PaymentRequest.show() - Web APIs 编辑

Secure context

This feature is available only in secure contexts (HTTPS), in some or all supporting browsers.

The PaymentRequest interface's show() method instructs the user agent to begin the process of showing and handling the user interface for the payment request to the user.

For security reasons, the PaymentRequest.show() method can't just be initiated at any time. It may only be called while handling events that represent user interactions, such as click, keyup, or the like.

Only one payment request can be in the process of being handled at once, across all documents. Once one PaymentRequest's show() method has been called, any other call to show() will by rejected with an AbortError until the returned promise has been concluded, either by being fulfilled with a PaymentResponse indicating the results of the payment request, or by being rejected with an error.

Note: In reality, despite the fact that the specification says this can't be done, some browsers, including Firefox, support multiple active payment requests at a time.

If your architecture doesn't necessarily have all of the data ready to go at the moment it instantiates the payment interface by calling show(), specify the detailsPromise parameter, providing a Promise that is fulfilled once the data is ready. If this is provided, show() will not allow the user to interact with the payment interface until the promise is fulfilled, so that data can be updated prior to the user engaging with the payment process.

Processing the result and, if necessary, calling PaymentResponse.retry() to retry a failed payment can all be done either asynchronously or synchronously, depending on your needs. For the best user experience, asynchronous solutions are typically the best way to go. Most examples on MDN and elsewhere use async/await to wait asynchronously while results are validated and so forth.

Syntax

paymentPromise = paymentRequest.show(detailsPromise);

Parameters

detailsPromise Optional

An optional Promise that you can provide if your architecture requires that the payment request's details need to be updated between instantiating the payment interface and the user beginning to interact with it. The promise should resolve with a PaymentDetailsUpdate object containing the updated information.

Return value

A Promise that eventually resolves with a PaymentResponse. The promise is resolved when the user accepts the payment request (such as by clicking a "Pay" button in the browser's payment sheet).

Exceptions

AbortError

The returned promise rejects with an AbortError if the user agent is already showing a payment panel. Only one payment panel may be visible at a time across all documents loaded by the user agent.

The promise is also rejected with AbortError if the user cancels the payment request.

InvalidStateError

The promise rejects with an InvalidStateError if the same payment has already been shown for this request (its state is interactive because it is being shown already).

NotSupportedError

The promise rejects with a NotSupportedError if the user agent does not support the payment methods specified when the PaymentRequest constructor was called.

SecurityError

The promise rejects with a SecurityError if the call to show() was not in response to a user action, such as a click or keyup event. Other reasons a SecurityError may be thrown are at the discretion of the user agent, and may include situations such as too many calls to show() being made in a short time or show() being called while payment requests are blocked by parental controls.

Usage notes

The most common patterns for using show() involve either the async/await syntax or the use of show().then().catch() to handle the response and any possible rejection. Those look like this:

async/await syntax

Using await to wait for a promise to be resolved makes it possible to write the code to handle payments particularly cleanly:

async function processPayment() {
  try {
    const payRequest = new PaymentRequest(methodData, details, options);

    payRequest.onshippingaddresschange = ev => ev.updateWith(checkAddress(payRequest));
    payRequest.onshippingoptionchange = ev => ev.updateWith(checkShipping(payRequest));

    const response = await payRequest.show();
    await validateResponse(response);
  } catch(err) {
    /* handle the error; AbortError usually means a user cancellation */
  }
}

In this code, the methods checkAddress() and checkShipping(), respectively, check the shipping address and the shipping option changes and supply in response either a PaymentDetailsUpdate object or a promise to return one; this object contains the fields in the PaymentResponse which have been or need to be changed.

The validateResponse() method, below, is called once show() returns, in order to look at the returned response and either submit the payment or reject the payment as failed:

async function validateResponse(response) {
  try {
    if (await checkAllValues(response)) {
      await response.complete("success");
    } else {
      await response.complete("fail");
    }
  } catch(err) {
    await response.complete("fail");
  }
}

Here, a custom function called checkAllValues() looks at each value in the response and ensures that they're valid, returning true if every field is valid or false if any are not. If and only if every field is valid, the complete() method is called on the response with the string "success", which indicates that everything is valid and that the payment can complete accordingly.

If any fields have unacceptable values, or if an exception is thrown by the previous code, complete() is called with the string "fail", which indicates that the payment process is complete and failed.

Instead of immediately failing, you could choose to call retry() on the response object to ask the user agent to try to process the payment again; this should usually only be done after the user has made any needed corrections to the order.

Starting the payment process, in the end, is as simple as calling the processPayment() method.

then/catch syntax

You can also use the older promise-based approach to work with payments, using the then() and catch() functions on the promise returned by show():

function processsPayment() {
  const payRequest = new PaymentRequest(methodData, details, options);

  payRequest.onshippingaddresschange = ev => ev.updateWith(checkAddress(payRequest));
  payRequest.onshippingoptionchange = ev => ev.updateWith(checkShipping(payRequest));

  payRequest.show()
    .then(response => validateResponse(response))
    .catch(err => handleError(err));
}

This is functionally equivalent to the processPayment() method using the await syntax.

function validateResponse(response) {
  checkAllValues(response)
    .then(response => response.complete("success"))
    .catch(response => response.complete("fail"));
}

You could even have checkAllValues() be a synchronous function, although that may have performance implications you don't want to deal with:

function validateResponse(response) {
  if (checkAllValues(response) {
    response.complete("success");
  } else {
    response.complete("fail");
  }
}

See the article Using promises for more information if you need more information about working with promises.

Examples

In the following example, a PaymentRequest object is instantiated before the show() method is called. This method triggers the user agent's built-in process for retrieving payment information from the user. The show() method returns a Promise that resolves to a PaymentResponse object when the user interaction is complete. The developer then uses the information in the PaymentResponse object to format and send payment data to the server. You should send the payment information to the server asynchronously so that the final call to paymentResponse.complete() can indicate the success or failure of the payment.

button.onclick = async function handlePurchase() {
  // Initialization of PaymentRequest arguments are excerpted for the sake of
  // brevity.
  const payment = new PaymentRequest(methods, details, options);
  try {
    const response = await payment.show();
    // Process response here, including sending payment instrument
    // (e.g., credit card) information to the server.
    // paymentResponse.methodName contains the selected payment method
    // paymentResponse.details contains a payment method specific response
    await response.complete("success");
  } catch (err) {
    console.error("Uh oh, something bad happened", err.message);
  }
}

The following example shows how to update the payment sheet as it's being presented to the end-user.

async function requestPayment() {
  // We start with AU$0 as the total.
  const initialDetails = {
    total: {
      label: "Total",
      amount: { value: "0", currency: "AUD" },
    },
  };
  const request = new PaymentRequest(methods, initialDetails, options);
  // Check if the user supports the `methods`
  if (!await request.canMakePayment()) {
    return; // no, so use a web form instead.
  }
  // Let's update the total as the sheet is shown
  const updatedDetails = {
    total: {
      label: "Total",
      amount: { value: "20", currency: "AUD" },
    },
  };
  const response = await request.show(updatedDetails);
  // Check response, etc...
}

document.getElementById("buyButton").onclick = requestPayment;

Specifications

Browser compatibility

BCD tables only load in the browser

See also

• Payment Request API
• Using the Payment Request API
• PaymentRequest.abort()
• PaymentRequest.retry()
• PaymentRequest.complete()
• PaymentResponse