No tests yet, looking for concept ACKs, or to see what people would want for this.
It's essentially mirror functionality of #7558.
I am not sure how mempool interacts with this(I'm guessing I should check if it's in block/mempool like abandontransaction does) but the usage is intended for in-block transactions.
While similar, this functionality is more meant to pair with #7558, which isn't applicable to abandontransaction. If you have a transaction in your wallet that is confirmed, yet you do not want to "account for" as having the funds, there isn't a way to do this. Primarily I think this could happen for instance with an erroneous application of #7558 .
I understand the application is different, but there's a risk in having vaguely similar functionality in different RPC calls. Before adding abandontransaction we had a long discussion about what would be safe as well versus unsafe RPC call behavior for this purpose, so it already is a compromise.
Idea: If it is meant as antidote to #7558, why not call it removeprunedfunds and make it only work on funds thus added?
Understood and agreed. I might want to merge the two PRs then, and tag transactions in such a way to ensure that property.
I don't have a strong opinion about this functionality. I do think it is quite different from abandontransaction though
Well if really they're sufficiently different hat people won't get confused, I'm ok with adding this... But based on the names at least I would get confused as API user, I think. But that may be a documentation issue.
Closing in favor of merging the two PRs, with more aligned names and description.