<UkoeHB> "Although from one point of view,..." <- Any change to cryptography falls into this category. But for some reason bulletproofs+ are not protected by experimental flag. Inconsistency is everywhere

<luigi1112> "let's improve it" <- The best time is before or after hardfork ?

Ooo, being angry doesn't get us anywhere.

You know the answer to that question is "as soon as someone improves it"

<luigi1112> "on a related note, improving..." <- `dedicated_maintainer = random.choice(monerokon.members)` would it work ?

Propose a solution or a direction towards one

The opposite doesnt help. Its like telling a fish that it is wet.

"monero-project/monero #7760#issuecomment-1144626397", "Just that it turned out to be a dead end, only speaks for him, that he can think self-critically, and not that I'm not helping." incompetent developer justify blind approve by saying that PR author can think self-critically; facepalm

.merges

is there a Monero node service??

i need to get UTXOs

MeowingCat: there are remote nodes volunteered by community members (and possibly malicious actors)

or you could be your own node...

i will setup a node but i need one to test rn

im working on building and signing transactions

Some good resources here MeowingCat community.rino.io

looooooking thank youuuuuu

.merge+ 8358 8357 8356

Added

ohhh damnn

this is a node

i hope there is an RPC server like BTC

isn't there a Monero RPC server i can use???

MeowingCat: you have to be more precise with what you are looking for

there is an RPC service in Monero client

buttt is there an online service that i can use rn

i didn't setup my own Monero node yet

i need to get Monero UTXOs (on testnet)

HenryHollingwort, is there an online service for tesnet???

run the wallet rpc server locally and connect to testnet daemon right?

use testnet faucet if you want txmr

but i need to download all blockchain

no

ohh

how??

There's a big list of nodes here, including stagenet and testnet: monero.fail

wallet rpc server ---internet---> [daemon { blockchain} ] ---> peers

ummm

you run wallet rpc locally and point it at a public (or local) daemon

ohh got it

okiiiii

Hi MeowingCat ,

We have #monero channel for this :)

Hi everyone! How would you go about integrating a front end application with Monero? Monero doesn’t have an API, right?

It does have an API: getmonero.org/resources/developer-guides/wallet-rpc.html

There's also a number of higher level APIs, see the monero-ecosystem github user. Also, "monero integrations" for really high level (as in, plugins for a number of online shop software).

<sech1> "It does have an API: www..." <- Thanks. I guess I missed it.

<freedomf8ter[m]> "Hi everyone! How would you go..." <- if you want to get started quickly you can just fork this repo. I collected some scripts there to play around with the wallet rpc github.com/spirobel/monero-playground

you can also take a look at moneropay.eu people work on making the wallet rpc more stable and accessible.

but probably still good to play with the wallet rpc so you get a feel for it. github.com/monero-ecosystem/monero-javascript is also very good.

and also maybe read cryptogrampy s hotshop source code if you are into typescript. github.com/CryptoGrampy/HotShop

weird. CI is failing to compile my last push to #7774 but it compiles fine on my ubuntu laptop

<freedomf8ter[m]> "Hi everyone! How would you go..." <- spirobel: already linked to my HotShop project, but it's essentially an in-browser view-only web-wallet that talks directly to public (web-compatible) Monero nodes.

can anyone tel me Monero's transaction struct

struct monero_tx {

char* m_address;

uint64_t m_amount;

monero_dest* m_destinations;

char* m_payment_id;

priority m_priority;

uint32_t m_ring_size;

uint64_t m_fee;

uint32_t m_account_index;

uint32_t m_subaddress_indices[];

uint64_t m_unlock_height;

bool m_can_split;

bool m_relay;

char* m_note;

char* m_recipient_name;

uint64_t m_below_amount;

bool m_sweep_each_subaddress;

char* m_key_image;

}

(sorry for multiline) is this correct??

this is what i understand from Monero source

that's what pastebins are for. if it's in the source, then it is correct.

i have to be sure :(

because im still dealing to build a correct transaction

wrap your code in ``` ``` or use pastebin

this thing loooks like something different

something like a command for another interface

chesterfield[m], okiiiiii

no, never do multiline pastes, quoted or not. just use pastebin.

For some reason `-DCMAKE_EXPORT_COMPILE_COMMANDS=ON` isn't exporting anything for the tests folder... anyone have any clues what I should do? Apparently it used to work, but who knows what I did...

i just need memory representation of a transaction

like C types or byte lengths and damn endiannes

MeowingCat: for serialization?

in this context serialization means encoding the transaction as hex

i think

i will serialize it and send to RPC server

after signing

im making a C# library for our thing also

just transaction building/signing is left

has anyone tried to call libwallet API from Rust using CXX ?

please anyone tell me Monero's transaction struct :(((((((((((((((((((((((((((((((((((((((((((((((((((

MeowingCat: you probably want to check cryptonote_basic/cryptonote_basic.h class transaction;

UkoeHB, looooooking thank youuuu <3

i really don't understand why this source is like this

everything could be more understandable

with just C

just a struct and serializing it is very simple

using C++ was definitely a poor choice

:((((((

but it's what we inherited from cryptonote

no one has the time to rewrite it all in plain C

i understand Monero software has tons of stuff and using tons of libraries for them and all of things is very complicated and toooooo mcuh work to implement

but C++ and ridicilious design patters makes everything worse

yep. totally agreed.

im really happy to see an intelligent person

ok found my problem, generating the build tree with `cmake -B build -S . -D_....` was the wrong choice, I just need to rely on `make`...

UkoeHB, there is no any transaction struct here :(((((

is `transaction` class the thing that is being serialized????

and holds things about the transaction?*

yep pretty much

has to all be crunched by a bunch of libboost serialization code

PITA to pot to other languages

port*

UkoeHB, this is making me cry

just 5-10 lines

that tells the memory representation of a transaction

MeowingCat: dealing with monero code requires time and patience, no way around it

We really need a proper protocol spec, written in BNF

might be worth putting up a bounty for that

If you do, it should include a requirement for tests that pass every tx in the chain.

makes sense

Well, not every one since it'll have to be done manually I guess. But a sample of known txes that exercise various stuff.

moneromooo: easy enough `bool validate_tx_test(const cryptonote::transaction &tx) { return true; }` :p

however, it would be good to get a parallel tx validator that can be manually validated

i derived things

isn't that python project already doing that?

pub/sec spend/view keys, pub address

MeowingCat: you can look at the json representation to get started, although idk about the wire format

does RPC service accept JSON?

i will just UTXOs

get* UTXOs*

build and sign transaction and broadcast

sounds like you're reimplementing libwallet

yess im making a C# library

i will use that in the app im making

im using C# because Godot has Mono and it is cross-platform with Godot

developer.bitcoin.org/reference/transactions.html,

i need an explaination like thissssss

hyc: yeah the python library is moneroinflation.com, although there is no spec to match against

MeowingCat: sounds like something you could work on

:((((((((((((((

please someone stopp C+++++++ and design patterns shits

CI macos build fails due to not finding OpenSSL

CI windows build fails due to no def of std::numeric_limits ?

not giving me warm fuzzies here

MeowingCat: for the next gen protocol, I'm trying to do things a little more cleanly (not that it helps you :p) github.com/UkoeHB/monero/blob/9e185…c/seraphis/txtype_squashed_v1.h#L70

loooooking

UkoeHB, is this a transaction structure?

yes

m_input_images, m_outputs, m_balance_proof, m_image_proofs, m_membership_proofs, m_supplement, m_fee, m_tx_semantic_rules_version

i need to build the transaction with these???

what does "input_images" mean?

For seraphis you would, it's experimental and WIP. The current protocol has similar components.

is "m_outputs" UTXO inputs????

Input images are masked versions of the real inputs you are spending (including the key image). Outputs are the literal outputs - the new enotes (aka UTXOs) created by the tx.

monero has a lot of cryptography, so there is a lot more in a monero tx than a bitcoin tx

you're probably better off using a foreign function interface and calling monero libwallet from your app

would be a waste of time to implement all of the crypto mechanisms from scratch in C#

yeah, there's a good reason no one has every written a full third-party monero wallet

ever*

even if it was easy to do it wouldn't be a good idea. would make fingerprinting possible.

most of that can be avoided with strict semantic rules

although decoy selection is a weakness

decoy selection, fees, and tx extra

<hyc> "even if it was easy to do it..." <- I disagree. Any activity which exposes weaknesses is a great idea. Such endevors would hasten the development of a more secure specification.

That... is a fair point.

Where it gets dicey is for monerod.

Shoutout to dangerousfreedom here :D

I think that's a good sweet spot.

yeah I suppose if we actually had detailed protocol specs it would be viable

so that is clearly the next step before one can safely create a new implementation

This week I almost had a heart attack. I had scanned 200k borromean signatures and suddenly one failed. I tried to rewind the blockchain to print exactly what was happening but it was faster to start from 0, when I reached my desired tx, I printed every point multiplication and realized it was an error in my code. The scalar reading was not reading the same value as it was in the blockchain. I still don't know why because it has

worked for hundred of thousands of txs... I will take a look another time :p

lol

and with some languages (e.g. javascript) having completely braindead data types and arithmetic behaviors, even a precise spec may not be enough

or you have to define the spec down to the behavior of basic arithmetic ops

MeowingCat: on deeper thought, fully reimplementing monero txs in C# is a huge job. You'd need to implement both CLSAG and bulletproofs(+), in addition to output scanning, tx building logic (there are non-trivial details e.g. decoy selection, fee calculations, when and how to make special change outputs...), etc. Wrapping monero libwallet seems like your best bet.

:((((((((((((((((((

i will get UTXOs with sending private view key to our node

You didn't want to read Zero to Monero, but it is a long book because there are a lot of complicated things to think about.

does that need something that i can't get from RPC service???

i think the important question is thissss

does building a transaction need something that i can't get from RPC server?

like monero-lws? github.com/vtnerd/monero-lws

You can do pretty much everything from RPC. It depends how low level you want to go. See the functional tests for what can be done. This is all done via RPC.

But "biulding a transaction" here is "please send N to address A".

the only thing i can send to RPC server is private view key

The RPC server would have your secret keys in the first place.

But I concur with "Wrapping monero libwallet seems like your best bet".

MeowingCat: moneroecosystem.org

you mean i can't build a transaction with things from Monero's RPC server with just sending private view key???

Correct. If you could build a valid tx without the private spend key, I think it might lead to pessimal action on chain...

nooooooooo

i don't mean that :((((((((((((((

i just need to build a transaction without sending private spend key anywhere

i can send private view key to RPC server

and get UTXOs

and whatever i need

then build and sign a transaction

Wait. I might have assumed too much:

just like MetaMask

MeowingCat: did you hear me? To build a transaction, you need to implement all the monero cryptography. This is months of work.

In the case I envision you in, you'd have two pieces of software:

- monero-wallet-rpc

- your own software

monero-wallet-rpc would (1) have the spend key and (2) build transactions

The side which builds the tx obviously needs the spend secret key.

UkoeHB, whyyyyy it is just a simple thing what do i need else than UTXOs???

cryptography

Now, what was suggested above (wrapping wallet2) means those who pieces of software would merge into one.

all this stuff: github.com/monero-project/monero/tree/master/src/ringct

Then you would not need to send the spend secret key anywhere, since you'd make the tx. I'm sorry if I'm getting the wrong idea, it's not qiute clear to me which case you're in.

But when I said "Correct. If you could build a valid tx without the private spend key, I think it might lead to pessimal action on chain...", I meant "if another piece of software is gonna build the tx, it's gonna need the spend key".

I was assuming you'd talk to monero-wallet-rpc here, which needs the spend key (and already has it).

yesss

It's the same for bitcoin fwiw. I assume. You need the spend key to create a tx.

Well, the key. They have just one for view and spend AIUI.

i mean the client appp im making has private spend key and needs to build and sign a transaction

So you have two main solutions:

how much different a Monero transaction than Bitcoin transaction??

1: run monero-wallet-rpc on the side, which has the spend and view secret keys, and you direct it to create txes

MeowingCat: the difference is enormous

2: add FFI to wallet2 and call those directly from your process, keeping spend and view secret keys in that process

doesn't monero-wallet-rpc thing need to send private spend key to RPC server?

MeowingCat: in monero, the RPC server is just a wallet interface

monero-wallet-rpc IS the RPC server.

yesssss it is just an interface to use Monero appppp

Well, there areo two RPC servers. The wallet's and the node's. The node will never need your keys.

That might be the cause of a misunderstanding :)

and in this context the thing that builds and signs transactions is Monero software

"Monero software" is not precise enough here.

i mean private spend key can't get out from user's wallet

It does not need to leave monero-wallet-rpc's process.

really???????

The process that creates the tx needs it. No other one does.

im confuseddd

:((((((((((((((

i need to drink cola

I think the confusion might have been from me assuming you had some software which had the spend secret key, *plus* moneor-wallet-rpc, which will also need to know it.

But your software does not have to know it, unless you want it to know it for other reasons (eg, creating it in the firts place).

the wallet im making has all keys

and should not send private keys except private view key

you mean monero-wallet-rpc thing doesn't need to send private spend key and can send transactions?????????????????????????

Then the process creating that wallet will know the keys. If it's monero-wallet-rpc, you're good.

No, I am sauing the opposite :S

im confusedddddddddd

the process is

you mean processsssss???????

Mayyyybeeeeeee

im deriving keys in C#

lmao

i can just pass them to monero-wallet-rpc thing

and do you mean that can send transaction without sending private spend key anywhere??????

You will have to, unless you reimplement most of wallet2. BUT:

Can you get away with letting monero-wallet-rpc create hte keys for you ?

ohhhh

If you can, then only one process will ever know the keys: monero-wallet-rpc.

why does it matter??????

private keys are in the processssssss

i meannn

Which processssss ?

:(((((((((((((((((((((

client process

!!!!!!!!!1

user's wallet appppp

That's your C# code ?

yesssssss

just like MetaMask

(Your keyboard's fucked btw)

it is same

(or, if not your keyboard, some wetware)

Anyway, to recap:

wdymm??

1: whatever creates the txes must have the spend secret key

2: if you want just one proicess to have the spend secret key, it'll have to be hte same process that creates the txes

3: so either FFI wallet2 to merge C# and wallet2, or have your C# process ask monero-wallet-rpc to create the keys itself

Is that clear ?

i mean does wallet-rpc thing create transactions and sign with private spend key or send transaction information and private spend key for the RPC server to build and sign?

monero-wallet-rpc creates transactions and signs them with the private spend key. monero-wallet-rpc is a RPC server and so doesn't need to send itself the keys, it already has them.

If by "the RPC server" you mean monerod, the monerod does not need the secret keys. Ever.

ohhh

i thought

you mean

this thing with monero-wallet-rpc

this thing doesn't do that right?

I do not understand that question, it is too vague.

(rephrase with more specificity)

okiii

does this lib: github.com/monero-ecosystem/csharp-monero-rpc-client

build and sign a transaction without sending private spend key??

I do not know that lib, but from the name it's probably something that sends RPC. As such, it would not have access to the secret keys.

That is my ASSUMPTION from the name only.

i think it is sending private spend key :((((((((

If my above assumption is correct (it might not), then it would not build and sign a tx, it would ask a running monero-wallet-rpc client to do so.

If my assumption is correct, it would not send the secret spend key. monero-wallet-rpc would have it already in the first place.

(and the process using github.com/monero-ecosystem/csharp-monero-rpc-client would NOT have it)

Again, this is all based on my ASSUMPTION based on the name of that project.

If someone actually knows that project, they can speak with authority.

:(((((

thank you alllll

Also, as said above, you can bypass all this by calling wallet2 directly (that's what I'm doing with Townforge, the wallet that uses the urho3d game engine).

It's simpler, only one process has the keys.

But if you don't, you can still ensure only one process (monero-wallet-rpc) has the keys. You just need to ask monero-wallet-rpc to create the keys for you, instead of creating them with your C# code.

In all cases, monerod never sees those keys.

I know that lib, github.com/monero-ecosystem/csharp-monero-rpc-client. It's just monero-wallet-rpc calls turned into C# calls. A mere wrapper.

I use that extensively for my web wallet

Like we have other such wrappers e.g. for Python

Or Java. Or JavaScript.

<hyc> "using C++ was definitely a..." <- not using rust was a poor choice :(

if the rewrite-in-c crowd gains more traction that the rewrite-in-rust crowd, i'm suggesting rewrite-in-elixir just so no one's happy with what language we're working with

:P

All jokes aside, I do believe C++ was the right choice for a piece of code like this. I believe C++ can be much easier to wield as Monero needs than C, even if C is simpler in a lot of regards. I'm also scared to think of the amount of memory we'd be dealing with in C. While I understand memory can be, and is, used properly, I legitimately believe a C implementation would have less developers participating.

I dunno, GW-BASIC has the advantage of being pretty old, so lots of time for review. It's probably the safest.

Let's rewrite it in GW-BASIC.

concept ack

can you have a prototype by monday?

Fuck no.

preferably with seraphis so we don't add delays there, and also with quantum security because might as well if we're rewriting

boooo

GW-BASIC is such an old, safe language, it needs careful implementation. 2062 earliest.

But direct and honest :p Thanks for always being yourself, mooo :)

I'll need a salary till then ofc.

and we can't expect results before then, so obviously it should be with tenure to ensure the 2035 community doesn't try to revoke it

I'll go beyond poke and hope, and enter to the poke and peek and hope epoch...

Is it possible to add more settings for localnode in advanced mode like in/out_peers , blocklist, db sync mode ? Same way like blockchain location is there

New users don’t find it easy to specify startup flags

All of these already exist as command line parameters. If you mean adding console commands, in/out peers exist, not the others.

IIRC blocklist was a bit annoying (I looked). db sync mode seems doable fairly easily.

Blockchain location, not without a restart.

is the inventor of this thing an alien?

yes

goooooooooooooooooooood

why everything about Monero (except itself) can't compile on anywhere????

Monero's itself is interesting tooooo

im getting 1982319823918 warnings

and totally not portable or usable with anything :(((

Impressive. It'd take more than a terabyte even with one byte per warning.

:(((

<moneromooo> "All of these already exist as..." <- Yep, but normal users can’t use cli that well

Sorry for posting here, was supposed to post in the gui channel

Ah, right. Dunno about the GUI then, sorry.

/mnt/c/proje/mymonero-core-cpp/contrib/monero-core-custom/epee/src/wipeable_string.cpp:147:67: error: ‘::max’ has not been declared; did you mean ‘std::max’?

that numeric_limits error sounds like what hyc was talking about

if this code is wrong how does it work in somewhere

that code is put of date at least

i will make a kilobytecoin

#include <numeric> IIRC for numeric_limits.

my private version is up to date but i'm a hairs width away from being able to publish it

should be <limits> (which is in that file) en.cppreference.com/w/cpp/types/numeric_limits

MeowingCat: are you compiling master or a release version?

master of mymonero-core-cpp

let me try a release

Then why are you complaining about the monero project if you're compiling anothre related project ? :D

<rbrunner> "I use that extensively for my..." <- Link to this wallet?

:(((((

because Monero is making me cry

no doc

and devs who are saying me implementing a tx building and signing is impossible

Well, it's Esperanto for "In your face"...

MeowingCat: perhaps i'll give you access to my version of monero core cop and core custom

cpp

it is written by an alien and impossible to implement in another platform

i solved the problem if embedding

and something like that :(

i wrote the code repo youre using lol

Wow. Those devs are ignorant then. It's possible. Proof by example: wallet2.

idk why you're missing a std

mooo he's talking about without including the entire world

and without libunbound lolz

why wallet"2"????

what does that mean???

MeowingCat: watch my talk lol

endogenic: I'm only being sarcastic here :)

endogenic, sorry

im crying

haha he doesnt know that mooo

it's ok meowingcat

endo's here

there are interesting issues

I half realize that.

i really wonder about

can i build it for Android, Linux and iOS

MeowingCat: m.youtube.com/watch?v=DY0iE0cBXbc

yes you really can

if you use my codebase

tho well for linux you will possibly want to put multithreading back in idk

what is epee library???

that's really just commented

MeowingCat: epee is a type of sword

im really curious about these things

you might call it a tool

or a util

el oh el

endogenic: that's the idea :D

trolling now

man i wish i could release my work wtf

i mean.. i guess i can now

> what is epee library???

In the author's own words: "epee - is a small library of helpers, wrappers, tools and and so on, used to make my life easier."

well, almost

one last thing

legal related

give it 1-3 days

Someone who knows about git grep ^

Learning about git grep was a life changing expereince

"to make my life easier"

but not necessarily simpler

git bisect was mine. Thank, whoever you are...

(but git bisect would be mostly pointless without ccache (which has works for years fwiw), so thank you whoever too)

endogenic, can i cross-compile it on Linux?

sure

damnit

dammit*

i really want to give you my updated docs

it details exactly what i changed

error :(

youre getting tons of weird errors

chances are this is upstream of "my" code

that is

in your cmake lists or in your source code's global includes

it is 1.1.2 branch

nothing changed :(

on Linux

ok just fyi that code is way old and my access to it was revoked

ummmmm

my open source PR feedback comments on it were literally censored

i tried to use libwalet.a thing before

i cant vouch for it, idk what changed but

but Monero source is really interesting

include paths are not relative

and more things

yeah you could call it interesting lol

check out mymonero core cpp cmakelists.txt

how i include headers and sources

i dont use all of the monero embedded cmakelisfs

though i would like to

i'm gonna bounce to another location now. feel free to leave questions

okiii

<endogenic> "man i wish i could release my..." <- Plz sir

rbrunners webwallet is, i believe: monerotech.info/Wallet

Y hold back code you’ll eventually release? 👀

plowsof[m]: Why aren’t there any wallets that use wallet rpc?

what is making building transactions complicated this much?

anyone can't implement the same thing in another language

an alien wrote it and we have to use it somehow

endogenic, you have a private repo??

chesterfield[m]: Looking for a technical answer here… seems like a selfhosted wallet rpc would be a very low bandwidth solution for a wallet on a phone

if i was making an exchange it would be way easier

because damn build target would be just Linux

chesterfield[m]: do you mean connect to a private server running wallet rpc, and the phone wallet is just an interface for that?

UkoeHB: Exactly

probably doesn't exist because it isn't commercially viable

i will rewrite Monero in future

in C

zero issues on building

zero warnings

maybe after that i can make a Kilobytecoin

MeowingCat: yes, a few of them, some very useful stuff

them, -> them.

github.com/rohanrhu can you give me access???

MeowingCat: rewriting C is the idea of one of my team mates as well

in C

libmonero

C++ is dumb

starting as an alternate implementation would be fine

C++ is not dumb

it is dumbbbbb just making everything problematic

chesterfield[m]: is that not better solved by the light wallet scanning server?

either way if you want to use the RPC interface then I wrote a server that pretends to be wallet rpc server and is backed by a light scanning server

actually i can even use CLI with just Monero executable

or run RPC server and communicate

butttt iOS doesn't allow that

what do you actually want to do MeowingCat

wallet

he just wants to compile monero core crypto

BTC, ETH and Monero

MeowingCat: if you share your repo with me then i can at least look and see why you are struggling to compile mymonero-core-cpp

but Monero is not like BTC and ETH :(

i willl

endogenic: the light wallet server that doesn’t support sub addresses and has more identifiable transactions?

it will be open source

it does support subaddrs

i've had that support built for quite a while now >:)

it's also very very cool

jberman and i shared info on a spec for it

it's pretty simple

i incluuded damn stdexcept in cryptonote_config.h and it is building now

i hope it won't stop somewhere

as for privacy tradeoffs, well, you just run your own server

and the txs afaik are not more identifiable beyond that

it's supposed to be "native" monero

MeowingCat: woooo

idk how Monero UTXOs work

yeah i noticed another missing include myself

well, team member did

<memory> in wipeable_string.h

gotta do a PR

and other things buttttt everyone is saying building Monero txs is extremely complicated

not sure how it happened for him and not me though to be honest

MeowingCat: it's not complicated if you know how

why there are codde errors in release??? :(((((((((((

because only a couple people actually try to embed monero code

well, couple libraries

ohh another error

feel free to DM me fwiw

maybe i should make a new IRC room tho

if this is annoying

i don't think so because here is silent

this room is more-so for dev on monero code itself

rather than solving build errors in a third party's lib

in/on/using

endogenic: When pr to lws ?

not gonna pr to lws

but gonna pr to meta

to lws spec

i have a few of them

soon(tm) trying to do a couple other things for the moment

nothing wrong with C++, it's a tool that can be misused like any other tool (plenty of trash C out there)

losing a team member (family emergency) so i have to try to do a couple things asap

included limits

now another error

im feeling like shit how this is a release :(((((((((((((((((

share your (c)makefile plox

:)

that makefile builds test targets

not your application target

are you trying to use it?

it's good feedback, if you are

means i gotta document that better

just doing cmake and cmake --build

oh, in that repo?

yessss

oh ok good to know, 1sec

in build direcotry

this is a build script fwiw

yessss

yeah idk why that's erroring for you. i havent tried that repo recently. i also am probably on a different dev env than you

i mean, i can tell you

actually nevermind

i was mistaken about what i was just thinking lol

maybe something else needs to be included in find_package(…) under your dev system

<endogenic> "but gonna pr to meta" <- Does this mean there will be code available for people to run an lws-like thing on their own or not? I can’t tell if you’re going to release code or a custodial solution

chesterfield[m]: there already exists a monero-lws impl but i think justin hasn't released his code for implementing subaddrs on it yet since he was waiting on people to OK his/our meta spec PR proposing the shape of subaddrs on the LWS API

as for me we've built internally a new LWS implementation from scratch and it already does subaddrs

and new client code that fully supports it, which will be open source

github.com/vtnerd/monero-lws

Cool so you’re doing a closed source custodial implementation with an open source client api?

it's not custodial of the spend key

it's just for user convenience. i am highly prioritizing client side scanning after that

Closed source non-custodial. Got it 👌

for the server yeah

it's all just to support something else i'm building lol

endogenic: Why so much secrecy on what you are building 😅

oh yeah sorry

i forgot to reply to that

Monero could really use a good open source light wallet server. Hope you reconsider releasing the code for it.

endogenic: So the specs will be presented and a open source implementation will be done on lws? There is a do not merge pr for subadress, is that waiting this specs ?

there are the typical reasons why companies like Apple would keep something secret before they release it, and then there are other reasons right now specific to my situation related to precautions I've been taking to protect myself, the code itself, and protect the emergence of a project that's being built on the code that I want the community to be able to have. finally for the scanning server i investigated foreign incorp.

so that if we held viewkeys for any new users that the jurisdiction would provide optimal insulation against e.g. dragnet surveillance

i spent a lot of money out of pocket researching that already tbh :P

it's going to have to wait, though, partially due to being slightly resource cautious presently

even though i found the answer

chesterfield[m]: i said one already exists

vtnerd is an excellent engineer

i dont know how i could improve vtnerd as an engineer lol .. put it to you like that.

he's like 9 or 10/10

nikg83[m]: the spec has existed for many years

it's part of monero meta

Lessen his focus on speeding up stuff that doesn't show up in the profile ^_^

and many servers implement it now

moneromooo: lol

(but generally, I agree he's pretty good)

does libwallet.a have things to build and sign transactions???

<endogenic> "there are the typical reasons..." <- Centralized light wallet servers are bad for Monero (until we get seraphis limited view keys)

sure if you dont trust the operator and if the setup is insecure

No, they’re bad period.

proof?

Proof that an entity with a desirable product that has thousands of view keys is bad for privacy of the system as a whole?

yeah i mean what does that have to do with centralized scanners. sounds like a monero flaw to me

whatever happened to pragmatism

you guys remember that?

whatever that was

anyway, guess who funded vtnerd/monero-lws initially?

libwallet.a does have stuff to build and sign transactions.

I don’t have a problem with lws, but I do have a problem with trusted entities holding thousands of view keys

you have a problem with untrusted entities

people will choose to use a centralized server like choosing venmo or paypal

If you’re not allowing people to selfhost your lws solution, you’re undermining system privacy

if it compromises others' privacy then you need to realize you're saying people shouldnt hack your vulnerable system

monero has many such attack surfaces

endogenic: hence the need for decentralised lws, you might feel users need another mymonero but it won’t be that well accepted

related to ring sigs

RINO has entered the chat

nikg83[m]: there already exists decentralized scanners , what is not getting through to you about that

my infra is for private customers

and yes it can do multisig

guys, drifting off-topic for this channel

indeed

thanks