fix(EMI-3047): Express checkout bugs

[erikdstock]

The type of this PR is: fix

This PR solves EMI-3047 by fixing several express checkout issues:

• It removes a non-null assertion that was breaking shipping address setting and the entire subsequent flow.
• It makes the order mutation and submission flow match what we have in our normal checkout, avoiding the need to fork our logic in exchange in https://github.com/artsy/exchange/pull/2909#discussion_r2732197818
• It encapsulates logic for the callbacks passed to the express checkout element to account for differences between apple and google pay.

Testable at https://express-checkout-test.artsy.net

cc @artsy/emerald-devs

[relativeci]

#6272 Bundle Size — 9.2MiB (+4.52%).

cf24853(current) vs 9eecfa0 main#6268(baseline)

Important

Bundle introduced 10 duplicate packages – View changed duplicate packages

Warning

Bundle introduced 43 new packages: @artsy/xapp, node-uuid, crypto-browserify and 40 more – View changed packages

Bundle metrics   9 changes 4 regressions

	Current #6272	Baseline #6268
Initial JS	3.5MiB(+12.77%)	3.11MiB
Initial CSS	0B	0B
Cache Invalidation	50.45%	0%
Chunks	99(-1%)	100
Assets	102(-0.97%)	103
Modules	5806(+2.93%)	5641
Duplicate Modules	560(+0.36%)	558
Duplicate Code	3.88%(-1.27%)	3.93%
Packages	308(+20.78%)	255
Duplicate Packages	41(+32.26%)	31

Bundle size by type   2 changes 1 regression 1 improvement

	Current #6272	Baseline #6268
JS	9.07MiB (+4.61%)	8.67MiB
Other	139.59KiB (-0.78%)	140.68KiB

Bundle analysis report Branch erik.emi3047-submit-payment-meth... Project dashboard

---

^{Generated by RelativeCI Documentation Report issue}

[erikdstock]

both of our set_ helpers return this shape (maybe it should be a type) but now the act of setting a shipping address automatically triggers the next step for google pay because it does not automatically do so.

[starsirius]

🎉