launchPurchaseFlow for a subscription will return an error, but still charge the user for the subscription.
See original GitHub issuePreviously I had been running subscription purchases through the IABHelper.java’s - launchPurchaseFlow
. I did not realize that this was actually running a purchase for ITEM_TYPE_INAPP
items. Regardless, this used to function correctly, and the subscriptions were being purchased correctly.
This week I got a report from my client that there was an issue with subscription purchases. It seems that now purchasing a subscription while using launchPurchaseFlow will return an error (“IAB returned null purchaseData or dataSignature”), despite the user actually being charged for the subscription. I now understand that I should be calling launchSubscriptionPurchaseFlow
(and have updated my project accordingly), however, this seems like a bug (having an error returned while still being charged for the item) that should be addressed.
I realize that this may actually be a bug on the server side of the equation, but figured I would post here to see if this has been documented already, before creating an issue on issuetracker.
Issue Analytics
- State:
- Created 6 years ago
- Comments:9 (2 by maintainers)
Top GitHub Comments
These samples and moderation here are a mess. Why do the Google employees here not know and recognize the TrivalDrive v1 example anymore? It was the recommended implementation for years and until a few weeks ago! It is like it vanished from the hive minds memory.
We are trying to give you money Google!
The official docs recommended IABHelper.java.
@gorgon I don’t get your confusion. In his very first line @ssawchenko says
IABHelper.java
. He clearly is using the ‘old’ TrivalDrive v1. I assume you didn’t start working on this last week.It is like #45 In June: @dec1: ‘there is an NPE bug in the TrivalDrive example’ @netomarin: ‘we will look into it’ August: @netomarin: ‘still looking into it’ End August: @gorgon ‘It is not a bug. You need to do a retry as shown in TrivalDrive v2’ October: @netomarin: ‘It is not a bug in the sample.’
What is going on? It is a bug in the V1 sample. Retrying the BillingClient is no solution. But I guess pretending the recommended method of the last years doesn’t exsist, is easier.
Why is this so frustrating.
Please refer to ClassyTaxy And Trivial Drive 2.
Closing as obsolete.
Please reopen if you think this issue is still relevant.