Welcome to Firefox Extension Workshop, a launchpad for building Firefox extensions! π
If you would like to update content or other resources on Firefox Extension Workshop, please refer to contributing.md
These instructions get a copy of the project up and running on your local machine for development and testing purposes.
For notes on how to deploy the project on a live system, see Deployment.
yarn install
To start local development, run:
yarn start
βΉοΈΒ NOTE: Running locally will show unpublished content that uses the published: false
convention in frontmatter. Content with published: false
will not be available on staging or production.
Command | Description |
---|---|
yarn start |
Starts eleventy and includes unpublished content. |
yarn build:production |
Builds the site for production. |
yarn build:unpublished |
Builds the site for production with unpublished content. |
yarn clean |
Clears the output directory. (You probably won't need to use this manually.) |
The site is built with Eleventy, a NodeJS-based static site generator.
The site works in slightly different ways depending on whether you're running the site for local development or building the site for production.
When you run yarn start
the CSS and JS is built in parallel with the eleventy build. Once up and running both eleventy and the JS and CSS build scripts watch for changes. When something changes the site is re-built.
In development Eleventy knows nothing about the CSS and JavaScript builds. For automatic reloading of the JS and CSS, each script uses a fetch to the public API to tell browserSync there is new code and it reloads it for you.
Building for production is slightly different. The Eleventy process and the JS and CSS builds happen in series. Then a 3rd asset-pipeline
process initiates and takes the the built content from ./build
directory and runs it through various optimizations.
During these optimizations, the following takes place:
- Binary files are versioned with hashes in the file names.
- References to these file in CSS and JS are updated.
- CSS and JS are minified.
- The HTML is processed to update the references to the assets new hash-based filenames.
All of this means that we can serve the site with far-future Expires
headers. If the resource is in the browser's cache, the browser won't even make a request for it. To break the cache, the resource's URL needs to change. When something is updated and the script is re-run, the hash in the filename will change, so the new filename won't be cached and the browser will know to fetch it. This helps the site be fast.
Whilst the asset-pipline
script is custom, it leverages a lot of existing libs where possible, these include Terser, postHTML, postCSS, and various plugins.
It's likely that some day, 11ty will have its own mechanism for wrangling assets. At that point, this will no longer be required.
For the asset-pipeline
script to do its thing, all you need to do is refer to all assets with a path beginning with /assets/
. If you do that, everything else is handled for you β¨
This site has three templates:
- A full-width page
- A sidebar "page" for documentation
- A Content Guidelines page
extensionworkshop.com
βββ bin
β βββ asset-pipeline # The asset build script
β βββ build-script # The JS build script
β βββ build-styles # The CSS build script
β
βββ build # Where eleventy builds the site to
β
βββ dist # Where production builds are built
β
βββ libs
β βββ markdown.js # The markdown renderer instance and plugins
β βββ slugify.js # The central slug function
β βββ templates.js # The liquidjs template instance
β
βββ screenshots # Screenshots used in README.md
β
βββ src
β βββ assets # Assets (CSS, JavaScript, fonts and images)
β βββ content # Content (Markdown and JS (generated))
β βββ data # Data files (JSON)
β βββ includes # Components (Liquid)
β βββ layouts # Layout templates
β
βββ tests # Test files run by jest `yarn test`.
β
βββ eleventy.config.js # Eleventy configuration
βββ .eleventyignore # Files ignored by Eleventy
βββ .gitignore # Files not tracked by Git
βββ .stylelintrc # Stylelint configuration
βββ .prettierrc # Prettier config
βββ .prettierignore # Files ignored by prettier
βββ .eslintrc # eslint config
βββ .eslintignore # Files ignored by eslint
βββ package.json # Node.js package manifest
βββ yarn.lock # Package manager lock file
βββ README.md # This file
- Add the image files to
src/assets/img/
- In your page, link to images using this page structure:
You can reference images with the full path from the assets/
directory (e.g, /assets/img/image.png
).
Here's an example in markdown
:
![Remembear subtitle screenshot](/assets/img/remembear-subtitle.png "Remembear subtitle text")
For a note, use the markdown syntax extensions as follows. (These markdown extensions are supplied by a plugin to the markdown renderer.)
::: note
This is a note
:::
For an alert, use the following:
::: note alert
This is an alert
:::
- Open
data/pages.json
. - Add a node with appropriate attributes, in the appropriate location, for the new page. See below: Understanding the
pages.json
structure. - Create a new page, nested inside a folder struture that matches the URL path. For example, for permalink
/documentation/develop/best-practices-for-collecting-user-data-consents/
, you would create a file calledbest-practices-for-collecting-user-data-consents.md
and place it indocumentation βΆοΈ develop
. - For reference on how to create a page, review the
sidebar-master-template.md
file, which lists all available modules. Some notes:published: false
will withhold this content from staging and production. To publish content, remove this line.skip_index: true
is used for pages that shouldn't be indexed for search results.- When creating page sections that should be listed in the table of contents, add an
id
attribute to the section container that matches thesubpageitems
entry added topages.json
. If your layout requires several sections for one table of contents entry, nest your sections inside a containing element which has theid
attribute. - Rule for creating section
id
s: use theh2
title of the section, converted to lowercase, spaces replaced with dashes, all non-alphanumeric characters removed. For example, the sectionh2
title "Know your privacy settings" would be converted toknow-your-privacy-settings
for the sectionid
. - The first section following the "Page Hero" module should be the "Table of Contents" module:
modules/column-w-toc.html
.
- Each page has a
title
andurl
attribute.
βΉοΈΒ NOTE: Theurl
attribute must exactly match thepermalink
attribute of the page's front matter (including leading and trailing slashes). - Pages may also have a
subpageitems
node for sections within the page to be referenced in the table of contents for that page:- Each
subpageitem
node has atitle
andid
attribute. The value ofid
matches theid
attribute of the section container.
(βΉοΈΒ NOTE:id
s must be added to the containing element, rather than the heading element, of the section. This ensures that highlighting for the section remains active, even when the section title is out of view.)
- Each
- Overview pages have
category
nodes for each of their contained (sub)categories
. - Categories have a
category
attibute (which denotes the category title), and apages
attribute (which lists sub-pages of the overview page). - The Documentation Topics section pages are nested inside a
subfolderitems
node, which creates the dropdown panel.
General overview of the pages.json
layout:
[
{
"title": "Documentation Topics",
"subfolderitems": [
{
"title": "Develop",
"url": "/documentation/develop/",
"subpageitems": [
{
"title": "Firefox Tools",
"id": "firefox-tools"
}
],
"categories": [
{
"category": "Getting Started",
"pages": [
{
"title": "Firefox Workflow Overview",
"url": "/documentation/develop/firefox-workflow-overview/"
}
]
}
]
},
{
"title": "Publish",
"url": "/documentation/publish/",
"subpageitems": [
{
"title": "Get your extension signed",
"id": "get-your-extension-signed"
}
]
},
{
"title": "Manage",
"url": "/documentation/manage/",
"subpageitems": [
{
"title": "Stay informed when Firefox changes",
"id": "stay-informed-when-firefox-changes"
}
]
},
{
"title": "Enterprise",
"url": "/documentation/enterprise/",
"subpageitems": [
{
"title": "Section Title",
"id": "introduction"
}
]
},
{
"title": "Themes",
"url": "/documentation/themes/",
"subpageitems": [
{
"title": "What themes are",
"id": "what-themes-are"
}
]
}
]
},
{
"title": "Extension Basics",
"url": "/extension-basics/",
"subpageitems": [
{
"title": "Getting started",
"id": "getting-started"
}
]
},
{
"title": "Community",
"url": "/community/",
"subpageitems": [
{
"title": "Who is part of the community?",
"id": "who-is-part-of-the-community"
}
],
"categories": [
{
"category": "About the Community",
"pages": [
{
"title": "",
"url": ""
}
]
}
]
}
]
- Create new file
- Add frontmatter (see example below)
- Copy 'modules' needed from
content-guidelines/master-template.md
and paste in new file - Save as markdown:
content-guidelines/page-name.md
---
layout: guides
title: Page Name
permalink: /content-guidelines/page-name/
published: false
---
βΉοΈΒ NOTE: published: false
will withhold this content from staging and production. To publish content, remove this line.
Go to data/content-guidelines-pages.json
and add a new entry for your page:
{
"title": "Page Name",
"url": "/content-guidelines/page-name/",
"draft-label": true
}
If you don't want the page to be labelled as a draft (such as and when it's ready), remove "draft-label": true
from the relevant entry in data/content-guidelines-pages.json
.
Tags should aim to follow the AMO calendar format: YYYY.MM.DD
with an optional
-x
suffix for cherry-picking.
The site is auto-deployed on commits to master
to https://extensionworkshop.allizom.org/. You can check the version with the stage version link.
Tags matching ^20\d{2}\.\d{2}\.\d{2}(?:-\d+)?$
regular expression will be deployed to https://extensionworkshop.com/. You can check the version on production with the production version link.
A good example tag for a production deploy would be 2022.03.03
.