sarang moneromooo UkoeHB Any idea if this would be possible with Triptych, and could be included without major impacts in a HF with Triptych?

Would love to see easier support for L2s moving forward (as well as more private atomic swaps) and if this change doesn't have major negative side effects it may be worth exploring.

What is "this" ?

Sorry, maybe the above long text didn't come through from Lucas

I'll pastebin it

(Debian pastebin won't let me paste the text, says "do not spam"): paste.centos.org/view/ba754067 moneromooo

centos.org always fails to load here, something to do with their SSL cert I think.

Hmm, any other pastebin that works for you?

paste.ubuntu.com

pastebin.com usually works, but sometimes cloudflare blocks it.

<utxobr[m] "my node seems to keep dropping a"> You sure it's not restarting/getting restarted by something?

Holy hell why so many outs

This is independent of triptych really.

Lets take this to #monero:matrix.org

utxobr: ^

<moneromooo "This is independent of triptych "> I figured, but wondering if we should include this change *with* Triptych as long as it doesn't interfere or cause other issues.

Probably not.

I think a different output selection scheme will have to be done first. Binning and/or deterministic rings.

Due to complexity, or some issue that it would cause? I'm not really well versed in how this would affect... anything, honestly.

Deterministic rings would seem to be iincompatible with this.

Isn't binning a needed thing for Triptych either way?

Mostly because we've not done this for years now, why rush it right now ? :)

I don't think anyone found any fatal flaw (beyond the extra network usage) before.

But it's always stayed at the "look at it later" stage.

Binning or similar would be almost mandatory for triptych.

I think their idea would be an ok change for the current protocol, but it is incompatible with binning/deterministic ring selection where the relative position of outputs must be known in advance. If a chained tx can be identified with heuristics (e.g. by looking at the fee or analyzing ring member distribution), then the 'guess-newest' heuristic will reliably identify the true spend.

If we really wanted to support chaining tx, we should define a tx input type that uses a 1-member ring sig. Not sure how much traction that would get.

from Libera:

22:18 <ruisiang> Hello everyone

22:19 <ruisiang> Don't know it's the right place for this, but I wrote a merchant api wrapper for monero-wallet-rpc at github.com/RuiSiang/monero-merchant

22:19 <ruisiang> Hope to get some feedback or suggestions. Thanks

wtf

oh it's a spammer

rasenqan != rasengan