開發者會議紀錄 DevMeeting 20170806

  • 關於環簽交易使用的交易輸入預設值將有更動
    • 已經有普遍的共識將由規劃中的3改為5,不過改用8或10的方案也同時在討論中
    • 也許要找社群裡的數學家Surae跟(或)Sarang來提供點意見比較好
  • 測試網路(Testnet)的硬分叉將在2017的八月七號禮拜一進行
  • GUI最新資訊
    • Android和iOS系統的官方GUI已經在封閉/公開測試
    • 而iOS的部分則因為app store帳號問題而延宕了
    • Android的版本則因linux box空間不足而耽擱了。1.5個月前的版本在此: https://github.com/monero-project/monero-core/pull/780
    • 正在分析iOS的測試狀況,至於MyMonero,基本上是個”超想趕快把這玩意上架”的狀況。
  • monerod的記憶體流失(memory leak)的問題已經在著手處理

以上翻譯來自於紀錄原文: https://monerobase.com/wiki/DevMeeting_2017-08-06

rehrar> The usual: https://github.com/monero-project/meta/issues/99
rehrar> Btw, pony has asked me to make the meeting issues from now on cuz he forget. :)
hyc> ok cool
Fymiywtfbtcxmr> 1st Issue should be min ring size
Fymiywtfbtcxmr> Time to decide before code freeze
gingeropolous> From my viewpoint, the issue stands at this: there’s consensus to move it to ringsize 5, and there’s some support to make it more than that. But I don’t think there’s any scientific reason for any particular number greater than 5 scoobybejesus ([email protected]) has joined
hyc> right.
johnalan> yes 5 is optimal
hyc> I expressed a preference for 8 because I like powers of 2.
johnalan> lol
gingeropolous> so the question is whether we should go greater than 5 just based on a gut feeling that more is better
gingeropolous> or wait until we have some valid reason with some maths backing it
rehrar ([email protected]/web/freenode/ip. has joined rehrar has quit (Client Quit)
ArticMine> The case for more than 5 is two mix sets of 5 recent and random
hyc> I think the question is what does it do to min tx size
hyc> and min tx fee
ArticMine> Yes that is very important
endogenic> so, more research?
rehrar> ^
dEBRUYNE> hyc: increasing mixin from 2 to 10 increases the tx size with ~ 1 kB
dEBRUYNE> assuming 2 in 2 out transactions
hyc> yeah, 5 for now. more research for additional changes.
gingeropolous> should we bump the wallet default?
dEBRUYNE> Perhaps leave it at 4 for uniformity?

endogenic> fwiw we use 9 in new mymonero apps currently (ringsize of 10)
endogenic> but we also turned off our tx fees due to ringct tx size et al
endogenic> temporarily…
hyc> endogenic: have you also changed output selection algo?
endogenic> in what way?
ArticMine> We have to raise min blocksize if we go over 15k tx size
Fymiywtfbtcxmr> Ring size 10 for uniformity
Fymiywtfbtcxmr> All apps and wallets just like mymonero
endogenic> personally i would love to hear Surae’s input on this as well regarding whether we should do this now before making confirmations
Fymiywtfbtcxmr> Tx size not that much larger scoobybejesus has quit (Quit: ZNC 1.7.x-git-735-879c813 - http://znc.in)
endogenic> heck it would be nice to hear from Sarang as well :P
rehrar> Well friends, in the interest of the meeting, can we go to item 2? :) Brief review of what’s been completed since the previous meeting.
Fymiywtfbtcxmr> Range proof reductions coming
rehrar> Oops. After this discussion is done.
endogenic> hyc feel free to ping me later
Olufunmilayo ([email protected]/olufunmilayo) has joined
vtnerd> hyc endogenic : mymonero does not use the wallet selection algo currently moneuser555_ has quit (Quit: Page closed)
vtnerd> its just random selection over the set of possible outputs
endogenic> yes i can confirm that
ArticMine> JollyMort[m] did the simulations but as I recall we are ok if the 2/2 tx size is below 15K
ArticMine> for min blocksize
ArticMine> So ring 10 may work
moneromooo> mixin 14 would still be ok, 19 not. I have no data for 15-18.
gingeropolous> so r we increasing it or waiting for more research?
scoobybejesus ([email protected]) has joined
ArticMine> I am ok up to 10
moneromooo> I’m keeping it to 5 unless a number of people like smooth, ArticMine, luigi1111, surae, knaccc think it’s best to increase now.
ArticMine> Sounds good to me
hyc> cool
moneromooo> Maybe we make a vote with those people :)
gingeropolous> yeah. i propose we push this to core team for final decision. seems to fit the bill of that scenario in the governance structure thingy
endogenic> vtnerd: any thoughts?
Jaquee> sounds good. when is testnet HF?
moneromooo> Tomorrow. If pony merges in time ^_^
Jaquee> :D
vtnerd> on mandatory ring size ? not really
endogenic> kk
knaccc> moneromooo I’m fine with 5 until further research is done
vtnerd> other than I recall at least 4 being useful, based on the way research paper I saw
Fymiywtfbtcxmr> Pony drinking wine
luigi1111> I vote 5 because reasons
moneromooo> And that guy has 1111 votes ^
hyc> I vote 8 because we need drama and dissent.
moneromooo> And because winning a vote with 100% would feel a bit too dictatorship.
moneromooo> I vote for someone to summon the pony.
ArticMine> At the point 5 and continue the research
ArticMine> this
hyc> yeah herch ([email protected]/web/freenode/ip. has joined
endogenic> rehrar?
rehrar> Ye?
endogenic> may be time for next item
vtnerd> also moving mymonero to the standard wallet output selection needs a todo somewhere
rehrar> K. Well I say we jump to 2 as I said above. :) Brief review of what’s been completed since the previous meeting. I’d love to hear from Jaquee and even you regarding mymonero
endogenic> issue exists on mymonero-app-js
endogenic> at least in terms of using the official lib
vtnerd> the mymonero front is probably a crappy but decent place to mark that
vtnerd> endogenic : ok good
endogenic> def important to me
pigeons> maybe moneromooo repeat the list of PRs you’ve asked fluffypony to merge?
Fymiywtfbtcxmr has quit (Ping timeout: 260 seconds)
Jaquee> i’m currently rebasing the huge GUI lightwallets/mobile PR to exclude the lightwallets parts
moneromooo> I don’t have a backlog.
Jaquee> because #2109 (the wallet2 parts of the lightwallet support) wont make it to the release.
Jaquee> but i still want to have the mobile and remote node parts from that PR in next release.
Jaquee> I’d say both ios and android version is ready for alpha/beta testing. But it’s being blocked by lack of app store accounts.
Jaquee> At least the ios version.
Jaquee> For android we could just build apk on buildbot, but unfortunately there’s not enough space on the linux box for that. Pigeons is working on that afaik.
moneromooo> I’m wary of the changes that affect the normal wallet tx making operations. The rest could be merged if needed.
Jaquee> yeah. i’m totally ok with that.
Jaquee> but hopefully we don’t need to wait another 6 months for next release after hf
rehrar> We can launch a little campaign for Android testers? Get the Community workgroup to front that, no?
moneromooo> If it’s like last time, we’ll have to wait just a few days ^_^
Jaquee> Fluffy said a couple of months back that he was working on app store accounts for us, but i don’t know what the status is. Anyone from core team have more info on that?
pigeons> android versions. you need one for each architecture
endogenic> Jaquee he was traveling and needed info which was at home in SA
endogenic> got back a day ago i think
Jaquee> endogenic: that was for mymonero accounts afaui. are we using the same for monero apps+
Jaquee> ?
endogenic> nah
endogenic> separate, as i suggested
johnalan> I have an accoutn and can add 100 team members
rehrar> Endogenic, you wanna give an update on your fun things?
johnalan> you’re welcome to use it
endogenic> unless you guys really want to hop on the MEA bandwagon
hyc> for android, you can get by with just 32bit binaries
johnalan> but for prod, better to have MEA account
endogenic> everyone should be using MEAT tokens Ggjffg47844 ([email protected]/web/freenode/ip. has joined TinusMars has quit (Ping timeout: 260 seconds)
endogenic> lol johnalan MEA was just a joke :P
johnalan> oh I know
endogenic> oh hehe
johnalan> :)
Jaquee> johnalan: biggest issue is trust. core team need to build and distribute scoobybejesus has quit (Remote host closed the connection)
johnalan> yes with Apple code signing
hyc> not sure if we care to support android x86. that userbase is tiny.
johnalan> you need an account
Jaquee> otherwise noone will download =)
Jaquee> hopefully …
johnalan> look at how Signal etc do it
johnalan> they have the code online open src
pigeons> so mainly armv8?
johnalan> but you have to trust the binary on the store
johnalan> not much you can do
endogenic> johnalan: Apple’s acting as the authority there
johnalan> @endogenic on the my Monero iOS app, are you close to beta test?
endogenic> certainly are
luigi1111> fluffypony fluffypony fluffypony
johnalan> great!
endogenic> fluffypony fluffypony fluffypony
luigi1111> \
fingers crossed*
moneromooo> `fluffupoony `fluffupoony `fluffupoony
endogenic> lol
ArticMine> The pony is in the “land of the free temporarily” scoobybejesus ([email protected]) has joined
endogenic> snoozetown, south africa
ArticMine> He is back
scoobybejesus has quit (Remote host closed the connection) fzerosum has quit (Quit: Leaving) duil ([email protected]/web/freenode/ip. has joined
hyc> we we can still release iOS without waiting for apple account https://inojb.net/
ArticMine> On jailbreak?
johnalan> yuck
johnalan> not keen
hyc> no jailbreak required
ArticMine> html5? scoobybejesus ([email protected]) has joined
johnalan> I think it’s OTA for their ‘store’/repo
johnalan> and then you install thru that
Jaquee> lol. that page doesn’t look serious
Jaquee> have you tried it hyc?
johnalan> but after all the work for iOS, it’s paramount to have it on the store natively
hyc> there are many other alternatives http://www.iphonetopics.com/how-to-install-cydia-without-jailbreak-iphone-ipad/
hyc> and no, I haven’t tried it, I own no Apple crap.
johnalan> but we’d still need to distrubute a binary through Cydia
johnalan> it wouldn’t be on the ‘App Store’
johnalan> so that’s where it would be suboptimal
endogenic> anything else completed since last mtg?
johnalan> people need to be able to take a new iphone, open the store and type ‘Monero’
hyc> that can come later. if we want people to beta test now, this is the fast route.
johnalan> fair enough I guess
johnalan> we can just use test flight
endogenic> hockeyapp is good too
endogenic> but then it’s a certs question again
johnalan> you can invite 1K testers
johnalan> 10K
johnalan> https://developer.apple.com/testflight/
Jaquee> yeah. hockey/testflight would be nice
endogenic> 1 million testers \
puts pinky to mouth*
johnalan> lol
johnalan> anyway we need to stay tight to the iOS ecosphere on this
johnalan> native native native
moneromooo> You have a limit to the number of testers when you code an apple program ?
johnalan> Yes, but it’s very elegant
endogenic> moneromooo: not exactly
johnalan> and when your app is on the store
serhack> Hello Febo ([email protected]) has joined
endogenic> anyone can install the app on their device
serhack> Is meeting finished?
Jaquee> no
hyc> stil; requires a corporate registration first
serhack> Nice
johnalan> you can move through betas simultaneosly
endogenic> serhack: nah we need to go over code/ticktets/Q&A
johnalan> @hyc yup thats true
endogenic> and confirm next mtg
serhack> Okay
endogenic> moneromooo: https://developer.apple.com/testflight/
endogenic> it used to be 100 lol
endogenic> and they didn’t used to have testflight
moneromooo> Sorry, I was not really interested, just dumbfounded.
johnalan> its very cool
moneromooo shuts up
johnalan> @endogenic would you use testflight or hock for the mymonero ios app?
endogenic> either. honestly want to just launch the damn thing
Jaquee> ^^
Jaquee> so, fluffy is back, and hopefully we can have app store accounts soon^rm?
hyc> PRs (17:53:15) moneromooo: 2132 2141 2147 2148 2151 2153 2182 2188 2189 2191 2195 2196 2197 2198 2200 2201 2203 2206 2210 2211 2214 2215 2216 2219 2220 2222 2223 2226 2230 2231 2232 2234 2236 2239 2242 2243 2244
Jaquee> to merge?
endogenic> soon™
moneromooo> Those were the easy merges.
herch> for what it’s worth, I have doubt if apple will ever approve the app. I wish I am wrong here.
endogenic> herch: i’m not worried about that
serhack> Hyc: a lot
endogenic> pretty sure they would rather have official apps than scam apps
endogenic> after all their priority is user experience
endogenic> plus we heard through Jaxx that they contacted an apple rep who said it would be ok
serhack> +1 endogenic, scam users will try to scam us with their fucking apps
endogenic> but then again do we listen to them? :P
Jaquee> these are also “easy” merges #2254 #2247 #2233 #2138
rehrar> Post #25 this week: Hey so I downloaded freewallet and put all my XMR on there. It’s gone. Help?
Jaquee> endogenic: lol. i wouldn’t trust jaxx on that information
endogenic> Jaquee: well /you/ know that… :)
Jaquee> :D
Jaquee> anyway. let’s move on
hyc> I missed a question back there - on Android we could release ARMv7 (32bit) wallet. it would work fine on ARMv8 (64bit).
pigeons> sounds good
hyc> majority of ARMv8 devices on the market today are still running 32bit OS.
serhack> Amazing
serhack> Is there any apk that I can test? I have an android smartphone
hyc> I think you have to compile for yourself for now
endogenic> i actually sent a self signed apk to nm90 a long time ago. ppl could still test like that i bet
endogenic> if you trust them… :P
Jaquee> i also have one
serhack> :0
Jaquee> https://github.com/monero-project/monero-core/pull/780
serhack> Thanks jaquee
Jaquee> 5 minutes left… i’m wondering about that memory leak in monerod. do we have any clue yet?
moneromooo> I don’t think there’s really one.
moneromooo> (modulo small bits and bobs I already fixed)
hyc> schrodingers leak
Jaquee> really? my daemon keeps on OOM crashing
Jaquee> on a 2GB vps
moneromooo> Well, run it with mleak, and kill it with SIGPROF from time to time, and send me those logs.
moneromooo> mleak is on hyc’s github repo.
Jaquee> ah. i’ll do that
moneromooo> But for me, memory usage was constant modulo short term noise.
hyc> https://github.com/hyc/mleak
Jaquee> thanks
moneromooo> hyc: the RSS differences are from the OS deciding when to swap in/out the mmapped pages, right ?
moneromooo> (for lmdb, I mean)
hyc> yeah
hyc> the SHARED number should be relatively constant, growing when the map is resized
hyc> RSS should fluctuate with other memory pressure
hyc> s/SHARED/SHM/
moneromooo> Jaquee: sorry, missing a step: a couple seconds after SIGPROF, do: mdump ml.{data,info} > “monerod.log-`date" moneromooo> Jaquee: sorry, missing a step: a couple seconds after SIGPROF, do: mdump /path/to/monerod ml.{data,info} > "monerod.log-\date`“
serhack> The app is nice but I noticied a lot of graphical glitch
serhack> UX must be a prority
hyc> probably OpenGL crap
rehrar> Jaquee, did you see that UX guys’ thing?
Jaquee> no?
rehrar> Sec
serhack> @krokbaltsar is a UX designer
rehrar> Monero UX - Persona and Mobile Wallet Concept https://www.reddit.com/r/Monero/comments/6qkw8e/monero\_ux\_persona\_and\_mobile\_wallet\_concept/
rehrar> This one.
serhack> https://www.gustafgarnow.com/monerowallet/ ajsantos ([email protected]/ajsantos) has joined
serhack> ^^ best design
serhack> Oh, maybe same person @rehrar ;)
rehrar> Anyways, it’s past time.
Ggjffg47844 has quit (Ping timeout: 260 seconds)
rehrar> Confirm next meeting time?
rehrar> Hopefully FP will come to that one.
hyc> 2 weeks
hyc> (TM)
rehrar> Boom. Done.