[17:15:19] does dandelion need help testing? i can't quite understand the tests. i can probably fire up up to 6 nodes at a time [17:22:23] transaction propagation and confirmation can be tested [17:22:48] that means, transactions are included in blocks and also arrive to nodes before the block is confirmed [17:25:45] do i just start up some nodes and watch how the blocks propagate? what about the test you wrote? [17:26:14] that's what the test checks. that the transaction is propagated in all the cases [17:27:09] a test would be, if you have 6 nodes, which are connected consecutively (node 1 to node 2, node 2 to node 3, etc) [17:28:22] send transactions from one to other nodes and then generating blocks from other nodes check that they are added [17:29:11] okay i'll see what i can do [17:47:16] does the consecutive connection need to make a circle? 1 > 2 > 3 > 1 [17:47:44] that'd be valid [17:47:57] it's a possible real world scenario [18:13:05] 1. Resistance to active probing: Stem: 0 --> 1 --> 2 --> 0 where each node has argument "-dandelion=1" Probe: TestNode --> 0 Node 0 generates a Dandelion transaction "tx": 1.0 BTC from Node 0 to Node 2 TestNode immediately sends getdata for tx to Node 0 Assert that Node 0 does not reply with tx [18:13:13] how do i getdata for tx? [18:13:32] you can't do it from the wallet, it's a protocol msg [18:14:16] okay, i have 6 nodes connected now, you said send a tx from 1 to any node and just confirm the transaction can arrive? [18:15:05] yep [19:29:26] transactions seem to broadcast normally, but it can take more than a minute for the receiving node to see the transaction [19:30:04] yes that's normal [19:31:22] because of the network topology [19:50:11] is there a way to monitor mempool live? [19:52:22] getrawmempool iirc [19:53:02] but i want it to be live so i don't have to keep typing it in (up and enter i mean) [19:54:11] ah, not that i know [21:25:14] @aguycalled how long will a transactino stay in mempool in devnet before it's dropped? [21:41:06] days i think [21:47:13] @salmonskinroll it's technically possible for it to stay forever [21:47:27] If forexample the node that broadcast it keeps broadcasting it again [21:47:38] @aguycalled Am I correct in this assumption? [21:53:21] gotcha, also how to i check the connect node ip in CLI? [21:54:46] nvm getpeeroinfo [22:03:23] what's a normal timeframe for dandelion broadcast delay between two nodes? [22:07:27] @mxaddict Yes transactions can stay forever As long the utxo don't change it should be valid [22:50:08] @aguycalled let me know if my description on github can be understood. i'm a little fried up from nailing it down [22:56:15] im starting to think i might be good at finding bugs 😅 [23:20:19] ill read it now [23:27:49] @salmonskinroll when your setup is 1>2>3>4>5>6>1 and you shutdown lets say 6. is 5 only connected to 4, and 1 to 2? [23:28:43] yes [23:30:30] reading this maybe helps to understand it https://github.com/bitcoin/bips/blob/master/bip-0156.mediawiki [23:30:46] the node selects an inbound peer randomly to broadcast the transaction to [23:31:16] if there's only one inbound peer, it's always going to be the same (thats also why the transaction goes from 6 to 1) [23:31:34] and why the transaction takes longer if the inbound connection is terminated [23:32:06] in my case, if there is no inbound peer, transactions actually don't go through [23:32:16] but i guess it works as designed if that's how it's designed [23:32:43] to prevent that scenario, there's a timeout so if the transaction did not broadcast after x seconds through dandelion it goes the normal path and its broadcasted to every peer [23:33:08] gotcha, what's that x? [23:33:21] i waited more than 5 minutes [23:33:31] even longer, can't recall [23:34:00] should be less than 1 minute [23:34:02] https://cdn.discordapp.com/attachments/416000318149754881/627256662473834499/unknown.png [23:34:11] 10 seconds plus a random number which in average is 20 [23:34:36] yeah i waited minutes for many transactions [23:36:39] theres a dedicated mempool for dandelion txs before they get to be normal txs [23:36:43] one can see it with getstempoolinfo [23:40:06] maybe that helps to describe what happens [23:41:34] im getting now in bed but i can help u further tomorrow with this [23:47:45] Sounds good.