gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
drakonis has quit [Ping timeout: 252 seconds]
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos-chat
jasongrossman has quit [Read error: Connection reset by peer]
drakonis has joined #nixos-chat
drakonis_ has quit [Ping timeout: 250 seconds]
<steveeJ> I still don't quite understand how things from nixpkgs/master land in a channel eventually. what is the ETA of https://github.com/NixOS/nixpkgs/pull/56839 landing on nixos-19.03?
<{^_^}> #56839 (by avnik, 1 week ago, merged): Fix cairocffi build
<joepie91> steveeJ: it's my understanding that a channel advances when the build of a branch/release 'settles', ie. all of the required builds on the build server have completed and are available
<steveeJ> I just noticed there's are release branches on nixpkgs too. so the manual process is to merge the changes from master into nixos-19.03 and then the CI/CD system will do the rest?
<joepie91> steveeJ: yep; and there's a separate nixpkgs-channels repo that tracks the channels
<steveeJ> I've found the latter a while ago ;-)
<steveeJ> joepie91: thanks for your response!
<joepie91> steveeJ: you can find the build status etc. if you're interested at https://hydra.nixos.org
<joepie91> though the UI will take a little getting used to probably :)
drakonis has quit [Ping timeout: 252 seconds]
jasongrossman has joined #nixos-chat
drakonis has joined #nixos-chat
iqubic has quit [Ping timeout: 264 seconds]
drakonis has quit [Ping timeout: 250 seconds]
drakonis_ has joined #nixos-chat
drakonis has joined #nixos-chat
drakonis_ has quit [Ping timeout: 258 seconds]
drakonis1 has joined #nixos-chat
<manveru> steveeJ: http://howoldis.herokuapp.com/ :)
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
jasongrossman has joined #nixos-chat
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
pie__ has joined #nixos-chat
pie___ has quit [Ping timeout: 272 seconds]
jasongrossman has joined #nixos-chat
jasongrossman has quit [Remote host closed the connection]
drakonis1 has quit [Quit: WeeChat 2.3]
jasongrossman has joined #nixos-chat
jasongrossman has quit [Remote host closed the connection]
jasongrossman has joined #nixos-chat
jasongrossman has quit [Remote host closed the connection]
endformationage has quit [Quit: WeeChat 2.4]
jasongrossman has joined #nixos-chat
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
<sphalerite> grmph. Trying to set up a mini PC without a monitor is a pain
jasongrossman has joined #nixos-chat
<andi-> why? Prepare system closure in VM, copy/install on the real disk? :-)
<sphalerite> andi-: yeah pretty much what I did (except no VM), but it's not booting and I can't tell why
<andi-> serial?
<sphalerite> don't have the kit
<andi-> if you think the kernel is loaded alreayd try `netconsole`
<sphalerite> no, I'm pretty sure it's a firmware issue
<sphalerite> something stupid like the system clock not being set
<andi-> grml, 19.03 broke ipv6-only containers back in october...
<srhb> andi-: I think we're too lax on requiring tests to pass..
<andi-> yeah
<srhb> I've seen a lot of breakage that only happens because we're not making the test set big enough..
<andi-> I am writingone now
<srhb> Doesn't it already exist?
<srhb> Just not in the tested set.
<andi-> it isn't tested
<andi-> there is a very simple container test that I've seen that tests veth pairs
<srhb> Ah, right.
<ivegotasthma> nix repo has 6666 commits in it, check'em
<srhb> andi-: Still, it's not even in the tested set.
<andi-> it tests ipv6 with a weird netmask that nobody would ever use for addressing in ipv6 (/7)
<srhb> ivegotasthma: I noticed that too :-P
<srhb> \m/
<andi-> srhb: yeah, I'll think about adding it there
<srhb> andi-: Thanks :)
<srhb> Similarly, ldap appears to be quite probably broken since january. The test revealed it, but it's not in the tested set.
<srhb> Sadface.
<srhb> The simple ipv6 containers test is also broken.
<andi-> :/
<srhb> command `ping -n -c 1 fc00::1' did not succeed (exit code 1)
<andi-> yeah
<ivegotasthma> srhb: https://i.imgur.com/phSxidM.jpg
<srhb> Considering our advanced CI capabilities, we're missing some grand opportunities. :D
<srhb> ivegotasthma: lol
<andi-> https://s.h4ck.space/broken-sh.svg i made that for my new version of broken.sh but fits nicely :-)
<srhb> augh, that's painful :P
<andi-> I have the feeling that in general we start the containers too early and networking fails due to yet unknown issues
<andi-> srhb: mind having a look at this change? https://github.com/andir/nixpkgs/commit/04af592789ae53d7ca184587a0349daffb61196c I have a few appointments today and might not get to PR that until tonight.
<srhb> Sure
<andi-> Thanks
<srhb> andi-: I *think* that's fine, but it's not really my area of expertise :-)
<andi-> srhb: it basically just adjusts the previous changes... What I do not like and that I still want to understand is if we really need all of those bash-if-else constructs while we know very well what was configured via the module system.. probably related to imperative containers?
<srhb> andi-: Hmm, possibly.
<srhb> I guess I'll have to check the history
<andi-> Yes
<srhb> In that case, yeah, I think what you're doing is a reasonable fix for now, but I agree with your assesment.
<srhb> It's all a bit fragile.
<srhb> And by a bit I mean a lot
__monty__ has joined #nixos-chat
<flokli> srhb: weird that the ldap test is broken. It did work on my machine...
* flokli is confused
<flokli> Maybe an older iteration of the PR, and there was a typo sneaked in?
<srhb> flokli: I am confused too. It looks like that change to runtime dir is the culprit. It doesn't _feel_ like a typo, more that it just shouldn't work..
<srhb> I might be completely wrong though, any insights welcome :-)
<srhb> As far as I can see, nslcd fails to start.
srhb has quit [Quit: ZNC 1.7.1 - https://znc.in]
srhb has joined #nixos-chat
<flokli> srhb: I'll try to take a look at it later. Need to fix some electric cabling foo in a basement, wish me luck :-D
<srhb> flokli: Don't die. :-)
<flokli> I'll try not to
Myhlamaeus has quit [Quit: WeeChat 2.2]
Myhlamaeus has joined #nixos-chat
drakonis1 has joined #nixos-chat
Myhlamaeus has quit [Quit: WeeChat 2.2]
Myhlamaeus has joined #nixos-chat
drakonis has quit [Remote host closed the connection]
goibhniu has joined #nixos-chat
<gchristensen> anyone here familiar with mastadon?
<manveru> gchristensen: more or less
goibhniu has quit [Ping timeout: 246 seconds]
<__monty__> Mastodon's kinda weird. Servers can choose not to accept messages from certain other servers. Sounds like a good idea, freedom and everything. In practice it apparently means you might be missing parts of conversations and not know.
<manveru> sounds just like real life :P
<gchristensen> haha
<gchristensen> is there an advantage to going with an existing server, over running one's own?
<manveru> but anyway, the idea behind it was that everyone who cares enough about that issue can run their own instance...
<manveru> the main advantage is that you don't have to run it yourself i guess, the server is still not packaged because it's so hard to do :|
<__monty__> manveru: Yeah, that's not the problem. The problem is the UX. You might see bits of the conversation and not know you're missing out on other bits.
<gchristensen> being federated, it feels there is great opportunity cost to picking an instance whereas there isn't anything about being on twitter dot com
<manveru> __monty__: can't say i disagree, but not sure i'd come up with a much better solution myself
<__monty__> I agree, it's a hard problem.
<manveru> it's kinda like shadowbanning
<__monty__> Irc is federated too and doesn't have that problem at least as seriously. It's more all-or-nothing.
<gchristensen> not *really* federated.
<manveru> IRC requires much closer collaboration i think
<manveru> you can't just federate with freenode just for fun
<gchristensen> freenode is more sharded than federated
<manveru> it's more of a load-balancer :)
<__monty__> It's closed federation but that doesn't make it not federation.
<__monty__> I don't like how "federation" has become a buzzword btw.
<gchristensen> well anyway the commitment to a domain has me hesitant to get using it
<manveru> i just chose one mostly based on interests :)
<manveru> it's useful to bootstrap some random contacts you can see in your feed at the beginning
<gchristensen> aye
<joepie91> gchristensen: there's a lengthy thread about the problem of continued availability of instances and account migration: https://github.com/tootsuite/mastodon/issues/177
<{^_^}> tootsuite/mastodon#177 (by hach-que, 2 years ago, open): Support account migration
<joepie91> as of yet unsolved
<gchristensen> another reason I don't feel comfortable investing in an account on mastadon.social
<gchristensen> I'm interested in giving it a go, but don't want to pay $7/mo to play around.
<joepie91> gchristensen: huh? $7/mo?
<gchristensen> masto.host :)
<joepie91> oh, right. not sure I see much point in instance hosting, to be honest
<joepie91> seems to kind of defeat the point
<gchristensen> well I'd like to keep my account with me even if my instance host dies
<joepie91> sure, but that's more an addressing issue than anything else, and there's no reason that can't be solved without dedicated instances
<joepie91> by for example supporting domain aliasing within an instance
<andi-> those instance hosters are often not about running you instance (that is definitly a use case) but about building your community with a set of rules etc.. So not just federation is a thing for mastodon but also following some kind of guidelines that you / your community builds.
<gchristensen> as long as the name of the instance is part of your identity, I disagree :/
<flokli> srhb: Back from cablework. No electric shocks, still alive :-)
<andi-> I was trying to say that there is more then keeping an identity. There is multiple scenarios where federation is deliberatly not wanted. As for the "replace twitter.com" case keeping you identity is ofc the more important criteria. I tried packaging pleroma a few times but gave up on some of the dependencies.. For the time being I am using an instance that is run by admins I trust. I prefer using that
<andi-> over twitter since it aleast doesn't put everything in the hand of one entity. One can even argue that using Mastodon is less private or less privacy friendly as federation inherently needs more trust (the way it is done).
<joepie91> gchristensen: what I'm saying is that solving the addressing problem would be perfectly possible by having people use their own domain, but pointing it at a third-party instance; such that the identity is tied to one's own domain name, and that identity can be ported to a different instance (in terms of references, not in terms of data) by pointing the domain name to a different instance
<gchristensen> sure
<joepie91> this doesn't require every domain to have its own dedicated instance; just for instances to be capable of handling >1 domains
<gchristensen> I guess I'll wait for that before trying it, then
<joepie91> yeah, that'd be a totally reasonable decision IMO :P
<__monty__> Hmm, another consequence of the censoring stuff is you can only ever migrate to a server with *identical* or more likely less stringent censoring. Otherwise you lose history.
goibhniu has joined #nixos-chat
<joepie91> jD91mZM2: how come rnix' InheritFrom node type has a `from` method for obtaining the (optional) source of the inherited identifiers, but not something like an `identifiers` method to get a list of all the identifiers? just calling .children() wouldn't suffice since that might still contain an InheritFrom entry
<joepie91> err, sorry, Inherit node type*
goibhniu has quit [Ping timeout: 272 seconds]
goibhniu has joined #nixos-chat
goibhniu has quit [Ping timeout: 272 seconds]
goibhniu has joined #nixos-chat
<manveru> man, handling rebar3 really took some effort, but i think i got most of the edge-cases figured out...
<manveru> was like a week-long game of whack-a-mole :P
<sphalerite> Does anyone know what I should use as a swap device nowadays?
endformationage has joined #nixos-chat
<gchristensen> I've got a regular partition, setup under luks
<sphalerite> I mean what sort of physical device
<sphalerite> my nano pi only has 4GB of RAM, which can get a bit tight when building stuff and using zfs
<sphalerite> (also I wonder why there don't seem to be any rk3399 device with more than 4GB RAM)
<samueldr> sphalerite: rk3399 can't do more than 4GB
<sphalerite> oh damn
<sphalerite> why not?
<samueldr> that's why
<samueldr> I'm not an engineer! (and I didn't ask)
<sphalerite> :D
<samueldr> but the libre.computer peep said so and they are trustworthy for that
goibhniu has quit [Ping timeout: 245 seconds]
drakonis1 has quit [Quit: WeeChat 2.3]
pie_ has joined #nixos-chat
<pie_> did there ever end up being an rfc or github issue or something for extending package meta info with other conventions? currently im thinking of stuff like having a "packager documentation" field or something like that, where i can put references i used to package something
<pie_> well, I made this I guess https://github.com/NixOS/nixpkgs/issues/57833
<{^_^}> #57833 (by deliciouslytyped, 1 minute ago, open): Additional meta-information conventions
drakonis has joined #nixos-chat
aszlig has quit [Quit: Kerneling down for reboot NOW.]
<joepie91> I am... confused. my partially-completed Nix parser in JS seems about equally fast as rnix in Rust?
* joepie91 is surely measuring something wrong here...
aszlig has joined #nixos-chat
<emily> V8 is magic.
<emily> (also I think one of the rust parser libraries is pathetically slow or something?)
<joepie91> emily: this one is using rowan
__monty__ has quit [Quit: leaving]