Supercharge your Electron apps with Rust

Electron has become the premier technology for creating cross-platform desktop applications. It is used in platforms as well-known and popular as Atom, VS Code, Spotify, Slack, Discord, and many more.

JavaScript is a slow language, but in this tutorial, we’ll demonstrate how you can use Rust to speed up an Electron app for a better user experience. Rust is a low-level language with high-level ergonomics. It’s blazing fast with a very low memory footprint.

Setting up an Electron project

We’ll use electron-react-boilerplate to create a boilerplate for our demo application. Start by cloning the GitHub repo.

git clone --depth 1 --single-branch https://github.com/electron-react-boilerplate/electron-react-boilerplate.git md_notes

Install all the dependencies using npm or Yarn.

npm run dev

We’ll create a note-taking application with support for markdown. The markdown parser will be written in Rust. We’ll use a variety of methods to interface between Electron and Rust.

Adding UI to your Electron project

We’ll use Material-UI to get ready-made components for our application.

import React, { useState, useEffect } from 'react';
import { makeStyles, createStyles, Theme } from '@material-ui/core/styles';
import Grid from '@material-ui/core/Grid';
import Paper from '@material-ui/core/Paper';
import { TextField, Typography, ListItemText, Button } from '@material-ui/core';
import List from '@material-ui/core/List';
import ListItem from '@material-ui/core/ListItem';
import { remote } from 'electron';
// import fs and path nodejs core module for wrting and reading files
const fs = remote.require('fs');
const path = remote.require('path');
const useStyles = makeStyles((theme: Theme) =>
  createStyles({
    root: {
      flexGrow: 1,
      width: '100vw',
    },
    paper: {
      width: '100%',
      minHeight: '100vh',
      padding: '5px',
    },
    control: {
      padding: theme.spacing(2),
    },
    textArea: {
      width: '100%',
      height: '100%',
    },
    heading: {
      textAlign: 'center',
    },
  })
);
const NOTES_FOLDER = 'notes';
const save = (name: string, data: any) => {
  fs.writeFileSync(
    path.join('.', NOTES_FOLDER, `${name}${Date.now()}.json`).toString(),
    JSON.stringify(data)
  );
};
export default function SpacingGrid() {
  const [editorState, setEditorState] = useState('');
  const [previewState, setPreview] = useState('');
  const [notes, setNotes] = useState([]);
  useEffect(() => {
    // we need to add code to parse md and set it to preview
  }, [setPreview, editorState]);
  useEffect(() => {
    const note = fs
      .readdirSync(path.join('.', NOTES_FOLDER))
      .map((file: string) => {
        return JSON.parse(
          String(fs.readFileSync(path.join('.', NOTES_FOLDER, file)))
        );
      });
    setNotes(note);
  }, []);
  const [title, setTitle] = useState('');
  const classes = useStyles();
  return (
    <Grid container className={classes.root} spacing={2}>
      <Grid item xs={12}>
        <Grid container justify="center" spacing={2}>
          <Grid item xs={9}>
            <Paper className={classes.paper}>
              <Grid container justify="space-between">
                <Grid item>
                  <Typography variant="h3" align="center">
                    Add Notes
                  </Typography>
                </Grid>
                <Grid item>
                  <Button
                    variant="outlined"
                    onClick={() => {
                      save(title, {
                        title,
                        createdAt: Date.now(),
                        body: editorState,
                      });
                    }}
                  >
                    Save
                  </Button>
                </Grid>
              </Grid>
              <Grid
                container
                direction="column"
                justify="space-between"
                spacing={2}
              >
                <Grid item>
                  <TextField
                    label="Title"
                    value={title}
                    onChange={(e) => setTitle(e.target.value)}
                    variant="outlined"
                    className={classes.textArea}
                  />
                </Grid>
                <Grid item>
                  <TextField
                    multiline
                    label="Content"
                    value={editorState}
                    onChange={(e) => setEditorState(e.target.value)}
                    variant="outlined"
                    className={classes.textArea}
                    rows={20}
                  />
                </Grid>
                <Grid item>
                  <div dangerouslySetInnerHTML={{ __html: previewState }} />
                </Grid>
              </Grid>
            </Paper>
          </Grid>
          <Grid item xs={3}>
            <Paper className={classes.paper}>
              <List>
                {notes.map((note, i) => (
                  <ListItem
                    key={`key-${i}`}
                    onClick={() => {
                      setEditorState(note.body);
                      setTitle(note.title);
                    }}
                  >
                    <ListItemText
                      primary={note.title}
                      secondary={new Date(note.createdAt).toLocaleString('in')}
                    />
                  </ListItem>
                ))}
              </List>
            </Paper>
          </Grid>
        </Grid>
      </Grid>
    </Grid>
  );
}

This is boilerplate code for Electron applications. Now we’ll implement a parse function to convert markdown to HTML.

We’ll explore various methods of using Rust for parsing markdown, weighing the pros and cons of each. We’ll also examine use cases for each method.

Using WebAssembly in Electron

At its core, an Electron app is just a Chrome browser and Node.js. We can leverage WebAssembly support in Chrome and Node.

WebAssembly is a binary instruction format that runs inside a browser or a stack-based virtual machine. We’ll compile our markdown parser written in Rust in WebAssembly.

Creating a Rust markdown parser in WebAssembly

Rust has first-class support for WebAssembly. It uses LLVM to compile for WebAssembly as a target. \wasm-pack\ is a tool the Rust community created to help developers write Node modules in Rust using WebAssembly.

Using wasm-pack

wasm-pack is a tool for creating, building, and publishing Node modules written in WebAssembly.

curl https://rustwasm.github.io/wasm-pack/installer/init.sh -sSf | sh

wasm-pack provides an installer script for Linux and macOS. Once wasm-pack is installed on your system, you can create a boilerplate for the WebAssembly module using wasm-pack new.

This will create a new directory named md-parser with boilerplate.

wasm-pack new md_webassembly


├── Cargo.toml
├── LICENSE_APACHE
├── LICENSE_MIT
├── README.md
├── src
│   ├── lib.rs
│   └── utils.rs
└── tests
    └── web.rs

The lib.rs is the main file for the Wasm library. A close look at lib.rs reveals that it exports the greet function called by JavaScript.

// utils set a panic hook so it is console in js
mod utils;
use wasm_bindgen::prelude::*;
// When the `wee_alloc` feature is enabled, use `wee_alloc` as the global
// allocator.
#[cfg(feature = "wee_alloc")]
#[global_allocator]
static ALLOC: wee_alloc::WeeAlloc = wee_alloc::WeeAlloc::INIT;

// function defination for alert. It is dom api for generating a dialog
#[wasm_bindgen]
extern {
    fn alert(s: &str);
}

// wasm_bingen is a pro macro which transfor return type and arguments to the function
#[wasm_bindgen]
pub fn greet() {
// calling alert
    alert("Hello, md-webassembly!");
}

wasm-bidgen is a crate developed by the Rust community that makes it easier create Wasm modules. Wasm doesn’t support strings or array data type; it only supports integers and floats. So it needs the address of the string on the JavaScript heap to use access string. wasm-bindgen adds all that boilerplate code to our functions.

For markdown parsing, we’ll use the comrak crate, which supports GitHub-flavored markdown.

[package]
name = "md-webassembly"
version = "0.1.0"
authors = ["anshul <[email protected]>"]
edition = "2018"

# libtype must be cdylib so it can be dynamically linked by electron
[lib]
crate-type = ["cdylib", "rlib"]

[features]
default = ["console_error_panic_hook"]

[dependencies]
wasm-bindgen = "0.2.63"
comrak = "0.8.2"
console_error_panic_hook = { version = "0.1.6", optional = true }
wee_alloc = { version = "0.4.5", optional = true }

[dev-dependencies]
wasm-bindgen-test = "0.3.13"

[profile.release]
# Tell `rustc` to optimize for small code size.
opt-level = "s"

# avoid wasm-opt error while build
[package.metadata.wasm-pack.profile.release]
wasm-opt = false

Using the comrak crate

comark provides markdown_to_html functionality, which parses markdown to HTML. We’ll create a parse function that takes a markdown string as an argument and then returns the HTML string.

mod utils;
use comrak::{markdown_to_html, ComrakOptions};
use wasm_bindgen::prelude::*;
#[cfg(feature = "wee_alloc")]
#[global_allocator]
static ALLOC: wee_alloc::WeeAlloc = wee_alloc::WeeAlloc::INIT;

// take md string return html string
#[wasm_bindgen]
pub fn parse(md: String) -> String {
    markdown_to_html(md.as_str(), &ComrakOptions::default())
}

We are all set to use our rust markdown parser. For compiling use wasm-pack build . build command will create a wasm module with typescript types.

Using WebAssembly in Electron

We can use our Rust markdown parser in both browser-side and Node.js APIs in Electron.

Both Node and Chromium support WebAssembly out of the box. Moreover, the boilerplate used to bootstrap Electron also supports importing wasm modules by default. We can directly import Wasm code without adding any boilerplate.

  useEffect(() => {
    import('../../md-webassembly/pkg/md_webassembly').then((module) =>
      setPreview(module.parse(editorState))
    );
  }, [setPreview, editorState]);

You should use dynamic import so the UI rendering are not blocked while the Wasm module is instantiated. webpack will take care of rest. This method is also useful if you want to support a web app.

Using NAPI in Electron

Napi is a C header that communicates with Node.js. This method is a bit difficult to implement.

We’ll use node-bindgen to create a Rust library that can be used in Node.js directly.

Install nj-cli using cargo.

cargo install nj-cli

We can now create a boilerplate for the Node module, then import it into an Electron application.

Let’s start by creating a simple Rust project.

cargo new md_napi --lib

Now we need to configure cargo to compile it in a dynamic library that can be used by Node.js. You should use this method when the application requires a thread, SIMD, or processor-specific code.

[package]
name = "md_napi"
version = "0.1.0"
authors = ["anshul <[email protected]>"]
edition = "2018"

[lib]
# compile to .so or .dl file based on OS
crate-type = ["cdylib"]

[dependencies]
# Node.js NAPI helper library
node-bindgen = { version = "2.1.1" }
# Parse MarkDown
comrak = "0.8.2"

[build-dependencies]
node-bindgen = { version = "2.1.1", features = ["build"] }

We can export a function in the lib.rs file.

use node_bindgen::derive::node_bindgen;
use comrak::{markdown_to_html, ComrakOptions};

#[node_bindgen]
fn parse(md: String) -> String {
  markdown_to_html(md.as_str(), &ComrakOptions::default())
}

Now we can build the library using the nj-cli build command. It will create a dist directory with index.node. This file is directly required in Electron.

import { remote } from 'electron';
const mdModule = remote.require('../md_napi/dist/index.node');

console.log((mdModule.parse('# Heading'))

In the Electron app, we’ll simply replace the WebAssembly module imports.

 useEffect(() => {
    setPreview(mdModule.parse(editorState));
  }, [setPreview, editorState]);

Other libraries

There are a few additional libraries you can use to create add-ons. The libraries listed below use V8 APIs or NAPIs.

• Neon
• napi-rs
• nodejs-sys

Comparision

Both of the methods described above have pros and cons. The right choice for a particular project depends on the use case and requirements. Use the chart below as a quick reference guide to help you select the most appropriate approach.

Feature	NAPI/add-on	WebAssebmly
Threads	Libuv threads	Web workers
SIMD	Full support	Partial support
Web browser	Unsupported	Supported
C FFI	Supported	Unsupported
External Crates	Full support	May or may not support

Conclusion

JavaScript is a slow language. Therefore, Electron apps tend to be slow. The trick is to offload CPU-intensive tasks to other, faster languages. Whereas JavaScript is well-suited to UI control, Rust is better equipped to handle those complex calculations. If you’ve completed this tutorial, you now have a pair of tried-and-true methods for using Rust to speed up Electron apps under your belt.