mu/apps/mu.subx

# The Mu computer's level-2 language, also called Mu.
# http://akkartik.name/post/mu-2019-2
#
# To run:
#   $ ./ntranslate init.linux 0*.subx apps/mu.subx

# A sketch of planned data structures. Still highly speculative.
== data

# A program is currently a linked list of functions
Program:  # (address function)
  0/imm32

# A function consists of:
#   name: (address string)
#   inputs: (address var-type)  # tbd
#   outputs: (address var-type)  # tbd
#   body: (address block)
#   next: (address function)
Function-next:
  0x10/imm32
Function-size:
  0x14/imm32/20

# A block is a list of statements:
#     statements: (address statement)

# A statement can be either a regular statement consisting of:
#     name: (address string)
#     inputs: (address var)
#     outputs: (address var-r)
# or a variable declaration on the stack:
#     name: (address string)
#     type: (address type-sexpr)
# or a regular statement writing to a single new variable in a register:
#     name: (address string)
#     inputs: (address var)
#     output: var-r
# or a block of statements:
#     statements: (address statement)

# Kinds of local variable declarations:
#   var f : (array foo 10)
#   var f/ecx : int <- copy 0
# Variables live in either the stack or a register.
# Variables in the stack are auto-initialized.
#   (This is non-trivial for arrays, and arrays inside structs... We'll see.)
# Variables in register need a real instruction.

# var is a variable declaration. e.g. `foo: (array int 3)`
#   name: (address string)
#   type: (address type-sexpr)

# var-r is a variable declaration in a register. e.g. `foo/eax: (array int 3)`
#   name: (address string)
#   type: (address type-sexpr)
#   reg: int [0..7]

# type-sexpr is a tree of type identifiers. e.g. (array (address int) 3)
# either
#   id: type-identifier
# or
#   car: (address type-sexpr)
#   cdr: (address type-sexpr)

# Still todo:
#   global variables
#   heap allocations (planned name: 'handle')
#   user-defined types: 'type' for structs, 'choice' for unions
#   short-lived 'address' type for efficiently writing inside nested structs

== code

Entry:
    # . prologue
    89/<- %ebp 4/r32/esp
    (new-segment Heap-size Heap)
    # if (argv[1] == "test') run-tests()
    {
      # if (argc <= 1) break
      81 7/subop/compare *ebp 1/imm32
      7e/jump-if-lesser-or-equal break/disp8
      # if (argv[1] != "test") break
      (kernel-string-equal? *(ebp+8) "test")  # => eax
      3d/compare-eax-and 0/imm32
      74/jump-if-equal break/disp8
      #
      (run-tests)
      # syscall(exit, *Num-test-failures)
      8b/-> *Num-test-failures 3/r32/ebx
      eb/jump $mu-main:end/disp8
    }
    # otherwise convert Stdin
    (convert-mu Stdin Stdout)
    (flush Stdout)
    # syscall(exit, 0)
    bb/copy-to-ebx 0/imm32
$mu-main:end:
    b8/copy-to-eax 1/imm32/exit
    cd/syscall 0x80/imm8

convert-mu:  # in : (address buffered-file), out : (address buffered-file)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    #
    (parse-mu *(ebp+8))
    (check-mu-types)
    (emit-subx *(ebp+0xc))
$convert-mu:end:
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

test-convert-empty-input:
    # empty input => empty output
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # setup
    (clear-stream _test-input-stream)
    (clear-stream _test-input-buffered-file->buffer)
    (clear-stream _test-output-stream)
    (clear-stream _test-output-buffered-file->buffer)
    #
    (convert-mu _test-input-buffered-file _test-output-buffered-file)
    (flush _test-output-buffered-file)
    (check-stream-equal _test-output-stream "" "F - test-convert-empty-input")
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

test-convert-function-skeleton:
    # empty function decl => function prologue and epilogue
    #   fn foo {
    #   }
    # =>
    #   foo:
    #     # . prologue
    #     55/push-ebp
    #     89/<- %ebp 4/r32/esp
    #     # . epilogue
    #     89/<- %esp 5/r32/ebp
    #     5d/pop-to-ebp
    #     c3/return
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # setup
    (clear-stream _test-input-stream)
    (clear-stream _test-input-buffered-file->buffer)
    (clear-stream _test-output-stream)
    (clear-stream _test-output-buffered-file->buffer)
    #
    (write _test-input-stream "fn foo {\n")
    (write _test-input-stream "}\n")
    # convert
    (convert-mu _test-input-buffered-file _test-output-buffered-file)
    (flush _test-output-buffered-file)
#?     # dump _test-output-stream {{{
#?     (write 2 "^")
#?     (write-stream 2 _test-output-stream)
#?     (write 2 "$\n")
#?     (rewind-stream _test-output-stream)
#?     # }}}
    # check output
    (check-next-stream-line-equal _test-output-stream "foo:"                  "F - test-convert-function-skeleton/0")
    (check-next-stream-line-equal _test-output-stream "# . prologue"          "F - test-convert-function-skeleton/1")
    (check-next-stream-line-equal _test-output-stream "55/push-ebp"           "F - test-convert-function-skeleton/2")
    (check-next-stream-line-equal _test-output-stream "89/<- %ebp 4/r32/esp"  "F - test-convert-function-skeleton/3")
    (check-next-stream-line-equal _test-output-stream "# . epilogue"          "F - test-convert-function-skeleton/4")
    (check-next-stream-line-equal _test-output-stream "89/<- %esp 5/r32/ebp"  "F - test-convert-function-skeleton/5")
    (check-next-stream-line-equal _test-output-stream "5d/pop-to-ebp"         "F - test-convert-function-skeleton/6")
    (check-next-stream-line-equal _test-output-stream "c3/return"             "F - test-convert-function-skeleton/7")
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

test-convert-multiple-function-skeletons:
    # multiple functions correctly organized into a linked list
    #   fn foo {
    #   }
    #   fn bar {
    #   }
    # =>
    #   foo:
    #     # . prologue
    #     55/push-ebp
    #     89/<- %ebp 4/r32/esp
    #     # . epilogue
    #     89/<- %esp 5/r32/ebp
    #     5d/pop-to-ebp
    #     c3/return
    #   bar:
    #     # . prologue
    #     55/push-ebp
    #     89/<- %ebp 4/r32/esp
    #     # . epilogue
    #     89/<- %esp 5/r32/ebp
    #     5d/pop-to-ebp
    #     c3/return
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # setup
    (clear-stream _test-input-stream)
    (clear-stream _test-input-buffered-file->buffer)
    (clear-stream _test-output-stream)
    (clear-stream _test-output-buffered-file->buffer)
    #
    (write _test-input-stream "fn foo {\n")
    (write _test-input-stream "}\n")
    (write _test-input-stream "fn bar {\n")
    (write _test-input-stream "}\n")
    # convert
    (convert-mu _test-input-buffered-file _test-output-buffered-file)
    (flush _test-output-buffered-file)
#?     # dump _test-output-stream {{{
#?     (write 2 "^")
#?     (write-stream 2 _test-output-stream)
#?     (write 2 "$\n")
#?     (rewind-stream _test-output-stream)
#?     # }}}
    # check first function
    (check-next-stream-line-equal _test-output-stream "foo:"                  "F - test-convert-multiple-function-skeletons/0")
    (check-next-stream-line-equal _test-output-stream "# . prologue"          "F - test-convert-multiple-function-skeletons/1")
    (check-next-stream-line-equal _test-output-stream "55/push-ebp"           "F - test-convert-multiple-function-skeletons/2")
    (check-next-stream-line-equal _test-output-stream "89/<- %ebp 4/r32/esp"  "F - test-convert-multiple-function-skeletons/3")
    (check-next-stream-line-equal _test-output-stream "# . epilogue"          "F - test-convert-multiple-function-skeletons/4")
    (check-next-stream-line-equal _test-output-stream "89/<- %esp 5/r32/ebp"  "F - test-convert-multiple-function-skeletons/5")
    (check-next-stream-line-equal _test-output-stream "5d/pop-to-ebp"         "F - test-convert-multiple-function-skeletons/6")
    (check-next-stream-line-equal _test-output-stream "c3/return"             "F - test-convert-multiple-function-skeletons/7")
    # check second function
    (check-next-stream-line-equal _test-output-stream "bar:"                  "F - test-convert-multiple-function-skeletons/10")
    (check-next-stream-line-equal _test-output-stream "# . prologue"          "F - test-convert-multiple-function-skeletons/11")
    (check-next-stream-line-equal _test-output-stream "55/push-ebp"           "F - test-convert-multiple-function-skeletons/12")
    (check-next-stream-line-equal _test-output-stream "89/<- %ebp 4/r32/esp"  "F - test-convert-multiple-function-skeletons/13")
    (check-next-stream-line-equal _test-output-stream "# . epilogue"          "F - test-convert-multiple-function-skeletons/14")
    (check-next-stream-line-equal _test-output-stream "89/<- %esp 5/r32/ebp"  "F - test-convert-multiple-function-skeletons/15")
    (check-next-stream-line-equal _test-output-stream "5d/pop-to-ebp"         "F - test-convert-multiple-function-skeletons/16")
    (check-next-stream-line-equal _test-output-stream "c3/return"             "F - test-convert-multiple-function-skeletons/17")
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

parse-mu:  # in : (address buffered-file)
    # pseudocode
    #   var curr-function = Program
    #   var line : (stream byte 512)
    #   var word-slice : slice
    #   while true                                  # line loop
    #     clear-stream(line)
    #     read-line-buffered(in, line)
    #     if (line->write == 0) break               # end of file
    #     while true                                # word loop
    #       word-slice = next-word-or-string(line)
    #       if slice-empty?(word-slice)             # end of line
    #         break
    #       else if slice-starts-with?(word-slice, "#")  # comment
    #         break                                 # end of line
    #       else if slice-equal(word-slice, "fn")
    #         var new-function : (address function) = new function
    #         populate-mu-function(in, new-function)
    #         *curr-function = new-function
    #         curr-function = &new-function->next
    #       else
    #         abort()
    #
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
    50/push-eax
    51/push-ecx
    52/push-edx
    57/push-edi
    # var line/ecx : (stream byte 512)
    81 5/subop/subtract %esp 0x200/imm32
    68/push 0x200/imm32/length
    68/push 0/imm32/read
    68/push 0/imm32/write
    89/<- %ecx 4/r32/esp
    # var word-slice/edx : slice
    68/push 0/imm32/end
    68/push 0/imm32/start
    89/<- %edx 4/r32/esp
    # var curr-function/edi : (address function) = Program
    bf/copy-to-edi Program/imm32
    {
$parse-mu:line-loop:
      (clear-stream %ecx)
      (read-line-buffered *(ebp+8) %ecx)
      # if (line->write == 0) break
      81 7/subop/compare *ecx 0/imm32
      0f 84/jump-if-equal break/disp32
#?       # dump line {{{
#?       (write 2 "line: ^")
#?       (write-stream 2 %ecx)
#?       (write 2 "$\n")
#?       (rewind-stream %ecx)
#?       # }}}
      { # word loop
$parse-mu:word-loop:
        (next-word-or-string %ecx %edx)
        # if slice-empty?(word-slice) break
        (slice-empty? %edx)
        3d/compare-eax-and 0/imm32
        0f 85/jump-if-not-equal break/disp32
        # if (*word-slice->start == "#") break
        # . eax = *word-slice->start
        8b/-> *edx 0/r32/eax
        8a/copy-byte *eax 0/r32/AL
        81 4/subop/and %eax 0xff/imm32
        # . if (eax == '#') break
        3d/compare-eax-and 0x23/imm32/hash
        0f 84/jump-if-equal break/disp32
        # if (slice-equal?(word-slice, "fn")) parse a function
        {
          (slice-equal? %edx "fn")
          3d/compare-eax-and 0/imm32
          0f 84/jump-if-equal break/disp32
          # var new-function/eax : (address function) = populate-mu-function()
          (allocate Heap *Function-size)  # => eax
          (populate-mu-function-header %ecx %eax)
          (populate-mu-function-body *(ebp+8) %eax)
          # *curr-function = new-function
          89/<- *edi 0/r32/eax
          # curr-function = &new-function->next
          8d/address-> *(eax+0x10) 7/r32/edi
          e9/jump $parse-mu:word-loop/disp32
        }
        # otherwise abort
        e9/jump $parse-mu:abort/disp32
      } # end word loop
      e9/jump loop/disp32
    } # end line loop
$parse-mu:end:
    # . reclaim locals
    81 0/subop/add %esp 0x214/imm32
    # . restore registers
    5f/pop-to-edi
    5a/pop-to-edx
    59/pop-to-ecx
    58/pop-to-eax
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

$parse-mu:abort:
    # error("unexpected top-level command: " word-slice "\n")
    (write-buffered Stderr "unexpected top-level command: ")
    (write-buffered Stderr %edx)
    (write-buffered Stderr "\n")
    (flush Stderr)
    # . syscall(exit, 1)
    bb/copy-to-ebx  1/imm32
    b8/copy-to-eax  1/imm32/exit
    cd/syscall  0x80/imm8
    # never gets here

# errors considered:
#   fn foo { {
#   fn foo { }
#   fn foo { } {
#   fn foo  # no block
populate-mu-function-header:  # first-line : (address stream byte), out : (address function)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
    50/push-eax
    51/push-ecx
    57/push-edi
    # edi = out
    8b/-> *(ebp+0xc) 7/r32/edi
    # var word-slice/ecx : slice
    68/push 0/imm32/end
    68/push 0/imm32/start
    89/<- %ecx 4/r32/esp
    # save function name
    (next-word *(ebp+8) %ecx)
    (slice-to-string Heap %ecx)  # => eax
    89/<- *edi 0/r32/eax
    # assert that next token is '{'
    (next-word *(ebp+8) %ecx)
    (slice-equal? %ecx "{")
    3d/compare-eax-and 0/imm32
    74/jump-if-equal $populate-mu-function-header:abort/disp8
    # assert that there's no further token
    {
      # word-slice = next-word(line)
      (next-word *(ebp+8) %ecx)
      # if (word-slice == '') break
      (slice-empty? %ecx)
      3d/compare-eax-and 0/imm32
      75/jump-if-not-equal break/disp8
      # if (slice-starts-with?(word-slice, "#")) break
      # . eax = *word-slice->start
      8b/-> *edx 0/r32/eax
      8a/copy-byte *eax 0/r32/AL
      81 4/subop/and %eax 0xff/imm32
      # . if (eax == '#') break
      3d/compare-eax-and 0x23/imm32/hash
      74/jump-if-equal break/disp8
      # otherwise abort
      eb/jump $populate-mu-function-header:abort/disp8
    }
$populate-mu-function-header:end:
    # . reclaim locals
    81 0/subop/add %esp 8/imm32
    # . restore registers
    5f/pop-to-edi
    59/pop-to-ecx
    58/pop-to-eax
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

$populate-mu-function-header:abort:
    # error("function header not in form 'fn <name> {'")
    (write-buffered Stderr "function header not in form 'fn <name> {' -- '")
    (rewind-stream *(ebp+8))
    (write-stream 2 *(ebp+8))
    (write-buffered Stderr "'\n")
    (flush Stderr)
    # . syscall(exit, 1)
    bb/copy-to-ebx  1/imm32
    b8/copy-to-eax  1/imm32/exit
    cd/syscall  0x80/imm8
    # never gets here

# errors considered:
#   { abc
populate-mu-function-body:  # in : (address buffered-file), out : (address function)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
    50/push-eax
    51/push-ecx
    52/push-edx
    53/push-ebx
    # var line/ecx : (stream byte 512)
    81 5/subop/subtract %esp 0x200/imm32
    68/push 0x200/imm32/length
    68/push 0/imm32/read
    68/push 0/imm32/write
    89/<- %ecx 4/r32/esp
    # var word-slice/edx : slice
    68/push 0/imm32/end
    68/push 0/imm32/start
    89/<- %edx 4/r32/esp
    # var open-curly-count/ebx : int = 1
    bb/copy-to-ebx 1/imm32
    { # line loop
$populate-mu-function-body:line-loop:
      # if (open-curly-count == 0) break
      81 7/subop/compare %ebx 0/imm32
      0f 84/jump-if-equal break/disp32
      # line = read-line-buffered(in)
      (clear-stream %ecx)
      (read-line-buffered *(ebp+8) %ecx)
      # if (line->write == 0) break
      81 7/subop/compare *ecx 0/imm32
      0f 84/jump-if-equal break/disp32
      # word-slice = next-word(line)
      (next-word %ecx %edx)
      # if slice-empty?(word-slice) continue
      (slice-empty? %ecx)
      3d/compare-eax-and 0/imm32
      75/jump-if-not-equal loop/disp8
      # if (slice-starts-with?(word-slice, '#') continue
      # . eax = *word-slice->start
      8b/-> *edx 0/r32/eax
      8a/copy-byte *eax 0/r32/AL
      81 4/subop/and %eax 0xff/imm32
      # . if (eax == '#') continue
      3d/compare-eax-and 0x23/imm32/hash
      74/jump-if-equal loop/disp8
      {
        # if slice-equal?(word-slice, "{") ++open-curly-count
        {
          (slice-equal? %ecx "{")
          3d/compare-eax-and 0/imm32
          74/jump-if-equal break/disp8
          43/increment-ebx
          eb/jump $curly-found:end/disp8
        }
        # else if slice-equal?(word-slice, "}") --open-curly-count
        {
          (slice-equal? %ecx "}")
          3d/compare-eax-and 0/imm32
          74/jump-if-equal break/disp8
          4b/decrement-ebx
          eb/jump $curly-found:end/disp8
        }
        # else break
        eb/jump $populate-mu-function-body:end/disp8
      }
      # - check for invalid tokens after curly
$curly-found:end:
      # second-word-slice = next-word(line)
      (next-word %ecx %edx)
      # if slice-empty?(second-word-slice) continue
      (slice-empty? %ecx)
      3d/compare-eax-and 0/imm32
      0f 85/jump-if-not-equal loop/disp32
      # if (slice-starts-with?(second-word-slice, '#') continue
      # . eax = *second-word-slice->start
      8b/-> *edx 0/r32/eax
      8a/copy-byte *eax 0/r32/AL
      81 4/subop/and %eax 0xff/imm32
      # . if (eax == '#') continue
      3d/compare-eax-and 0x23/imm32/hash
      0f 84/jump-if-equal loop/disp32
      # abort
      eb/jump $populate-mu-function-body:abort/disp8
    } # end line loop
$populate-mu-function-body:end:
    # . reclaim locals
    81 0/subop/add %esp 0x214/imm32
    # . restore registers
    5b/pop-to-ebx
    5a/pop-to-edx
    59/pop-to-ecx
    58/pop-to-eax
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

$populate-mu-function-body:abort:
    # error("'{' or '}' should be on its own line, but got '")
    (write-buffered Stderr "'{' or '}' should be on its own line, but got '")
    (rewind-stream %ecx)
    (write-stream 2 %ecx)
    (write-buffered Stderr "'\n")
    (flush Stderr)
    # . syscall(exit, 1)
    bb/copy-to-ebx  1/imm32
    b8/copy-to-eax  1/imm32/exit
    cd/syscall  0x80/imm8
    # never gets here

check-mu-types:
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    #
$check-types:end:
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

emit-subx:  # out : (address buffered-file)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
    50/push-eax
    51/push-ecx
    57/push-edi
    # edi = out
    8b/-> *(ebp+8) 7/r32/edi
    # var curr/ecx : (address function) = Program
    8b/-> *Program 1/r32/ecx
    {
      # if (curr == NULL) break
      81 7/subop/compare %ecx 0/imm32
      0f 84/jump-if-equal break/disp32
      (write-buffered %edi *ecx)
      (write-buffered %edi ":\n")
      (emit-subx-prologue %edi)
      (emit-subx-epilogue %edi)
      # curr = curr->next
      8b/-> *(ecx+0x10) 1/r32/ecx
      e9/jump loop/disp32
    }
$emit-subx:end:
    # . restore registers
    5f/pop-to-edi
    59/pop-to-ecx
    58/pop-to-eax
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

emit-subx-prologue:  # out : (address buffered-file)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    #
    (write-buffered *(ebp+8) "# . prologue\n")
    (write-buffered *(ebp+8) "55/push-ebp\n")
    (write-buffered *(ebp+8) "89/<- %ebp 4/r32/esp\n")
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

emit-subx-epilogue:  # out : (address buffered-file)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    #
    (write-buffered *(ebp+8) "# . epilogue\n")
    (write-buffered *(ebp+8) "89/<- %esp 5/r32/ebp\n")
    (write-buffered *(ebp+8) "5d/pop-to-ebp\n")
    (write-buffered *(ebp+8) "c3/return\n")
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return