devtools.inspectedWindow.eval()
Executes JavaScript in the window that the devtools are attached to.
This is somewhat like using tabs.executeScript()
to attach a content script, but with two main differences:
First, the JavaScript can use a set of special commands that browsers typically provide in their devtools console implementation: for example, using "$0" to refer to the element currently selected in the Inspector.
Second, the JavaScript you execute can see any changes made to the page by scripts that the page loaded. This is in contrast to content scripts, which see the page as it would exist if no page scripts were loaded. However, note that the isolation provided by content scripts is a deliberate security feature, intended to make it harder for malicious or uncooperative web pages to confuse or subvert WebExtensions APIs by redefining DOM functions and properties. This means you need to be very careful if you waive this protection by using eval()
, and should use content scripts unless you need to use eval()
.
The script is evaluated by default in the main frame of the page. The script must evaluate to a value that can be represented as JSON (meaning that, for example, it may not evaluate to a function or an object that contains any functions). By default, the script does not see any content scripts attached to the page.
You can't call eval()
on privileged browser windows such as "about:addons".
You can optionally provide an options
parameter, which includes options to evaluate the script in a different frame or in the context of attached content scripts. Note that Firefox does not yet support the options
parameter.
The eval()
function returns a Promise
that resolves to the evaluated result of the script or to an error.
Helpers
The script gets access to a number of objects that help the injected script interact with the developer tools. The following helpers are currently supported:
Syntax
let evaluating = browser.devtools.inspectedWindow.eval(
expression, // string
options // object
)
Parameters
expression
-
string
. The JavaScript expression to evaluate. The string must evaluate to an object that can be represented as JSON, or an exception will be thrown. For example,expression
must not evaluate to a function. options
Optional-
object
. Options for the function (Note that Firefox does not yet support this options), as follows:frameURL
Optional-
string
. The URL of the frame in which to evaluate the expression. If this is omitted, the expression is evaluated in the main frame of the window. useContentScriptContext
Optional-
boolean
. Iftrue
, evaluate the expression in the context of any content scripts that this extension has attached to the page. If you set this option, then you must have actually attached some content scripts to the page, or a DevTools error will be thrown. contextSecurityOrigin
Optional-
string
. Evaluate the expression in the context of a content script attached by a different extension, whose origin matches the value given here. This overridesuseContentScriptContext
.
Return value
A Promise
that will be fulfilled with an array
containing two elements.
If no error occurred, element 0 will contain the result of evaluating the expression, and element 1 will be undefined
.
If an error occurred, element 0 will be undefined
, and element 1 will contain an object giving details about the error. Two different sorts of errors are distinguished:
-
errors encountered evaluating the JavaScript (for example, syntax errors in the expression). In this case, element 1 will contain:
- a boolean property
isException
, set totrue
- a string property
value
, giving more details.
- a boolean property
-
other errors (for example, an expression that evaluates to an object that can't be represented as JSON). In this case, element 1 will contain:
- a boolean property
isError
, set totrue
- a string property
code
containing an error code.
- a boolean property
Browser compatibility
BCD tables only load in the browser
Examples
This tests whether jQuery is defined in the inspected window, and logs the result. Note that this wouldn't work in a content script, because even if jQuery were defined, the content script would not see it.
function handleError(error) {
if (error.isError) {
console.log(`DevTools error: ${error.code}`);
} else {
console.log(`JavaScript error: ${error.value}`);
}
}
function handleResult(result) {
console.log(result);
if (result[0] !== undefined) {
console.log(`jQuery: ${result[0]}`);
} else if (result[1]) {
handleError(result[1]);
}
}
const checkJQuery = "typeof jQuery !== 'undefined'";
evalButton.addEventListener("click", () => {
browser.devtools.inspectedWindow.eval(checkJQuery).then(handleResult);
});
Helper examples
This uses the $0
helper to set the background color of the element that's currently selected in the Inspector:
const evalButton = document.querySelector("#reddinate");
const evalString = "$0.style.backgroundColor = 'red'";
function handleError(error) {
if (error.isError) {
console.log(`DevTools error: ${error.code}`);
} else {
console.log(`JavaScript error: ${error.value}`);
}
}
function handleResult(result) {
if (result[1]) {
handleError(result[1]);
}
}
evalButton.addEventListener("click", () => {
browser.devtools.inspectedWindow.eval(evalString).then(handleResult);
});
This uses the inspect()
helper to select the first <h1> element in the page:
const inspectButton = document.querySelector("#inspect");
const inspectString = "inspect(document.querySelector('h1'))";
function handleError(error) {
if (error.isError) {
console.log(`DevTools error: ${error.code}`);
} else {
console.log(`JavaScript error: ${error.value}`);
}
}
function handleResult(result) {
if (result[1]) {
handleError(result[1]);
}
}
inspectButton.addEventListener("click", () => {
browser.devtools.inspectedWindow.eval(inspectString).then(handleResult);
});
Note:
This API is based on Chromium's chrome.devtools
API.