Acknowledgements

The beginning of your journey with Web Components.

Foreword

A brief introduction to the set of specifications called Web Components. We'll introduce you to concepts such as Shadow DOM, HTML templates, Custom Elements v1 and uncover typical use cases for developing Web Components.

Introduction to Web Components

We'll introduce you to the mockups in Figma, the development environment used throughout the book, and other conventions used in this book.

Conventions Used In The Book

A guide for installing various technologies needed prior to coding the exercises in this book.

Setting Up The Development Environment

We'll look at what you'll learn throughout Fullstack Web Components.

Module 1 Summary

Introduction

In Part 1, you'll code a form using several Web Components specifications.

Specification

Part One

By developing the CardComponent, you'll get hands on experience with autonomous custom elements.

Autonomous custom elements

Before you can develop the CardComponent, you'll need to scaffold the component in Storybook.

Template Slots

Shadow DOM provides encapsulation for styling and template. You'll style CardComponent with CSS in the context of the autonomous custom element's Shadow DOM.

Styling CardComponent

Write integration and snapshot tests for CardComponent using Cypress and Storybook Storyshots.

Testing the card

You built an autonomous custom element with Shadow DOM, HTML templates, and ES Modules. Let's reflect on that.

Chapter One Summary

Getting to Know Web Components

Form-associated custom elements allow custom elements to participate in HTML forms. From validation to submission, you'll code a text input end-to-end.

Form-associated custom elements

Scaffold a form-associated custom element named TextInputComponent in Storybook.

TextInputComponent

You'll learn what distinguises a form-associated custom element from an autonomous custom element, the formAssociated property and ElementInternals interface.

Form Associated and ElementInternals

Form-associated custom elements don't come with a pattern that validates an entire form, so you'll implement a pattern that allows form validation.

Handling Validation

Using observedAttributes and attributeChangedCallback to listen and react to changes in custom element attributes.

Listening for attribute changes

You'll style the various interaction states of the TextInputComponent form-associated custom element.

User feedback

We've created a problem by encapsulating the HTMLInputElement by Shadow DOM. The HTMLInputElement API cann't be accessed by the outside world. In this section, you'll fix that by implementing getters and setters that emulate the HTMLInputElement on TextInputComponent.

Emulating HTMLInputElement

You'll update the Story and provide the missing APIs to TextInputComponent that allow the component to fully interact with a HTML form.

Making a form

Reviewing the work completed on the form-associated custom element TextInputComponent.

Chapter Two Summary

You'll code ButtonComponent, a customized built-in element that inherits accessibility characteristics from HTMLButtonElement.

Customized built-in elements

Scaffold the customized built-in element ButtonComponent with Storybook.

Scaffolding ButtonComponent

Customized built-ins can't benefit from the encapsulation provided by Shadow DOM, so you'll have to find another way to style ButtonComponent.

Styling ButtonComponent

Revist the HTML form by making the user experience more accessible.

Making the form accessible

How customized built-in elements can make custom elements more accessible.

Chapter Three Summary

You'll learn best practices for developing a UI Library with Web Components by coding a micro-library, developing two custom elements with the micro-library, and figuring out how to package the UI library for distribution.

UI Library

Part Two

You'll abstract common patterns intro reusable code in a micro-library powered by TypeScript Decorators.

Micro-library

How to code an interface for Web Components with TypeScript decorators that enables declaration of the selector, styles, and template.

Class Decorator

You'll code a method decorator named Listen that binds event listeners to class methods.

Method Decorator

In this chapter you learned how to code a micro-library with TypeScript decorators that streamlines Web Component development.

Chapter Four Summary

You'll investigate multiple methods for components to communicate by coding an editable table.

Component Communication

How to use mock data in Storybook to code Web Components.

Mocking Table Data in Storybook

You'll learn about three methods for custom elements to communicate, attribute drilling, CustomEvent, and BroadcastChannel.

Coding Compound Components

You'll learn about how to create a separation of concerns between the many custom elements that comprise the editable table.

Making the Table Editable

You learned about how to create a separation of concerns between various custom elements in the same user experience flow.

Summary

Compound Custom Elements

You'll learn how to code a common UX pattern (dialogs) using HTML templates and custom elements.

HTML Templates

You'll mock the DialogComponent in Storybook.

Mocking the Dialog in Storybook

You'll make a configurable dialog with custom elements.

DialogComponent

You'll scaffold two custom elements that leverage DialogComponent to display modals and tooltips.

Modal and Tooltip

You'll fix an issue faced in the last section by learning how to query for elements across Shadow boundaries.

Querying DOM Across Shadow Boundaries

You'll control the ordering and event cycle of dialogs with a custom class named DialogStack.

DialogStack

Summarizing the work completed in DialogComponent, ModalComponent, TooltipComponent, and DialogStack.

Chapter Six Summary

Powering Dialogs with HTML Templates

You'll learn best practices for organizing and distributing Web Component UI libraries.

UI Libraries

You'll learn brest practices for distributing Web Components by formatting the code with named exports, configuring the build with Rollup, providing a standard package format, and writing a custom Node.js script that enables you to publish packages.

Distribution

Custom Elements Manifest provides a wealth of information about each component. After generating a manifest, it will provide documentation for each component in Storybook.

Documentation

You'll learn how to setup continuous integration for the UI library with Github Actions.

Continuous Integration

Let's reflect on the best practices you learned for distributing UI libraries filled with Web Components.

Chapter Seven Summary

Distributing Custom Element Libraries

You'll integrate the UI library with a single-page application and code the end-to-end user experience with custom elements.

Application

Part Three

You'll provide a separation of concerns with custom elements in a single page application.

Application Development with Custom Elements

You'll scaffold the landing page with an autonomous custom element named MainView.

Landing Page

You'll code a global header with autonomous custom elements named AppHeader.

Header

You'll code another autonomous custom element that handles cookie authorization named CookieFooter.

CookieFooter

You'll lazyload an image with customized built-ins.

Background

You developed the landing page of the application, learning how to architect a separation of concerns between the UI components.

Chapter Eight Summary

To handle multiple views in the single page application, you'll need to implement a router that can navigate between each view.

Routing Custom Elements in a SPA

You'll code the Router class that handles the creation and removal of views in the application.

Routing

You'll code the necessary components and services required in the login view.

Login View

Integrate TableCardComponent into a third view of the single page application, the dashboard.

Dashboard View

You coded a single page web application with custom elements and a custom router.

Chapter Nine Summary

You'll server-side render the views of the application with Declarative Shadow DOM and @lit-labs/ssr

Server-Side Rendering Custom Elements

You'll declare a server-side renderable Declarative Shadow DOM HTML template for each component.

Declarative Shadow DOM Templates

To get custom elements to work in Node.js, you'll need to polyfill DOM APIs. Luckily, @lit-labs/ssr provides a tool that enables just that.

Shim

You'll develop Express middleware that server-side renders Declarative Shadow DOM HTML Templates with @lit-labs/ssr.

Express Middleware

You rendered custom elements server-side using Declarative Shadow DOM and @lit-labs/ssr.

Chapter Ten Summary

Server-side Rendering with Declarative Shadow DOM

The conclusion of the book Fullstack Custom Elements.

Conclusion

Fullstack Web Components

Enter your name and email address below, then click the **"Send my free chapter"** button to get started

There are several books about Web Components, but none go into how to integrate Web Components in a typical enterprise web development workflow. This book is designed to help you learn Web Components and integrate it into your work.

In this book, we cover the basics of Web Components, how to code complex user interfaces with custom elements in Storybook, how to test and maintain UI libraries built with custom elements, and how to build web applications entirely with Web Components and browser specifications. It comes with complete code examples. We walk through the entire process step-by-step.

It's taught by Stephen Belovarich, a software engineer and web developer. Stephen has provided what he's learned about UI libraries and web applications while working at Ubiquiti Networks, Norton, Nike, NBC Universal, and Workday, that can benefit other developers, students, and anyone making a website.

The Complete Guide to Building UI Libraries with Web Components

In this book, we will learn Web Components through the lens of a JavaScript developer.

The basics of working with and manipulating Web Components

The best practices for maintaining a UI library made of custom elements

How to create a micro-library for Web Components with TypeScript Decorators

What problems Web Components can solve in everyday development

How to build web applications entirely from Web Components and JavaScript

 can't simply query for a HTML template using 

 method doesn't cross the Shadow DOM boundaries imposed by the multiple instances of 

 could be an autonomous custom element that doesn't use Shadow DOM. We could use other methods from the 

 package so modals get the appropriate styling, like we have done with customized built-in elements in previous chapters. If we choose this approach, we lose the main benefit of Shadow DOM: the encapsulated style and template. Without Shadow DOM, 

Another approach that would allow us to retain Shadow DOM in 

 is to code a new utility that can query for elements while piercing the Shadow boundary. For that, we'd need a 

 that can walk the tree of nodes that make up DOM and if an element has a 

 and walk that tree of nodes, trying to query for the element we need to select.

 already that could help walk a tree of nodes. 

 is a specialized method for setting up a 

 loop to iterate through every node in a tree of nodes, starting at the 

 doesn't by default cross the shadow boundary, so we'll have to code some custom logic to ensure that happens, but 

 does provide an extremely performant means to walk a tree of nodes.

const walk = document.createTreeWalker(
  element ? element : (document.body as Element),
  NodeFilter.SHOW_ELEMENT,
  {
    acceptNode(node) {
      return NodeFilter.FILTER_ACCEPT;
    },
  },
  false
);
while (walk.nextNode()) {}

In this section, you'll move back to the library code in the 

 should get called until the last node in the tree has been parsed or a match is found via 

 will be exported from the library and imported into the 

touch packages/common/src/template/query.ts

, so we'll need to pass the starting point for the query in the second argument named 

This algorithm needs to be recursive because there will be instances where the 

 will reach a shadow boundary and stop, but we want to run the algorithm again, starting at the 

 and walking the tree of nodes from there, querying for the element that should match the 

 that takes similar arguments. This time the second argument represents a match, which only happens once, but we need to pass it recursively and check if there's a match at the start of the 

, a starting point to query for the element. The arguments get flipped here because 

 to satisfy the type definitions for now.

export function query(selector: string, root?: HTMLElement): Element {
  function __query(
    selector: string,
    match: Element,
    element?: Element
  ): Element {
    return match;
  }
  return __query(selector, root);
}

The code we'll write goes inside of the recursive 

 to find the element and set the value of 

 here, we'll want to skip to the next line. On the next line, use a conditional to check if 

if (!element && document.querySelector(selector)) {
  match = document.querySelector(selector);
}

if (match) {
  return match;
}

, we want to instantiate a new instance of 

 and passing all the arguments required by 

. The first argument is the root node to start walking, which should be the 

 is a way to filter types of nodes. We'll use the 

. The next argument is a filter for the nodes that pass the first filter, it tells when 

, a boolean that indicates if an entire sub-tree should be discarded when a reference is also discarded. For more information about 

const walk = document.createTreeWalker(
  element ? element : (document.body as Element),
  NodeFilter.SHOW_ELEMENT,
  {
    acceptNode(node) {
      return NodeFilter.FILTER_ACCEPT;
    },
  },
  false
);

The above code will cause a TypeScript error 

. This is due to the TypeScript DOM library marking the fourth argument as deprecated. W3C 

 and browsers implement the argument. This argument, also known as 

 because we're only concerned with the current node as the algorithm walks the tree, not the entire sub-tree. To discard the TypeScript error, place a 

 comment in the line above the fourth argument.

const walk = document.createTreeWalker(
  element ? element : (document.body as Element),
  NodeFilter.SHOW_ELEMENT,
  {
    acceptNode(node) {
      return NodeFilter.FILTER_ACCEPT;
    },
  },
  // @ts-ignore
  false
);

if ((walk.currentNode as Element).querySelector(selector)) {
  match = (walk.currentNode as Element).querySelector(selector);
}

Write another conditional that checks if the 

if ((walk.currentNode as HTMLElement).shadowRoot) {
  __query(
    selector,
    match,
    (<unknown>(walk.currentNode as HTMLElement).shadowRoot) as Element
  );
}

 at the very end. That's all there is to coding an algorithm that recursively walks a tree of nodes and queries for a single 

, but what if you wanted to replicate the functionality of 

---
title: Querying DOM Across Shadow Boundaries
description: You'll fix an issue faced in the last section by learning how to query for elements across Shadow boundaries.
isPublicLesson: true
---

## Querying DOM Across Shadow Boundaries

There's a problem in the `DialogComponent` that causes the `TooltipComponent` not to appear when `DialogComponent` is declared in the context of the `ModalComponent` Shadow DOM. Since the instances of `DialogComponent` is embedded in Shadow DOM, `DialogComponent` can't simply query for a HTML template using `document.querySelector`. The `querySelector` method doesn't cross the Shadow DOM boundaries imposed by the multiple instances of `ModalComponent`.

There is one way we could fix this. `ModalComponent` could be an autonomous custom element that doesn't use Shadow DOM. We could use other methods from the `@in/common` package so modals get the appropriate styling, like we have done with customized built-in elements in previous chapters. If we choose this approach, we lose the main benefit of Shadow DOM: the encapsulated style and template. Without Shadow DOM, `querySelector` would be able to function per normal.

Another approach that would allow us to retain Shadow DOM in `ModalComponent` is to code a new utility that can query for elements while piercing the Shadow boundary. For that, we'd need a `function` that can walk the tree of nodes that make up DOM and if an element has a `shadowRoot`, recursively call the `function` and walk that tree of nodes, trying to query for the element we need to select.

Luckily, a method exists on the `document` already that could help walk a tree of nodes. [`document.createTreeWalker`](https://developer.mozilla.org/en-US/docs/Web/API/Document/createTreeWalker) is a specialized method for setting up a `TreeWalker Object`. While holding a reference to the `TreeWalker`, you can use a `while` loop to iterate through every node in a tree of nodes, starting at the `document.body` or any other node, i.e. the `shadowRoot`. `createTreeWalker` doesn't by default cross the shadow boundary, so we'll have to code some custom logic to ensure that happens, but `TreeWalker` does provide an extremely performant means to walk a tree of nodes.

```typescript
const walk = document.createTreeWalker(
  element ? element : (document.body as Element),
  NodeFilter.SHOW_ELEMENT,
  {
    acceptNode(node) {
      return NodeFilter.FILTER_ACCEPT;
    },
  },
  false
);
while (walk.nextNode()) {}
```

In this section, you'll move back to the library code in the `@in/common` package and code a new `function` named `query` which returns a recursive `function` that gets called every time the `TreeWalker` encounters a node that has a `shadowRoot`. The recursive `function` should get called until the last node in the tree has been parsed or a match is found via `querySelector`. `query` will be exported from the library and imported into the `DialogComponent` just like any other `function` from the `@in/common` package.

Make a new file named `query.ts` in the directory `packages/common/src/template`.

```bash
touch packages/common/src/template/query.ts
```

In `query.ts` export a new `function` named `query`. This function will work much like `querySelector`, in that the first argument takes the `selector`. Unlike `querySelector` which exists on the prototype of every `Element`, `query` is a "pure" `function`, so we'll need to pass the starting point for the query in the second argument named `root`. `query` returns an `Element`, the element we are selecting with the `selector`.

This algorithm needs to be recursive because there will be instances where the `TreeWalker` will reach a shadow boundary and stop, but we want to run the algorithm again, starting at the `Shadow Root` and walking the tree of nodes from there, querying for the element that should match the `selector`. Return another `function` named `__query` that takes similar arguments. This time the second argument represents a match, which only happens once, but we need to pass it recursively and check if there's a match at the start of the `function`. The third argument is another `Element` which we'll use similarly to the `query function`, a starting point to query for the element. The arguments get flipped here because `element` isn't required in the initial call to `__query`. Declare the `__query function` here in the context of `query` and make sure it returns the `Element` to satisfy the type definitions for now.

```typescript
export function query(selector: string, root?: HTMLElement): Element {
  function __query(
    selector: string,
    match: Element,
    element?: Element
  ): Element {
    return match;
  }
  return __query(selector, root);
}
```

The code we'll write goes inside of the recursive `__query function`. At the start of this `function` check if you can just use `querySelector` to find the element and set the value of `match` with what is returned by `querySelector`. If `element` is not `undefined` here, we'll want to skip to the next line. On the next line, use a conditional to check if `match` is not `undefined` and return `match`. This will short circuit the recursive `function` if an element is found.

```typescript
if (!element && document.querySelector(selector)) {
  match = document.querySelector(selector);
}

if (match) {
  return match;
}
```

If `match` is `undefined`, we want to instantiate a new instance of `TreeWalker` by setting a new `const` named `walk` and passing all the arguments required by `document.createTreeWalker`. The first argument is the root node to start walking, which should be the `document.body` if `element` is `undefined`. The next argument optional for `createTreeWalker` is a way to filter types of nodes. We'll use the `NodeFilter.SHOW_ELEMENT` which filters only `Element` nodes and discards other nodes like `Comment` or `DocumentFragment`. The next argument is a filter for the nodes that pass the first filter, it tells when `TreeWalker` should accept a node. We need `FILTER_ACCEPT` here. Set the fourth argument to `false`, a boolean that indicates if an entire sub-tree should be discarded when a reference is also discarded. For more information about `TreeWalker`, [view the documentation on MDN](https://developer.mozilla.org/en-US/docs/Web/API/Document/createTreeWalker).

```typescript
const walk = document.createTreeWalker(
  element ? element : (document.body as Element),
  NodeFilter.SHOW_ELEMENT,
  {
    acceptNode(node) {
      return NodeFilter.FILTER_ACCEPT;
    },
  },
  false
);
```

The above code will cause a TypeScript error `Expected 1-3 arguments, but got 4.`. This is due to the TypeScript DOM library marking the fourth argument as deprecated. W3C [documents the fourth argument](https://www.w3.org/TR/DOM-Level-2-Traversal-Range/traversal.html#Traversal-TreeWalker) and browsers implement the argument. This argument, also known as `expandEntityReferences`, should be set to `false` because we're only concerned with the current node as the algorithm walks the tree, not the entire sub-tree. To discard the TypeScript error, place a `@ts-ignore` comment in the line above the fourth argument.