luigi1111w: Did we decide on v0.16? If yes, you can branch whenever you have time

unclear

Hi guys, i am trying to compile cakeWallet from source but there is no documentation on how to do that, how do i compile monero api libs and boost and all that, is anyoane here from Cake Wallet project?

you can ask on their telegram

what channel?

Thank you @selsta

->#monero

||

Hello

selsta, luigi1111w: My vote is for v0.16, as it will probably incentivize people to upgrade, thereby strengthening dandelion++ for the network

+1

does 16 mean a new name?

Lithium Luigi.

Though we missed the boat for 0.11.11.

lol

it requiring a new name doesn't stop anything, just something that we need to figure out before actual release

I would stick to major version = hard fork = new name

ErCiccione[m]: releasing it under v0.15.1.0 will mean less people upgrading and less people using Dandelion

if we do less hardforks in the future then we shouldn’t get stuck on the same name

Fair enough. But in that case i wouldn't use another name or people will definitely think major version + new name = hard fork

There will be confusion either way, but i think using the same name would limit it

we can decide that later

meeting this sunday?

18:16 <ErCiccione[m]> meeting this sunday? <-- yes

Reposting stuff I said in private convo. Just my 2c on version:

Historically there have typically been hardfork protocol updates together with major versions. Which explains why it's tempting to see them as interdependent. But afaik this is not a given, this was never explicitly agreed or communicated or consciously planned that we should keep doing these two together forever.

As the codebase/community/protocol/etc are maturing we are going to do less frequent major versions, that's understood. But we will probably also do less frequent protocol upgrade requiring hardfork. So that question of bumping major version or not might become a regular one

I see the "marketing" value of a switch to 0.16, which better reflects the improvements introduced with a lot of hard work. And helps getting people to run new version.

I think the incompatible-protocol-or-not rule is too limiting to use as definition for what constiutes something major. We might have minor improvement to Monero coming requiring a hardfork, and on the other hand massive improvements that are backward compatible without a technical hardfork.

It's not the end of the world if we go for 0.15.1. But we'll struggle more to get news outlets to write about update, Dandelion and all the improvements made. And generally people to grasp and recognize the importance of the work done.

to play devils advocate, i don't think we should concern ourselves with what news outlets will write about monero - i think we should focus on what makes the most sense for the ecosystem. though as i write that, i could consider that the news is part of the ecosystem because it gets the word out blah blah

or we could just put a cork in it and bump the ringsize by 2, citing "moores law" and whatever the network bandwidth law is

don’t want to do a hard fork with CLSAG upcoming

else we will possible have 2 hardforks in 3 months

also looks like CCS for CLSAG audit will be up soon

well if clsag is coming soon (tm) we might also want to consider not having a highly recommended upgrade followed up a necessary upgrade

why not?

upgrade fatigue?

"these monero devs just always making me update, grinds my gears!"

3 months between major updates is not too bad

3 months between hardforks would be bad

can we monitor versions in the network? Would be interesting to see how a "highly recommended but not necessary" update behaves compared to a "necessary" update

if I would guess people most of the time only update if something breaks

do nodes advertise their sw version? I thought they still did

only RPC version IIRC

sw version only when unrestricted but I could be misremembering

Someone forcefully asked for a non RPC version RPC IIRC. Kind of a selfish reason IIRC (so I only connect to the version I like).

Kinda made a middle ground by tagging release, I believe. If that ever went in.

RPC version still changes often so you can still get human version by mapping though.

That's RPC though. For P2P, you'd have to fingerprint.

and I suppose the idea was to prevent p2p fingerprinting

IIRC, it had to do with Cake Wallet and another "light" wallet that had trouble with incompatabilties b/w wallet and remote node during the upgrade period

It was a third party wallet, yes, dunno which one. But the point of the RPC version is to know what incompatabilties you face.

Not 100% foolproof of course, I often forget to bump the damn thing even though I added it...

But this system means someone suddenly finds themselves unable to use their node for no good reason.

Anyway. Kinda minor I guess.

I'm seeing all sorts of dandelion-related errors on compile with most recent master... anyone else seeing any?

logs?

Preparing now

I added some unrelated changes and want to confirm a fresh build without them

Ah, I think I see the problem, and it's entirely mine :)