Overview and Logs for the Dev Meeting Held on 2017-05-21

Posted by: dEBRUYNE / fluffypony

May 21, 2017

Overview

An overview can be found on MoneroBase.

Logs

<endogenic> we have to do the dev meeting really quietly so the pony can get some sleep
<moneromooo> 1. Greetings
<moneromooo> 2. Brief review of what's been completed since the previous meeting
<moneromooo> 3. Code + ticket discussion / Q & A
<moneromooo> 4. Any additional meeting items
<moneromooo> 5. Confirm next meeting date/time
<moneromooo> (no relay bot btw)
<moneromooo> So, since last time, I'm not sure what's been merged. Not much I think ? There was the key image flaw made public too.
<moneromooo> I'm working on abstracting network addresses. Mostly works, but serialization is being a massive pain.
<moneromooo> Anyone else want to say anything about what they're working on ?
<vtnerd> eventually I hope to have the amd64-51-30k speedups in wallet2.cpp (assuming fluffy approves)
<vtnerd> and improve the hex -> bin performance would be nice too (not nearly ready for a merge though)
<moneromooo> What is that short for ?
<vtnerd> oh sorry, the ASM for wallet scanning
<moneromooo> What are the 51 and 30k short for ?
<vtnerd> its already written so that it can be removed by cmake, and does not touch the consensus code at all
<vtnerd> 51-bit limbs and 30k tables
<vtnerd> 30k table singular, sorry
<moneromooo> Anyone else doing something nice ? hyc, LMDB wallet code maybe ? :)
<moneromooo> Nobody else ? tewinget ?
<vtnerd> everyone is out spending bytecoins
<moneromooo> Alright. Ticket discussion. The only thing I know is people reporting sync stopping. Some of them are user error, some are corrupt db, but I think there may be something else. I'd need to see good logs for both sides though.
<moneromooo> Anyone wants to say anything about pending bugs ?
<iDunk> Testing related… your 1982 works nicely, with no apparent issues after several days. That on a daemon with none to very light load, fwiw.
<moneromooo> Cool, thanks.
<moneromooo> About pending bugs: we're still waiting for finalization of the disposable address theory, IIRC RandomRun and knacc were discussing a change recently.
<moneromooo> knaccc: any comment about this ?
<iDunk> I'm syncing on Windows right now, albeit with non-default settings. Will report if it gets wedged.
<vtnerd> change to which portion, disposable addresses or sub-addresses?
<vtnerd> or I guess you specifically said disposable address
<moneromooo> I don't remember. It's 1753.
<moneromooo> I did, but not on purpose :)
<moneromooo> I have a feeling it'll be a fast meeting.
<moneromooo> luigi1112: anything new about smaller range proofs ?
<endogenic> he's out spending bytecoin too
<jollymort> happened to me today (sync); it just hanged; killed the process and restarted monerod; continued seamlessly
<jollymort> can re-sync with different loglvl to see if it happens again
<moneromooo> Next time, don't kill it before you've looked at the thread stack traces.
<moneromooo> Anything else someone wants to say in the meeting ?
<rehrar> Kovri test site? :D Not super dev-y, but nobody else says anything.
<moneromooo> Kovri meeting is in 48 minutes.
<rehrar> there is no kovri meeting today
<moneromooo> aw… then feel free.
<moneromooo> anonimal: around ?
<luigi1112> moneromooo sorry no. will update soooooon
<moneromooo> ty
<moneromooo> Hmm.
<luigi1112> next meeting will be exciting :-)
<moneromooo> Well, no test site either then.
<unknownids> oh baby i know what that means
<unknownids> i think
<moneromooo> Next meeting: in two weeks, same time.
<moneromooo> Done. The only good meetings are short meetings.
<luigi1112> am I doing this right
<luigi1112> ann of ann
<jollymort> mooo, how to use stacktrace?
<moneromooo> Oh he's back now :)
<moneromooo> jollymort: gdb /path/to/monerod pidof monerod
<moneromooo> thread apply all bt
<jollymort> so i can do it while it's running
<moneromooo> Yes.
<jollymort> no need to start it in some special way?
<jollymort> ok
<jollymort> i'll set loglvl to max also
<jollymort> and start sync from 0
<moneromooo> I wouldn't.
<jollymort> ok, so just restart and wait for it to hang?
<moneromooo> 3 and 4 are toop spammy for being much use in general usage.
<moneromooo> Yes.
<moneromooo> rehrar: free free if you have something to say/show.
<rehrar> Anyways, Monero test site here: http://45.32.161.183/ mobile responsive now and a brand new mobile navigation.
<rehrar> Kovri test site: http://45.32.171.42/ completely integrated into the Jekyll. The custom CSS framework was ported without issue. This bodes well for my further work on getmonero.org.
<moneromooo> Looks clean.
<Jaquee> hi. a bit late to meeting. i'm working on lightwallet integration in cli and gui
<Jaquee> or wallet2 mostly
<moneromooo> You mean mymonero ?
<rehrar> We scrubbed it down.
<Jaquee> yeah. and openmonero
<Jaquee> should work with both backends
<Jaquee> and should be switchable from full wallet to light wallet. vtnerd. would you consider implementing input/output import in mymonero api?
<Jaquee> https://github.com/moneroexamples/openmonero/issues/21
<Jaquee> would make it possible to switch from full wallet to light wallet without a rescan
<endogenic> not clear how the mobile experience would go
<endogenic> just an initial impression
<Jaquee> ?
<Jaquee> experience of what?
<endogenic> as in how user gets wallet file on their mobile and how they pick it
<vtnerd> I don't know what you mean Jacquee
<vtnerd> you want to import an existing wallet quickly from a local DB to a mymonero backend ?
<vtnerd> and also the other direction ?
<luigi1112> that would be cool. maybe :-)
<Jaquee> yeah. if i have a full wallet on my desktop and want to use the same wallet on my mobile. I could convert my full wallet to lightwallet by sending my inputs/outputs to mymonero/openmonero backend. And then login as usual on my phone
<Jaquee> with seed
<Jaquee> without paying the 1 xmr rescan fee
<Jaquee> that was my idea
<Jaquee> so the magic happens in desktop wallet.
<Jaquee> on mobile you only login as usual
<vtnerd> yeah, but its doable but I would have to think about it a bit
<vtnerd> the backend would still want to verify correctness, so its not stomping on someone elses coins, etc
<vtnerd> but it would be doing much less work
<Jaquee> ok. yeah backend could verify with view key
<Jaquee> that's probably best
<vtnerd> yes. but it would only be looking at the set of outputs specified in the RPC call which will be much faster
<Jaquee> exactly
<endogenic> Jaquee: ic
<vtnerd> yeah so the import out will be much easier, and you can almost do it with unspent_outs call
<vtnerd> *export I think is what normal people say
<Jaquee> yeah. i'm almost done with that
<Jaquee> i'm trying to make wallet2 fully compatible so i can switch between light wallet and using a normal node
<vtnerd> well I meant but the export from the backend side, but ok interesting that you've already started the wallet side
<vtnerd> yup, understood
<zueore> how monero is configured to only cpu/gpu mining?
<Jaquee> so, to summarize. if i create the mymonero wallet first i can already switch between full and light wallet mode. This import call i'm suggesting also makes it possible to switch from an old full wallet never initialized on mymonero.
<vtnerd> ok


Post tags : Dev Diaries, Monero Core, Cryptography