Overview and Logs for the Dev Meeting Held on 2017-02-19

Сообщение от: dEBRUYNE / fluffypony

Logs

<fluffypony> 1. Greetings
<fluffypony> 2. Brief review of what's been completed since the previous meeting
<fluffypony> 3. Code + ticket discussion / Q & A
<fluffypony> 4. Any additional meeting items
<fluffypony> 5. Confirm next meeting date/time
<hyc> hola
<fluffypony> so greetings
<fluffypony> hi!
<ArticMine> hi
<tewinget> (here)
<vtnerd> also present
<moneromooo> hi
<tewinget> I'll be sorta afk for the next 5ish minutes, but I'm around.
<fluffypony> ok
<fluffypony> 2. Brief review of what's been completed since the previous meeting
<fluffypony> so
<fluffypony> second meeting of the year
<fluffypony> and we've been pushing ahead solidly
<fluffypony> we've had a bunch of PRs by newcomers
<jollymort> once you go solid state you never go back
<fluffypony> including tpltnt, and IPGlider has also pushed a few PRs
<fluffypony> we switched to EasyLogging++, which is a pretty big change
<fluffypony> and MoroccanMalinois made Android builds happen
<fluffypony> tdprime also pushed their first PR
<fluffypony> and then the usual rash of PRs from moneromooo, vtnerd, hyc, NanoAkron, and Jaquee
<fluffypony> (I've probably missed someone)
<pigeons> reveler with the background mining
<hyc> revler
<fluffypony> yes thanks pigeons
<fluffypony> oh and pigeons had a PR too
<jollymort> kenshi84 disposable addresses
<fluffypony> and kenshi84
<fluffypony> ok - anything else major I missed that happened in the last two weeks before we move on to 3?
<moneromooo> All the new one time address stuff from knaccc, randomrun, kenshi, jollymort.
<knaccc> yes subaddresses are back!
<moneromooo> And luigi.
<fluffypony> moneromooo: I was going to get to that in 3
<hyc> ok, sounds like we move on to 3
<fluffypony> since it's in the MRL repo
<fluffypony> 3. Code + ticket discussion / Q & A
<jollymort> I wasn't following that discussion, focused on study on fees atm
<fluffypony> ok so there have been a few long-form discussions going on in various issues
<jollymort> also the one for ring size
<fluffypony> yeah the ring size one is one I wanted to bring up
<fluffypony> I think the discussion has mostly been positive, nobody's gotten crazy out-of-hand or anything
<fluffypony> it is a complex topic
<fluffypony> and I think that we have enough time to figure out a good route forward
<fluffypony> does anyone have an objection to it continuing in the GitHub issue?
<hyc> seems like all the context is there
<hyc> so it should continue there
<jollymort> I feel like it's more suitable for an issue under MRL, any thoughts of moving it there (if possible)?
<jwinterm> I haven't been following the issue too closely, but is there still some sentiment building around fixing ring size for all txs?
<fluffypony> jollymort: I think the GH issue is fine, it can kinda be anywhere, but if we're going to turn that into a publication that explores the various options and reasons for recommendations then it would develop as PRs to the research-lab repo
* jwinterm goes to github
<jollymort> I meant, keep it as a GH issue, but under another repo so it doesn't get buried under all code/bug related issues
<jwinterm> as someone not following the issue, it does kind of get lost in the noise with 164 open issues
<jwinterm> #1673?
<fluffypony> yeah it would be nice if GH let you subscribe to just a single issue
<fluffypony> I'm not opposed to moving it to the research-lab repo
<fluffypony> how would we do that tho
<hyc> no idea
<pero> someone creates a synopsis of where the discussion is at currently in the new ticket and links back to older ticket
<hyc> yeah, new text
<hyc> with @mentions of original participants
<fluffypony> ok I'll suggest it in the thread
<fluffypony> then on the topic of discussion
<fluffypony> I'd like to encourage us to also take some Q&A / discussion items to StackExchange where we can
<fluffypony> and to redirect people who open issues to just ask a question to StackExchange
<hyc> hm, I would expect that SE is for "settled" questions
<fluffypony> SE is a great place for canonical information that is updated over years
<jollymort> I'd just like to add that SE is not really a format for discussions, more for things with an actual answer existing
<fluffypony> hyc: nope, anyone can edit an accepted answer with new information
<jollymort> like hyc said
<jollymort> there's SE chat, though - which nobody uses
<fluffypony> jollymort: some of the questions on GitHub issues are perfect for SE
<jollymort> sure
<fluffypony> https://github.com/monero-project/monero/issues/1751 as an example
<hyc> monero clone?
<fluffypony> hyc: probably, I thought that too
<fluffypony> but that's a good question for SE
<fluffypony> which also has a larger group of answer-ers than the GH repo
<taushet> it is already answered though, sort of http://monero.stackexchange.com/questions/2886/how-can-i-create-a-new-monero-genesis-block
<fluffypony> taushet: yes but SE has tools to close as a duplicate and redirect them to the answer
<fluffypony> and moderators can do that without us needing to
<taushet> fluffypony - agreed. also the 'issue' was not so much an issue with the code as much as it was a question but a user/tinkerer who could not get something to work
<taushet> anyway
<fluffypony> yeah exactly
<Slack> <nanoakron> What if someone went through and opened parallel SE questions for all those types of question, redirected the original asker, then we shut the issue?
<hyc> sounds good
<fluffypony> @NanoAkron yes that's exactly my recommendation
<Slack> <nanoakron> Ok I might see what I can do if I get any free time tonight
<fluffypony> then you even get SE karma or points or rep or whatever it's called
<Slack> <nanoakron> Woo!
<fluffypony> gamification ftw!
<fluffypony> ok anything else under Q&A ?
<hyc> specific tickets?
<hyc> like 0MQ PR?
<fluffypony> yep I believe tewinget said he had to check if it was working with RingCT
<fluffypony> tewinget: ^^
<Slack> <nanoakron> Any thoughts about 0.10.2?
<fluffypony> @NanoAkron yes
<fluffypony> we've been discussing it
<fluffypony> because it will coincide with beta 2 of the GUI
<Slack> <nanoakron> Oh nice
<fluffypony> as we're marrying daemon / GUI versions
<Slack> <nanoakron> Makes good sense. Will #1746 be addressed too?
<jollymort> do you intend to code in stuff for the next HF in 0.10.2?
<Slack> <nanoakron> I.e. Auto starting daemon
<fluffypony> there are a few things that need to be done in the daemon / GUI before the next release
<tewinget> sry
<fluffypony> jollymort: no
<fluffypony> we only have to finalise that by like July
<tewinget> just got my second monitor back from a friend, was setting it up real quick.
<jollymort> ok, thanks
<fluffypony> @NanoAkron I don't see why we can't make sure 1746 is sorted, Jaquee any thoughts ?
<tewinget> so Snipa was kind enough to chuck a battery of tests at my zmq branch, which is great. It seems there are a couple things I need to look at, which is expected, but his tests seem rather comprehensive, so once those are passing it should be good to go.
<moneromooo> Does this keep a backward compat layr for the current JSON API ?
<tewinget> moneromooo: currently it neither replaces nor modifies any current RPCs
<Slack> <nanoakron> Esp since the number of rpc commands has increased
<fluffypony> moneromooo: long term yes - the current JSON API will be in its own binary, like monero-rpc-server, and that will use 0MQ to communicate with the daemon
<tewinget> but also short-term yes because I haven't done anything to the existing RPCs
<Slack> <nanoakron> I heard it would be passing plaintext commands/JSON and not binary. Or am I mistaken?
<tewinget> nanoakron: that is correct, everything is marshalled via json
<tewinget> this is so that higher-level languages have no problem using the RPC
<pigeons> or jaxx :P
<tewinget> as the (de)serialization takes no time at all compared to the computations/fetching
<tewinget> pigeons: hope springs eternal
<hyc> this is for wallet-style client RPCs only then
<Slack> <nanoakron> Doesn't that mean that there are now two sets of commands to maintain in sync - JSON-RPC (won't be deprecated) and JSON-over-ZMQ?
<fluffypony> JSON-RPC will be deprecated for the daemon
<fluffypony> we won't add new RPC commands to it
<fluffypony> JSON RPC for the wallet will continue to evolve and exist
<fluffypony> because web apps rely on it
<fluffypony> communication with the daemon will be relegated to 0MQ only
<moneromooo> !bookie no-json-rpc-added-ever-again yes no
<Slack> <nanoakron> But in JSON format
<moneromooo> aw…
<fluffypony> (eventually)
<fluffypony> @NanoAkron yes
<fluffypony> so existing apps that interact with the daemon, eg. pool software, can continue by adding a 0MQ library
<fluffypony> and modifying any calls that have changed
<tewinget> (which won't be many, there were just a few that seemed silly in some ways, and changed accordingly, but none of that is set in stone)
<fluffypony> anything else?
<fluffypony> or I guess we can include that in the next item on the agenda
<fluffypony> 4. Any additional meeting items
<Slack> <nanoakron> Neigh
<fluffypony> lol
<fluffypony> we should use HAY and NEIGH instead of ACK and NACK
<Slack> <nanoakron> Lol
<tewinget> I'll keep updating over the next couple days, fwiw. Gotta get with Snipa to see if he can make a couple of modifications to the tests for me to make issues track-down-able, but he's afk until tomorrow.
<moneromooo> Hmm, range sig reduction… multisig… fee formula change…
<Slack> <nanoakron> Yes
<pigeons> Snipa: are these tests in your github?
<fluffypony> oh I have an item for brief discussion
<jollymort> it's not just the fee, penalty needs tweaking
<fluffypony> as everyone knows, the dynamic block adjuster isn't adjusting very well since the txs became larger
<pero> snipa is afk until tmrw iirc
<jollymort> either by increasing the min. blocksize, or having a transition formula
<fluffypony> does everyone think we should leave it as-is until September, with the occasional backlog
<fluffypony> or should we have some intermediary hard fork ?
<ArticMine> We may need one
<hyc> If we have a fix now, would be nice to deplot it sooner
<hyc> deploy
<Slack> <nanoakron> But it has to be smart enough to account for potential future changes in range proof and therefore Tx size
<jollymort> it accounts :)
<pero> september is a long time away
<Slack> <nanoakron> As well as ring size. Txes will become much more standardised in size and non-parametrically distributed
<fluffypony> ok I think that's reasonable consensus, as soon as we have something workable we'll put it out to the community as a hard fork and see how they feel
<Slack> <nanoakron> So medians etc may not make statistical sense
<DaveyJones> you could HF at around the time when originally the ringct hf was supposed to happen
<ArticMine> nanoakron We are looking at a fall in tx size?
<Slack> <nanoakron> Hopefully. Size would fall will range proof improvements, but distribution of sizes would flatten with ring size standardisation. Parametric statistics would no longer apply.
<fluffypony> DaveyJones: that's in March, too soon for a planned fork
<Slack> <nanoakron> So adjusting based on moving medians would be meaningless. We'd need to deploy alternate statistical tests.
<moneromooo> Can you explain that ?
<fluffypony> ok
<fluffypony> anything else before we close the meeting ?
<fluffypony> (we can discuss specifics after the meeting)
<Slack> <nanoakron> Even now with a mix of rct and non-rct transactions the median is meaningless because the size distribution is non parametric
<jollymort> it's some typical size which is most important, currently at 13kB
<hyc> calculate two separate medians. one for rct and one for non-rct.
<jollymort> doesn't matter if it's +/- 1kB
<Slack> <nanoakron> It's instead bimodal
<hyc> sounds like we're done with the meeting side of things then
<jollymort> I mean, if you roll out some change to TX format, you already know the next typical size it will cause
<fluffypony> last item is the next meeting time
<jollymort> and you will need to HF anyway, so all you'd need to do is adjust one parameter for the dynamic blocksize/fee
<fluffypony> 2 weeks from now
<fluffypony> March 5th
<hyc> cool
<fluffypony> I will be on a plane to London, but should have wifi and should be able to attend the meeting
<fluffypony> thanks guys


Post tags : Dev Diaries, Monero Core, Cryptography, ZeroMQ