Overview and Logs for the Dev Meeting Held on 2016-10-02

Posted by: dEBRUYNE / fluffypony

October 02, 2016


An overview can be found on Hello Monero


<fluffypony> Hi all
<fluffypony> starting meeting bot
<moneromooo> I'm none of them.
<dEBRUYNE> I am here
<fluffypony> moneromooo: I know you're here :)
<fluffypony> ok meeting bot is up
<dnaleor> watching
<fluffypony> so
<meeting-bot> [anonimal] #kovri-dev too?
<fluffypony> kovri-dev is roped in too
<othe> k
<meeting-bot> [i2p-relay] {-moneromooo} o
<fluffypony> this is our first post-0.10.0 meeting
<fluffypony> the 0.10.0 release went fairly smoothly
<fluffypony> apart from the boost oddities
<fluffypony> on the Boost serialisation stuff
<fluffypony> it'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
<fluffypony> since I can secure local machines far more easily than Internet-facing machines at a DC
<fluffypony> in the interim, if anyone is struggling with it the easiest thing for them to do is recover their wallet from the seed / keys
<moneromooo> And keep the old cache if they have tx keys they want to keep.
<fluffypony> ^^
<fluffypony> I'd also like to welcome all the new contributors
<fluffypony> even if it's just correcting a spelling error, all contributions are valuable
<fluffypony> and very much appreciated
<fluffypony> one thing I would like to encourage with the new contributors is to submit your GPG key via PR
<fluffypony> and side-channel it to myself or moneromooo or someone so we can independently verify the correct key goes in
<fluffypony> they live in utils/gpg_keys/
<meeting-bot> [anonimal] and if moneromooo says "ok" in your PR, that's a GOOD thing!
<fluffypony> and then once you've done that you can GPG sign your commits with -S
<fluffypony> eg. git commit -S -am "meaningful commit message"
<moneromooo> Only if it's at the start.
<patthehuman> Is there anything that needs to be made for iOS
<fluffypony> lol anonimal
<fluffypony> patthehuman: it's an open-source project, so if you want to build something for iOS then please do
<fluffypony> no need to ask permission or anyway
<fluffypony> I'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)
<patthehuman> sure, i guess i'm wondering more along the lines of what needs to be built
<patthehuman> can you outline how that would work?
<fluffypony> and then an old iPhone or iPod Touch would work as a full node on wifi
<fluffypony> well we have native ARMv8 builds
<fluffypony> and interacting with the daemon over RPC isn't hard
<ArticMine> Why not just go with Cydia on jailbroken iOS?
<fluffypony> but I have no idea if an iOS app would let you arbitrarily launch a process
<endogenic> ArticMine: too few jailbreak their devices, no?
<fluffypony> ArticMine: absolutely - would be nice to be able to launch it as an app tho
<patthehuman> generally apps get removed from the app store if they are simple "remotes"
<endogenic> fluffypony: iOS does not allow you to spawn processes… no NSTask etc equivalent
<fluffypony> ah
<TedTheFicus> ArcticMine: I think Cydia is a good plan B. Not many non tech people are going to have jail broken phones
<patthehuman> right, they have a list of closed API's that will get you banned
<fluffypony> does Objective-C let you also run native C / C++ code?
<endogenic> fluffypony: yep as long as you can compile for an ARM target
<endogenic> compile it*
<MK__> ArticMine : A good idea and a lot of IOS devices have Jailbreak , Remember that XMR are used in the DNM as well
<fluffypony> yeah we have ARMv8 support across the board
<endogenic> Objective-C is a strict superset of C, so any C is good, and C++ can be compiled too
<ArticMine> Net applications Android 69.18% IOS 25.02% market share
<fluffypony> anyway we're getting side-tracked a little - patthehuman feel free to play around with it, if you feel like it
<patthehuman> sure
<fluffypony> so
<fluffypony> ringCT is live in testnet, and more testing would be appreciated
<fluffypony> mWo12's testnet block explorer is helpful
<fluffypony> but performance testing is also immensely useful
<fluffypony> see what cracks under pressure
<fluffypony> we have a short window until the January hard fork, so we need to hammer testnet as much as possible
<trustedsetup> is there a testnet faucet somewhere? or is mining or irc begging for testnet xmr recommended?
<fluffypony> just ask me and I'll send testnet XMR your way
<patthehuman> are there any automation processes that can hammer on testnet?
<patthehuman> i have access to a lot of r510 servers that i could potentially mirror some script to hammer on it
<fluffypony> patthehuman: you could pretty much just write a bash script to send to yourself once a second
<fluffypony> and cycle it that way
<patthehuman> cool
<fluffypony> and then see how your testnet node(s) handle catch ups, and if they keep up with testnet when blocks are bigger
<fluffypony> we also have our new buildbots ticking along nicely
<fluffypony> so we'll be killing off Travis at some point in the coming weeks
<fluffypony> build bot output has been relegated to #monero-bots
<fluffypony> and that channel is relayed to irc2p
<meeting-bot> [anonimal] Thanks pigeons
<meeting-bot> [anonimal] monero-build.i2p is also online
<fluffypony> yeah pigeons has done great work
<fluffypony> at the moment we're building for a ton of platforms
<fluffypony> including macOS 10.9, 10.10, 10.11
<fluffypony> so we should pick up PRs that break compilation more rapidly
<fluffypony> how we handle testing is a bit harder
<fluffypony> especially since some of the tests take several hours to run
<fluffypony> my current leaning is towards nightly builds + tests
<fluffypony> (of master)
<fluffypony> that way we'll catch tests that are broken by any merged PRs
<moneromooo> Daily core_tests would be useful.
<fluffypony> performance_tests would also be useful
<fluffypony> that way we can track anything that has a huge impact on performance
<moneromooo> As long as the outcome can be seen without too many hoops (ie, javacrsipt)
<fluffypony> moneromooo: we'll probably just grab the output, parse it, and shove it in a database
<fluffypony> then we can create a profiler for the site without too many issues
<fluffypony> on the PR side
<fluffypony> has anyone looked at 1082?
<patthehuman> no my apologies for being new but can you elaborate on what 1082 is
<fluffypony> or actually moneromooo: can you give everyone a brief overview of what 1082 does
<fluffypony> oh sorry patthehuman - PR = pull request
<fluffypony> so PR 1082 = https://github.com/monero-project/monero/pull/1082
<patthehuman> yeah im familiar with PR's (worst part of my work day lol)
<moneromooo> Ah, as the comment says, really.
<moneromooo> It just tries to avoid the case where someone sends money just after receiving it.
<moneromooo> That'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."
<trustedsetup> where did the 25% come from? 25% seems somewhat arbitrary. did MRL have input on this number?
<moneromooo> They're aribtrary.
<fluffypony> trustedsetup: the MRL is of the opinion that we're never going to find a "perfect" distribution, and that distribution should be re-evaluated regularly
<luigi1112> Will look at it
<fluffypony> 25% would only be a single output at minimum mixin
<trustedsetup> ok thanks
<moneromooo> It's actually 25%, except if that gives 0, in which case it uses 1.
<fluffypony> moneromooo: and it's 25% including the "real" output, right?
<moneromooo> Yes.
<moneromooo> See line 2744 in wallet2.cpp
<fluffypony> kk
<fluffypony> as to the other open PRs
<fluffypony> most of them are not merged yet due to their being unreviewed
<fluffypony> I try give PRs a little while before reviewing them myself, otherwise I end up being the only reviewer, which is bad for security obvs
<fluffypony> bear in mind that a review is not in-depth line-by-line analysis
<patthehuman> pr is just a quick overview
<moneromooo> I'd hope the review does look at all lines.
<fluffypony> yup
<fluffypony> it's a sanity check, and a check for obvious screw-ups, and a check for snuck-in backdoors
<fluffypony> moneromooo: the key there was in-depth, not line-by-line ;)
<moneromooo> OK, that's fine.
<fluffypony> medusa_: are you around?
<fluffypony> ok in the absence of medusa_ being around, dEBRUYNE have you been following Ilya's progress on medusa_'s issues?
<dEBRUYNE> Yeah, he has fixed all issues opened by medusa_ as far as I know
<fluffypony> ok great
<medusa_> yes im here
<dEBRUYNE> Except -> https://github.com/monero-project/monero-core/issues/29
<dEBRUYNE> but that's more of a feature, which should be implemented later
<fluffypony> oh cool - medusa_ how are you finding it now that most of the issues have been fixed?
<dEBRUYNE> could*
<TedTheFicus> MK_ + 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
<fluffypony> medusa_: there's a PR that's supposed to fix that
<fluffypony> I haven't merged it yet, but it's gone through review
<dEBRUYNE> Ilya merged jacquee's PR as well
<dEBRUYNE> he 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
<dEBRUYNE> Beta binaries will also bring more testers, who possibly could notice something we might have overlooked
<fluffypony> ok
<TedTheFicus> Im down to test once the Betas are out
<fluffypony> we'll need a point release of 0.10 to go with it
<fluffypony> so 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
<moneromooo> It creates a keys file, right ?
<fluffypony> medusa_: well that's a pretty big issue :-P
<medusa_> yes
<fluffypony> ah ok
<fluffypony> so monero-wallet-cli could be shipped with it for recovery
<TedTheFicus> Good idea
<moneromooo> Well, 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
<dEBRUYNE> It's just that only in the wizard the seed is shown once
<dEBRUYNE> oh wait, you mean restore with the .keys file?
<fluffypony> yes I meant recovery as in "recover my seed from the .keys file"
<ArticMine> Recovery from seed is the issue with the GUI?
<dEBRUYNE> ah gotcha
<dEBRUYNE> No ArticMine, there isn't a window yet to see your seed
<dEBRUYNE> after the initial wizard
<fluffypony> ArticMine: no - it just doesn't display the seed again after the wizard
<fluffypony> and given how many MyMonero support emails I get where people didn't write down their seed…
<dEBRUYNE> Should be fairly trivial to add though
<fluffypony> ok so that's about it from my side - tewinget isn't around to give us a 0MQ update
<fluffypony> hyc I don't think has started tinkering with the walletDB stuff
<fluffypony> also 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
<dEBRUYNE> fluffypony: 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
<dEBRUYNE> Like I said, should be fairly trivial to add
<fluffypony> dEBRUYNE: good idea - open an issue for it and let Ilya do it asap :)
<fluffypony> ok so we have 7 minutes before the Kovri meeting, any other things to discuss?
<dEBRUYNE> fluffypony: will do
<dEBRUYNE> NoodleDoodle isn't here right?
<moneromooo> Who wants to volunteer to review some patches from time to time ? :)
<dEBRUYNE> moneromooo: Similiarly, would you be able to glance over / review the trezor XMR code?
<moneromooo> Where is it ?
<fluffypony> on NoodleDoodle's computer
<dEBRUYNE> https://github.com/NoodleDoodleNoodleDoodleNoodleDoodleNoo
<dEBRUYNE> ^ moneromooo
<dEBRUYNE> he has some commits in his monero repository and in trezor-xmr
<fluffypony> :-P
<i2p-relay> {-anonimal} moneromooo: I review many of them but I don't spend enough time with the codebase to ack/nack
<moneromooo> anonimal, thanks :)
<moneromooo> dEBRUYNE: Do you know which one of the three repos is the right one ?
<moneromooo> xmr, common, mcu. xmr seems likely to be one at least.
<dEBRUYNE> oh trezor-xmr
<dEBRUYNE> and monero
<dEBRUYNE> So his commits here -> https://github.com/NoodleDoodleNoodleDoodleNoodleDoodleNoo/monero/commits/master?author=NoodleDoodleNoodleDoodleNoodleDoodleNoo
<moneromooo> OK, 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
<dEBRUYNE> trezor-mcu / trezor-common has no commits from NoodleDoodle moneromooo
<moneromooo> medusa_: 1082 and 1121 could do with some testing if you feel like it.
<moneromooo> And 1140 :)
<moneromooo> 1082 changes the way fake outs are selected
<cryptotekk> wow in this pace i see GUI by tonight lol
<moneromooo> 1121 replaces the sweep_unmixable code to be more stable and, well, better
<moneromooo> 1141 adds cold wallet signing
<medusa_> oh i can test 1141
<medusa_> i still have the setip from the –offline thing
<moneromooo> 1140, sorry. Off by one…
<fluffypony> oh no off by one bug!
<medusa_> will have a look, thanks
<moneromooo> Mac, Linux, and Plan9.
<cryptotekk> android please
<liberte> lol
<fluffypony> hokay
<fluffypony> that's the end of that

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