<reeemuru[m]> "jberman: sure no prob. will..." <- this is the most interesting one: mymonero/mymonero-core-cpp #36

this one is less interesting, just outdated code that doesn't match wallet2: mymonero/mymonero-core-cpp #35

<jberman[m]> "this is the most interesting one..." <- Interesting indeed. I have some old k-means models code in R. I will try repurpose / running on some transactions and see if I can provide any useful info.

Haven't really dug into monero tx structure but looking forward to try and break stuff.

on the topic of rpc vs lws, the rpc system is a functional subset of the lws API afaik

meaning an rpc server can be implemented as backed by a lws scanner

which, shortly before i decided i had to leave mymonero, i implemented here, mostly as something that would be very easily marketable to exchanges who're already coding to the RPC api

what i dont like about the rpc server as i recall is that it holds a wallet open - suppose some code forgets to match an open call with a close call or say the consumer crashes and can't close the wallet,.. that'd be somewhat bad

of course there are ways to resolve that. you could put some sort of challenge token on each call

i've developed a more privacy preserving authentication system for my own systems that i'll be sharing hopefully soon, probably not a bad time to mention it here, which is pretty straightforward.. it's just an interactive process where a server gives a client an ephemeral, fairly random token that the client must sign with its spend key

yeah current wallet-rpc is pretty braindead

the problem is using http-based RPC in the first place, since HTTP is inherently a stateless protocol

i notice that mymonero says the system was never implemented but i suspect it's being sold

it should use an inherently stateful session-oriented protocol

yeah i hear you hyc

a wallet client should open a connection and use it for the duration, then when the connection closes all of that session is disposed of

overall the lws API itself can take some significant improvements which we also designed way back when but somehow they never got rolled out

including a websocket API which has implicit sessions

the websocket client is shown in that repo linked above

so, i'm definitely going to be bringing that back

when the time comes..

reeemuru[m]: i'm curious about what you're working on and would be interested to see if it would be beneficial for us to collaborate

I guess there's no reason we can't write a brand new rpc system from the ground up. it's not like wallet-rpc itself is widely used

yeah lots of this stuff just needs a ground up overhaul

and again I'd model it after LDAP protocol - an explicit login op to start a session, one session per connection, one connection per session

my C++ developer team mate who's going to be introducing a CCS proposal for some wallet2 pure function factors with me got covid so we're slightly delayed on that

but i think the factoring would be a big help in the direction of just overhauling all this stuff

there's no reason e.g. monero-lws should be implementing its own scanning code afaik

indeed

nor any reason, as Justin is seeing, for the lws client to implement its own tx creation entrypoints

all that serialization and data structures can be shared

yeah that's prime code-reuse stuff there

<endogenic> "reeemuru: i'm curious about what..." <- endogenic: i'm spending most of my time [here](github.com/hyahatiph-labs/infosec) and monero codebase. trying to learn as much as i can

looks cool

overlaps with my work somewhat

hah those are cool

oops mistell

for the rpc would it possible to have the request wrapped in an asynchronous thread which also wraps `open_wallet` and `close_wallet` around every `method`. (Would need to send username and password on each request?) Maybe use a new field in the json request context to extend the existing logic for these new requests. As far as syncing clients could call `refresh` at a threshold of blocks behind in background task or scheduler.

Just tossing ideas out.

I still need to look through the rpc code though

that would really suck since the entire wallet cache needs to be decrypted on each open

in the past we've discussed perf problems for monero wallets at busy exchanges

their wallet caches are huge

the biggest improvement in that case would be rewriting the wallet code to store state incrementally in LMDB, instead of slurping it all up at open time and writing it all out at save/close time

I finished implementing the DB encryption hooks in LMDB, for the wallet to use. but nobody has taken up the task of rewriting the wallet to use LMDB

hyc, about this statement of yours some hours ago: "I guess there's no reason we can't write a brand new rpc system from the ground up. it's not like wallet-rpc itself is widely used"

This assessment surprises me a little. I would have guessed that by now the RPC interface is "entrechend" in tons of places

Mostly where anything has to run in an automated way, means most exchanges

Or for most if not all those shopping system plugins

I suppose. but as I already mentioned, the existing wallet code sucks for exchanges

it needs an overhaul anyway

... and not to speak about my personal favorite subject right now, the coming of Seraphis, which again needs plenty of overhauls as it seems

I see scenarious to arrive at a Seraphis wallet that don't look crazy or dumb already at first glance that would lead to moving wallet2 out of the way and start afresh on a green field

wallet2_and_a_half?

jberman[m]: ^

shout out

Well, we need at least wallet4, because there is already wallet3: tewinget is very busy with a full wallet rewrite over at Oxen

With a wallet cache as a Sqlite database. Take that, hyc :)

i think by that standard we're already on like wallet9 or something lol

(thanks to selsta for that link, was interesting to see what they are up to)

like this?

looks pretty good though

not that i have really looked at it in any detail

but it makes one wonder why tewinget didn't stick with monero

hmmmm

i dont understand proof of stake though

" but it makes one wonder why tewinget didn't stick with monero": gras, other side, greener :)

pretty sure he got quite a lot of dough from his ccs proposal though, in case you didn't know

i mean ffs proposal

Yeah, and why not. If they are doing good work for good currency. Which they do I would guess as they are still active member of the dev team.

tewinget? hasn't been active on monero in ages

and using sqlite for a wallet? blech

Yes, that's the point, as far as I know tewinget left Monero for Loki, now Oxen, years ago already and is working there probably on pay

Yes, sqlite complete with triggers to cover some consistency requirements :)

(Not that I recommend that, just for entertainment, as we don't seem to disturb too much research here right now :)

that's not my point

(your subsequent reply)

if he was getting $$ here then why on earth did he have the reaction he did

makes you wonder

it's not an isolated incident

it's sad, anyway

but PoS tho lol wut

endogenic: Loki did lot of presale/mine and made good money to fund devs

Also mindset’s change, so who knows what triggered the move

i'm sure some people know

Any idea if dr. Sarang will ever join mrl again ? or Cypherstack?

yep

thank you for asking, though

endogenic: More bright minds the better

i'm just wondering though, what is it about monero that causes highly devoted people's sentiments to turn bitter or harsh?

yes that's true

endogenic: Community driven projects will have different views imho, hopefully individuals researchers/devs look above it

I dont think that was ever an issue for anyone used to publicly visible and critiqueable work

something else is going on

we probably have to start with the assumption that whatever it is, is not fully visible right now

for example, why would someone feel more safe working on a different project?

it's not just the money

in any case, regardless, i think we could do better for our contributors

i remember people used to debate whether or not to fund the Noethers loil

lol

can you imagine? after some of these other CCS proposals are getting easily funded now

endogenic: Maybe, keeping them anonymous would be better

there really were people who used to quibble over whether we should fund the Noethers

if they were even real

yes nikg83[m] that is one aspect

endogenic: That was absurd indeed

but it has downsides

knowing these people for who they really are we see how absurd it is that we would ever do anything other than rally virtually our entire community around supporting them

There are always people. Everywhere. People doing, writing, saying dumb things. Just look at Monero's subreddit. They don't have to be our reference.

they're not our reference, but i dont understand why you're apologizing for them

The world is like it is. Not a apology for anything.

the world doesnt have to be like it is

it wasnt always like this here either

i'm doing what i can but am a bit behind, so i dont have my solutions to show for it yet. guess i should get back to it :)

If you have work that gives you pleasure, by all means continue.

lol

^ my coping mechanism

Yeah, mine as well.

endogenic: why don't we ask the people that left to come and tell us what we can do to improve.

we know what we can do

i've also suggested it

but it may take a specific person to do it

so i am working on it

midipoet: from what i saw the people that left, wanted to leave, but never really openly explained to the community the reason. the only person that was really open about wanting to leave was surae

maybe i missed the convos, but that is how i saw it.

well if i have my hand on a hot stove then i would *want* to pull it away

bad example maybe

how about something my hand is supposed to actually be on

but which malfunctions

endogenic: yeah, that's a great metaphor, but explains jack

no it's not a great metaphor

i would rather actually hear it from a person

oh i'm not a person now lol

i saw it with my own eyes

i mean, there are the mediums for this sort of talk

and i'm telling you what i saw

you're right

monerotalk, community talks, reddit threads.

idk why you pinged me

they can teach us

otherwise we are just guessing

no we're not

stop lying lol

lying includes saying what you dont know as if you do

or listening to your preaching (just as bad)

i should go

indeed

<endogenic> "if they were even real" <- Judging by what happened with the monero telegram channel, I think there are a lot of people getting paid to derail the convo.... (full message at libera.ems.host/_matrix/media/r0/do…8abff90b42410943b56b48a4836013c4703)

right?

people dont even have to be paid

> The first few days was non stop Nazi Jew spam < really?

it doesn't have to be feds. could be rival projects trying to discredit the network.

or even just individuals trying to discredit the network. it doesn't take much to be shit spreader.

midipoet: It was *a lot*.

And real people that would argue for hours

Ive personally argues with pirate chain ppl, who clearly show up to start shit

w[m]: like, i have sort of just realised that cryptocurrency is adversarial in general. everybody is thinking they are playing a zerosum game, with the winner measured by price. it's easier (and more advantageous) to talk shit about another project, than to try and foster friendships/connections. i mean, we do it as well for most of the transparent networks

It is kind of zero sum for specific use cases

IMO

endogenic: check out @hyc's idea of a background sync cache here: monero-project/monero #8082

it doesnt matter who it is. anyone who disrupts the activists within the community who are trying to wake it up to how it must evolve is doing it a disservice .. and we can't just go on forever saying that the people who left were weaklings or had personality disorders

Like z-cash vs monero

jberman[m]: ty

And check out my comment here, suggesting that it be implemented virtually identically to the lws scanner in an attempt to avoid leaks to 3rd party remote nodes: monero-project/monero #8082#issuecomment-1007921314

I think view only wallet scanning could be implemented like that as well

So in reality, I think the best path on that idea lands us with an lws scanner inside the code base anyway

I don't think it would be a huge challenge to wrap a light wallet API around that background sync cache, where the consumer of that API uses it to sync upon re-visiting, same as a light wallet client

Unless there is objection to that comment I made, and it should really just be coded assuming honest nodes

ah bloom filter rears its head again

vtnerd had that idea in the past iirc.. maybe 4-5 yrs ago

endogenic: > we can't just go on forever saying that the people who left were weaklings or had personality disorders < when have people said this????

i'm sorry but social cooling is now in effect

please return to your quarters

i think we should take this to another chan in any case