< tryphe_>
kallewoof, i think the avoid_reuse wallet flag should have more obvious documentation somewhere (unless i'm missing something) as i've been recently chatting with people who want this feature but i've been forgetting to tell them about it until i recently realized your commit #13756 exists in the codebase. the question is, where to put it? maybe an extended description of the complete functionality of "setwalletflag" flags would do better,
< tryphe_>
as there's only a very terse example of the usage, but no extended description anywhere in any "help" commands/arguments.
< tryphe_>
i feel like "finding" this feature should be much more obvious/intuitive
< tryphe_>
it looks like the description in the release notes would do the job (and tell the user they need to -rescan also), just not sure of the correct place for it
< tryphe_>
also slightly related, when users look to enable features, they do ./bitcoind -help, ./bitcoin-cli -help, but not necessarily ./bitcoin-cli help, so they might not even stumble on setwalletflag. but not sure how to effectively get people to do this :(
< luke-jr>
tryphe_: note the "avoid_reuse" flag does not in fact avoid reuse :x
< tryphe_>
luke-jr, ah, how so?
< luke-jr>
would have to review the code to explain accurately
< tryphe_>
i've never really used the flag, fwiw, but the idea is nice
< tryphe_>
luke-jr, it looks like it doesn't avoid reuse, but avoids spending outputs from keys that were already "spent from"?
< luke-jr>
something along those lines
< tryphe_>
not sure if that's what you meant
< luke-jr>
IIRC the main thing is that if multiple such outputs exist, it tries to spend them all at once
< tryphe_>
ahh okay, yeah the functionality is not really explicit to the definition
< luke-jr>
although that may be default behaviour now too\
< tryphe_>
ahh yeah i saw that as well also, in another PR
< tryphe_>
-avoidpartialspends is always enabled when avoid_reuse is enabled
< tryphe_>
i think it would also be useful to note that in some documentation as well (avoidpartialspends documentation links to avoid_reuse but there's not really any documenation of avoid_reuse to explain that it enables avoidpartialspends)
< tryphe_>
so you wouldn't know about it unless | grepping some help output
< tryphe_>
also, i guess the implication of a "dusting" attack here is that you would only avoid the dusting attack completely if the address was already marked dirty before the dusting happens
< luke-jr>
eh?
< tryphe_>
luke-jr, you might sweep up the dust the first time an address is marked dirty, but never after that, right?
< luke-jr>
tryphe_: what's the problem?
< luke-jr>
sweeping up the dust is the right thing to do there
< tryphe_>
ah nm, you're right
< tryphe_>
confusing myself :p
< fanquake>
sipa: probably time to block tnawanna
< fanquake>
grubles: there’s no particularly easy way
< kallewoof>
tryphe_: i'm all for making it more known, but not sure where to do so. perhaps add an entry about it to the bitcoin wiki might be a good idea
< wumpus>
it's not bad to have competing approaches and have a discussion about what is the best one, but it does make me wonder what makes this particular thing so interesting/important to work on :)
< wumpus>
and "why does everyone want it differently"
< wumpus>
fanquake: yes that's a great overview
< wumpus>
I don't have a strong opinion on this besides, please don't add "WeakAssert"
< hebasto>
wumpus: the recent 19979 removes the need to choose the name at all
< yanmaani>
what's the process of suggesting a change to the protocol?
< yanmaani>
Do I just describe it informally on the mailing list and ask something like "Would you consider a BIP to this effect to be potentially suitable for inclusion in the Bitcoin protocol?"
< ryanofsky>
jonasschnelli: it just reads and writes bytes to anything. the current PRs open a unix socket not a TCP socket, but is easy to change
< ryanofsky>
ssh and socat commands are also happy to forward traffic between tcp sockets and unix sockets (maybe netcat too, not sure)
< jnewbery>
Perhaps ryanofsky, aj and hebasto could get together and figure out which is the best way? They seem to care about this more than other people and I think other reviewers are probably going to be disinclined to weigh in or review 4 different PRs.
< bitcoin-git>
[bitcoin] prayank23 closed pull request #19945: Remove connect_nodes global and Replace connect_nodes(self.nodes[a], b) with self.connect_nodes(a, b) (master...master) https://github.com/bitcoin/bitcoin/pull/19945