User:Zarutian/Smallcaps

From Erights

Latest revision as of 12:14, 9 September 2010

Notes of what I have gleaned from the incomplete spefication and implementation of Smallcaps.

Smallcaps is a bytecode based format of executable code.

The opcodes are:

 Hex Memnomic         Effects/Description
 
 0x01 OP_DUP           [x],[] => OP_DUP => [x, x],[]
 0x02 OP_POP           [x],[] => OP_POP => [],[]
 0x03 OP_SWAP          [x, y],[] => OP_SWAP => [y, x],[]
 0x04 OP_ROT           [x, y, z],[] => OP_ROT => [y, z, x],[]
 0x05 OP_RETURN        [x],[] => OP_RETURN
 
 0x08 OP_JUMP          [],[] => OP_JUMP(label) => [],[]
                       label := ?Waterken Wholenum?
 0x09 OP_BRANCH        [ootEjector, flag],[] => OP_BRANCH => [],[]
 0x0A OP_CALL_ONLY     [recip, args...],[] => OP_CALL_ONLY(verb, arity) => [],[]
 0x0B OP_CALL          [recip, args...],[] => OP_CALL(verb, arity) => [result],[]
                       verb := ?Waterken UTF8 string?
                       arity := ?Waterken Wholenum?
 
 0x0E OP_EJECTOR_ONLY  [],[] => OP_EJECTOR_ONLY(label) => [ejector],[handler] 
                       when ejector is invoked: [...],[...] => ejector(_) => [],[]
 0x0F OP_EJECTOR       [],[] => OP_EJECTOR(label) => [ejector],[handler]
                       when ejector is invoked: [...],[...] => ejector(result) => [result],[]
 0x10 OP_TRY           [],[] => OP_TRY(label) => [],[handler]
                       when handler is invoked: [...],[...] => handler(arg) => [arg],[]
                       when handler is dropped: [...],[...] => handler.drop(_) => [...],[...]
 0x11 OP_UNWIND        [],[] => OP_UNWIND(label) => [],[handler]
                       when handler is invoked: [...],[...] => handler(arg) => [rethrower(arg)],[]
                       when handler is dropped: [...],[...] => handler.drop(pc) => [returner(pc)],[]
 0x12 OP_END_HANDLER   [],[handler] => OP_END_HANDLER => [],[]
 
 0x16 OP_WHOLE_NUM     [],[] => OP_WHOLE_NUM(wholeNum) => [wholeNum],[]
 0x17 OP_NEG_INT       [],[] => OP_NEG_INT(wholeNum) => [-wholeNum],[]
 0x18 OP_FLOAT64       [],[] => OP_FLOAT64(float64) => [float64],[]
 0x19 OP_CHAR          [],[] => OP_CHAR(chr) => [chr],[]
 0x1A OP_STRING        [],[] => OP_STRING(str) => [str],[]
 0x1B OP_TWINE         [],[] => OP_TWINE(twine) => [twine],[]
 0x1C OP_TRUE          [],[] => OP_TRUE => [true],[]
 0x1D OP_FALSE         [],[] => OP_FALSE => [false],[]
 0x1E OP_NULL          [],[] => OP_NULL => [null],[]
 0x1F OP_SCOPE         [],[] => OP_SCOPE => [scope],[]
 0x20 OP_OBJECT        [ivars..., auditors...],[] => OP_OBJECT(wholeNum numOfFields, wholeNum scriptIndex) => [object],[]
                       metanote: hypothesis: OP_OBJECT takes these immitiate args:
                                   wholenum of ivarAuditorpairs to pop of the stack
                                   some sort of reference to code that handles calls and sends to the object
 0x21 OP_LIST_PATT     [optEjector, specimen],[] => OP_LIST_PATT(n) => [optEjector, specimen[n-1]..., optEjector, specimen[0]]
                       metanote: is n a wholenum?
 0x22 OP_CDR_PATT
 
 0x28 OP_NOUN
   +addrMode
 0x30 OP_SLOT
   +addrMode
 0x38 OP_ASSIGN
   +addrMode
 0x40 OP_BIND
   +addrMode
 0x48 OP_BIND_SLOT
   +addrMode

dash pointed me to http://bazaar.launchpad.net/~washort/ecru/trunk/annotate/head:/doc/smallcaps.txt which is also a bit stale and to http://bazaar.launchpad.net/~washort/ecru/trunk/annotate/head:/esrc/com/twistedmatrix/ecru/compiler.emaker

http://bazaar.launchpad.net/~washort/ecru/trunk/annotate/head%3A/src/vm.c

See also http://epan.cvs.sourceforge.net/viewvc/epan/epan/lib/smallcaps/org/erights/e/elang/smallcaps/SmallcapsEncoder.java?view=markup

And http://epan.cvs.sourceforge.net/viewvc/epan/epan/lib/smallcaps/org/erights/e/elang/smallcaps/SmallcapsOps.java?view=markup

  # A Tcl procedure that decodes smallcaps bytecode stream
  proc decode_smallcaps {input} {
    set output [list]
    set op [read_byte]
    switch -exact -- $op {
      "\x01" { lappend output [list OP_DUP] }
      "\x02" { lappend output [list OP_POP] }
      "\x03" { lappend output [list OP_SWAP] }
      "\x04" { lappend output [list OP_ROT] }
      "\x05" { lappend output [list [list OP_UNDERDUP OP_RETURN]] }
      "\x06" { lappend output [list [list OP_RETURN {}]] }
      "\x07" { lappend output [list unused_bytecode 1] }
      "\x08" { lappend output [list OP_JUMP [read_wholeNum]] }
      "\x09" { lappend output [list OP_BRANCH] }
      "\x0a" { lappend output [list OP_CALL_ONLY [read_string] [read_wholeNum]] }
      "\x0b" { lappend output [list OP_CALL [read_string] [read_wholeNum]] }
      "\x0c" { lappend output [list OP_SEND_ONLY [read_string] [read_wholeNum]] }
      "\x0d" { lappend output [list OP_SEND [read_string] [read_wholeNum]] }
      "\x0e" { lappend output [list OP_EJECTOR_ONLY [read_wholeNum]] }
      "\x0f" { lappend output [list OP_EJECTOR [read_wholeNum]] }
      "\x10" { lappend output [list OP_TRY [read_wholeNum]] }
      "\x11" { lappend output [list OP_UNWIND [read_wholeNum]] }
      "\x12" { lappend output [list OP_ENDHANDLER] }
      "\x13" { lappend output [list unused_bytecode 2] }
      "\x14" { lappend output [list unused_bytecode 3] }
      "\x15" { lappend output [list unused_bytecode 4] }
      "\x16" { lappend output [list OP_WHOLENUM [read_wholeNum]] }
      "\x17" { lappend output [list OP_NEG_INT [read_wholeNum]] }
      "\x18" { lappend output [list OP_FLOAT64 [read_float64]] }
      "\x19" { lappend output [list OP_CHAR [read_char]] }
      "\x1a" { lappend output [list OP_STRING [read_string]] }
      "\x1b" { lappend output [list OP_TWINE ???unknown_arity???]] }
      "\x1c" { lappend output [list OP_TRUE]] }
      "\x1d" { lappend output [list OP_FALSE]] }
      "\x1e" { lappend output [list OP_NULL]] }
      "\x1f" { lappend output [list OP_SCOPE]] }
      "\x20" { lappend output [list OP_OBJECT [read_wholeNum] [read_wholeNum]] }
      "\x21" { lappend output [list OP_LIST_PATT [read_wholeNum]] }
      "\x22" { lappend output [list OP_LIST_PATT_ONLY [read_wholeNum]] }
      "\x23" { lappend output [list OP_CDR_PATT [read_wholeNum]] }
      "\x24" { lappend output [list OP_CDR_PATT_ONLY [read_wholeNum]] }
      "\x25" { lappend output [list unused_bytecode 5]] }
      "\x26" { lappend output [list unused_bytecode 6]] }
      "\x27" { lappend output [list unused_bytecide 7]] }
    }

    return $output
  }