00:01:34 Then with `monero-wallet-cli --restore-from-seed --offline --stagenet --generate-new-wallet Monero/wallets/stagenetOffline/stagnetOffline` I provide the same seed phrase, no offset, no password. 00:02:32 `monero-wallet-rpc --offline --stagenet --wallet-file Monero/wallets/stagenetOffline/stagnetOffline --rpc-bind-port=18099 --rpc-login offline:offline --prompt-for-password` also no password 00:04:30 but now via rpc 'query_key', {'key_type': 'mnemonic'}, the key (seed phrase) is suddenly: ostrich lesson highway whole jester juvenile sack citadel yearbook slackens viking liar rustled natural mouth gained jargon girth neither verification vipers large toolbox utensils mouth 00:05:51 well, until here I would say something went wrong on what I did - but, let's compare now the spend_key and view_key, will simple put it in the same format like up, `spend_key, view_key` 00:06:31 d38a52e481072c048921e9d5a6fcbf8c54366dfcda8d83e7719e43024b80c50d, 343a63ab0ff6a82b58226bbe6777946dbeb3a09f86fdb49000a398fb07c2910e 00:09:15 when I not completely mistaken then spend_key and view_key exactly the same, I created also from the python `seed` the data for a view only wallet on feather to test something and the feather wallet is in sync with the other monero-wallet-rpc (hotwallet/view only) which I used to make transactions. 00:12:39 So, why the heck I get now an other seed phrase, except from `rustled` to `vipers`? WTF is going on? 00:13:26 My guess is that somewhere in that process the wallet wasn’t initialized in a deterministic way 00:13:51 Similar thing happened to my “vanity” address that I use for receiving funds 00:16:52 same wallet, when I restore with that seed :o 00:18:27 now I'm even more puzzled :D Why I need always encounter things like that, have already no time. 00:20:09 preland: I have seen this conversation, and when I remember it should then only differ in the last words (except the last word). But only 8 words are the same and in the middle. 00:24:13 created also a new wallet from the monero-python seed, get again the same wallet, all transactions, but also when I look for the seed in setting it changed to: 'ostrich lesson highway whole jester juvenile sack citadel yearbook slackens viking liar rustled natural mouth gained jargon girth neither verification vipers large toolbox utensils mouth' 00:31:29 and in python the seed object is the same, except for the seed.phrase and seed.hex (binary represantation of the seedphrase) :/ And all that only because I wanted to seed if I could get the seed.phrase or seed.hex after I applied a password :D 02:21:25 Got a `bad_alloc` error when running monerod, should I post the stack trace here? 02:22:45 And it is using 7gb of ram... 03:42:11 Yeah put it in a pastebin 04:30:19 vthor: " So, why the heck I get now an other seed phrase, except from rustled to vipers? WTF is going on?" The "exquisite" seed is not reduced. The person coming up with it probably just produced 256 random bits and translated those into Monero seed words. But Monero seed don't fully use all the space that 256 bits span up, thus such "overly large" numbers get reduced: 04:30:20 https://crypto.stackexchange.com/questions/99465/what-is-the-order-of-the-generator-point-g-9-in-curve25519 05:18:46 hey rbrunner, thank you for the answer :) So the monero-python library is the culprit, how I anyway need to modify th lib to sign_transfer, I can also directly correct that, don't like esoteric outcomes :D, do you have quick pseudocode I can still understand tonight, or I better make this when I can concentrate. I almost think I wrote that in polyseed-python, but not sure and not able to think so deep at the moment :/ Will make a nap, first, maybe it helps. 06:38:43 https://paste.debian.net/1321803/ 06:38:44 Left monerod running out of curiosity, it is at 18.2gb usage now 08:06:00 <0​xfffc:monero.social> recanman: by any chance do you get this OOM regularly? if yes, can you make this change: https://paste.debian.net/1321811/ and put the stacktrace on pastebin 08:10:22 <0​xfffc:monero.social> that change should go here: https://github.com/monero-project/monero/blob/cc73fe71162d564ffda8e549b79a350bca53c454/CMakeLists.txt#L344 16:51:45 Seed( hexlify( hexlify(ed25519.scalar_reduce(unhexlify(seed.hex.encode())))).decode() ).phrase == Seed( hexlify( hexlify(ed25519.scalar_reduce(unhexlify(offline_wallet.seed().hex.encode())))).decode() ).phrase => True 16:51:58 but I would have expected that: 16:52:23 Seed( hexlify( hexlify(ed25519.scalar_reduce(unhexlify(seed.hex.encode())))).decode() ).phrase == offline_wallet.seed().phrase => True 16:52:32 but that is not the case. 16:53:26 rbrunner: you have the needed knowledge to indicate me where to search further? 18:31:00 That is a lot of nested hexlify/unhexlifly 18:55:52 vthor: All a bit confusing :) But to me this somehow looks as if you managed to produce a wallet with a spend private key value that is not reduced. Sounds improbable, because as far as I know you can't get a non-reduced value past the Monero software, it always reduces. 18:57:03 Anyway, that would be the first thing I would check: If you reduce the spend private key of the wallet, the value should not change. If it does, something bad happened. 19:32:22 preland, that is because the key produced from the urandom or seed phrase get stored as a hexlified bytes, but for ed25519.scalar_reduce you need to feed bytes and when you want create a seed phrase from the bytes you need to feed a hexlifed string of the 32 bytes. 19:35:37 well, the 32 bytes which are related to the seed phrase is not the spending key... the spending key I get through seed.secret_spend_key(). But the weird behaviour is that I can convert it back and forth with monero-python, but if I use the seed phrase with monero-wallet-* I can create the exact same wallet, but get then another seed phrase instead - which represents exact the same wallet. 19:39:15 And I had never hit this issue if I had not thought, when I anyway need the monero-wallet-cli on the pi zero, I could abuse it directly to support passwords for monero strandard seed phrase, and can then maybe get the not encrypted seed back to generate the different fingerprint, but now I see that I am stupid, I can simply use the spend key or the public address (more sane) for the fingerprint instead of the bytes from the seed. 19:44:00 So, it still bothers me, don't like things not fitting, but in reality I can ignore it because it should not matter - except if some user will create a seed on the pi zero, one day he puts the seed in monero-gui and is seeing like me suddenly another seed phrase as the one with which he created the wallet - it sends you some chills the spine down... But for now I will put it in my todo list and ignore it for now. How I anyway think in a more distant future it 19:44:01 would make more sense to change from Python to C++ - only for practice :D - well not really, but I think it is the best thing to stay on par with new development - on the other side I prefer to have different implementations, because then you see suddenly stuff like that - what I don't like :D 19:48:42 Now there is only one worry, I was convinced it I use the key from poolyseed and I converted it with seed to a monero standard phrase it would be the the same like a wallet on feather with polyseed - but now I assume that the key I get from polyseed is not the same als unhexlify(seed.hex), but should be the same as seed.secret_spend_key(), if this is the case i would need to calculate myself, the public address and view key from that spend key and import the 19:48:43 wallet this way in monero-wallet-rpc :/ 19:49:23 Fuck I feel raped, now I have 5 rabbit holes open, machine is cooking and my stress level beyond 1000% :D 20:42:54 This is how I always feel when doing anything coding wise 20:43:16 Next time I get to my computer I should set a timer for how long it takes before I find a bug or issue that no one has ever found before 20:48:43 :D 20:52:32 This is why I hate dependencies, with perfectionism this is playing russian roulet with a pistol instead of a revolver :D but I still can laugh, and from all work in the world I like coding almost like woodwork or metal work, because you can bend reality also a lot, with a piece of would or a block of aluminium you are bound to physics, with coding you can just in any moment decide you don't like the rules applying to you and you change reality (maybe I'm a 20:52:33 bit crazy) 23:40:04 Might have figured out the reason why: I was using a Copy-on-write filesystem (btrfs). I removed COW according to the Arch Linux wiki and I'm getting way better speed.