Chrome incompatibilities
Extensions built with WebExtension APIs are designed to be compatible with Chrome and Opera extensions. As far as possible, extensions written for those browsers should run on Firefox with minimal changes.
However, there are significant differences between Chrome, Firefox, and Edge. In particular:
- Support for JavaScript APIs differs across browsers. See Browser support for JavaScript APIs for more details.
- Support for
manifest.json
keys differs across browsers. See the "Browser compatibility" section in themanifest.json
page for more details. - JavaScript APIs:
- In Firefox: JavaScript APIs are accessed under the
browser
namespace. - In Chrome and Edge: JavaScript APIs are accessed under the
chrome
namespace. (cf. Chrome bug 798169)
- In Firefox: JavaScript APIs are accessed under the
- Asynchronous APIs:
- In Firefox: Asynchronous APIs are implemented using promises.
- In Chrome and Edge: Asynchronous APIs are implemented using callbacks. (cf. Chrome bug 328932)
The rest of this page summarizes these and other incompatibilities.
JavaScript APIs
chrome.* and browser.* namespace
- In Firefox: The equivalent APIs are accessed using the
browser
namespace.js
browser.browserAction.setIcon({path: "path/to/icon.png"});
- In Chrome: Extensions access privileged JavaScript APIs using the
chrome
namespace.js
chrome.browserAction.setIcon({path: "path/to/icon.png"});
Callbacks and promises
- In Firefox: Asynchronous APIs use promises to return values instead.
js
function logCookie(c) { console.log(c); } function logError(e) { console.error(e); } let setCookie = browser.cookies.set( {url: "https://developer.mozilla.org/"} ); setCookie.then(logCookie, logError);
- In Chrome: Asynchronous APIs use callbacks to return values, and
runtime.lastError
to communicate errors.js
function logCookie(c) { if (chrome.runtime.lastError) { console.error(chrome.runtime.lastError); } else { console.log(c); } } chrome.cookies.set( {url: "https://developer.mozilla.org/"}, logCookie );
Firefox supports both the chrome and browser namespaces
As a porting aid, the Firefox implementation of WebExtensions supports chrome
, using callbacks, as well as browser
, using promises. This means that many Chrome extensions will just work in Firefox without any changes.
Note: However, this is not part of the WebExtensions standard and may not be supported by all compliant browsers.
If you choose to write your extension to use browser
and promises, then Firefox also provides a polyfill that will enable it to run in Chrome: https://github.com/mozilla/webextension-polyfill.
Partially supported APIs
The page Browser support for JavaScript APIs includes compatibility tables for all APIs that have any support in Firefox. Where there are caveats around support for a given API item, this is indicated in these tables with an asterisk "*" and in the reference page for the API item, the caveats are explained.
These tables are generated from compatibility data stored as JSON files in GitHub.
The rest of this section describes compatibility issues that are not already captured in the tables.
Notifications API
For notifications.create()
, with type "basic"
:
- In Firefox:
iconUrl
is optional. - In Chrome:
iconUrl
is required.
When the user clicks on a notification:
- In Firefox: The notification is cleared immediately.
- In Chrome: This is not the case.
If you call notifications.create()
more than once in rapid succession:
- In Firefox: In Firefox, the notifications may not display at all. Waiting to make subsequent calls until within the
chrome.notifications.create()
callback function is not a sufficient delay to prevent this.
Proxy API
Firefox's proxy
API followed a completely different design from Chrome's Proxy API.
- In Firefox: An extension can register a PAC file.
- In Chrome: An extension can register a PAC file, but can also define explicit proxying rules.
Because this API is incompatible with Chrome's proxy
API, the Firefox proxy API is only available through the browser
namespace.
Tabs API
When using tabs.executeScript()
or tabs.insertCSS()
:
- In Firefox: Relative URLs passed are resolved relative to the current page URL.
- In Chrome: These URLs are resolved relative to the extension's base URL.
To work cross-browser, you can specify the path as an absolute URL, starting at the extension's root, like this:
/path/to/script.js
When querying tabs by URL tabs.query()
:
- In Firefox: Extensions must have the
"tabs"
permission. - In Chrome: Extensions do not need the
"tabs"
permission, but only tabs whose URLs match the extension's host permissions will be included in the results.
When calling tabs.remove()
:
- In Firefox: The
tabs.remove()
promise is fulfilled after thebeforeunload
event. - In Chrome: The callback does not wait for
beforeunload
.
WebRequest API
- In Firefox:
- Requests can be redirected only if their original URL uses the
http:
orhttps:
scheme. - The
activeTab
permission does not allow intercepting network requests in the current tab. (See bug 1617479) - Events are not fired for system requests (for example, extension upgrades or search bar suggestions).
- From Firefox 57 onwards: Firefox makes an exception for extensions that need to intercept
webRequest.onAuthRequired
for proxy authorization. See the documentation forwebRequest.onAuthRequired
.
- From Firefox 57 onwards: Firefox makes an exception for extensions that need to intercept
- If an extension wants to redirect a public (e.g., HTTPS) URL to an extension page, the extension's
manifest.json
file must contain aweb_accessible_resources
key with the URL of the extension page.-
Note: Any website may then link or redirect to that URL, and extensions should treat any input (POST data, for example) as if it came from an untrusted source, just as a normal web page should.
-
- Some of the
browser.webRequest.*
APIs allow returning Promises that resolveswebRequest.BlockingResponse
asynchronously.
- Requests can be redirected only if their original URL uses the
- In Chrome: Only
webRequest.onAuthRequired
supports asynchronouswebRequest.BlockingResponse
via supplying'asyncBlocking'
.
Windows API
- In Firefox:
onFocusChanged
of thewindows
API, will trigger multiple times for a given focus change.
Unsupported APIs
DeclarativeContent API
- In Firefox: Chrome's declarativeContent API has not yet been implemented. In addition, Firefox will not be supporting the
declarativeContent.RequestContentScript
API (which is rarely used, and is unavailable in stable releases of Chrome).
Miscellaneous incompatibilities
URLs in CSS
- In Firefox: URLs in injected CSS files are resolved relative to the CSS file itself.
- In Chrome: URLs in injected CSS files are resolved relative to the page they are injected into.
Support for dialogs in background pages
web_accessible_resources
- In Firefox: Resources are assigned a random UUID that changes for every instance of Firefox:
moz-extension://«random-UUID»/«path»
. This randomness can prevent you from doing a few things, such as add your specific extension's URL to another domain's CSP policy. - In Chrome: When a resource is listed in
web_accessible_resources
, it is accessible aschrome-extension://«your-extension-id»/«path»
. The extension ID is fixed for a given extension.
Manifest "key" property
- In Firefox: Since Firefox uses random UUIDs for
web_accessible_resources
, this property is unsupported. - In Chrome: When working with an unpacked extension, the manifest may include a
"key"
property to pin the extension ID across different machines. This is mainly useful when working withweb_accessible_resources
.
Content script HTTP(S) requests
- In Firefox: When a content script makes an HTTP(S) request, you must provide absolute URLs.
- In Chrome: When a content script makes a request (for example, using
fetch()
) to a relative URL (like/api
), it will be sent tohttps://example.com/api
.
Content script environment
- In Firefox: The global scope of the content script environment is not strictly equal to
window
(Firefox bug 1208775). More specifically, the global scope (globalThis
) is composed of standard JavaScript features as usual, pluswindow
as the prototype of the global scope. Most DOM APIs are inherit from the page throughwindow
, through Xray vision to shield the content script from modifications by the web page. Content scripts may encounter JavaScript objects from its own global scope or Xray-wrapped versions from the web page. - In Chrome: The global scope is
window
and the available DOM APIs are generally independent of the web page (other than sharing the underlying DOM). Content scripts cannot directly access JavaScript objects from the web page.
Executing code in web page from content script
- In Firefox:
eval
runs code in the context of the content script, andwindow.eval
runs code in the context of the page. See Usingeval
in content scripts. - In Chrome:
eval
andwindow.eval
always runs code in the context of the content script, not in the context of the page.
Sharing variables between content scripts
- In Firefox: You cannot share variables between content scripts by assigning them to
this.{variableName}
in one script and then attempting to access them usingwindow.{variableName}
in another. This is a limitation created by the sandbox environment in Firefox. This limitation may be removed, see Firefox bug 1208775.
Content script lifecycle during navigation
-
In Firefox: Content scripts remain injected in a web page after the user has navigated away, however, window object properties are destroyed. For example, if a content script sets
window.prop1 = "prop"
and the user then navigates away and returns to the pagewindow.prop1
is undefined. This issue is tracked in Firefox bug 1525400. To mimic the behavior of Chrome, listen for the pageshow and pagehide events. Then simulate the injection or destruction of the content script. - In Chrome: Content scripts are destroyed when the user navigates away from a web page. If the user then returns to the page through history, by clicking the back button, the content script is injected into the web page again.
"per-tab" zoom behavior
- In Firefox: The zoom level persists across page loads and navigation within the tab.
- In Chrome: Zoom changes are reset on navigation; navigating a tab will always load pages with their per-origin zoom factors.
manifest.json keys
The main manifest.json
page includes a table describing browser support for manifest.json
keys. Where there are caveats around support for a given key, this is indicated in the table with an asterisk "*" and in the reference page for the key, the caveats are explained.
These tables are generated from compatibility data stored as JSON files in GitHub.
Native messaging
Connection-based messaging arguments
On Linux and Mac: Chrome passes one argument to the native app, which is the origin of the extension that started it, in the form: chrome-extension://«extensionID/»
(trailing slash required). This enables the app to identify the extension.
On Windows: Chrome passes two arguments:
- The origin of the extension
- A handle to the Chrome native window that started the app
allowed_extensions
- In Firefox: The manifest key is called
allowed_extensions
. - In Chrome: The manifest key is called
allowed_origins
instead.
App manifest location
- In Chrome: The app manifest is expected in a different place. See Native messaging host location in the Chrome docs.
App persistance
- In Firefox: When a native messaging connection is closed, Firefox kills the subprocesses if they do not break away. On Windows, the browser puts the native application's process into a Job object and kills the job. Suppose the native application launches other processes and wants them to remain open after the native application is killed. In that case, the native application must use
CreateProcess
, instead ofShellExecute
, to launch the additional process with theCREATE_BREAKAWAY_FROM_JOB
flag.
Data cloning algorithm
Some extension APIs allow an extension to send data from one part of the extension to another, such as runtime.sendMessage()
, tabs.sendMessage()
, runtime.onMessage
, the postMessage()
method of runtime.port
, and tabs.executeScript()
.
- In Firefox: The Structured clone algorithm is used.
- In Chrome: The JSON serialization algorithm is used. It may switch to structured cloning in the future (issue 248548).
The Structured clone algorithm supports more types than the JSON serialization algorithm. A notable exception are (DOM) objects with a toJSON
method. DOM objects are not cloneable nor JSON-serializable by default, but with a toJSON()
method, these can be JSON-serialized (but still not cloned with the structured cloning algorithm). Examples of JSON-serializable objects that are not structured cloneable include instances of URL
and PerformanceEntry
.
Extension that rely on the toJSON()
method of the JSON serialization algorithm can use JSON.stringify()
followed by JSON.parse()
to ensure that a message can be exchanged, because a parsed JSON value is always structurally cloneable.