weh stands for WebExtensions Helper.
This toolkit speeds up browser add-ons development by providing a number of facilities for WebExtensions-based (Firefox, Chrome, Opera and Edge) extensions.
This is not a framework in the sense that the developer does not have to embrace all the provided utilities and there is not many architectural constraints to follow in order to take benefit of the tool.
The build system generates automatically a directory you can directly install into your browser, compiling automatically CoffeeScript, TypeScript and JSX to Javascript, Sass, Less and Stylus to CSS, optionally concatenating and minifying scripts, styles and HTML.
weh also provides some libraries that goes into your addon to ease a number of common tasks like managing preferences and two-way communications between the extension background and its user interface content pages, providing a way for the end-user to customize any string in the add-on user interface. Developing the user interface using ReactJS is also simplified but you may choose not to use this library.
In addition, an inspector application (under the form of a weh-based extension) is provided to monitor other weh extensions in real-time.
weh-generated extensions are compatible with Firefox, Chrome, Opera and Edge. You should of course maintain this compatibility in the code you add to your project.
npm install -g weh gulp
weh init --prjdir myextension
You can now install your skeleton extension from the myextension/build
directory as described
here.
npm install -g gulp
git clone https://github.com/mi-g/weh.git
cd weh
npm install
npm link
You can now move away from the weh directory.
To create a new extension project:
weh init --prjdir myextension
You now have a myextension
folder. The myextension/src
sub-directory is the place where your add-on specific
code goes. After running weh init
, the directory contains a simple skeleton code that demonstrates preferences edition. This code is to be modified
to do what your extension is supposed to do.
The myextension/build
contain an add-on ready to be installed into your browser.
To build and maintain the add-on:
cd myextension
weh
You will notice that the last weh
command does not return. It is in watch mode, meaning whenever you make a change into the myextension/src
directory, those changes are rebuild into myextension/build
. If you do not want this behaviour and prefer running the build command manually,
add --no-watch
to the command line.
Run weh help
to see more command line options.
- on Firefox: visit
about:debugging
, click Load Temporary Addon, select themyextension/build/manifest.json
file - on Chrome: visit
chrome://extension
, check Developer mode, click Load unpacked extension, select themyextension/build
directory - on Opera: visit
about:extension
, click Developer mode, Load unpacked extension, selectmyextension/build
directory - on Edge: (tested with insider Edge version 39.14959) click the 3 dots icon at the top right, select Extensions, click Load extension, select
myextension/build
directory
weh expects all project-specific code to be put into the src
sub-directory:
src/manifest.json
: your add-on's manifestsrc/**/*.html
: those files are processed, so resources like js and css (and other supported languages) are learned and processed to the build directory.src/locales
: files are copied tobuild/_locales
src/manifest.json
: file is copied tobuild
src/**/_assets/
: files and directories are processed (for language compilation) and copied to the build directory (without the_assets
part)etc/jsbanner.txt
: file that you can optionnally create to setup a header in the JS files.
Note that a .js
or css
file that would be located in src/
but not referenced from manifest.json
nor any .html
file won't be copied to the build directory. If you want this file in the build, you must put it in a _assets
sub-directory of src/
.
Also note that you can change the src
directory by specifying a directory path with the --srcdir
option.
You don't need to do anything special to make weh background libraries available to your add-on. Just declare your own modules in manifest.json
:
"background": {
"scripts": [
"background/main.js"
]
}
When the add-on is built, weh background modules will be added automatically. If the --prod
option is provided, weh and custom files
will be concatenated and minified.
Content scripts and styles included in manifest.json
are processed the same way:
"content_scripts": [{
"js": [ "script1.js", "script2.js" ],
"css": [ "style1.css", "style2.css" ]
}]
For content files, you must load a number of scripts and styles into your HTML files. For instance:
<!DOCTYPE html>
<html>
<head>
<!-- build:css settings.css -->
<link href="vendor/bootstrap.css" type="text/css" rel="stylesheet">
<link href="myextension-ui.css" type="text/css" rel="stylesheet">
<!-- endbuild -->
</head>
<body class="container-fluid">
<div id="root"></div>
<!-- build:js vendor-bundle.js -->
<!-- weh:js weh-all -->
<!-- endbuild -->
<!-- build:js myextension-ui.min.js -->
<script src="myextension-ui.js"></script>
<!-- endbuild -->
</body>
</html>
When the extension is built, <!-- weh:js weh-all -->
is first replaced by weh scripts inclusion. Then, the sections between
<!-- build:type file -->
and <!-- endbuild -->
are processed to be concatenated into a single file (per section) and minified
if you specify the --prod
for production mode.
Weh obviously supports Javascript (.js
file extension) for scripts and Cascading Style Sheets (.css
extension),
but you can also use other languages:
- scripts: JSX (
.jsx
), Typescript (.ts
), Coffee (.coffee
) - styling: Sass (
.scss
), Less (.less
), Stylus (.styl
)
Whether in manifest.json
or .html
files, just use the file as if it was Javascript or CSS:
<!-- build:css settings.css -->
<link href="my-styles.scss" type="text/css" rel="stylesheet">
<link href="my-styles2.less" type="text/css" rel="stylesheet">
<!-- endbuild -->
...
<!-- build:js myextension-ui.min.js -->
<script src="myextension-ui.jsx"></script>
<script src="myextension-ui2.ts"></script>
<!-- endbuild -->
Weh will then take care of calling the appropriate processors and renaming the files inside manifest.json
and
.html
files, so that the browser will be able to run this code.
All files with a .ejs
are processed first by an EJS processor. For instance, a file named myscript.js.ejs
will
be transformed to myscript.js
before being processed. You can specify one or several JSON files to provide data
for the EJS resolution using the --ejsdata
option.
The EJS pre-processing occurs in a first place, so a file named myscript.ts.ejs
will first be EJS-processed, then
compiled using Typescript, and will endup in the build directory as myscript.js
.
Any text file in the src
directory can be processed with EJS, not only js and css-like.
Pre-processing is useful if you want to generate different builds from the same source code.
Preferences are to be formally defined in order to be used in your add-on. An example of preferences description could be:
weh.prefs.declare([{
name: "myparam_string",
type: "string",
defaultValue: "Default value",
maxLength: 15,
regexp: "^[a-zA-Z ]+$"
},{
name: "myparam_integer",
type: "integer",
defaultValue: 42,
minimum: -10,
maximum: 100
},{
name: "myparam_float",
type: "float",
defaultValue: 3.14159,
minimum: 1.5,
maximum: 10.8
},{
name: "myparam_boolean",
type: "boolean",
defaultValue: true
},{
name: "myparam_choice",
type: "choice",
defaultValue: "second",
choices: [{
name: "First choice",
value: "first"
},{
name: "Second choice",
value: "second"
},{
name: "Third choice",
value: "third"
}]
}]);
For each parameter, you must provide at least name
, type
and defaultValue
. type
must be one of string
, integer
, float
, boolean
or
choice
. A specific preference parameter can then be accessed, as read or write, through weh.prefs["parameter name"]
.
You can install preferences listeners using weh.prefs.on(whatToWatch,callback)
and uninstall listeners using weh.prefs.off
with the same parameters. whatToWatch
uses a dotted notation. For instance, listening to ""
, "a"
, "a.b"
or "a.b.c"
will trigger the callback whenever
parameter a.b.c
is modified. Note that the preferences listeners are available from both background and local content.
You should also define a couple of human viewable strings associated to each parameter in locales/<locale>/messages.json
:
weh_prefs_label_<parameter name>
defines a label for the parameterweh_prefs_description_<parameter name>
defines an optional longer description for this parameter
Example (locales/en_US/messages.json
):
"weh_prefs_label_myparam_string": {
"message": "String parameter"
},
"weh_prefs_description_myparam_string": {
"message": "Only letters and spaces, 20 characters max"
},
You can define a number of constraints to your preferences. This is useful with the settings user interface provided by weh.
maxLength
: (typestring
,integer
andfloat
) the number of characters in the inputregexp
: (typestring
) a regular expression the string must matchminimum
: (typeinteger
andfloat
) the minimum acceptable valuemaximum
: (typeinteger
andfloat
) the maximum acceptable valuechoices
: (typechoice
) the set of possible choices to appear in a select input. This is array of either:- object containing fields
value
(the actual preference value) andname
(what is to be displayed to the user) - string representing the actual preference value. The label to be displayed for this choice is searched in
locales/<locale>/messages.json
asweh_prefs_label_<parameter name>_option_<parameter value>
- object containing fields
Note that the preferences definition can be declared or updated at any time. This is useful if, for instance, you don't the list of choices in advance.
From the content side, you should create an HTML file in your source code to represent the settings page:
<!DOCTYPE html>
<html>
<head>
<!-- build:css settings.css -->
<link href="bootstrap.css" type="text/css" rel="stylesheet">
<link href="styles.css" type="text/css" rel="stylesheet">
<!-- endbuild -->
</head>
<body class="container-fluid">
<div id="root"></div>
<!-- build:js vendor-bundle.js -->
<!-- weh:js weh-all -->
<!-- endbuild -->
<!-- build:js settings-bundle.js -->
<script src="settings.jsx"></script>
<!-- endbuild -->
</body>
</html>
And a settings JSX file like this:
function Prefs() {
return (
<WehParams>
<WehVersion/>
<WehParamSet wehPrefs={["myparam_string","myparam_integer","myparam_float","myparam_boolean","myparam_choice"]}>
<WehParam/>
</WehParamSet>
</WehParams>
)
}
ReactDOM.render (
<div>
<h1 className="text-center">{weh._("settings")}</h1>
<br/>
<Prefs/>
</div>,
document.getElementById('root')
)
weh will automatically create a dynamic form to allow modifying settings based on the preferences definition.
weh simplifies communications between background and content by handling automatically the message ports.
Let's see an example (src/background/main.js
):
weh.ui.update("default",{
onMessage: function(message) {
switch(message.type) {
case "open-settings":
weh.ui.close("default");
weh.ui.open("settings");
break;
}
}
});
weh.ui.update("settings",{
type: "tab",
contentURL: "content/settings.html",
});
Here we defined 2 user interface pages. The first one has a special name default
meaning this is the popup declared in manifest.json
. When
this popup sends a message with the type
field equal to open-settings
, a tab is opened to URL content/settings.html
. If a tab is already
open to this URL, it is activated instead.
In the content popup code, when the Setting button/link is pressed, we just do:
weh.post({
type: "open-settings"
});
To send asynchronous messages to the content pages, just do:
weh.ui.post(panelName,message)
Where panelName
is the name of a page your previously defined with weh.ui.update
and message
is any JSONizable javascript object.
You can also call weh.ui.post(message)
to send the same message to all open pages.
When defining a content page, you can, in addition to catching messages with onMessage
, define callbacks for page opening with onShow
or closing
with onHide
.
The callback parameters are:
onMessage(message,postFn,panelName)
withmessage
is the object sent from the content,postFn
a function that can be used to send back messages directly andpanelName
the name of the panel (as defined inweh.ui.update
) that sent the message. So doingpostFn({...})
is equivalent toweh.ui.post(panelName,{...})
.
From the content side, you can listen for incoming background messages with weh.on(callback)
and weh.off(callback)
to remove the listerner.
If you go for ReactJS for building your user interface, you can defines components like this:
class MyComponent extends React.Component {
constructor(props) {
super(props);
weh.react.attach(this,this.onWehMessage);
}
onWehMessage(message) {
/* do something wxith the incoming message */
}
}
weh takes care of adding/removing the listener when the component is mounted/unmounted and delivering the message to the onWehMessage
method.
The preferences and localization API are available for background and local content. Some weh
API methods are only available from the local content:
post(message)
: post a message to the background.message
is any JSONizable object.postLocal(message)
: emulate a message received from the background.message
is any object.on([type,]callback)
: install a handler to receive background messages. Iftype
is used, the callback will be invoked only if the received message contains a fieldtype
with the same value.off([type,]callback)
: uninstall the callback from a previous call toon(...)
.copyToClipboard(data,mimeType)
: copy data to the clipboard.mimeType
default isplain/text
.setPageTitle(newTitle)
: update or create the<title>
element in the HTML<head>
section.
The weh toolkit includes an extension called weh-inspector which allows to:
- monitor messages between the background and UI
- read/write addon preferences
- read add-on storage
As of now, this tool is only available on Chrome and Opera, as Firefox currently lacks some APIs for the tool to operate.
The weh-inspector is available as a template in the weh toolkit. As such, you can install it with weh init --template weh-inspector --prjdir weh-inspector
and then load the generated extension into the browser like any regular weh addon.
weh provides some utilities for dealing with locales.
Instead of browser.i18n.getMessage()
, you should use weh._()
, with the same parameters:
- it's shorter
- it automatically turns character
'-'
into'_'
in string tags while leaving a warning in the console - more important: it allows overwriting some or all locale strings. Whenever a call is made to
weh._()
, the library first searches for a storage-based translation for this tag. If not found, it uses the default string defined in_locales/<locale>/messages.json
. By default, weh provides a user interface page for the user to edit locale strings. It is up to the add-on developer to write the code to centralize the user-generated translations on a server, so that it can be shared amongst all users.