Use propertyBag in doPayment

[mattwire]

Overview

This fixes the test failure in #17886. If a propertyBag is passed into doPayment you get a PHP notice:
Deprecated function PropertyBag array access for core property 'amount', use getAmount().

Currently the eventcart PR is the only place that passes a propertyBag into doPayment so it triggers this notice.

Before

PHP notice because a property is not accessed via it's standard method.

After

Property is not accessed via it's standard method.

Technical Details

Comments

[civibot]

(Standard links)

• If this is your first pull-request for CiviCRM, please browse CONTRIBUTING.md for information about the development and testing processes.
• If you are reviewing this pull-request, you may wish to consult the test sites and the Review Standards (long template, short template).

[eileenmcnaughton]

Yep - this makes sense & I think test cover is adequate on this

[eileenmcnaughton]

@mattwire I merged this but I'm having misgivings now - we need to make sure $params['amount'] works since it's not just the CRM_Core_Payment class that will be accessing $params['amount'] - but likely a whole swag of processors 'out there'. Given that most known flows have passed $params['amount'] to date it's implictlly part of the contract. We need to ensure that it continues to work without e-notices (& if that is broken in event cart we actually need to fix it to pass out $params['amount']) - or something that can be accessed in that way without notices

[mattwire]

This PR won't break anything because we are not passing the propertyBag on anywhere - but the eventcart implementation demonstrates the issues that payment processors that haven't updated to use propertyBag will face.
All payment processors written/maintained by MJW natively support propertyBag and do the recommended propertybag::cast($params) at the start of doPayment - that's the only way to guarantee what your input is. If you look at https://lab.civicrm.org/dev/financial/-/issues/133#note_38422 you'll see we discussed the need for a way to allow processors to transition.

[eileenmcnaughton]

@mattwire yeah - I agree this won't break anything - it just removes a guard rail because we can't change event cart to pass out propertyBag unless it will work with all processors & this removes one of the places where we might have detected that it didn't work.

I see you are still thinking that supportsPropertyBag might be part of the transition - I don't think that applies to parameters - only to the possibility that they might be using array functions that would break