monero-site/_posts/2019-07-28-logs-for-the-dev-meeting-held-on-2019-07-28.md
el00ruobuob 45832b719e
Meeting logs for the week of 2019-07-22 to 28
+ newline pb solved on i2p meeting
2019-07-30 08:23:51 +02:00

15 KiB

layout title summary tags author
post Overview and Logs for the Dev Meeting Held on 2019-07-28 Development status, Code & ticket discussion, Moving off GitHub, and miscellaneous
dev diaries
core
crypto
el00ruobuob / rehrar

Logs

<rehrar> Alright, it's time to start ladies and gentlemen.
<rehrar> 1. Greetings
<rehrar> Anybody out there?
<rbrunner> Nope. All people on holiday.
<rbrunner>
<vtnerd__> yes
<rehrar> well, the three of us can have a grant ol time then
<rehrar> just pinging hyc moneromooo dsc_ selsta and dEBRUYNE also
<rehrar> maybe a luigi1111 and fluffypony in the mix too, who knows
<rehrar> either way
<rehrar> 2. What's been completed since last meeting?
<dsc_> dont ping me broooo
<rehrar> _csd
<rehrar> there, I took it back
<dEBRUYNE> CLI v0.14.1.2 has been released
<dEBRUYNE> Which is mostly bug fixes I guess
<dEBRUYNE> mooo is holding a vacation, but he provided me a list of stuff that is ready to be merged
<dEBRUYNE> So I am working with luigi to get the merge queue relatively empty
<kinghat> 🙏
<rehrar> ah, did not know mooo was on vacation
<rehrar> I hope he ignores my ping then. Sorry!
<dEBRUYNE> For the GUI, we tagged 0.14.1.2 and we're waiting on pony to finish the builds
<dsc_> I'm more than half way into integrating i2p-zero and tor into GUI
<dEBRUYNE> Main new feature is optimized Tails support
<dEBRUYNE> Perhaps dsc_ can share a bit more detail
<dsc_> Sure thing.
<rehrar> dsc_: that's actually super exciting
<dsc_> So Tails integration is a collection of improvements that improve user experience for our Tails users
<dsc_> You would be suprised how big the audience is there
<dsc_> So, starting from 0.14.1.2 it will be very easy to use GUI on Tails
<dsc_> As for i2p-zero, I need a couple of days then that's done too
<dsc_> I suggest to package i2p-zero with the GUI release
<dsc_> Which can pose problems for our reproducible build efforts
<dEBRUYNE> I don't see much harm in packaging it with the GUI, especially if the user first has to check a checkbox to turn it on
<dsc_> but that's a problem for later
<dsc_> That also
<rbrunner> Shouldn't the CLI pack it also then?
<dEBRUYNE> I think most CLI users are able to download the package themselves
<dEBRUYNE> And get it running (especially given the instructions present on Github)
<dEBRUYNE> By contrast, GUI users need a checkbox and not tons of steps
<rbrunner> Ok, makes sense, more or less ...
<dsc_> The fact we package i2p-zero is a convienence feature. I'm also 'integrating' support for Tor but we wont be packaging that. That's basically running Tor and using socks proxy at :9050
<rehrar> we GUI people are scrubs
<xmrmatterbridge> <xmr-romine> is there already some doc about running monerod over tor ?
<dsc_> https://github.com/monero-project/monero/blob/master/ANONYMITY_NETWORKS.md
<dEBRUYNE> https://github.com/monero-project/monero/blob/master/ANONYMITY_NETWORKS.md & https://github.com/monero-project/monero#using-tor
<xmrmatterbridge> <xmr-romine> thanks guys
<rehrar> hyc any update on randomx audit?
<rehrar> The last one?
<dsc_> One more thing regarding i2p-zero.. It can take up to 1-2 minute(s) for the local socks proxy to become 'active' which means, if the user has i2p-zero enabled and starts up the GUI to create a quick tx ... might have to wait a minute or two
<dsc_> This is related to how i2p-zero works
<dsc_> so beware of that limitation
<rbrunner> So it's i2p-two, from "two minutes"?
<kinghat> can it check if i2p is currently running?
<dsc_> rbrunner: haha
<rehrar> it's how i2p works in general, no? Getting bootstrapped into the network, I mean.
<dsc_> rehrar: most likely
<dsc_> kinghat: it can
<rehrar> under bad circumstances it can take even longer
<rehrar> I wonder if there can be some kind of notification to let users know this
<vtnerd__> tor should have similar issue iirc, building circuits isn't immediate
<kinghat> also, are we still going to do the beta builds for the gui?
<dsc_> vtnerd__: Sure, but not 1-2 minutes?
<dsc_> kinghat: yes
<kinghat> 👌
<vtnerd__> I think it might be longer than that, at least it seemed like forever yesterday when I was testing
<vtnerd__> actually no, I've run tails once before, it takes some amount of minutes, just don't remember the exact time frame
<dsc_> vtnerd__:
<dsc_> $ time (sudo service tor restart && sleep 2 && torify curl icanhazip.com)
<dsc_> 171.25.193.77
<dsc_> real0m2.759s
<vtnerd__> if your not using outbound connections tor destroys the circuits and goes into an idle mode too, which isn't a problem if "noise" is enabled (because its always sending)
<rehrar> there goes dsc_ bragging about his internet again
<dsc_> yes, my internet.
<vtnerd__> it keeps state in a file that it might re-use, so you'd have to a cold restart
<dsc_> vtnerd__: fair enough
<rehrar> alright, anything else?
<dEBRUYNE> vtnerd__: I was wondering if you intend to pick this up somewhere in the future? https://github.com/monero-project/monero/pull/2317
<vtnerd__> see https://github.com/monero-project/supercop/pull/2
<vtnerd__> this one and the next one after that PR is a bit rough because its x86 ASM
<rehrar> that was a hilarious conversation to witness
<rehrar> person 1: i was wondering if git link
<rehrar> person 2: Well you see it's like git link
<dEBRUYNE> Thanks
<vtnerd__> Im also not certain that having a separate repo for this is beneficial, but I guess some other project could make use of the custom hooks
<vtnerd__> my original code should need some re-working since the device/ledger stuff got slammed in, but I look at recently and I think its possible to rebase/update that without crying
<vtnerd__> Im also not certain whether that supercop repo should be an external git thing, or something that you manually have to do
<vtnerd__> I'm thinking the latter, since people are blasted without another submodule
<vtnerd__> *are not
<dEBRUYNE> I see, seems like a thing that warrants more in depth discussion in a future dev meeting
<rehrar> With more attendees?
<rehrar> serious question
<rbrunner> And maybe some more background info to prepare
<rehrar> who should be present for such a discussion?
<rehrar> we can add it to next meeting as an item and try to ping the people to be there
<vtnerd__> the background is either the supercop code goes as a submodule and it must be synced by every person
<vtnerd__> or only those that which to accelerate wallet scanning drop the repo into a folder and toggle a cmake bit
<rbrunner> So that's not something for the normal, general release then?
<vtnerd__> its not needed for testing, unless you need to test the feature specifically
<vtnerd__> no, likely it would be in the normal general release, but I suppose it doesn't have to be
<vtnerd__> this came up because ryo recently added support for basically the same thing, and it should cut wallet-scanning time in half
<vtnerd__> although its probably dependent on your HD speeds too
<rehrar> alright, any more discussion on this?
<hyc> re: randomx audit, Quarkslab are finalizing their report, said we may get it by middle of next week
<rehrar> nice!
<rehrar> does anyone have anything else that they want to discuss?
<hyc> since they've been opening github issues as they arose, there's not likely to be anything in the report we haven't already seen
<ErCiccione> yep! https://github.com/monero-project/meta/issues/236#issuecomment-515669422
<ErCiccione> GitHub is starting to censor people coming from countries under US sanctions, we should definitely migrate away
<ErCiccione> i don't know how the core team feel about it
<rehrar> do people here have any thoughts about the Github/Gitlab thing, especially in light of the recent Github things?
<vtnerd__> those countries listed probably require some vpn/tor/i2p opsec to contribute anyway ... ? I'm certainly not attempting to stop a switch to a self-hosted gitlab, but meh
<rbrunner> If such a switch was so easy it would probably have happened already, I think, after the Microsoft acquisition ...
<vtnerd__> or maybe not, but they would have to after that hammer that was thrown
<rbrunner> With all submodules still on GitHub, for example
<vtnerd__> the support team for the gitlab instance are probably the most relevant for this
<ErCiccione> rbrunner: the point was to test monero-site on the self-hosted gitlab first, and then see what to do with the other repos. But has been almost an year now
<hyc> I've always favored moving to self-hosted gitlab and away from github
<rbrunner> Yeah, I remember. But I think the site is a particularly simple case
<vtnerd__> I mean most devs should have a copy of the code and history luckily, but its the discussions that can get purged if not backed up properly
<hyc> but this particular issue only affected private, paid github accts
<hyc> (i.e., github closing off accts from sanctioned countries)
<ErCiccione> hyc: i guess that depends by how they are forced to enforce the sanctions
<hyc> https://twitter.com/natfriedman/status/1155533738278699008
<hyc> https://twitter.com/natfriedman/status/1155311121038864384
<hyc> they're splitting some hairs I guess. trade sanctions only affect commerce. if you paid for a private acct, you're affected.
<rbrunner> Has any other notable cryptocurrency team already made such a switch?
<hyc> if you're on a public/free acct, you're not affected
<rehrar> rbrunner: define notable?
<rbrunner> I mean, some "interesting" team. Top 100 coin, maybe? Sizeable community, I mean
<ErCiccione> thanks for the links. it's still a very precarious situation.
<ErCiccione> rbrunner: sia
<rbrunner> And complexity of the code
<hyc> ErCiccione: agreed. as I already said, I'm in favor of moving off.
<kinghat> i think ppl are worried about contributions after moving off github. i dont think it would make a difference.
<hyc> again, just my perspective: the OpenLDAP Project maintains our own repo. we only mirror to github, because it seems you're not a real project these days if you're not on github
<hyc> out github mirror is read-only
<hyc> our*
<ErCiccione> kinghat: i agree.
<dsc_> I keep repeating myself but I question people whose first thought is 'gitlab' when thinking of a git alternative
<dsc_> There is more! Noooo.. Yes!!
<dsc_> Consider it...
<hyc> gitea?
<dsc_> for example, yep
<hyc> I don't really care which alternative is used. all I care is self-hosted.
<ErCiccione> iirc gitea was considered
<hyc> you could make an argument that being off github adds a barrier to entry - people don't like having to create accounts just to interact with another project
<rehrar> In favor of gitea
<dsc_> Even more so when considered gitlab is a company, enterprise, they have funding, maybe they have shareholders - and capitalism is evil. Therefor, gitea
<rehrar> hyc which is why I am arguing for a federated git ecosystem
<rehrar> open issues on another git based project management site from yours, etc
<ErCiccione> let's not forget that we already have half repos on github and half on gitlab. That's not ideal
<hyc> rehrar: agree, federated identity would be preferable
<rehrar> let's build it
<rehrar> we can have it out by Wednesday at the latest
<hyc> +1
<dsc_> ok sounds good
<ErCiccione> i want 2
<rehrar> that'll take an extra day
<hyc> this could be a long conversation, but totally - we need to rethink the notion of accounts and identities
<rehrar> But for reals though this is an ongoing discussion and...
<rehrar> ^ hyc
<rehrar> we have the issue open about this that ErCiccione
<rehrar> although that one is about moving to gitlab, I think
<rehrar> so maybe another one can be opened for discussion on which platform?
<dsc_> +1 for centralized authentication service
<dsc_> (hosted by monero)
<rbrunner> Oh the added complexity
<dsc_> rbrunner: for whom?
<rbrunner> The system in general, I mean.
<dsc_> It could simplify things
<hyc> not quite centralized authent. just public key based.
<rbrunner> Something more that can go wrong
<ErCiccione> rehrar it is yes, as hyc says for me the important is to move away from a centralized place. I think gitlab is the best choice because we have been testing it for an year, but sure let's discuss alternatives
<hyc> we could use monero wallet addresses as our identities
<rehrar> hyc I've been thinking of using Monero signatures as password alternatives
<dsc_> thats a cool idea
<rehrar> public key as ID
<rehrar> signature from private key as password
<kinghat> does that work with subaddys or am i missing something?
<rehrar> but then losing your seed is even a bigger deal
<rehrar> or getting it stolen
<hyc> good point
<rehrar> as long as there are standard ways of resetting or something
<rehrar> it's no more annoying than losing or getting your regular password stolen
<hyc> having a reset facility implies a centralized admin authority
<dsc_> Call the Monero helpdesk and provide them your miaden name, your first car brand and your date of birth
<hyc> ^
<rehrar> enforced 2/3 MFA then
<hyc> should we wrap up the mtg before continiung this conv?
<rehrar> but yeah, we'll call it here
<rehrar> unless anyone has any last minute things?
<vtnerd__> I hope to finally issuing some PRs related to the i2p/tor white noise stuff, took some time to architect this bt hopefully. nothing else needs to be reported by that really other than look for more stuff to review
<rehrar> cool deal!
<rehrar> well everyone, thanks for making the time today.
<rehrar> You're all free to disperse. Coffee on the little table outside the meeting room.