Added BackendErrorCode.purchasedProductMissingInAppleReceipt#2033
Conversation
See https://github.com/RevenueCat/khepri/pull/4716 This makes that error `ErrorCode.invalidReceiptError` instead of `.unknownBackendError`. I've also added #2032 to improve error messages when these errors are unknown in the future.
| case .cannotTransferPurchase: | ||
| return .receiptAlreadyInUseError | ||
| case .invalidReceiptToken: | ||
| case .invalidReceiptToken, .purchasedProductMissingInAppleReceipt: |
There was a problem hiding this comment.
Nitpick, these multi-cases seem to be set in multiple lines here. Just to be consistent:
| case .invalidReceiptToken, .purchasedProductMissingInAppleReceipt: | |
| case .invalidReceiptToken, | |
| .purchasedProductMissingInAppleReceipt: |
There was a problem hiding this comment.
Also, not sure if we have tests for these... If we do, it might be good to add one, if not, I guess something to think about.
There was a problem hiding this comment.
What type of test would you write, other than simply duplicating this logic?
There was a problem hiding this comment.
Yeah I think it would be good to have a test that checks the mapping from backend code to purchases code for each of these values. As you said, it's basically duplicating the logic, so not a huge benefit, but might catch some issues if someone changes the mapping at some point and moves one of the error codes accidentally. That said, I don't think this is needed for this PR
There was a problem hiding this comment.
Like you said yourself, that would just be duplicating the logic. This is purely data-driven, so the data is the test as well.
for each of these values.
In order to do this cleanly, we would need to have a mapping for each of the errors, instead of duplicating each test. Then iterate the mapping and assert it's correct... but wait, that mapping is the data we're testing, so there would be no valid in that duplication.
There was a problem hiding this comment.
Well I agree that it's redundant, but there is some security in redundancy. In any case, I agree the chances of this being changed mistakenly are minimal and if we did this, we would probably have to duplicate logic for tests in a bunch of places so I'm ok with leaving this.
**This is an automatic release.** ### Bugfixes * `ISO8601DateFormatter.withMilliseconds`: fixed iOS 11 crash (#2037) via NachoSoto (@NachoSoto) * Changed `StoreKit2Setting.default` back to `.enabledOnlyForOptimizations` (#2022) via NachoSoto (@NachoSoto) ### Other Changes * `Integration Tests`: changed weekly to monthly subscriptions to work around 0-second subscriptions (#2042) via NachoSoto (@NachoSoto) * `Integration Tests`: fixed `testPurchaseWithAskToBuyPostsReceipt` (#2040) via NachoSoto (@NachoSoto) * `ReceiptRefreshPolicy.retryUntilProductIsFound`: default to returning "invalid" receipt (#2024) via NachoSoto (@NachoSoto) * `CachingProductsManager`: use partial cached products (#2014) via NachoSoto (@NachoSoto) * Added `BackendErrorCode.purchasedProductMissingInAppleReceipt` (#2033) via NachoSoto (@NachoSoto) * `PurchaseTesterSwiftUI`: replaced `Purchases` dependency with `SPM` (#2027) via NachoSoto (@NachoSoto) * `Integration Tests`: changed log output to `raw` (#2031) via NachoSoto (@NachoSoto) * `Integration Tests`: run on iOS 16 (#2035) via NachoSoto (@NachoSoto) * CI: fixed `iOS 14` tests Xcode version (#2030) via NachoSoto (@NachoSoto) * `Async.call`: added non-throwing overload (#2006) via NachoSoto (@NachoSoto) * Documentation: Fixed references in `V4_API_Migration_guide.md` (#2018) via NachoSoto (@NachoSoto) * `eligiblePromotionalOffers`: don't log error if response is ineligible (#2019) via NachoSoto (@NachoSoto) * Runs push-pods after make-release (#2025) via Cesar de la Vega (@vegaro) * Some updates on notify-on-non-patch-release-branches: (#2026) via Cesar de la Vega (@vegaro) * Deploy `PurchaseTesterSwiftUI` to TestFlight (#2003) via NachoSoto (@NachoSoto) * `PurchaseTesterSwiftUI`: added "logs" screen (#2012) via NachoSoto (@NachoSoto) * `PurchaseTesterSwiftUI`: allow configuring API key at runtime (#1999) via NachoSoto (@NachoSoto)
See https://github.com/RevenueCat/khepri/pull/4716
This makes that error
ErrorCode.invalidReceiptErrorinstead of.unknownBackendError.I've also added #2032 to improve error messages when these errors are unknown in the future.