submitpackage require at least two transactions
#31085
I want to call submitpackage and I don’t want to have to check if I have more than one transaction to do it. I’m lazy but that restriction seems arbitrary.
No response
No response
No response
No response
It won’t hurt to broadcast a single transaction if it passes consensus and policy checks via submitpackage RPC.
From the package relay BIP and the policy documentation, I see that a package is defined as a child transaction with its unconfirmed ancestors, topologically sorted.
I think the intent of the check is to be strict and ensure that what’s being broadcast is a package. As such RPC behavior is correct, as an individual transaction doesn’t qualify as a package based on this definition.
I want to call
submitpackageand I don’t want to have to check if I have more than one transaction to do it. I’m lazy but that restriction seems arbitrary.
I had similar thoughts when reviewing submitpackage a while ago. Forcing the user to use different RPCs based on the number of transactions in a package feels arbitrary and not ergonomic to me. A package with size 1 should be a valid package and accepted as such in the relevant RPCs (slowly obsoleting their single-tx alternatives over time, but that’s a different discussion).
I see that a package is defined as a child transaction with its unconfirmed ancestors, topologically sorted.
A tx without unconfirmed ancestors would make for a valid package according to that definition, though.
Concept ACK.
