Field "cart_id" on "Order" is not being resolved when using "setPaymentMethodAndPlaceOrder"

[pmzandbergen]

Description
The (deprecated) setPaymentMethodAndPlaceOrder mutation is broken: [object] (GraphQL\\Error\\Error(code: 0): Cannot return null for non-nullable field \"Order.cart_id\".

This is caused by:

adyen-magento2/etc/schema.graphqls

Line 57 in 4b98628

cart_id: String! @doc(description: "Cart ID.")

This field is not returned by \Magento\QuoteGraphQl\Model\Resolver\SetPaymentAndPlaceOrder.

Fix
You either need to add your own resolver (which could simply be using the input parameters) or use a plugin on the existing (Magento Core) resolver:

etc/graphql/di.xml

<?xml version="1.0"?>
<config xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
        xsi:noNamespaceSchemaLocation="urn:magento:framework:ObjectManager/etc/config.xsd">
    <type name="Magento\QuoteGraphQl\Model\Resolver\SetPaymentAndPlaceOrder">
        <plugin name="Adyen\Payment\Plugin\SetCartIdOnSetPaymentMethodAndPlaceOrderMutation"
                type="Adyen\Payment\Plugin\SetCartIdOnSetPaymentMethodAndPlaceOrderMutation"/>
    </type>
    <!-- ... -->
</config>

<?php

declare(strict_types=1);

namespace Adyen\Payment\Plugin;

use Magento\Framework\GraphQl\Config\Element\Field;
use Magento\Framework\GraphQl\Schema\Type\ResolveInfo;
use Magento\QuoteGraphQl\Model\Resolver\SetPaymentAndPlaceOrder;

class SetCartIdOnSetPaymentMethodAndPlaceOrderMutation
{
    public function afterResolve(
        SetPaymentAndPlaceOrder $subject,
        mixed $result,
        Field $field,
        mixed $context,
        ResolveInfo $info,
        array $value = null,
        array $args = null
    ) {
        return isset($args['input']['cart_id']) && is_array($result) ? array_merge_recursive($result, [
            'order' => [
                'cart_id' => $args['input']['cart_id'],
            ],
        ]) : $result;
    }
}

This enables users to use the (deprecated) setPaymentMethodAndPlaceOrder mutation which is a temporary fix for issue #2754

[candemiralp]

Hello @pmzandbergen,

Thank you for introducing this temporary fix to the issue described on #2754. We have decided not to implement it to the code base (as this mutation has been deprecated and our documentation does not recommend using it) but will keep the issue open for anyone who needs it.

The idea is solving the issue in an abstract way with our internal investigation. We will keep you updated.

Best Regards,
Can