Website | Blog | Source

Ecstasy is a new, general-purpose, programming language, designed for modern cloud architectures, and explicitly for the serverless cloud. Actually, to be completely honest, it's the most amazing programming language ever. No, really, it's that awesome.

The Ecstasy project includes a development kit (XDK) that is produced out of this repository, a programming language specification, a core set of runtime modules (libraries), a portable, type-safe, and verifiable Intermediate Representation (IR), a proof-of-concept runtime (with an adaptive LLVM-based optimizing compiler in development), and a tool-chain with both Java and Ecstasy implementations being actively developed.

The Ecstasy language supports first class modules, including versioning and conditionality; first class functions, including currying and partial application; type-safe object orientation, including support for auto-narrowing types, type-safe covariance, mixins, and duck-typed interfaces; complete type inference; first class immutable types; first class asynchronous services, including both automatic async/await-style and promises-based (@Future) programming models; and first class software containers, including resource injection and transitively-closed, immutable type systems. And much, much more.

Articles

InfoQ
TheServerSide
yC

Examples

HelloWorld

module HelloWorld
    {
    void run()
        {
        @Inject Console console;
        console.println("Hello World!");
        }
    }

FizzBuzz

module FizzBuzz
    {
    void run()
        {
        @Inject Console console;
        for (Int x : 1..100)
            {
            console.println(switch (x % 3, x % 5)
                {
                case (0, 0): "FizzBuzz";
                case (0, _): "Fizz";
                case (_, 0): "Buzz";
                case (_, _): x.toString();
                });
            }
        }
    }

Example1.x

module Example1
    {
    @Inject Console console;

    void run()
        {
        console.println("Hello world!");

        @Inject ecstasy.Timer timer;
        timer.reset();

        @Inject FileStore storage;
        console.println($"root={storage.root} capacity={storage.capacity}");

        @Inject Clock clock;
        console.println(clock.now);

        console.println($"Elapsed {timer.elapsed.milliseconds} ms");
        }
    }

Example3.x

module Example3
    {
    const Person(String name, Int age);

    interface AgeAware
        {
        Int age;
        }

    mixin AverageAgeCalculator
            into Array<AgeAware>
        {
        Int average()
            {
            Int sum = 0;
            for (AgeAware el : this)
                {
                sum += el.age;
                }
            return size == 0 ? 0 : sum/size;
            }
        }

    interface Extract<Element>
        {
        List insertAll(Int index, Iterable<Element> values);
        }

    void run()
        {
        @Inject Console console;

        // for exactly the same reason (Iterator is a producer now) the two below don't compile
        // Extract<AgeAware> array = new Array<Person>();
        // val calc = new @AverageAgeCalculator Array<Person>();

        val calc = new @AverageAgeCalculator Array<AgeAware>();

        calc += new Person("Bob", 31);
        calc += new Person("Sue", 25);

        console.println(calc.average());
        }
    }

XVM

XTC Virtual Machine?

Opcodes

// fundamentals
+NOP                                                ; traditional no-op (i.e. it does nothing)
+LINE_1                                             ; advance source code line counter by 1 line
+LINE_2                                             ; advance source code line counter by 2 line
+LINE_3                                             ; advance source code line counter by 3 line
+LINE_N     #lines                                  ; advance source code line counter by n lines
+ENTER                                              ; variable scope begin
+EXIT                                               ; variable scope end
+GUARD      #handlers:(TYPE, STRING, addr)          ; try { (+ catch type, var name, handler address) (implicit ENTER)
+GUARD_E    addr                                    ; } ... // ("E"=) end guarded block with a jump (implicit EXIT)
+CATCH                                              ; begin an exception handler (implicit ENTER and VAR_IN for exception)
+CATCH_E    addr                                    ; ("E"=) end an exception handler with a jump (implicit EXIT)
+GUARDALL   addr                                    ; try { (+ "finally" address (implicit ENTER, also intercepts boundary-crossing-jumps and returns)
+FINALLY                                            ; begin a "finally" handler (implicit EXIT/ENTER and VAR_I of type "Exception?")
+FINALLY_E                                          ; ("E"=) end  a "finally" handler (implicit EXIT)
+THROW      rvalue                                  ; raise exception

// function calls
// rvalue-fn, rvalue-param and lvalue-return support "local property" mode
// return values can specify the "next register" to allocate
// - first digit is # params: 0, 1, n, or "t"=tuple
// - second digit is # returns: 0, 1, n, or "t"=tuple
+CALL_00    rvalue-fn
+CALL_01    rvalue-fn, lvalue-return
+CALL_0N    rvalue-fn, #:(lvalue)
+CALL_0T    rvalue-fn, lvalue-treturn
+CALL_10    rvalue-fn, rvalue-param
+CALL_11    rvalue-fn, rvalue-param, lvalue-return
+CALL_1N    rvalue-fn, rvalue-param, #:(lvalue)
+CALL_1T    rvalue-fn, rvalue-param, lvalue-treturn
+CALL_N0    rvalue-fn, #:(rvalue)
+CALL_N1    rvalue-fn, #:(rvalue), lvalue-return
+CALL_NN    rvalue-fn, #:(rvalue), #:(lvalue)
+CALL_NT    rvalue-fn, #:(rvalue), lvalue-treturn
+CALL_T0    rvalue-fn, rvalue-tparams
+CALL_T1    rvalue-fn, rvalue-tparams, lvalue-return
+CALL_TN    rvalue-fn, rvalue-tparams, #:(lvalue)
+CALL_TT    rvalue-fn, rvalue-tparams, lvalue-treturn

// method calls ("NVOK"=invoke)
// REVIEW rvalue-target should not support lprops
// lvalue-param and lvalue-return support "local property" mode
// return values can specify the "next register" to allocate
// - first digit is # params: 0, 1, n, or tuple
// - second digit is # returns: 0, 1, n, or tuple
+NVOK_00    rvalue-target, METHOD
+NVOK_01    rvalue-target, METHOD, lvalue-return
+NVOK_0N    rvalue-target, METHOD, #:(lvalue)
+NVOK_0T    rvalue-target, METHOD, lvalue-treturn
+NVOK_10    rvalue-target, METHOD, rvalue-param
+NVOK_11    rvalue-target, METHOD, rvalue-param, lvalue-return
+NVOK_1N    rvalue-target, METHOD, rvalue-param, #returns:(lvalue)
+NVOK_1T    rvalue-target, METHOD, rvalue-param, lvalue-treturn
+NVOK_N0    rvalue-target, METHOD, #:(rvalue)
+NVOK_N1    rvalue-target, METHOD, #:(rvalue), lvalue-return
+NVOK_NN    rvalue-target, METHOD, #:(rvalue), #:(lvalue)
+NVOK_NT    rvalue-target, METHOD, #:(rvalue), lvalue-treturn
+NVOK_T0    rvalue-target, METHOD, rvalue-tparams
+NVOK_T1    rvalue-target, METHOD, rvalue-tparams, lvalue-return
+NVOK_TN    rvalue-target, METHOD, rvalue-tparams, #:(lvalue)
+NVOK_TT    rvalue-target, METHOD, rvalue-tparams, lvalue-treturn

// method and function binding
// - does NOT support "local property" mode nor "next register"  REVIEW why not? agreed that rvalue-target and rvalue-param must not support lprops
+BIND_M      rvalue-target, METHOD, lvalue-fn-result
+BIND_F      rvalue-fn, #params:(param-index, rvalue-param), lvalue-fn-result

// construct calls
 SYN_INIT                                                           ; default initializer
+CONSTR_0   CONSTRUCT                                               ; "super" construct call
+CONSTR_1   CONSTRUCT, rvalue-param
+CONSTR_N   CONSTRUCT, #params:(rvalue)
 CONSTR_T   CONSTRUCT, rvalue-tparams

// construction
// ("G"=) generic types; ("C"=) child classes;
// - rvalue parameters can be registers, constant values, or local-property indicators (PROPERTY)
// - lvalue will hold the newly constructed object
// - the lvalue can be a register or a local-property indicator (PROPERTY)
// - a new register of the constructed type is allocated if the lvalue is the next unallocated register
+NEW_0      CONSTRUCT, lvalue                                       ; simple "new"
+NEW_1      CONSTRUCT, rvalue-param, lvalue
+NEW_N      CONSTRUCT, #:(rvalue), lvalue
+NEW_T      CONSTRUCT, rvalue-tparams, lvalue
+NEWG_0     CONSTRUCT, TYPE, lvalue                                 ; generic-type "new"
+NEWG_1     CONSTRUCT, TYPE, rvalue-param, lvalue
+NEWG_N     CONSTRUCT, TYPE, #:(rvalue), lvalue
+NEWG_T     CONSTRUCT, TYPE, rvalue-tparams, lvalue
+NEWC_0     CONSTRUCT, rvalue-parent, lvalue                        ; virtual "new" for child classes
+NEWC_1     CONSTRUCT, rvalue-parent, rvalue-param, lvalue
+NEWC_N     CONSTRUCT, rvalue-parent, #:(rvalue), lvalue
 NEWC_T     CONSTRUCT, rvalue-parent, rvalue-tparams, lvalue
+NEWCG_0    CONSTRUCT, rvalue-parent, TYPE, lvalue                  ; virtual generic-type "new" for child classes
+NEWCG_1    CONSTRUCT, rvalue-parent, TYPE, rvalue-param, lvalue
+NEWCG_N    CONSTRUCT, rvalue-parent, TYPE, #:(rvalue), lvalue
 NEWCG_T    CONSTRUCT, rvalue-parent, TYPE, rvalue-tparams, lvalue
+NEWV_0     CONSTRUCT, rvalue-type, lvalue                          ; virtual "new"
+NEWV_1     CONSTRUCT, rvalue-type, rvalue-param, lvalue            ;  - CONSTRUCT may be on an interface (implies compile-time type)
+NEWV_N     CONSTRUCT, rvalue-type, #:(rvalue), lvalue              ;  - rvalue-type is the virtual type
 NEWV_T     CONSTRUCT, rvalue-type, rvalue-tparams, lvalue

// returns
// - rvalues support a local-property mode
+RETURN_0                                           ; return (no return value)
+RETURN_1   rvalue                                  ; return (single return value)
+RETURN_N   #vals:(rvalue)                          ; return (multiple return value)
+RETURN_T   rvalue-tuple                            ; return (multiple return from a tuple)

// next register allocation (variables, including refs)
// - does NOT support "local property" mode
+VAR        TYPE                                    ; uninitialized anonymous variable
+VAR_I      TYPE, rvalue-src                        ; ("I"=) initialized anonymous variable
+VAR_N      TYPE, STRING                            ; uninitialized ("N"=) named variable
+VAR_IN     TYPE, STRING, rvalue-src                ; ("I"=) initialized ("N"=) named variable
+VAR_D      TYPE                                    ; anonymous ("D"=) dynamic reference
+VAR_DN     TYPE, STRING                            ; ("N"=) named ("D"=) dynamic reference
-VAR_C      rvalue-ref                              ; ("C"=) capture variable: rvalue must be of type Ref or Var; local property mode is supported
-VAR_CN     rvalue-ref, STRING                      ; ("N"=) named ("C"=) capture variable: rvalue must be of type Ref or Var; local property mode is supported
+VAR_S      TYPE, #values:(rvalue)                  ; initialized anonymous ("S"=) Sequence variable
+VAR_SN     TYPE, STRING, #values:(rvalue)          ; initialized ("N"=) named ("S"=) Sequence variable
+VAR_T      TYPE. #values:(rvalue)                  ; initialized anonymous ("T"=) Tuple variable
+VAR_TN     TYPE, STRING, #values:(rvalue)          ; initialized ("N"=) named ("T"=) Tuple variable
+VAR_M      TYPE, #entries:(rvalue, rvalue)         ; initialized anonymous ("M"=) Map variable
+VAR_MN     TYPE, STRING, #entries:(rvalue, rvalue) ; initialized ("N"=) named ("M"=) Map variable

// "move" operations
// - rvalues do NOT support "local property" mode;
// - the lvalue can be a register or a local-property indicator (PROPERTY)
// - lvalue-dest can be a next unallocated register
// - for MOV the source could be unassigned
+MOV        rvalue-src, lvalue-dest                 ; copy source to destination ("protection" proxy may be generated)
+MOV_VAR    rvalue-src, lvalue-dest                 ; move ("R"=) Var-to-source to destination
+MOV_REF    rvalue-src, lvalue-dest                 ; move ("R"=) Ref-to-source to destination (read-only)
+MOV_THIS   #, lvalue-dest                          ; # (an inline unsigned byte) specifies the count of this-to-outer-this steps (0=this, 1=ImmediatelyOuter.this, etc.)
+MOV_THISA  #, lvalue-dest, A_*                     ; same as above with an additional access modifier (A_TARGET, A_PUBLIC, A_PROTECTED, A_PRIVATE)
+MOV_TYPE   rvalue-src, lvalue-dest                 ; place the type of the r-value into the l-value
+CAST       rvalue-src, lvalue-dest, TYPE           ; copy source to destination with cast

// Value comparisons (mostly Boolean checks)
// - rvalues support a local-property mode
// - lvalue return values can specify the "next register" to allocate
+CMP        rvalue1, rvalue2, lvalue-return         ; T <=> T -> Ordered
+IS_ZERO    rvalue-int, lvalue-return               ; T == 0 -> Boolean
+IS_NZERO   rvalue-int, lvalue-return               ; T != 0 -> Boolean
+IS_NULL    rvalue, lvalue-return                   ; T == null -> Boolean
+IS_NNULL   rvalue, lvalue-return                   ; T != null -> Boolean
+IS_EQ      rvalue, rvalue, lvalue-return           ; T == T -> Boolean
+IS_NEQ     rvalue, rvalue, lvalue-return           ; T != T -> Boolean
+IS_LT      rvalue, rvalue, lvalue-return           ; T < T -> Boolean
+IS_LTE     rvalue, rvalue, lvalue-return           ; T <= T -> Boolean
+IS_GT      rvalue, rvalue, lvalue-return           ; T > T -> Boolean
+IS_GTE     rvalue, rvalue, lvalue-return           ; T >= T -> Boolean
+IS_NOT     rvalue, lvalue-return                   ; !T -> Boolean
+IS_TYPE    rvalue, rvalue-type, lvalue-return      ; T instanceof Type -> Boolean
+IS_NTYPE   rvalue, rvalue-type, lvalue-return      ; !(T instanceof Type) -> Boolean
+IS_SVC     rvalue, lvalue-return                   ; (T is a service) -> Boolean
+IS_CONST   rvalue, lvalue-return                   ; (T is a const) -> Boolean
+IS_IMMUT   rvalue, lvalue-return                   ; (T is an immutable) -> Boolean

// jumps (all addresses are relative to the PC of the jump op)
// - rvalues support a local-property mode
+JMP        addr                                    ; unconditional relative jump
+JMP_TRUE   rvalue, addr                            ; jump if Boolean value is Boolean.True
+JMP_FALSE  rvalue, addr                            ; jump if Boolean value is Boolean.False
+JMP_ZERO   rvalue, addr                            ; jump if value is zero
+JMP_NZERO  rvalue, addr                            ; jump if value is NOT zero
+JMP_NULL   rvalue, addr                            ; jump if value is null
+JMP_NNULL  rvalue, addr                            ; jump if value is NOT null
+JMP_EQ     rvalue1, rvalue2, addr                  ; jump if value-1 is equal to value-2
+JMP_NEQ    rvalue1, rvalue2, addr                  ; jump if value-1 is NOT equal to value-2
+JMP_LT     rvalue1, rvalue2, addr                  ; jump if value-1 is less than value-2
+JMP_LTE    rvalue1, rvalue2, addr                  ; jump if value-1 is less than or equal to value-2
+JMP_GT     rvalue1, rvalue2, addr                  ; jump if value-1 is greater than value-2
+JMP_GTE    rvalue1, rvalue2, addr                  ; jump if value-1 is greater than or equal to value-2
+JMP_TYPE   rvalue, TYPE, addr                      ; jump if type of the value is “instanceof" specified type
+JMP_NTYPE  rvalue, TYPE, addr                      ; jump if type of the value is NOT “instanceof" specified type
+JMP_COND   COND, addr                              ; jump if the link-time condition evaluates to True
+JMP_NCOND  COND, addr                              ; jump if the link-time condition evaluates to False
+JMP_NFIRST addr                                    ; jump if this is NOT the first time the op has been executed
+JMP_NSAMPL rvalue, addr                            ; jump if this is NOT a selected sample based on the rvalue sample rate (a compile-time or run-time constant)
+JMP_INT    rvalue, offset, #:(addr), addr-default      ; if (value - offset) equals (0,1,2,...), jump to corresponding address, otherwise default
+JMP_VAL    rvalue, #:(CONST, addr), addr-default       ; if value equals a constant, jump to address, otherwise default
-JMP_VAL_N  #:(rvalue), #:(CONST, addr), addr-default   ; if values equals fields of a tuple constant, jump to address, otherwise default

// asserts
// (specifying CONSTRUCT as A_IGNORE implies a breakpoint only, and not a raised exception)
+ASSERT     rvalue, CONSTRUCT                       ; assert boolean (zero parameter exception constructor)
+ASSERT_M   rvalue, CONSTRUCT, STRING               ; assert with a ("M"=) message (one String parameter exception constructor)
+ASSERT_V   rvalue, CONSTRUCT, STRING, #vals(rvalue); assert with a message and ("V"=) values to print (one String parameter exception constructor)

// ("GP"=) general purpose register/local-property operations:
// - rvalues could be a register or a constant value or a local-property indicator (PROPERTY)
// - the lvalue(s) can be a register or a local-property indicator (PROPERTY)
// - a new register of type T is allocated if the lvalue is the next unallocated register
+GP_ADD      rvalue1, rvalue2, lvalue                ; T + T -> T
+GP_SUB      rvalue1, rvalue2, lvalue                ; T - T -> T
+GP_MUL      rvalue1, rvalue2, lvalue                ; T * T -> T
+GP_DIV      rvalue1, rvalue2, lvalue                ; T / T -> T
+GP_MOD      rvalue1, rvalue2, lvalue                ; T % T -> T
+GP_SHL      rvalue1, rvalue2, lvalue                ; T << Int -> T
+GP_SHR      rvalue1, rvalue2, lvalue                ; T >> Int -> T
+GP_USHR     rvalue1, rvalue2, lvalue                ; T >>> Int -> T
+GP_AND      rvalue1, rvalue2, lvalue                ; T & T -> T
+GP_OR       rvalue1, rvalue2, lvalue                ; T | T -> T
+GP_XOR      rvalue1, rvalue2, lvalue                ; T ^ T -> T
+GP_DIVREM   rvalue1, rvalue2, lvalue-quotient, lvalue-remainder; T /% T -> T, T
+GP_DOTDOT   rvalue1, rvalue2, lvalue                ; T .. T -> Range<T>   (inclusive end)
+GP_DOTDOTEX rvalue1, rvalue2, lvalue                ; T .. T -> Range<T>   (exclusive end)
+GP_NEG      rvalue, lvalue                          ; -T -> T
+GP_COMPL    rvalue, lvalue                          ; ~T -> T

// ("L"=) local property and ("P"=) property load/store operations
// - lvalue return values can specify the "next register" to allocate
// - for P_GET and P_REF, an rvalue-target of "black hole" indicates a constant property
+L_GET       PROPERTY, lvalue                        ; get local property (move local property value to a register or local property)
+L_SET       PROPERTY, rvalue                        ; set local property (move register value to local property)
+P_GET       PROPERTY, rvalue-target, lvalue         ; get property (move target property value to a register or local property)
+P_SET       PROPERTY, rvalue-target, rvalue         ; set property (move register value to target property)
+P_VAR       PROPERTY, rvalue-target, lvalue-dest    ; move Var-to-property to destination
+P_REF       PROPERTY, rvalue-target, lvalue-dest    ; move Ref-to-property to destination

// register/local-property inc/dec ("A"=after, aka "post" / "B"=before, aka "pre"):
// - the lvalue-target indicates the register or the local-property indicator (PROPERTY)
// - the lvalue can be a register or a local-property indicator (PROPERTY)
// - a new register of type T is allocated if the lvalue is the next unallocated register
+IP_INC      lvalue-target                           ; in-place increment; no result
+IP_DEC      lvalue-target                           ; in-place decrement; no result
+IP_INCA     lvalue-target, lvalue                   ; T++ -> T
+IP_DECA     lvalue-target, lvalue                   ; T-— -> T
+IP_INCB     lvalue-target, lvalue                   ; ++T -> T
+IP_DECB     lvalue-target, lvalue                   ; --T -> T

// ("IP"=) in-place ("___ and assign") ops:
// - the lvalue can be a register or a local-property indicator (PROPERTY)
// - the rvalue can be a register or a constant value or a local-property indicator (PROPERTY)
+IP_ADD      lvalue, rvalue2                         ; T += T
+IP_SUB      lvalue, rvalue2                         ; T -= T
+IP_MUL      lvalue, rvalue2                         ; T *= T
+IP_DIV      lvalue, rvalue2                         ; T /= T
+IP_MOD      lvalue, rvalue2                         ; T %= T
+IP_SHL      lvalue, rvalue2                         ; T <<= Int
+IP_SHR      lvalue, rvalue2                         ; T >>= Int
+IP_USHR     lvalue, rvalue2                         ; T >>>= Int
+IP_AND      lvalue, rvalue2                         ; T &= T
+IP_OR       lvalue, rvalue2                         ; T |= T
+IP_XOR      lvalue, rvalue2                         ; T ^= T

// ("P"=) property ("IP"=) in-place inc/dec:
// - the rvalue-target indicates the object on which to get/set the property value
// - the rvalue-target supports the "local property" mode
// - the lvalue return value can be a register or a local-property indicator (PROPERTY)
// - lvalue return values can specify the "next register" to allocate
+PIP_INC     PROPERTY, rvalue-target                 ; in-place increment; no result
+PIP_DEC     PROPERTY, rvalue-target                 ; in-place decrement; no result
+PIP_INCA    PROPERTY, rvalue-target, lvalue         ; T++ -> T
+PIP_DECA    PROPERTY, rvalue-target, lvalue         ; T-- -> T
+PIP_INCB    PROPERTY, rvalue-target, lvalue         ; ++T -> T
+PIP_DECB    PROPERTY, rvalue-target, lvalue         ; --T -> T

// ("IP"=) in-place ("___ and assign") ("P"=) property ops:
// - the rvalue-target indicates the object on which to get/set the property value
// - the rvalue-target supports the "local property" mode
// - the rvalue2 can be a register or a constant value or a local-property indicator (PROPERTY)
+PIP_ADD     PROPERTY, rvalue-target, rvalue2        ; T += T
+PIP_SUB     PROPERTY, rvalue-target, rvalue2        ; T -= T
+PIP_MUL     PROPERTY, rvalue-target, rvalue2        ; T *= T
+PIP_DIV     PROPERTY, rvalue-target, rvalue2        ; T /= T
+PIP_MOD     PROPERTY, rvalue-target, rvalue2        ; T %= T
+PIP_SHL     PROPERTY, rvalue-target, rvalue2        ; T <<= Int
+PIP_SHR     PROPERTY, rvalue-target, rvalue2        ; T >>= Int
+PIP_USHR    PROPERTY, rvalue-target, rvalue2        ; T >>>= Int
+PIP_AND     PROPERTY, rvalue-target, rvalue2        ; T &= T
+PIP_OR      PROPERTY, rvalue-target, rvalue2        ; T |= T
+PIP_XOR     PROPERTY, rvalue-target, rvalue2        ; T ^= T

// ("I"=) index-based (tuple and single-dimensioned array) ops
// - all rvalues support the "local property" mode
// - lvalue return values can specify the "next register" to allocate
+I_GET       rvalue-target, rvalue-ix, lvalue        ; T = T[ix]
+I_SET       rvalue-target, rvalue-ix, rvalue        ; T[ix] = T

// ("I"=) index-based ("IP"=) in-place ops:
+IIP_INC     rvalue-target, rvalue-ix                ; ++T[ix] (no result)
+IIP_DEC     rvalue-target, rvalue-ix                ; --T[ix] (no result)
+IIP_INCA    rvalue-target, rvalue-ix, lvalue        ; T[ix]++ -> T
+IIP_DECA    rvalue-target, rvalue-ix, lvalue        ; T[ix]-— -> T
+IIP_INCB    rvalue-target, rvalue-ix, lvalue        ; ++T[ix] -> T
+IIP_DECB    rvalue-target, rvalue-ix, lvalue        ; --T[ix] -> T
+IIP_ADD     rvalue-target, rvalue-ix, rvalue2       ; T[ix] += T
+IIP_SUB     rvalue-target, rvalue-ix, rvalue2
+IIP_MUL     rvalue-target, rvalue-ix, rvalue2
+IIP_DIV     rvalue-target, rvalue-ix, rvalue2
+IIP_MOD     rvalue-target, rvalue-ix, rvalue2
+IIP_SHL     rvalue-target, rvalue-ix, rvalue2
+IIP_SHR     rvalue-target, rvalue-ix, rvalue2
+IIP_USHR    rvalue-target, rvalue-ix, rvalue2
+IIP_AND     rvalue-target, rvalue-ix, rvalue2
+IIP_OR      rvalue-target, rvalue-ix, rvalue2
+IIP_XOR     rvalue-target, rvalue-ix, rvalue2

// ("M"=) multi-dimensioned array ops
 M_GET       rvalue-target, #:(rvalue-ix), lvalue    ; T = T[ix*]
 M_SET       rvalue-target, #:(rvalue-ix), rvalue    ; T[ix*] = T
 M_VAR       rvalue-target, #:(rvalue-ix), lvalue    ; Var<T> = &T[ix*]
 M_REF       rvalue-target, #:(rvalue-ix), lvalue    ; Ref<T> = &T[ix*]

// ("M"=) multi-dimensioned array ("IP"=) in-place ops:
 MIP_INC     rvalue-target, #:(rvalue-ix)            ; ++T[ix] (no result)
 MIP_DEC     rvalue-target, #:(rvalue-ix)            ; --T[ix] (no result)
 MIP_INCA    rvalue-target, #:(rvalue-ix), lvalue    ; T[ix]++ -> T
 MIP_DECA    rvalue-target, #:(rvalue-ix), lvalue    ; T[ix]-— -> T
 MIP_INCB    rvalue-target, #:(rvalue-ix), lvalue    ; ++T[ix] -> T
 MIP_DECB    rvalue-target, #:(rvalue-ix), lvalue    ; --T[ix] -> T
 MIP_ADD     rvalue-target, #:(rvalue-ix), rvalue2
 MIP_SUB     rvalue-target, #:(rvalue-ix), rvalue2
 MIP_MUL     rvalue-target, #:(rvalue-ix), rvalue2
 MIP_DIV     rvalue-target, #:(rvalue-ix), rvalue2
 MIP_MOD     rvalue-target, #:(rvalue-ix), rvalue2
 MIP_SHL     rvalue-target, #:(rvalue-ix), rvalue2
 MIP_SHR     rvalue-target, #:(rvalue-ix), rvalue2
 MIP_USHR    rvalue-target, #:(rvalue-ix), rvalue2
 MIP_AND     rvalue-target, #:(rvalue-ix), rvalue2
 MIP_OR      rvalue-target, #:(rvalue-ix), rvalue2
 MIP_XOR     rvalue-target, #:(rvalue-ix), rvalue2



pre-defined variables/constants:
index  r/w  name            type        description
-----  ---  --------------  ----------  ------------------------------------------------
-0x01  r/w  (n/a)           auto        frame-local stack; auto type: each pop/read *MUST* correspond 1:1 with a previous push/write
-0x02   w   (n/a)           any         black hole (basically the /dev/null of registers)
-0x03   r   this:target                 the "this" that was used to invoke the method                             (only in virtual context)
-0x04   r   this:public                 type is meta.PublicType                                                   (only in virtual context)
-0x05   r   this:protected              type is meta.ProtectedType                                                (only in virtual context)
-0x06   r   this:private                type is meta.PrivateType                                                  (only in virtual context)
-0x07   r   this:struct     Struct      type is meta.StructType                                                   (only in virtual or construct context)
-0x08   r   this:service	Service     the run-time service within which the currently executing code is running (any context permitted)
-0x09   r   super           Function    type is the appropriate compile-time generic form                         (only in virtual context)
-0x0A       (n/a)                       *reserved
-0x0B       (n/a)                       *reserved
-0x0C       (n/a)                       *reserved
-0x0D       (n/a)                       *reserved
-0x0E       (n/a)                       *reserved
-0x0F       (n/a)                       *reserved
-0x10       (n/a)                       *reserved


Tags: language   vm   compiled   services   cloud   distribution  

Last modified 01 July 2021