10:45:46 This can still be solved. 16:13:24 @xmrusercz:matrix.org: What do you refer to? What can be solved? 16:58:05 Meeting in about 1 hour 18:00:02 Meeting time. Hello! https://github.com/monero-project/meta/issues/1465 18:00:10 Hey 18:00:10 Hello 18:00:36 waves 18:01:28 Ok, let's start with the reports from last week 18:01:37 Me: finishing touches for the I2P SAM patch (hopefully), random patches, working on a potential spec for signing with hardware wallets 18:01:57 mostly worked on rebasing the wallet-cli and wallet-rpc PRs onto polyseed (and all the other stuff that got merged recently) 18:02:17 @jpk68:matrix.org: What would that spec specify? 18:02:38 Howdy 18:02:58 me: we launched the FCMP++ & Carrot beta stressnet v3, and I'm preparing all upstream FCMP++ consensus integration PR's at the moment (FCMP++ verify tooling PR is here: https://github.com/monero-project/monero/pull/11407 ) 18:03:30 Nice how it all comes together bit by with with FCMP++. Fascinating to watch 18:03:42 *bit by bit 18:03:50 rbrunner7: It would expose a generic USB interface (or some other transport) for communication between signing devices and wallet software. It's based on vtnerd's CBOR {en,de}coder (#10659) 18:04:32 I may be able to upload a draft gist soon. Messages are described with CDDL, written up in a Markdown document 18:04:38 Me: working on monero-compat https://github.com/jeffro256/monero-compat. Got v0.18.0.0 plus compilation fixes to compile, fixing GitPython bugs, etc. Trying to get to the point where the test cases would have caught https://github.com/monero-project/monero/pull/11412 18:04:57 Ah, I see, so that hardware wallet manufacturers would hopefully come towards us a good part of the way, instead of us implementing complex protocols in our codebase 18:05:23 I suppose, yes, though it doesn't solve the issue of having to use Trezor's new protocol 18:05:28 https://github.com/bitcoin-core/hwi 18:05:34 ^ Design is somewhat based on this 18:06:36 Certainly worth an attempt, although we probably have nowhere near the weight of Bitcoin ... 18:06:38 In practice, you could quite easily make a Python script that uses Trezor's hardware wallet libraries + Protobuf, and then have it act as a middleman (that is, speaking "monero-hwi" to the wallet, and Protobuf/THP to the device) 18:07:07 Oops late, but present 18:07:18 Ok :) 18:08:24 I suspect most hardware manufacturers are going to say use our existing API and deal with it 18:08:44 Like some game of poker 18:09:01 It would be great to not have to deal with Abseil and all of that other BS that comes with Protobuf 18:09:10 Not to mention the new Trezor protocol abomination 18:09:12 Most = 3, at the moment, as far as I know, I mean hardware wallets with XMR support 18:09:35 Or did I miss some that joined lately? 18:09:37 Yes. I wrote a protobuf read/write that works except for vector/arrays. 18:10:07 And implements the new encoder/decoder you mentioned 18:10:23 As tobtoht remarked a few weeks/months ago, hardware wallet vendors have too much control of our dependencies 18:10:27 I agree with this 18:10:32 @vtnerd: Well the API with the HW device would remain the same, ostensibly we would add a bridge between HW-specific code and the core code 18:10:36 The trezor protocol is insane, I cant believe that got pushed out 18:11:26 They may have suspected LLMs would be able to help with reviewing it, or writing its code for other projects 18:11:57 Even with LLMs you are not forced to implement something really complex, seems to me 18:12:24 So the data flowing over the USB or Bluetooth would remain the same, so that API remains constant , but now there's data flowing between the core code and separate Monero-Trezor bridge process or something similar 18:12:29 @jeffro256: It might be good to have a separate 'core' repo for these scripts, with Python ostensibly being our best bet 18:12:35 TS and Rust are also available, however 18:12:50 The Monero-Trezor bridge could also be a shared library, although that has some security implications 18:13:21 You mean, Trezor or another manufacturer would be creating that library? 18:13:51 This could also be integrated into the GUI, if people are willing to bundle Python with it 18:14:39 The complexity would still be there, but least out of the core code, right? 18:14:47 *at least 18:15:15 Yes, unless vendors implement 'our' protocol 18:16:03 In my personal opinion, it might be prudent to just keep existing Ledger/Trezor code in the main repo, until it's no longer supported, without adding further support for new protocols 18:16:25 I don't think THP integration will be happening any time soon, and doing so would require a lot of code 18:17:01 Frankly, I currently don't see many other really feasable options anyway. 18:17:36 I have a rough plan for how that (THP integration) could be done (and it would actually not require any new dependencies), but it's like 2-3,000 lines of dense code using libsodium to implement parts of the Noise protocol, with handshakes and everything 18:17:45 I'm not experienced enough to be able to comfortably write that 18:18:42 Ok, probably this needs time to develop further, hopefully including some moves - finally - by the manufacturers 18:19:14 If we are through with the reports, and the hardware wallet discussion, I have a little question to the room 18:19:46 I was preparing a post for Polyseed and remembered we once had a quite long discussion about Polyseed, encryption and plausible deniability 18:20:08 I could not find it again, however, and want to ask whether somebody knows where it is 18:20:33 What I can say that this isn't it, it's already Tevador's reaction to that discussion: https://github.com/tevador/polyseed/issues/13 18:21:02 I think it was a GitHub discussion 18:21:41 Please post here if you know something, thanks! 18:22:09 Alright, something else for today's meeting, anybody? 18:23:06 Does not seem to be the case. So thanks everybody for attending, read you again next week! 18:24:10 Is there any chance it was on Reddit, this discussion? 18:24:20 Pretty sure that no 18:24:46 There once was a post with some interesting opinions titled "I don't like Polyseed" however 18:24:55 Thanks everyone. (sorry, I don't remember which discussion you're talking about) 18:25:09 Ahh, you are much too young for that :) 18:25:33 No, that was 2 years ago, at least, as can be seen from the date of the Polyseed issue 18:26:35 Maybe Google has not indexed it, or not anymore, or my search power slowly bleeds out of me ... 18:27:05 And no, I won't ask an AI, no way 19:14:47 FYI, Polyseed v3.0.1 has been released with 3 bug fixes in the library: https://github.com/tevador/polyseed/releases/tag/v3.0.1 19:27:59 Thanks for the info, tevador!