AssemblyScript defines a subset of TypeScript that it compiles to WebAssembly. The compiler itself is written in TypeScript and no binary dependencies are required to get started. Under the hood, it rewires TypeScript's compiler API to Binaryen's compiler backend.

Be warned that this is an early prototype and that it cannot do anything useful yet.

An initial playground to mess with the current state of the compiler is available at dcode.io/AssemblyScript.

export function add(a: int, b: double): short {
  return (a + (b as int)) as short;
}

Compiles to:

(module
 (type $iFi (func (param i32 f64) (result i32)))
 (memory $0 256)
 (export "memory" (memory $0))
 (export "add" (func $add))
 (func $add (type $iFi) (param $0 i32) (param $1 f64) (result i32)
  (return
   (i32.shl
    (i32.shr_s
     (i32.add
      (get_local $0)
      (i32.trunc_s/f64
       (get_local $1)
      )
     )
     (i32.const 16)
    )
    (i32.const 16)
   )
  )
 )
)

See the pre-configured example project for a quickstart.

The package also provides a convenient stand-alone loader component to run and work with compiled WebAssembly modules:

import load from "assemblyscript/loader"; // JS: var load = require("assemblyscript/loader").load;

load("path/to/module.wasm").then(module => {
  ...
});

An AssemblyScript program is valid TypeScript syntactically, but not necessarily semantically.

WebAssembly-specific types are obtained by referencing assembly.d.ts:

While generating a warning to avoid type confusion, the JavaScript types number and boolean resolve to double and bool respectively.

WebAssembly-specific operations are available as built-in functions that translate to the respective opcode directly:

• rotl(value: int, shift: int): int
  Performs the sign-agnostic rotate left operation on a 32-bit integer.
• rotll(value: long, shift: long): long
  Performs the sign-agnostic rotate left operation on a 64-bit integer.
• rotr(value: int, shift: int): int
  Performs the sign-agnostic rotate right operation on a 32-bit integer.
• rotrl(value: long, shift: long): long
  Performs the sign-agnostic rotate right operation on a 64-bit integer.
• clz(value: int): int
  Performs the sign-agnostic count leading zero bits operation on a 32-bit integer. All zero bits are considered leading if the value is zero.
• clzl(value: long): long
  Performs the sign-agnostic count leading zero bits operation on a 64-bit integer. All zero bits are considered leading if the value is zero.
• ctz(value: int): int
  Performs the sign-agnostic count tailing zero bits operation on a 32-bit integer. All zero bits are considered trailing if the value is zero.
• ctzl(value: long): long
  Performs the sign-agnostic count trailing zero bits operation on a 64-bit integer. All zero bits are considered trailing if the value is zero.
• popcnt(value: int): int
  Performs the sign-agnostic count number of one bits operation on a 32-bit integer.
• popcntl(value: long): long
  Performs the sign-agnostic count number of one bits operation on a 64-bit integer.
• abs(value: double): double
  Computes the absolute value of a 64-bit float.
• absf(value: float): float
  Computes the absolute value of a 32-bit float.
• ceil(value: double): double
  Performs the ceiling operation on a 64-bit float.
• ceilf(value: float): float
  Performs the ceiling operation on a 32-bit float.
• floor(value: double): double
  Performs the floor operation on a 64-bit float.
• floorf(value: float): float
  Performs the floor operation on a 32-bit float.
• sqrt(value: double): double
  Calculates the square root of a 64-bit float.
• sqrtf(value: float): float
  Calculates the square root of a 32-bit float.
• trunc(value: double): double
  Rounds to the nearest integer towards zero of a 64-bit float.
• truncf(value: float): float
  Rounds to the nearest integer towards zero of a 32-bit float.
• nearest(value: double): double
  Rounds to the nearest integer tied to even of a 64-bit float.
• nearestf(value: float): float
  Rounds to the nearest integer tied to even of a 32-bit float.
• min(left: double, right: double): double
  Determines the minimum of two 64-bit floats. If either operand is NaN, returns NaN.
• minf(left: float, right: float): float
  Determines the minimum of two 32-bit floats. If either operand is NaN, returns NaN.
• max(left: double, right: double): double
  Determines the maximum of two 64-bit floats. If either operand is NaN, returns NaN.
• maxf(left: float, right: float): float
  Determines the maximum of two 32-bit floats. If either operand is NaN, returns NaN.
• copysign(x: double, y: double): double
  Composes a 64-bit float from the magnitude of x and the sign of y.
• copysignf(x: float, y: float): float
  Composes a 32-bit float from the magnitude of x and the sign of y.
• reinterpreti(value: float): int
  Reinterprets the bits of a 32-bit float as a 32-bit integer.
• reinterpretl(value: double): long
  Reinterprets the bits of a 64-bit float as a 64-bit integer.
• reinterpretf(value: int): float
  Reinterprets the bits of a 32-bit integer as a 32-bit float.
• reinterpretd(value: long): double
  Reinterprets the bits of a 64-bit integer as a 64-bit double.
• current_memory(): int
  Returns the current memory size in units of pages. One page is 64kb.
• grow_memory(value: uint): int
  Grows linear memory by a given unsigned delta of pages. One page is 64kb. Returns the previous memory size in units of pages or -1 on failure.
• sizeof<T>(): uintptr
  Returns the byte size of the specified core or class type. Compiles to a constant.

The following built-in constants are present as immutable globals (note that optimizers might inline them):

• NaN: double
  NaN (not a number) as a 64-bit float.
• NaNf: float
  NaN (not a number) as a 32-bit float.
• Infinity: double
  Positive infinity as a 64-bit float.
• Infinityf: float
  Positive infinity as a 32-bit float.

By default (i.e. if the --nolib option isn't set), standard memory management routines based on dlmalloc and musl will be linked statically and exported to the embedder:

• malloc(size: uintptr): uintptr
  Allocates a chunk of memory of the specified size and returns a pointer to it.
• free(ptr: uintptr): void
  Frees a previously allocated chunk of memory by its pointer.
• memcpy(dest: uintptr, src: uintptr, size: uintptr): uintptr
  Copies data from one chunk of memory to another.
• memset(dest: uintptr, c: int, size: uintptr): uintptr
  Sets a chunk of memory to the provided value c. Usually used to reset it to all 0s.
• memcmp(vl: uintptr, vr: uintptr, n: uintptr): int
  Compares a chunk of memory to another. Returns 0 if both are equal, otherwise vl[i] - vr[i] at the first difference's byte offset i.

Linking in memory management routines adds about 11kb to a module. Once WebAssembly exposes the garbage collector natively, there'll be other options as well. Note that the new operator depends on malloc and will break when --nolib is specified (and no other malloc is present). Also note that calling grow_memory where malloc is present will most likely break malloc as it expects contiguous memory.

Type coercion requires an explicit cast where precision or signage is lost respectively is implicit where it is maintained. For example, to cast a double to an int:

function example(value: double): int {
  return value as int; // translates to the respective opcode
}

Global WebAssembly imports can be declared anywhere while WebAssembly exports are exported from the entry file (the file specified when calling asc or Compiler.compile). Aside from that, imports and exports work just like in TypeScript.

// entry.ts

import { myOtherExportThatDoesntBecomeAWebAssemblyExport } from "./imported";

declare function myImport(): void;

export function myExport(): void {
  myOtherExportThatDoesntBecomeAWebAssemblyExport();
}

Currently, imports can also be pulled from different namespaces by separating the namespace and the function with a $ character.

declare function Math$random(): double;

Naming a function start with no arguments and a void return type will automatically make it the start function that is being called on load even before returning to the embedder.

function start(): void {
  ...
}

• Compiler.compileFile(filename: string, options?: CompilerOptions): binaryen.Module | null
  Compiles the specified entry file to a WebAssembly module. Returns null on failure.

• Compiler.compileString(source: string, options?: CompilerOptions): binaryen.Module | null
  Compiles the specified entry file source to a WebAssembly module. Returns null on failure.

• Compiler.lastDiagnostics: typescript.Diagnostic[]
  Contains the diagnostics generated by the last invocation of compilerFile or compileString.

• CompilerOptions
  AssemblyScript compiler options.

  • uintptrSize: number
    Specifies the byte size of a pointer. 32-bit WebAssembly expects 4, 64-bit WebAssembly expects 8. Defaults to 4.
  • noLib: boolean
    Specifies whether the standard library (malloc, free, etc.) shall be excluded. Defaults to false.
  • silent: boolean
    Specifies that compilation shall be performed in silent mode without writing to stderr. Defaults to false.
  • treeShaking: boolean
    Disables built-in tree-shaking if set to false. Defaults to true.

import { Compiler, typescript } from "assemblyscript";

const module = Compiler.compileString(`
export function add(a: int, b: int): int {
  return a + b;
}
`, {
  uintptrSize: 4,
  silent: true
});

console.error(typescript.formatDiagnostics(Compiler.lastDiagnostics));
if (!module)
  throw Error("compilation failed");

module.optimize();

if (!module.validate())
  throw Error("validation failed");

const textFile = module.emitText();
const wasmFile = module.emitBinary();
...

The command line compiler asc works similar to TypeScript's tsc:

Syntax: asc [options] [file ...]

Options:
 -o, --out, --outFile   Specifies the output file name.
 -v, --validate         Validates the module.
 -O, --optimize         Runs optimizing binaryen IR passes.
 -t, --text             Emits text format instead of a binary.

That's it for now. Feel free to experiment. PRs welcome!

License: Apache License, Version 2.0