Installation

npm install @snelusha/balrun

CLI

npx @snelusha/balrun ./main.bal

Accepts a .bal file, a package directory, or . for the current package.

Usage

import { Ballerina } from "@snelusha/balrun";

const ballerina = new Ballerina({ colors: false });

// Returns null on success, or { error: "..." } on failure
const result = await ballerina.run("./main.bal", { colors: true });

Options passed to run() override the constructor defaults for that call only.

Options

colors

Diagnostics use ANSI colors by default. Pass colors: false to disable. The CLI auto-detects based on stderr.isTTY.

stdout / stderr

Redirect runtime output by passing any object that implements StreamWriter:

import type { StreamWriter } from "@snelusha/balrun";

const writer: StreamWriter = { write(chunk: string) {} };

Example:

import { Ballerina, type StreamWriter } from "@snelusha/balrun";

const buffer: string[] = [];
const writer: StreamWriter = { write(chunk) { buffer.push(chunk); } };

await new Ballerina({ stdout: writer, stderr: writer }).run("./main.bal");

fs

By default, Ballerina reads from disk via NodeFS. Swap it out by implementing the FS interface — useful for in-memory or virtual filesystems.

import { Ballerina, type FS } from "@snelusha/balrun";

class MemFS implements FS {
  // When running a single file, only `open` and `stat` are required.
  // When running a package, `readDir` is also required.
}

await new Ballerina({ fs: new MemFS() }).run("main.bal");

See examples/mem-fs for a full implementation.

wasmUrl / core

By default, Ballerina loads the bundled ballerina.wasm. Pass wasmUrl to load a different local path or HTTP(S) URL:

await new Ballerina({ wasmUrl: "https://example.com/ballerina.wasm" }).run("main.bal");

For custom loading, pass a BallerinaCore directly. WasmBridge.load() accepts a local path, URL, Response, or Promise<Response>:

import { Ballerina, WasmBridge } from "@snelusha/balrun";

const core = await WasmBridge.load(fetch("/ballerina.wasm"));
await new Ballerina({ core }).run("main.bal");

Acknowledgements

Built on ballerina-lang-go.