Puppeteer (@hint/connector-puppeteer)

A connector that uses puppeteer to communicate with the browsers in webhint.

Installation

This package is installed automatically when adding webhint to your project so running the following is enough:

1npm install hint --save-dev

To use it, activate it via the .hintrc configuration file:

1{
2    "connector": {
3        "name": "puppeteer"
4    },
5    ...
6}

Options

The set of settings supported by the Puppeteer connector are:

1{
2    "connector": {
3        "name": "puppeteer",
4        "options": {
5            "auth": AuthObject,
6            "browser": "Chrome|Chromium|Edge",
7            "headless": true|false,
8            "ignoreHTTPSErrors": true|false,
9            "puppeteerOptions": "object",
10            "waitUntil": "dom|loaded|networkidle0|networkidle2"
11        }
12    },
13    ...
14}

All properties of options are optional.

• auth: The credentials and elements to authenticate on a website. See next section for further details.
• browser (Chrome|Chromium|Edge): Tells the preferred browser to use. If unspecified webhint will look for a puppeteer installation before falling back to searching for an installed browser and fail if it does not find one. Keep in mind that not all browsers are available on all platforms and that you need to manually install either puppeteer or a browser for this connector to work.
• headless (boolean): Indicates if the browser should run in headless mode or not. It is true by default when running on CI or in WSL, false otherwise.
• ignoreHTTPSError (boolean): Indicates if errors with certificates should be ignored. Use this when checking self-signed certificates. It is false by default.
• puppeteerOptions (object): A set of launch options to pass to puppeteer. See the puppeteer launch options for more information.
• waitUntil (dom|loaded|networkidle0|networkidle2): Is the waiting strategy to decide when a page is considered loaded. See the puppeteer goto options to know more.

WSL support

To use this connector when running WSL you will have to install a chromium browser on your distro (e.g.: sudo apt-get install chromium-browser). Because by default WSL does not support graphics, the headless mode will be enabled by default. If you have an X Server working you will have to manually disable this option via the connector's options. E.g.:

1{
2    "connector": {
3        "name": "puppeteer",
4        "options": {
5            "headless": false
6        }
7    },
8    ...
9}

Website authentication

The puppeteer connector allows to authenticate on a website that supports Basic HTTP Authentication or:

• uses user/password (i.e.: no MFA or captcha).
• redirects to the login page and to the initial target after successful authentication.

For Basic Authentication the auth object properties are:

• user: a string with the user name to use
• password: a string with the password to use

E.g.:

1{
2    "user": "userName",
3    "password": "Passw0rd"
4}

Otherwise, auth properties are:

• user: the information needed to identify the input element via a query selector (e.g.: #login) to type the value for the username in (e.g.: username1).
• password: the information needed to identify the input element via a query selector (e.g.: #password) to type the value for the password in (e.g.: P@ssw0rd).
• next: the information needed to identify the input (or button) element via a query selector (e.g.: input[type="submit"]) to click to get to the next step of the authentication process. This is an optional property as not all services prompt first for the user name before asking for the password in the following screen. An example of such a service would be Azure Pipelines.
• submit: the information needed to identify the input (or button) element via a query selector (e.g.: input[type="submit"]) to click to submit the crendentials.

E.g.:

1{
2    "user": {
3        "selector": "string",
4        "value": "string"
5    },
6    "password": {
7        "selector": "string",
8        "value": "string"
9    },
10    "next": {
11        "selector": "string"
12    },
13    "submit": {
14        "selector": "string"
15    }
16}

User actions

Sometimes you might need the browser to interact in some way with the content before starting the analysis. For example, in the case of a SPA you might want to click in certain elements to get to the right state.

Sometimes, this actions need to be done before navigating to the page to analyze.

To achieve this, you can use "user actions". "User actions" are defined as follows:

1{
2    "connector": {
3        "name": "puppeteer",
4        "options": {
5            "actions": [
6                {
7                    "file": "pathToUserAction1.js",
8                    "on": "beforeTargetNavigation|afterTargetNavigation"
9                },
10                {
11                    "file": "pathToUserAction2.js",
12                    "on": "beforeTargetNavigation|afterTargetNavigation"
13                },
14                ...
15            ],
16            "actionsOptions": { },
17            ...
18        }
19    },
20    ...
21}

There's a property actions in the connector configuration that's an array of Action. You can define as many actions as you want.

An Action is an object with two properties:

• file: Absolute or relative path from the execution path to the file containing the action to execute.
• on: A string that indicates when the action needs to be executed:
  • beforeTargetNavigation: The action will be executed before navigating to the target. If you need to set up special headers you will have to do it at this moment.
  • afterTargetNavigation: The action will be executed after the target has been loaded. If the website is a SPA and you need to get to a certain state, this is the moment to use.

The file that contains the action needs to be written in JavaScript and export an object with an action property with the following signature:

1module.exports = {
2    action: async (page, options) => {
3        // your actions here
4    }
5};

The parameters the function receives are:

• page: The puppeteer Page with the tab used to navigate to the target. This gives you full control to do anything you need with the page (click, type, navigate elsewhere, etc.).
• options: The connector options. This allows you access to waitFor values and any other user configuration. If you need to pass anything specifically to the actions you can use options.actionOptions property to do so.

User action examples

The connector's authentication mechanisms rely on the user actions API. The following is the code for the Basic HTTP Auth (transpiled to JS):

1module.exports = {
2    action: async (page, config) => {
3        if (!config || !config.auth) {
4            return;
5        }
6
7        if (typeof config.auth.user !== 'string' || typeof config.auth.password !== 'string') {
8            return;
9        }
10
11        await page.authenticate({
12            password: config.auth.password,
13            username: config.auth.user
14        });
15    }
16};

Note: This user action uses options.auth which is already predefined. If your user action needs another type of user information you can use options.actionsOptions.

The following is an example of a user action that will click on an element configured via options.actionsOptions:

1{
2    "connector": {
3        "name": "puppeteer",
4        "options": {
5            "actions": [
6                {
7                    "file": "clickElement.js",
8                    "on": "afterTargetNavigation"
9                }
10            ],
11            "actionsOptions": {
12                "elementId": "#id"
13            }
14        }
15    },
16    ...
17}

1module.exports = {
2    action: async (page, config) => {
3        const selector = config.actionsOptions.elementId;
4
5        await page.click(selector);
6    }
7};

Please look at the source code of connector-puppeteer for other built-in actions.

Further Reading

• Connectors
• Puppeteer documentation