Seraphis wallet workgroup meeting #25 - Monday, 2023-06-05, 18:00 UTC about meta HOT 1 CLOSED

<rbrunner7[m]> Meeting time. Hello! https://github.com/monero-project/meta/issues/845
<Rucknium[m]> waves
<dangerousfreedom> Hello
<rbrunner7[m]> A somewhat modest meeting today as it seems ...
<rbrunner7[m]> Anyway, still something to report? Nothing from me this time.
<dangerousfreedom> Haha
<UkoeHB> Hi
<UkoeHB> Nothing from me
<dangerousfreedom> How far did you get with shalit in that interface that you both were trying to build?
<rbrunner7[m]> Who both?
<dangerousfreedom> You and shalit I guess
<rbrunner7[m]> Well, I support shalit when they have questions. But nothing on my own initiative, at least so far.
<rbrunner7[m]> In my understanding there is still a PR from your side to make, as a base for their work?
<rbrunner7[m]> At least that's my understanding from last week's meeting ...
<dangerousfreedom> I had an accident in my shoulder this week and havent been able to stay longer with my hands on the keyboard so I have been structuring my work this week actually. I believe I can make all the generation of the seraphis knowledge proofs just using the EnoteStore but I will have to use the node to do the same for the legacy proofs... so I have just made my TODO list more precise this week 
<dangerousfreedom> But yeah, I will make a PR for purposes of getting feedback this week
<rbrunner7[m]> +1
<dangerousfreedom> There is nothing ready so far but hopefully we will start a cycle of feedbacks/improvements
<dangerousfreedom> I will do it before Friday. Sorry.
<rbrunner7[m]> Isn't there some code that may be directly useful for what shalit is trying to implement? So the PR is for more than just review and merging?
<dangerousfreedom> The code is available on my branch, I sent here and also directly to shalit
<dangerousfreedom> He is using it for sure
<jberman[m]> hello, nothing from me either (was working on serai this past week).  After last week's discussions, I'm planning on trying out using libcurl in the async wallet scanner to get a feel for it as a substitution for the epee http client library. Timeboxing that to a couple hours of work
<rbrunner7[m]> Yeah, and that's not the way this should work, ideally ... frankly, with such criss-crossing of code we will have a mess in no time.
<rbrunner7[m]> Same argument as in last week's meeting, I know, I repeat myself.
<rbrunner7[m]> Or is the idea that shalit will PR that, not you?
<dangerousfreedom> He will eventualy PR the serialization of the EnoteStore at some point I hope
<rbrunner7[m]> With your code parts included then?
<rbrunner7[m]> As part of their PR I mean
<dangerousfreedom> But there will be endless clean up PRs I'm sure. So I will try to make the most basic until Friday and then we will see.
<rbrunner7[m]> Ok. I hope PR cleanup won't acutally be endless :)
<ghostway[m]> hello! sorry for being late, my key container, apart from tests, is done (I can't compile because of some weird stuff going on with dependencies and updates). expect a pr in the coming weeks
<Rucknium[m]> +1
<dangerousfreedom> rbrunner7[m]: Why would he do that? I will do it soon
<dangerousfreedom> ghostway[m]: Nice! But something to show or discuss so far?
<rbrunner7[m]> ghostway: You could consider a "Do not merge" PR at some earlier point, so people could have a first look and impression.
<rbrunner7[m]> I think right now we have non-zero chances that people wander into directions that make other people wonder what the heck is that, if you know what I mean
<ghostway[m]> yes, there is. what do you think about a KeyGuard (something like lock_guard), that unlocks the keys and, by reference, increases/decreases references and when it reaches 0, it locks the keys again?
<ghostway[m]> I'll make a draft pr in the coming days, anyway
<rbrunner7[m]> Out of fundamental misunderstanding, for example
<rbrunner7[m]> I never studied locking in detail in wallet2, maybe jberman or jeffro256 know details
<ghostway[m]> I already have written it, but it could be moot, considering you can just use references (which, with my familiarity with koe, seraphis is exclusively using them in this case)
<ghostway[m]> apparently wallet2 has that already, and is using that
<rbrunner7[m]> If we can have a first look at the code, maybe things will get a lot clearer
<jberman[m]> yep this is what wallet2 does here: https://github.com/monero-project/monero/blob/94e67bf96bbc010241f29ada6abc89f49a81759c/src/wallet/wallet2.cpp#L1063-L1115
<ghostway[m]> yep
<rbrunner7[m]> Yeah, but with some strange user interactivity thrown it, which maybe you can hardly generalize, no?
<ghostway[m]> can you rephrase?
<rbrunner7[m]> Ah, maybe I misread, this here is merely a variable access, not a call, right: tools::wallet2::AskPasswordToDecrypt
<rbrunner7[m]> So no sudden asking the user interactively for the password from deep inside wallet2
<ghostway[m]> yea. the key is passed to it
<rbrunner7[m]> Good
<rbrunner7[m]> Anything else to discuss right in this meeting?
<dangerousfreedom> Not from me.
<rbrunner7[m]> Btw, I see more and more references to Seraphis on the Monero subreddit. I am just not sure whether people are aware how far out the hardfork to Seraphis still is. But nice to see that we "got their attention" :)
<ghostway[m]> btw I don't really like the use of static there in wallet2 (unless we are sure there are no multiple threads, and then why would we want to use that mutex), but I also don't really like to pass a reference/pointer to the chacha_key multiple times, unless it doesn't allocate on copy. can you confirm?
<rbrunner7[m]> I am not the best with C++, but can't you avoid any copies with const&?
<rbrunner7[m]> At least that's what I always assumed when coding ...
<ghostway[m]> true, that was my first version, but it could look better
<ghostway[m]> I guess it's always good to be explicit. consider my question answered I guess
<rbrunner7[m]> Not sure what you mean, but maybe just let us have a look if you are ready
<rbrunner7[m]> Alright, thanks for attending, read you again next week at the latest.
<plowsof11[m]> +1
<ghostway[m]> +1
<dangerousfreedom> Thanks