Monero and Kovri Dev Meeting Note Highlights – 2016-10-02
Logs
351 lines
fluffyponyHi all
fluffyponystarting meeting bot
moneromoooI'm none of them.
dEBRUYNEI am here
fluffyponymoneromooo: I know you're here :)
fluffyponyok meeting bot is up
dnaleorwatching
fluffyponyso
botanonimal#kovri-dev too?
fluffyponykovri-dev is roped in too
othek
i2pmoneromoooo
fluffyponythis is our first post-0.10.0 meeting
fluffyponythe 0.10.0 release went fairly smoothly
fluffyponyapart from the boost oddities
fluffyponyon the Boost serialisation stuff
fluffyponyit's not really feasible to do a point release just for that just yet - in a few weeks I'll have local build infrastructure that will make releases a lot easier on me
fluffyponysince I can secure local machines far more easily than Internet-facing machines at a DC
fluffyponyin the interim, if anyone is struggling with it the easiest thing for them to do is recover their wallet from the seed / keys
moneromoooAnd keep the old cache if they have tx keys they want to keep.
fluffypony^^
fluffyponyI'd also like to welcome all the new contributors
fluffyponyeven if it's just correcting a spelling error, all contributions are valuable
fluffyponyand very much appreciated
fluffyponyone thing I would like to encourage with the new contributors is to submit your GPG key via PR
fluffyponyand side-channel it to myself or moneromooo or someone so we can independently verify the correct key goes in
fluffyponythey live in utils/gpg_keys/
botanonimaland if moneromooo says "ok" in your PR, that's a GOOD thing!
fluffyponyand then once you've done that you can GPG sign your commits with -S
fluffyponyeg. git commit -S -am "meaningful commit message"
moneromoooOnly if it's at the start.
patthehumanIs there anything that needs to be made for iOS
fluffyponylol anonimal
fluffyponypatthehuman: it's an open-source project, so if you want to build something for iOS then please do
fluffyponyno need to ask permission or anyway
fluffyponyI'd be interested to see if we could package a full node for iOS (without the wallet, lest it get removed from the app store)
patthehumansure, i guess i'm wondering more along the lines of what needs to be built
patthehumancan you outline how that would work?
fluffyponyand then an old iPhone or iPod Touch would work as a full node on wifi
fluffyponywell we have native ARMv8 builds
fluffyponyand interacting with the daemon over RPC isn't hard
ArticMineWhy not just go with Cydia on jailbroken iOS?
fluffyponybut I have no idea if an iOS app would let you arbitrarily launch a process
endogenicArticMine: too few jailbreak their devices, no?
fluffyponyArticMine: absolutely - would be nice to be able to launch it as an app tho
patthehumangenerally apps get removed from the app store if they are simple "remotes"
endogenicfluffypony: iOS does not allow you to spawn processes… no NSTask etc equivalent
fluffyponyah
TedTheFicusArcticMine: I think Cydia is a good plan B. Not many non tech people are going to have jail broken phones
patthehumanright, they have a list of closed API's that will get you banned
fluffyponydoes Objective-C let you also run native C / C++ code?
endogenicfluffypony: yep as long as you can compile for an ARM target
endogeniccompile it*
MK__ArticMine : A good idea and a lot of IOS devices have Jailbreak , Remember that XMR are used in the DNM as well
fluffyponyyeah we have ARMv8 support across the board
endogenicObjective-C is a strict superset of C, so any C is good, and C++ can be compiled too
ArticMineNet applications Android 69.18% IOS 25.02% market share
fluffyponyanyway we're getting side-tracked a little - patthehuman feel free to play around with it, if you feel like it
patthehumansure
fluffyponyso
fluffyponyringCT is live in testnet, and more testing would be appreciated
fluffyponymWo12's testnet block explorer is helpful
fluffyponybut performance testing is also immensely useful
fluffyponysee what cracks under pressure
fluffyponywe have a short window until the January hard fork, so we need to hammer testnet as much as possible
trustedsetupis there a testnet faucet somewhere? or is mining or irc begging for testnet xmr recommended?
fluffyponyjust ask me and I'll send testnet XMR your way
patthehumanare there any automation processes that can hammer on testnet?
patthehumani have access to a lot of r510 servers that i could potentially mirror some script to hammer on it
fluffyponypatthehuman: you could pretty much just write a bash script to send to yourself once a second
fluffyponyand cycle it that way
patthehumancool
fluffyponyand then see how your testnet node(s) handle catch ups, and if they keep up with testnet when blocks are bigger
fluffyponywe also have our new buildbots ticking along nicely
fluffyponyso we'll be killing off Travis at some point in the coming weeks
fluffyponybuild bot output has been relegated to #monero-bots
fluffyponyand that channel is relayed to irc2p
botanonimalThanks pigeons
botanonimalmonero-build.i2p is also online
fluffyponyyeah pigeons has done great work
fluffyponyat the moment we're building for a ton of platforms
fluffyponyincluding macOS 10.9, 10.10, 10.11
fluffyponyso we should pick up PRs that break compilation more rapidly
fluffyponyhow we handle testing is a bit harder
fluffyponyespecially since some of the tests take several hours to run
fluffyponymy current leaning is towards nightly builds + tests
fluffypony(of master)
fluffyponythat way we'll catch tests that are broken by any merged PRs
moneromoooDaily core_tests would be useful.
fluffyponyperformance_tests would also be useful
fluffyponythat way we can track anything that has a huge impact on performance
moneromoooAs long as the outcome can be seen without too many hoops (ie, javacrsipt)
fluffyponymoneromooo: we'll probably just grab the output, parse it, and shove it in a database
fluffyponythen we can create a profiler for the site without too many issues
fluffyponyon the PR side
fluffyponyhas anyone looked at 1082?
patthehumanno my apologies for being new but can you elaborate on what 1082 is
fluffyponyor actually moneromooo: can you give everyone a brief overview of what 1082 does
fluffyponyoh sorry patthehuman - PR = pull request
fluffyponyso PR 1082 = https://github.com/monero-project/monero/pull/1082
patthehumanyeah im familiar with PR's (worst part of my work day lol)
moneromoooAh, as the comment says, really.
moneromoooIt just tries to avoid the case where someone sends money just after receiving it.
moneromoooThat's a common enough case.
fluffypony"25% of the outputs are selected from the last 5 days (if possible), in order to avoid the common case of sending recently received outputs again. 25% and 5 days are subject to review later, since it's just a wallet level change."
trustedsetupwhere did the 25% come from? 25% seems somewhat arbitrary. did MRL have input on this number?
moneromoooThey're aribtrary.
fluffyponytrustedsetup: the MRL is of the opinion that we're never going to find a "perfect" distribution, and that distribution should be re-evaluated regularly
luigi1112Will look at it
fluffypony25% would only be a single output at minimum mixin
trustedsetupok thanks
moneromoooIt's actually 25%, except if that gives 0, in which case it uses 1.
fluffyponymoneromooo: and it's 25% including the "real" output, right?
moneromoooYes.
moneromoooSee line 2744 in wallet2.cpp
fluffyponykk
fluffyponyas to the other open PRs
fluffyponymost of them are not merged yet due to their being unreviewed
fluffyponyI try give PRs a little while before reviewing them myself, otherwise I end up being the only reviewer, which is bad for security obvs
fluffyponybear in mind that a review is not in-depth line-by-line analysis
patthehumanpr is just a quick overview
moneromoooI'd hope the review does look at all lines.
fluffyponyyup
fluffyponyit's a sanity check, and a check for obvious screw-ups, and a check for snuck-in backdoors
fluffyponymoneromooo: the key there was in-depth, not line-by-line ;)
moneromoooOK, that's fine.
fluffyponymedusa_: are you around?
fluffyponyok in the absence of medusa_ being around, dEBRUYNE have you been following Ilya's progress on medusa_'s issues?
dEBRUYNEYeah, he has fixed all issues opened by medusa_ as far as I know
fluffyponyok great
medusa_yes im here
dEBRUYNEExcept -> https://github.com/monero-project/monero-core/issues/29
dEBRUYNEbut that's more of a feature, which should be implemented later
fluffyponyoh cool - medusa_ how are you finding it now that most of the issues have been fixed?
dEBRUYNEcould*
TedTheFicusMK_ + Others who are wondering, the monero-core project that is being discussed now is the GUI
medusa_i think we need more feedback regarding the performance difference between gui wallet creation time and CLI wallet creation time
medusa_and i dont know of any major bugs that would be dangerous
fluffyponymedusa_: there's a PR that's supposed to fix that
fluffyponyI haven't merged it yet, but it's gone through review
dEBRUYNEIlya merged jacquee's PR as well
dEBRUYNEhe noticed a significant improvement
medusa_so my opinion is merge all to monero-core project master, test there again and if its good we build the bins
dEBRUYNE^ +1
dEBRUYNEBeta binaries will also bring more testers, who possibly could notice something we might have overlooked
fluffyponyok
TedTheFicusIm down to test once the Betas are out
fluffyponywe'll need a point release of 0.10 to go with it
fluffyponyso we should at least get through the current group of pending PRs before we do that
medusa_but we must communicate its for testing, since the seed is nowhere displayed after creation we dont want people to lose money
moneromoooIt creates a keys file, right ?
fluffyponymedusa_: well that's a pretty big issue :-P
medusa_yes
fluffyponyah ok
fluffyponyso monero-wallet-cli could be shipped with it for recovery
TedTheFicusGood idea
moneromoooWell, you do need the daemon anyway, don't you.
medusa_yes
dEBRUYNE<fluffypony> so monero-wallet-cli could be shipped with it for recovery <= It's able to recover seeds
dEBRUYNEIt's just that only in the wizard the seed is shown once
dEBRUYNEoh wait, you mean restore with the .keys file?
fluffyponyyes I meant recovery as in "recover my seed from the .keys file"
ArticMineRecovery from seed is the issue with the GUI?
dEBRUYNEah gotcha
dEBRUYNENo ArticMine, there isn't a window yet to see your seed
dEBRUYNEafter the initial wizard
fluffyponyArticMine: no - it just doesn't display the seed again after the wizard
fluffyponyand given how many MyMonero support emails I get where people didn't write down their seed…
dEBRUYNEShould be fairly trivial to add though
fluffyponyok so that's about it from my side - tewinget isn't around to give us a 0MQ update
fluffyponyhyc I don't think has started tinkering with the walletDB stuff
fluffyponyalso the forum - I know, I'm working on it, moved all broken deps into monero-project repos to better manage them and am fixing the last few niggly issues
dEBRUYNEfluffypony: re: GUI, preferably we would have a tab that displays viewkey/seed/spendkey, the tab could be named Private Keys or something, with a big fat warning label :P
dEBRUYNELike I said, should be fairly trivial to add
fluffyponydEBRUYNE: good idea - open an issue for it and let Ilya do it asap :)
fluffyponyok so we have 7 minutes before the Kovri meeting, any other things to discuss?
dEBRUYNEfluffypony: will do
dEBRUYNENoodleDoodle isn't here right?
moneromoooWho wants to volunteer to review some patches from time to time ? :)
dEBRUYNEmoneromooo: Similiarly, would you be able to glance over / review the trezor XMR code?
moneromoooWhere is it ?
fluffyponyon NoodleDoodle's computer
dEBRUYNE^ moneromooo
dEBRUYNEhe has some commits in his monero repository and in trezor-xmr
fluffypony:-P
i2panonimalmoneromooo: I review many of them but I don't spend enough time with the codebase to ack/nack
moneromoooanonimal, thanks :)
moneromooodEBRUYNE: Do you know which one of the three repos is the right one ?
moneromoooxmr, common, mcu. xmr seems likely to be one at least.
dEBRUYNEoh trezor-xmr
dEBRUYNEand monero
dEBRUYNESo his commits here -> https://github.com/NoodleDoodleNoodleDoodleNoodleDoodleNoo/monero/commits/master?author=NoodleDoodleNoodleDoodleNoodleDoodleNoo
moneromoooOK, I'll keep that in mind then.
dEBRUYNE& here -> https://github.com/NoodleDoodleNoodleDoodleNoodleDoodleNoo/trezor-xmr
medusa_i can not review the code, but i can test specific pull requests if you guys explain me what you changed
dEBRUYNEtrezor-mcu / trezor-common has no commits from NoodleDoodle moneromooo
moneromooomedusa_: 1082 and 1121 could do with some testing if you feel like it.
moneromoooAnd 1140 :)
moneromooo1082 changes the way fake outs are selected
cryptotekkwow in this pace i see GUI by tonight lol
moneromooo1121 replaces the sweep_unmixable code to be more stable and, well, better
moneromooo1141 adds cold wallet signing
medusa_oh i can test 1141
medusa_i still have the setip from the –offline thing
moneromooo1140, sorry. Off by one…
fluffyponyoh no off by one bug!
medusa_will have a look, thanks
moneromoooMac, Linux, and Plan9.
cryptotekkandroid please
libertelol
fluffyponyhokay
fluffyponythat's the end of that
Sources and notes
- Meeting log: Overview and Logs for the Dev Meeting Held on 2016-10-02, getmonero.org. The official post carried the raw log and pointed readers to Hello Monero for the overview.
- Original Hello Monero URL slug recovered from the Wayback Machine index; the original page content did not survive, so this reconstruction carries the log as archived by the Monero Project.