[01:01:08] * ColonelJ [ColonelJ!~ColonelJ@unaffiliated/colonelj] has quit (Quit: leaving).
[01:04:31] * bitmapper [bitmapper!~bitmapper@142.162.167.202] has quit (Ping timeout: 265 seconds).
[03:13:54] * Sgeo [Sgeo!~Sgeo@ool-18b982ad.dyn.optonline.net] has quit (Read error: Connection reset by peer).
[03:22:27] * Sgeo [Sgeo!~Sgeo@ool-18b982ad.dyn.optonline.net] has joined the channel.
[05:13:14] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has joined the channel.
[07:02:01] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has quit (Quit: My MacBook has gone to sleep. ZZZzzz…).
[07:50:46] * Sgeo [Sgeo!~Sgeo@ool-18b982ad.dyn.optonline.net] has quit (Read error: Connection reset by peer).
[08:46:52] <^alex> we have personal objections to most config management tools because they all work on the 'machines are livestock' model
[08:47:15] <^alex> whereas we run machines like pets
[09:29:31] * esph [esph!~weechat@unaffiliated/esph] has quit (Ping timeout: 246 seconds).
[10:58:43] * esph [esph!~weechat@unaffiliated/esph] has joined the channel.
[12:29:37] * ColonelJ [ColonelJ!~ColonelJ@unaffiliated/colonelj] has joined the channel.
[13:12:25] * esph [esph!~weechat@unaffiliated/esph] has quit (Ping timeout: 265 seconds).
[13:22:31] * esph [esph!~weechat@unaffiliated/esph] has joined the channel.
[14:08:26] <HackerFoo> Nix is definitely worth the time spent learning it.
[14:08:57] <HackerFoo> I'd say it improves software as much as Git/distributed VC did.
[14:10:20] <HackerFoo> On a Nix system, you can isolate changes in the whole system down to a single line of code, so that when something breaks or changes, you can determine with nearly 100% confidence what caused it.
[14:10:37] <HackerFoo> Or alternatively, you can make exactly one change at a time.
[14:11:01] <HackerFoo> This is important for high reliability systems, or where testing a change can take days.
[15:05:06] <^alex> our lack of interest in nix is because the backend implementation is a C++ monstrosity and the focus of develompment is on systemd linux, none of which are technologies we generally trust. 
[15:05:33] <^alex> nix is so close, yet so far from what we actually want
[15:06:04] <^alex> we also don't like git's design or implementation and use fossil, so, 
[15:07:04] <^alex> been building, in the back of our head, a OCaml-based config management tool structured conceptually after malware
[15:12:10] * bitmapper [bitmapper!~bitmapper@142.162.167.202] has joined the channel.
[15:22:55] * Sgeo [Sgeo!~Sgeo@ool-18b982ad.dyn.optonline.net] has joined the channel.
[16:37:55] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has joined the channel.
[16:42:04] <HackerFoo> What about Guix, then? I think the implementation language is mostly irrelevant unless you need to modify Nix. I've read through some of it and got a PR accepted; it's not bad.
[16:42:40] <HackerFoo> It's a lot easier to modify Nix than most other C++ code I've worked on.
[16:48:09] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has quit (Quit: My MacBook has gone to sleep. ZZZzzz…).
[17:28:45] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has joined the channel.
[17:37:43] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has quit (Ping timeout: 260 seconds).
[18:10:48] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has joined the channel.
[21:16:30] <doublec> ^alex: what does structured conceptually after malware mean?
[21:17:21] <doublec> When I ran NixOS the main issue I encountered was lack of timely security updates, unless you managed nixpkgs yourself.
[21:32:14] <^alex> doublec, you know how each CM tool has its metaphor? salt is a stack of grains, chef uses recipes and a kitchen metaphor, ansible uses a baseball metaphor?
[21:33:13] <^alex> the thing we're too ADHD to actually sit down and build is called `capsid` 
[21:33:42] <HackerFoo> doublec: It's better to think of NixOS as a distributed distro, in the way that Git allows distributed versioning.
[21:34:13] <HackerFoo> Managing the packages you care about most is part of it.
[21:34:40] <^alex> (a virus's shell)
[21:36:55] <doublec> HackerFoo: yeah, I ended up basically managing my own distro.
[21:37:33] <doublec> HackerFoo: things are better now though - I see they have automated ways of updating packages to get new versions instead of waiting for a maintainer.
[21:38:15] <^alex> in actual practice we have a shoddy collection of ansible scripts for a patchwork collection of half of our machines
[21:40:11] <^alex> shrug emoji