = ''
    call get_ch (unit_no, eof, ch)
    do while (.not. eof .and. ch /= newline_char)
       call strbuf%append (ch)
       call get_ch (unit_no, eof, ch)
    end do
    no_newline = eof .and. (strbuf%length() /= 0)
  end subroutine get_line_from_stream

  subroutine get_ch (unit_no, eof, ch)
    !
    ! Read a single code point from the stream.
    !
    ! Currently this procedure simply inputs ‘ASCII’ bytes rather than
    ! Unicode code points.
    !
    integer, intent(in) :: unit_no
    logical, intent(out) :: eof
    character(1, kind = ck), intent(out) :: ch

    integer :: stat
    character(1) :: c = '*'

    eof = .false.

    if (unit_no == input_unit) then
       call get_input_unit_char (c, stat)
    else
       read (unit = unit_no, iostat = stat) c
    end if

    if (stat < 0) then
       ch = ck_'*'
       eof = .true.
    else if (0 < stat) then
       write (error_unit, '("Input error with status code ", I0)') stat
       stop 1
    else
       ch = char (ichar (c, kind = ick), kind = ck)
    end if
  end subroutine get_ch

!!!
!!! If you tell gfortran you want -std=f2008 or -std=f2018, you likely
!!! will need to add also -fall-intrinsics or -U__GFORTRAN__
!!!
!!! The first way, you get the FGETC intrinsic. The latter way, you
!!! get the C interface code that uses getchar(3).
!!!
#ifdef __GFORTRAN__

  subroutine get_input_unit_char (c, stat)
    !
    ! The following works if you are using gfortran.
    !
    ! (FGETC is considered a feature for backwards compatibility with
    ! g77. However, I know of no way to reconfigure input_unit as a
    ! Fortran 2003 stream, for use with ordinary ‘read’.)
    !
    character, intent(inout) :: c
    integer, intent(out) :: stat

    call fgetc (input_unit, c, stat)
  end subroutine get_input_unit_char

#else

  subroutine get_input_unit_char (c, stat)
    !
    ! An alternative implementation of get_input_unit_char. This
    ! actually reads input from the C standard input, which might not
    ! be the same as input_unit.
    !
    use, intrinsic :: iso_c_binding, only: c_int
    character, intent(inout) :: c
    integer, intent(out) :: stat

    interface
       !
       ! Use getchar(3) to read characters from standard input. This
       ! assumes there is actually such a function available, and that
       ! getchar(3) does not exist solely as a macro. (One could write
       ! one’s own getchar() if necessary, of course.)
       !
       function getchar () result (c) bind (c, name = 'getchar')
         use, intrinsic :: iso_c_binding, only: c_int
         integer(kind = c_int) :: c
       end function getchar
    end interface

    integer(kind = c_int) :: i_char

    i_char = getchar ()
    !
    ! The C standard requires that EOF have a negative value. If the
    ! value returned by getchar(3) is not EOF, then it will be
    ! representable as an unsigned char. Therefore, to check for end
    ! of file, one need only test whether i_char is negative.
    !
    if (i_char < 0) then
       stat = -1
    else
       stat = 0
       c = char (i_char)
    end if
  end subroutine get_input_unit_char

#endif

end module reading_one_line_from_a_stream

module vm_reader
  use, intrinsic :: iso_fortran_env, only: error_unit
  use, non_intrinsic :: compiler_type_kinds
  use, non_intrinsic :: helpers
  use, non_intrinsic :: string_buffers
  use, non_intrinsic :: reading_one_line_from_a_stream

  implicit none
  private

  public :: vm_code_t
  public :: vm_t
  public :: read_vm

  !
  ! Arbitrarily chosen opcodes.
  !
  ! I think there should be a no-operation ‘nop’ opcode, to reserve
  ! space for later hand-patching. :)
  !
  integer, parameter, public :: opcode_nop = 0
  integer, parameter, public :: opcode_halt = 1
  integer, parameter, public :: opcode_add = 2
  integer, parameter, public :: opcode_sub = 3
  integer, parameter, public :: opcode_mul = 4
  integer, parameter, public :: opcode_div = 5
  integer, parameter, public :: opcode_mod = 6
  integer, parameter, public :: opcode_lt = 7
  integer, parameter, public :: opcode_gt = 8
  integer, parameter, public :: opcode_le = 9
  integer, parameter, public :: opcode_ge = 10
  integer, parameter, public :: opcode_eq = 11
  integer, parameter, public :: opcode_ne = 12
  integer, parameter, public :: opcode_and = 13
  integer, parameter, public :: opcode_or = 14
  integer, parameter, public :: opcode_neg = 15
  integer, parameter, public :: opcode_not = 16
  integer, parameter, public :: opcode_prtc = 17
  integer, parameter, public :: opcode_prti = 18
  integer, parameter, public :: opcode_prts = 19
  integer, parameter, public :: opcode_fetch = 20
  integer, parameter, public :: opcode_store = 21
  integer, parameter, public :: opcode_push = 22
  integer, parameter, public :: opcode_jmp = 23
  integer, parameter, public :: opcode_jz = 24

  character(8, kind = ck), parameter, public :: opcode_names(0:24) = &
       & (/ "nop     ",   &
       &    "halt    ",   &
       &    "add     ",   &
       &    "sub     ",   &
       &    "mul     ",   &
       &    "div     ",   &
       &    "mod     ",   &
       &    "lt      ",   &
       &    "gt      ",   &
       &    "le      ",   &
       &    "ge      ",   &
       &    "eq      ",   &
       &    "ne      ",   &
       &    "and     ",   &
       &    "or      ",   &
       &    "neg     ",   &
       &    "not     ",   &
       &    "prtc    ",   &
       &    "prti    ",   &
       &    "prts    ",   &
       &    "fetch   ",   &
       &    "store   ",   &
       &    "push    ",   &
       &    "jmp     ",   &
       &    "jz      " /)

  type :: vm_code_t
     integer(kind = rik), private :: len = 0_rik
     character(1), allocatable :: bytes(:)
   contains
     procedure, pass, private :: ensure_storage => vm_code_t_ensure_storage
     procedure, pass :: length => vm_code_t_length
  end type vm_code_t

  type :: vm_t
     integer(kind = rik), allocatable :: string_boundaries(:)
     character(:, kind = ck), allocatable :: strings
     character(1), allocatable :: data(:)
     character(1), allocatable :: stack(:)
     type(vm_code_t) :: code
     integer(kind = rik) :: sp = 0_rik
     integer(kind = rik) :: pc = 0_rik
  end type vm_t

contains

  subroutine vm_code_t_ensure_storage (code, length_needed)
    class(vm_code_t), intent(inout) :: code
    integer(kind = nk), intent(in) :: length_needed

    integer(kind = nk) :: len_needed
    integer(kind = nk) :: new_size
    type(vm_code_t) :: new_code

    len_needed = max (length_needed, 1_nk)

    if (.not. allocated (code%bytes)) then
       ! Initialize a new code%bytes array.
       new_size = new_storage_size (len_needed)
       allocate (code%bytes(0:(new_size - 1)))
    else if (ubound (code%bytes, 1) < len_needed - 1) then
       ! Allocate a new code%bytes array, larger than the current one,
       ! but containing the same bytes.
       new_size = new_storage_size (len_needed)
       allocate (new_code%bytes(0:(new_size - 1)))
       new_code%bytes(0:(code%len - 1)) = code%bytes(0:(code%len - 1))
       call move_alloc (new_code%bytes, code%bytes)
    end if
  end subroutine vm_code_t_ensure_storage

  elemental function vm_code_t_length (code) result (len)
    class(vm_code_t), intent(in) :: code
    integer(kind = rik) :: len

    len = code%len
  end function vm_code_t_length

  subroutine read_vm (inp, strbuf, vm)
    integer, intent(in) :: inp
    type(strbuf_t), intent(inout) :: strbuf
    type(vm_t), intent(out) :: vm

    integer(kind = rik) :: data_size
    integer(kind = rik) :: number_of_strings

    ! Read the header.
    call read_datasize_and_number_of_strings (inp, strbuf, data_size, number_of_strings)

    ! Allocate storage for data_size 32-bit numbers. Initialize them
    ! to zero, for no better reason than that C initializes global
    ! variables to zero.
    allocate (vm%data(0_rik:(4_rik * (data_size - 1))), source = achar (0))

    ! Allocate storage for indices/bounds of the strings to be loaded
    ! into the string storage space.
    allocate (vm%string_boundaries(0_rik:number_of_strings))

    ! Fill the strings storage and the string boundaries array.
    call read_strings (inp, strbuf, number_of_strings, vm)

    ! Read the program instructions.
    call read_code (inp, strbuf, vm)

    ! Allocate a stack. Let us say that the stack size must be a
    ! multiple of 4, and is fixed at 65536 = 4**8 bytes. Pushing a
    ! 32-bit integer increases the stack pointer by 4, popping
    ! decreases it by 4.
    allocate (vm%stack(0_rik:(4_rik ** 8)))
  end subroutine read_vm

  subroutine read_datasize_and_number_of_strings (inp, strbuf, data_size, number_of_strings)
    integer, intent(in) :: inp
    type(strbuf_t), intent(inout) :: strbuf
    integer(kind = rik), intent(out) :: data_size
    integer(kind = rik), intent(out) :: number_of_strings

    logical :: eof
    logical :: no_newline
    integer(kind = nk) :: i, j
    character(:, kind = ck), allocatable :: data_size_str
    character(:, kind = ck), allocatable :: number_of_strings_str
    integer :: stat

    call get_line_from_stream (inp, eof, no_newline, strbuf)
    if (eof) call bad_vm_assembly

    i = skip_whitespace (strbuf, 1_nk)
    i = skip_datasize_keyword (strbuf, i)
    i = skip_whitespace (strbuf, i)
    i = skip_specific_character (strbuf, i, ck_':')
    i = skip_whitespace (strbuf, i)
    j = skip_non_whitespace (strbuf, i)
    if (j == i) call bad_vm_assembly
    allocate (data_size_str, source = strbuf%to_unicode (i, j - 1))

    i = skip_whitespace(strbuf, j)
    i = skip_strings_keyword (strbuf, i)
    i = skip_whitespace (strbuf, i)
    i = skip_specific_character (strbuf, i, ck_':')
    i = skip_whitespace (strbuf, i)
    j = skip_non_whitespace (strbuf, i)
    if (j == i) call bad_vm_assembly
    allocate (number_of_strings_str, source = strbuf%to_unicode (i, j - 1))

    read (data_size_str, *, iostat = stat) data_size
    if (stat /= 0) call bad_vm_assembly
    read (number_of_strings_str, *, iostat = stat) number_of_strings
    if (stat /= 0) call bad_vm_assembly
  end subroutine read_datasize_and_number_of_strings

  subroutine read_strings (inp, strbuf, number_of_strings, vm)
    integer, intent(in) :: inp
    type(strbuf_t), intent(inout) :: strbuf
    integer(kind = rik), intent(in