Switch to Payment.create to fix issues with payment_instrument_id on payments

[mattwire]

@artfulrobot This fixes #63. I started debugging and then thought what happens if I switch to the new preferred api Payment.create instead of completetransaction. And by magic it works perfectly.

Payment.create calls completetransaction internally if required. As this is now our preferred, tested route for completing a transaction it makes sense to switch rather than investigate a broken "legacy" code-path.

[artfulrobot]

It looks like this change ends up with a different algorithm for setting membership dates - specifically that the membership date seems to include a time element now as well as a date - and the date's slightly off, too.

Can you update the failing test to fix this?

PHPUnit 6.5.14 by Sebastian Bergmann and contributors.

...........F...............                                       27 / 27 (100%)

Time: 20.34 seconds, Memory: 46.00MB

There was 1 failure:

1) GoCardlessTest::testWebhookPaymentConfirmationFirst
Failed asserting that two strings are equal.
--- Expected
+++ Actual
@@ @@
-'2020-02-01 00:00:00'
+'2020-02-03 17:59:25'

/buildkit/build/dmaster/sites/default/files/civicrm/ext/uk.artfulrobot.civicrm.gocardless/tests/phpunit/GoCardlessTest.php:785
/buildkit/extern/phpunit6/phpunit6.phar:570

FAILURES!
Tests: 27, Assertions: 120, Failures: 1.

[mattwire]

@artfulrobot This passes tests now - I've had to call contribution.create again to set some params - see comments in code