03:13:59 Has anyone ported the electron wallet to monero yet? 03:16:16 electrum? 03:16:27 electrons are a conspiracy.. pretty sure 03:16:44 i mean think about ir 03:16:47 it 03:21:04 yanmaani: as a user I perhaps do not want to do the work of verifying transactions myself, but nor do i want to offload that work to an unknown entity with uncertain incentives 03:21:30 but our world is already built on trust; i would be happy offloading that work to someone in my "community", for some definition of community 03:47:04 endogenic: stick a fork in your electrical socket to prove it 03:47:29 ok edison 04:40:08 I'm really new to how all this works so this might not make sense, but, with the way ring signatures work, would I have to send through several different addresses to obscure the trail or does it have the same affect if I were to keep sending to my own? I'm having trouble grasping the logic of this 04:42:21 burry: sending back to yourself is equivalent to sending to another of your addresses, so long as you aren't polluting the 'trail' with new outputs or creating trail 'loops' which both add a lot more complexity 06:52:47 Tari Labs compliance paper update from tari TG: "Very close here - we hope to confirm the publication date this week. Whitepaper + short form version in Bloomberg Law and a Harvard journal" 08:29:43 hello 08:32:34 what is the coinbase structure for v12 block version? 08:33:22 http://paste.debian.net/1158408/ is v1 version 08:57:48 what is the coinbase structure for v12 block version? 08:58:13 v1 is here http://paste.debian.net/1158408/ 08:58:21 thank u 09:03:28 banama: I don't think many Monero devs look at this on the byte level and have the info handy in this format. Personally, I would start with the C++ source, work out how that probably serializes and then check against an actual hexdump of a block 09:04:58 i check all hardfork changes but stil cant understand where code check for block version 09:05:24 or maybe is cutted in many pieces and i dont know enough c++ 09:06:47 this link explain enough https://monero.stackexchange.com/questions/2136/understanding-the-structure-of-a-monero-transaction?rq=1 10:12:35 bigslim[m], jbg: Yes, you can do that securely, but the problem is that if everyone does it, you will no longer have security. 10:12:43 so the devs make it harder for that reason 11:48:20 yanmaani: huh? 15:25:18 hola 18:27:49 blogged lately kayront 18:27:49 ? 18:28:38 not lately, not lately 18:28:58 lots of ideas flowing and might write something soon though. it's idle but not forgotten :) 18:29:03 nice of you to ask 18:29:25 haven't been on IRC much either, but every day brainstorming crypto stuff, hehe 18:29:29 how's it going over here? 18:35:17 i'm listening to schiff on the what bitcoin did podcast, deconstructing all of peter's arguments one by one would be a good writeup 18:35:20 dude's a true nonbeliever 18:37:16 i remember his debate with molyneux years ago 18:37:35 schiff was saying there's no inherent value, stef was saying the inherent value is the utility. essentially 18:39:34 he's using the same arguments everywhere 18:39:36 also on rogan 18:40:32 that's his ideology. Why would he change his position? Nobody can effectively refute them, as far as I can see, as they are based in belief (and not all incorrect either) 18:42:22 why would he change? why do people debate things? presumably to arrive at truth 18:44:07 come to think about it, maybe that's not so obvious nowdays 18:47:33 lol, they actually mention monero by name at the 1h mark 18:47:45 as usual in the context of providing cover to do something illegal 18:48:04 useful, when laws don't match common sense, but is it really so hard to understand that privacy in day to day things is also valuable, if not desirable? 19:40:43 anyone here running a full node on btrfs/zfs? if so how do you deal with the huge fragmentation in the db file causing lots and lots of i/o? 19:47:57 it's just what monerod does artefact 19:48:35 i got a giant l2arc nvme and now it's only a problem with a cold cache 19:49:38 i don't think it's zfs specific, though if you misconfigure zfs then that likely won't help the case 19:49:40 can't do that yet with btrfs sadly 19:49:56 it's just slow on mechanical disks in general, lots of random I/O 19:50:00 (monerod) 19:50:17 get more ram, also works 19:50:23 is it already on an ssd? then fragmentation should be ok 19:50:43 it's on spinning rust atm. i have a nvme drive but nowhere near enough free space for the chain 19:51:05 16g of ram so plenty of space for cached pages 19:51:26 just had a look at mine, and presently 80GB are allocated in l2arc 19:51:42 not all of it monero stuff, but it's by far the most active daemon in the whole system 19:51:58 i hope it slows down a lot once it's caught up with the network 19:52:39 well, zfs is copy on write, so the philosophy is that you have to live with it 19:53:18 imagine your node is fully synced already, and you connect your wallet to refresh - you haven't for a few days .. every transaction has decoys, and the decoys have to be looked up by monerod as it serves your wallet the data 19:53:58 that's my working theory as to why it's so slow in hdds. it's very perceptible with a cold cache when syncing wallets especially, i don't think it's coincidence 19:54:12 right. hopefully i don't have to do it very often 19:54:31 as long as you have enough free memory the os should cache as needed 19:54:51 when it has to fall back to reading the disk continuously, say goodbye to performance 19:54:58 (not a problem with ssds/nvmes ofc) 20:25:30 btrfs/zfs are extremely poor choices for storing the blockchain 20:25:57 they are copy-on-write filesystems, and LMDB is internally a copy-on-write dataase, so you're getting massively redundant writes 20:26:38 for LMDB the simpler the filesystem the better 20:26:58 the less the FS tries to do, the better... 20:33:56 nodatacow is always an option, but i'd rather not use it unless absolutely necessary 20:35:08 or put the blockchain DB on a separate filesystem. easiest is ext4 with journaling turned off 20:35:21 not feasible for practical reasons 20:35:29 with a couple small patches, the best option would be its own raw partition 20:35:42 (I have written the patches, but not merged and released) 20:36:23 how about patches that make the db dumber for smart fses :-) 20:38:48 sorry nope 20:39:23 you should read the usenix report from here https://twitter.com/Superfind/status/1287609039875121155 20:39:33 FSs all suck 20:39:48 LMDB's own data reliability mechanism works better than all of them 20:40:12 not sure you pasted the right link 20:40:21 bah 20:40:25 well, a filesystem isn't a rdbms. rdbmses aren't filesystems. 20:40:28 lemme try again 20:40:55 https://twitter.com/hyc_symas/status/1284689627295682563 20:41:36 LMDB is a DB engine, and can be used in an RDBMS but it is not an RDBMS itself 20:41:48 yes, i recall that fsync issue from postgres mailing lists too 20:41:52 and differences aside, filesystems should not allow data to be corrupted without warning 20:42:05 they assumed that if fsync failed , you could retry it, and if it succeeded, it meant that all the data was safe on the platter. turns out, no 20:42:10 it goes deeper than that. the psotgres thing was only one particular failure mode 20:42:22 and yet, the only filesystems afaik that do data checksumming are COW filesystems 20:42:39 anyway, i don't want to hijack the channel and start a debate, i just got here :-) 20:42:59 read the report, or at least the presentation slides 20:43:40 they claim two distinct failure modes for LMDB, one of which I'm pretty sure they're wrong, and the other is the fault of ext4's journal 20:43:55 which we can't do anything about, and we always tell our customers not to use that journaling mode anyway 20:47:40 what do they mean by “ext4 data mode” in the slides? data=writeback? 20:51:18 data=ordered 20:52:10 thanks 20:54:31 their writing is pretty unclear on that, if I were a reviewer of their work I would've pointed that out and asked for revision 20:54:59 there's no excuse for using fuzzy terminology in a tech paper with well established technical terms... 20:56:52 kayront: an ideology is a belief structure or system, not a search for truth. I can believe it's better to eat beans for breakfast, and you can believe it's better to eat bacon. There doesn't have to be a unified truth. 21:25:08 https://twitter.com/tomrobin/status/1287847373591531523 21:25:27 only morons use BTC ... 23:09:42 but muh network effect! 23:27:31 reading about lmdb. reminds me somewhat of judy arrays. very clever and very fast, but still rare in the field for some reason