Javascript implementation of CLVM (Chia Lisp VM)
npm install clvm-jsshell
npm install clvm
or
yarn add clvm
`Test
passes all the test equivalent to Python's original clvm.
You can compare test files for Javascript and Python
To run the test, execute the following command.
`shell
git clone https://github.com/Chia-Mine/clvm-js
cd clvm-jsnpm install
npm run test
or
yarn
yarn test
`
If you find something not compatible with Python's clvm, please report it to GitHub issues.Compatibility
This code is compatible with:
- of clvm
- Diff to the latest clvm
- 34f504bd0ef2cd3a219fea8ce6b15ff7684687fd of bls-signatures
- Diff to the latest bls-signaturesExample
`javascript
// in nodejs context
async function main(){
const clvm = require("clvm"); //
await clvm.initializeClvmWasm() is now always necessary.
// This loads clvm_wasm_bg.wasm.
// If you have a strong reason to use features of bls-signatures,
// you need to do await clvm.initialize() instead.
// clvm.initialize() loads both blsjs.wasm and clvm_wasm_bg.wasm.
await clvm.initializeClvmWasm();
const {SExp, KEYWORD_TO_ATOM, h, t, run_chia_program, Flag} = clvm;
const plus = h(KEYWORD_TO_ATOM["+"]); // byte representation of '+' operator
const q = h(KEYWORD_TO_ATOM["q"]); // byte representation of 'q' operator
const program = SExp.to([plus, 1, t(q, 175)]).as_bin().raw(); // (+ . (1 . (q . 175)))
const env = SExp.to(25).as_bin().raw();
const max_cost = BigInt(10000000);
const [cost, lazyNode] = run_chia_program(program, env, max_cost, Flag.allow_backrefs());
const result = new SExp(lazyNode);
let isEqual = result.equal_to(SExp.to(25 + 175));
console.log(isEqual: ${isEqual}); // 'isEqual: true'
isEqual = result.as_int() === (25 + 175);
console.log(isEqual: ${isEqual}); // 'isEqual: true'
}main().catch(e => console.error(e));
`More example with clvm_wasm
See this test case for clvm_wasmUse in browser
- Sample code - Use clvm-js with TypeScript, Webpack
- Sample code - Use clvm-js with TypeScript, React, ViteIf you'd like to run some javascript code which depends on
on browser,
you need to put clvm_wasm_bg.wasm and optionally blsjs.wasm to the same directory as the code who loads clvm.
Because most of BLS operations are now performed inside clvm_wasm_bg.wasm, in most cases you don't need blsjs.wasm.
├── ...
├── main.js # js file which clvm is compiled into
├── clvm_wasm_bg.wasm # copy it from npm_modules/clvm/browser/clvm_wasm_bg.wasm
└── (Optional) blsjs.wasm # copy it from npm_modules/clvm/browser/blsjs.wasm
If you use React with CRA(create-react-app), copy
blsjs.wasm and clvm_wasm_bg.wasm into folder. It automatically copies wasm file next to main js file.blsjs.wasm and clvm_wasm_bg.wasm into folder. IMPORTANT NOTE
When your code is loaded as a module, such as with
(common in React with Vite),
there is a path restriction for loading the WebAssembly (WASM) module.
See Known Issues. Also see code comment here Note1
Don't forget to wait
clvm.initializeClvmWasm().
clvm.initializeClvmWasm() only loads clvm_wasm_bg.wasm.
If you have a strong reason to use features of bls-signatures inside clvm-js, you need to wait
clvm.initialize() instead, since clvm.initialize() loads both blsjs.wasm and clvm_wasm_bg.wasm.
Note2
Redistributing your project with bundled blsjs.wasm and/or clvm_wasm_bg.wasm must be compliant with Apache2.0 License provided by Chia-Network
Note3
You may need blsjs.wasm if you want to run run_program, which has been deprecated as of clvm@3.0.0.$3
clvm-js uses BigInt. So if runtime environment does not support BigInt, clvm-js doesn't work as well.
If you transpile code using babel or something which uses babel (like create-react-app),
you need to tell the transpiler to optimize code only for the target browsers.
Just copy and paste below to your package.json and you can avoid a lot of runtime incompatibility issues.
`
"browserslist": [
"edge >= 79",
"firefox >= 68",
"chrome >= 67",
"safari > 14",
"opera >= 54",
"ios_saf >= 14.4",
"android >= 67",
"op_mob >= 48",
"and_chr >= 67",
"and_ff >= 68",
"samsung >= 9.2",
"node >= 10.4.0",
"electron >= 4.0.0"
]
Differences with Python's
Although I try hard to make it look like Python's , there are things users should be aware of.
I put the code which absorbs language incompatibility into src/__type_compaibility__.ts, so if you're interested take a look at it.$3
When you want to create a tuple, you need to write like this:
`javascript
const {t} = require("clvm");
const aTuple = t(1, 2);// Tuple is Array-Like object
aTuple[0] === 1; // true
aTuple[1] === 2; // true
// Tuple content cannot be changed
aTuple[0] = 99;
aTuple[0] === 99; // false
aTuple[0] === 1; // true
// Tuple accepts only 2 elements.
const tuple2 = t(1, 2, 3);
tuple2; // (1, 2)
// You can check if a variable is a tuple
const {isTuple, isList} = require("clvm");
isTuple([1, 2]); // false
isTuple(t(1, 2)); // true
isList([1, 2]); // true
isList(t(1, 2)); // false
`
Just add in front of tuple parenthesis (1, 2) and you get a tuple. $3
This is the most notable difference with Python.
I used to be a JavaScript developer for several years, and sometimes I heard Python is slow and JavaScript is fast.
But working on the project, I truly surprised that Python can handle byte data really well in ways:
- Python's bytes is immutable and can be used as a dict key.
- Python's bytes is fast and easy to write.
b'aaa' + b'bbb' == b'aaabbb', b'a' * 3 == b'aaa'
- Python's comparison is FAST.
If you are interested, compare the performance of test_very_long_blobs in tests/serialize_test.[ts|py]
See more details here`javascript
const {b} = require("clvm");
// Turns string to UTF-8 byte array.
b("abc"); // will be { Uint8Array(3) [97,98,99] }
b("あ"); // will be { Uint8Array(3) [227,129,130] }// If you want to do Byte comparison, use equal_to method.
b("abc").equal_to(b("abc")); // true
b("abc") === b("abc"); // false. Because it compares reference of Bytes instance.
// Initialize Bytes instance with hex string
const {h} = require("clvm");
h("0x616263"); // will be { Uint8Array(3) [97,98,99] }
h("616263"); // You can omit '0x' prefix
h("0x616263").equal_to(b("abc")); // true
// +: Bytes concat
b("a").concat(b("b")); // === b("ab")
// *: Bytes repeat
b("a").repeat(3); // === b("aaa")
// Bytes initialization
const {Bytes} = require("clvm");
uint8 = new Uint8Array([97, 98, 99]);
b1 = Bytes.from(uint8);
b1.equal_to(b("abc")); // true
// Initializing by Bytes.from copies value and cut reference apart.
uint8[0] = 0;
b1.at(0); // 97
// Initializing by new Bytes() just stores value and keep reference, for better performance
uint8 = new Uint8Array([97, 98, 99]);
b2 = new Bytes(uint8);
b2.equal_to(b("abc")); // true
uint8[0] = 0;
b2.at(0); // 0
// Check byte instance
const {isBytes} = require("clvm");
isBytes(b('aaa')); // true
isBytes(h('0xfe')); // true
isBytes(new Bytes()); // true
`$3
If you want to stringify or Bytes, just call x.toString() method.
`javascript
const {b, SExp, str} = require("clvm");
b("あ").toString(); // "b'\\xe3\\x81\\x82'"
SExp.to([1, [2, 3]]).toString(); // 'ff01ffff02ff038080'
str(SExp.to([1, [2, 3]])); // You can use str() function as well as Python by the way.
`$3
|Python|JavaScript|
|------|----------|
||-8n / 5n === -1n|
|-8 % 5 == 2|-8n % 5n === -3n|clvm license
clvm-js` is based on clvm with the