gchristensen changed the topic of #nixos-dev to: NixOS Development (#nixos for questions) | NixOS 19.09 is released! https://discourse.nixos.org/t/nixos-19-09-release/4306 | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | https://r13y.com | 19.09 RMs: disasm, sphalerite; 20.03: worldofpeace, disasm | https://logs.nix.samueldr.com/nixos-dev
ris has joined #nixos-dev
_ris has quit [Read error: Connection reset by peer]
das_j has joined #nixos-dev
Scriptkiddi has joined #nixos-dev
ajs124 has joined #nixos-dev
orivej has quit [Ping timeout: 268 seconds]
_ris has joined #nixos-dev
ris has quit [Ping timeout: 258 seconds]
<gchristensen> infinisil: turns out I'm at an infosec meetup whose topic is secrets management.
_ris has quit [Read error: Connection reset by peer]
_ris has joined #nixos-dev
Synthetica has quit [Quit: Connection closed for inactivity]
_ris has quit [Ping timeout: 272 seconds]
<infinisil> Oh neat
lovesegfault has quit [Quit: WeeChat 2.7]
lovesegfault has joined #nixos-dev
bhipple has joined #nixos-dev
<evils> what's the current argument w.r.t. asciidoc's toolchain's closure size?
<samueldr> with any tooling, the argument is that docbook's is extremely slim
<samueldr> which is important considering on nixos rebuild you build the manual
<infinisil> qyliss cut asciidoctor's closure size by a lot with #77149
<{^_^}> https://github.com/NixOS/nixpkgs/pull/77149 (by alyssais, 1 week ago, merged): Reduce AsciiDoctor closure size
<lovesegfault> What's the tool they used to gen Rust's docs?
<lovesegfault> mdbook?
<infinisil> And I believe pandoc's closure could also be reduced with some effort
<samueldr> yep, I was answering in a general manner why it matters
<evils> infinisil: k, the mention of "on master: 80% smaller" wasn't clear to me
<lovesegfault> This is what I'm referring to: https://doc.rust-lang.org/stable/book/
<lovesegfault> infinisil: pandoc has a zillion deps, no?
<infinisil> evils: I guess I should use the master number
<infinisil> lovesegfault: static compilation is one idea
<lovesegfault> Ah, yeah, it's mdBook (https://github.com/rust-lang/book)
orivej has joined #nixos-dev
<drakonis1> just to make sure i get this right, touching stdenv causes mass rebuilds?
<gchristensen> usually
<drakonis1> aight, because i'm looking into ways i can properly improve package metadata and oh man this is a a rube goldberg machine
<gchristensen> meta isn't part of the hash calculation
<drakonis1> right, good.
ajs124 has quit [Remote host closed the connection]
das_j has quit [Remote host closed the connection]
Scriptkiddi has quit [Remote host closed the connection]
Scriptkiddi has joined #nixos-dev
das_j has joined #nixos-dev
ajs124 has joined #nixos-dev
lovesegfault has quit [Quit: WeeChat 2.7]
<drakonis1> i'm spelunking through nixpkgs and oh man, customizing package inputs shouldn't be done when calling the package
<drakonis1> infinisil: what's this modules-ng repository you got there
bhipple has quit [Remote host closed the connection]
drakonis1 has quit [Quit: WeeChat 2.6]
<infinisil> ,tell drakonis1 modules-ng is some brainstorming for a better and faster nixos module system
<{^_^}> infinisil: I'll pass that on to drakonis1
Cale has quit [Ping timeout: 260 seconds]
Cale has joined #nixos-dev
Jackneill has joined #nixos-dev
Jackneill has quit [Max SendQ exceeded]
Jackneill has joined #nixos-dev
drakonis1 has joined #nixos-dev
FRidh has joined #nixos-dev
<drakonis1> apparently i have a tell
<{^_^}> drakonis1: 2 hours, 26 minutes ago <infinisil> modules-ng is some brainstorming for a better and faster nixos module system
<drakonis1> there it is
<niksnut> we should have a better module system as a language feature
<niksnut> doing it as a library feature is a bit of a dead end, since it will always be slow
<drakonis1> its currently missing a escape hatch for whole configuration files, rather than composed file plus extraconfig
<LnL> FRidh: I'll keep an extra eye on staging-next, but ping me if you notice any darwin shenanigans
drakonis1 has quit [Ping timeout: 272 seconds]
Synthetica has joined #nixos-dev
__monty__ has joined #nixos-dev
<Synthetica> ryantm: Do you intend to keep nixpkgs-update as a local batch command, or do you want to deploy it as a continuous service at some point?
<Synthetica> Oh, it's 3 am there, probably not awake 😅
<aanderse> is it possible to use mkRemovedOptionModule and mkChangedOptionModule on options defined under an attrsOf submodule ... ?
<yorick> Synthetica: the main thing stopping it is the logs I think
<yorick> Synthetica: so help wanted to write a better result aggregation thing so people can see why their packages can't auto-update
bennofs has quit [Remote host closed the connection]
bennofs has joined #nixos-dev
<infinisil> niksnut: I believe it's possible to make it fast and almost fully backwards compatible with the current module system
<infinisil> Which is probably pretty important, because otherwise we need to rewrite many many things
<infinisil> aanderse: Oh i forgot, these ones don't work by default under submodules, because they set `warnings = ...` which is usually not defined under submodules
<aanderse> infinisil: yeah in the past i've just left the options and then added warnings to the main module
<aanderse> but i thought hey maybe i should ask someone if there is a better way to do that before i comment
<infinisil> aanderse: You can add such a warnings option and then propagate all warnings to the outside
<infinisil> Grep for `warnings =`, some modules do that
<aanderse> infinisil: thanks, will grep
<infinisil> Ideall warnings and assertions would be integrated into the module system such that it's always an option
<aanderse> yeah that would be better
<infinisil> Hm actually that might be pretty simple, maybe i'll take a stab at it
<aanderse> infinisil++
<{^_^}> infinisil's karma got increased to 187
<infinisil> But also, i think warnings at least warnings should be silenceable, which might be nice to implement too, hmmyy
<infinisil> hmm..
<ryantm> Synthetica: I just started getting it onto the nix-community build server https://github.com/nix-community/infra/pull/6 and the next step will be cleaning that up some and fully automating it.
<{^_^}> nix-community/infra#6 (by ryantm, 2 days ago, merged): build01: add initial nixpkgs-update configuration
<ryantm> Synthetica: Why do you ask?
<Synthetica> ryantm: I was just wondering, since the current setup seemed suboptimal
<ryantm> Synthetica: Like how it runs in batches/spurts?
<Synthetica> ryantm: Yeah
<ryantm> I'd like it to run more continuously, also, I'd like some features of it to be exposed as command line tools for people making manual PRs. Those are my two top priorities with the project now.
<Synthetica> (Also, I wrote a script to fetch new releases from github instead of repology, if you're interrested https://github.com/Synthetica9/nixpkgs-update-github-releases )
<ryantm> Cool!
<ryantm> How long does it take to run?
<ryantm> Synthetica: I just tried running it and it failed: https://gist.github.com/ryantm/bf88b89f7eac126710a208b898b4d25f
<Synthetica> Strange
<ryantm> Running nix 2.3.1
<Synthetica> I'm runnning 2.3.2, perhaps there's a difference
<Synthetica> But I don't think there should be
<Synthetica> ryantm: Could you pull and try again?
noonien has quit []
noonien has joined #nixos-dev
<ryantm> Synthetica: same error with the new flag in there
<Synthetica> ryantm: What's the output of `nix-instantiate ./loadMetaFromPath.nix --arg url "https://github.com/nixos/nixpkgs/archive/master.tar.gz" --eval`?
<Synthetica> ryantm: Try again?
<Mic92> ryantm: Do you have a standalone-version one could use to just update one nixpkgs attribute to the latest version semi-automatically?
drakonis1 has joined #nixos-dev
<ryantm> Synthetica: Cool it worked until it ran out of API calls!
<ryantm> Mic92: Sort of but it is too hard to use.
<Synthetica> ryantm: Yeah, github without an api key is kinda stringent
<ryantm> Mic92: It depends on too much external stuff like github, hub, cachix
<Synthetica> ryantm: You can set API_TOKEN env var to <username>:<personal access token>
ixxie has joined #nixos-dev
FRidh has quit [Quit: Konversation terminated!]
rajivr___ has quit []
rajivr___ has joined #nixos-dev
arcnmx has joined #nixos-dev
<arcnmx> So I'm probably doing something very wrong but... having issues with the following no longer working on unstable-small:
<arcnmx> nix build '(with import <nixpkgs> { crossSystem = { system = builtins.currentSystem; useLLVM = true; }; crossOverlays = [ (self: super: { stdenv = super.overrideCC super.stdenv self.buildPackages.llvmPackages_latest.lldClangNoLibcxx; }) ]; }; stdenv)'
<arcnmx> which currently evaluates on nixos-unstable but not -small... I'm not sure what changed or how should I be doing this instead.
sdier has quit []
sdier has joined #nixos-dev
psyanticy has joined #nixos-dev
<arcnmx> oh sorry that's unnecessarily complicated, seems like the issue is just that (import <nixpkgs> { crossSystem = { system = builtins.currentSystem; useLLVM = true; }; }).stdenv won't work at all
ixxie has quit [Ping timeout: 268 seconds]
kalbasit has quit []
kalbasit has joined #nixos-dev
tilpner has quit [Quit: tilpner]
tilpner has joined #nixos-dev
ixxie has joined #nixos-dev
janneke_ has joined #nixos-dev
janneke has quit [Ping timeout: 260 seconds]
ixxie has quit [Ping timeout: 268 seconds]
drakonis1 is now known as drakonis
vdemeester has quit [Ping timeout: 246 seconds]
alunduil has quit [Ping timeout: 252 seconds]
kalbasit has quit [Ping timeout: 272 seconds]
tazjin has quit [Ping timeout: 272 seconds]
jared-w has quit [Ping timeout: 246 seconds]
sorear has quit [Read error: Connection reset by peer]
carter has quit [Ping timeout: 264 seconds]
c00w has quit [Ping timeout: 264 seconds]
aristid has quit [Ping timeout: 252 seconds]
feepo has quit [Ping timeout: 252 seconds]
Synthetica has quit [Read error: Connection reset by peer]
scott has quit [Ping timeout: 252 seconds]
teehemkay has quit [Ping timeout: 245 seconds]
chrisaw has quit [Ping timeout: 264 seconds]
manveru has quit [Ping timeout: 264 seconds]
davidtwco has quit [Ping timeout: 246 seconds]
teozkr has quit [Ping timeout: 246 seconds]
emily has quit [Ping timeout: 248 seconds]
psyanticy has quit [Ping timeout: 272 seconds]
srhb has quit [Ping timeout: 250 seconds]
psyanticy has joined #nixos-dev
tazjin has joined #nixos-dev
kalbasit has joined #nixos-dev
feepo has joined #nixos-dev
alunduil has joined #nixos-dev
davidtwco has joined #nixos-dev
chrisaw has joined #nixos-dev
scott has joined #nixos-dev
aristid has joined #nixos-dev
sorear has joined #nixos-dev
srhb has joined #nixos-dev
manveru has joined #nixos-dev
<infinisil> Well, I can't reproduce it, so I'll just restart it
psyanticy has quit [Read error: Connection reset by peer]
sorear has quit [Ping timeout: 268 seconds]
srhb has quit [Ping timeout: 245 seconds]
aristid has quit [Ping timeout: 260 seconds]
manveru has quit [Ping timeout: 272 seconds]
sorear has joined #nixos-dev
sorear has quit [Ping timeout: 260 seconds]
<drakonis> so, regarding metadata, how exactly should it be expressed, should it be extended to provide tagging
<infinisil> Hehe yeah also brought this up a couple times on IRC, I'd really like tagging to replace the weird pkgs/*/* structure we have
<drakonis> i'd like to just entirely do away with the current structure
<drakonis> are the tags going to be part of the library in this case?
<gchristensen> https://en.wikipedia.org/wiki/Taxonomy might have some interesting and related articles and publications
<drakonis> should assumptions be made when initially populating the tags? ie: packages built using python's build system naturally assumes it is a python package
<drakonis> i wanted to initially take debtags as a inspiration, as it is sufficiently fine grained for a initial attempt
<drakonis> typos not withstanding
<gchristensen> cool
<drakonis> then after implementing that the next step would be to allow querying the tags
<drakonis> maybe have module metadata to associate with the packages they depend on
<drakonis> this is specifically for modules that arent services
<drakonis> as these are normally self explanatory
<drakonis> tagging modules is also good, because it gets rid of the existing weird directory structure
<drakonis> although here its less urgent because it doesnt have a cascading pile of folders
srhb has joined #nixos-dev
<drakonis> i'll leave it here for now
<yorick> pkgs/arbitrary/arbitrary/stuff/default.nix
<drakonis> yes but that's not what i'm going for
justanotheruser has quit [Ping timeout: 268 seconds]
<drakonis> i'm going for a way to have additional package data, ie: find all packages that use ghc but aren't libraries, then get the overall metadata and its location in the tree
janneke_ is now known as janneke
_ris has joined #nixos-dev
justanotheruser has joined #nixos-dev
<pie_[bnc]> drakonis: i have
<pie_[bnc]> smol issue somewhere
<drakonis> do tell
<{^_^}> #57833 (by deliciouslytyped, 43 weeks ago, open): Additional meta-information conventions
<drakonis> huh, okay.
<drakonis> i didnt see that one
<pie_[bnc]> i think i had some small ideas but forgot to update the issue and then forgot too :/
<pie_[bnc]> basically thinking about all the random crap you want while using a package
<pie_[bnc]> id end up researching how librarians do stuff i guess
<pie_[bnc]> like, make sure informaiton links are discoverable in all directions
<drakonis> first things tho
<drakonis> if we do this, we're going to add a field for the cve name of the package
<pie_[bnc]> oh i should read what you actually said in here
<drakonis> i talked about a specific facet of what i want to do
justanotheruser has quit [Ping timeout: 260 seconds]
<drakonis> there's more pieces like cve names, a set of package licenses rather than a single one, show the available features
<pie_[bnc]> ( theres a lot of ideas possible in the "tags" space, a separate repo for tags, confirmed / tentative tags, lazy trees for searching (probably inefficient because of needing to process the whole set? / or precomputatiob) )
Synthetica has joined #nixos-dev
<pie_[bnc]> im not doing anything because i dont hav ethe capacity but ive wanted treating metadata as more first-class to be a thing for a while :P
<drakonis> i'd like to have metadata for defining package variants
<pie_[bnc]> i usually crosslink stuff i see on github but apparently i didnt in that issue, so either i didnt find anything or i forgot
<drakonis> for packages that are effectively "if you have this enabled, i'll build this extra binary"
<drakonis> rather than treating variations as entirely separate packages
<pie_[bnc]> sidenote; i figured the tag repo might want to be separate from nixpkgs since tags are probably going to have a lot of churn ((subtree or something?)
<drakonis> yes
<drakonis> it should be merged with nixpkgs every often though
<drakonis> squashed even
<drakonis> though that's why flakes exists now
<pie_[bnc]> i think ill only get around to flakes if it gets mainlined
<drakonis> tagging packages as "printer drivers" is also really good and ties up into other potential projects
<pie_[bnc]> drakonis: if there isnt one yet, i at least like to - make a tracking issue for this metadata stuff and crosslink things
<drakonis> its a big thing to be solved
<pie_[bnc]> stuff is hard to find in the issue tracker sometimes
<drakonis> the package plugin infrastructure would make use of that
<drakonis> get rid of foot shooting
<drakonis> this is a big undertaking
<drakonis> how hard is it to write a nlnet proposal for this?
<pie_[bnc]> i might be about to crash, irc is good but key stuff should either get mentioned in th eissue tracker or link irc logs - just so this stuff gets kept track of, i know i cant keep it in my head
<pie_[bnc]> drakonis: does nlnet accept random nixkgs refactorings? :P
<drakonis> they're accepting mobile-nixos and spectrumos, so yes
<gchristensen> it might take such a project under the search and discovery open call
<pie_[bnc]> drakonis: thats...differnt? :P
<drakonis> its changes definitely run deep though
<drakonis> they took ambrevar's project, which does things similarly to what i want
<pie_[bnc]> this stuff sounds like it could be handled incrementally at least
<pie_[bnc]> whats that
<samueldr> drakonis: quite reductive to say "they're accepting mobile nixos and spectrumos, so yes"
<drakonis> things to be learned from them.
<drakonis> its very reductive, yes.
<drakonis> but both end goals are bigger than mine
<samueldr> the grants are linked to specific greated goals
<samueldr> two of then are "PET: Privacy Enhancing Technologies" and "Discovery: Search and Discovery"
<pie_[bnc]> that guy is doing a lot of random guix stuff in that
<drakonis> i dont mean disrespect with that
<pie_[bnc]> also didnt know search and discovery was a thing
justanotheruser has joined #nixos-dev
<pie_[bnc]> tfw guix gets a gui installer before we do :P
<gchristensen> it would be cool if they got one!
<drakonis> they already have one
<gchristensen> nice
<drakonis> objective 4 is to further enhance the experience
<pie_[bnc]> oh
<gchristensen> that is awesome
<pie_[bnc]> hehe
<drakonis> rather, they have a TUI right now, up next is doing GUI
<qyliss> NLnet were strongly encouraging me to look at GUI installers
<qyliss> fwiw
<drakonis> nix definitely needs one
<gchristensen> maybe we can borrow what they've learned and built and apply it to NixOS
<qyliss> Person I spoke to offered to introduce me to the Calamares(?) developers
<pie_[bnc]> not a huge difference imo
<pie_[bnc]> i mean yeas, but no
<drakonis> there's a lot to learn from them
<pie_[bnc]> guix is the next big thing in a few years :P
<pie_[bnc]> feels like they have velocity
<drakonis> oh yes they do
<gchristensen> yeah, it'd be cool to have some of what they learn and build applied back to nixos
<drakonis> their implementation of home-manager is pretty interesting
<pie_[bnc]> i just keep thinking where do they get their velocity
<drakonis> guile
<pie_[bnc]> is that really it?
<drakonis> scheme, guile, knowing all the things that nix did wrong
<pie_[bnc]> not sure how much we can do much about that
<drakonis> guile is extremely hackable and easy to work with
<gchristensen> that is not it :P
<drakonis> also the people working on it are very dedicated
<gchristensen> not just it*
<qyliss> it's gotta be mailing lists :P
<gchristensen> they're good at marketing, branding, and positioning themselves in interesting spots, too
<drakonis> oh yes i was getting there
<gchristensen> qyliss: you're right, it is probably the mailing list
<drakonis> they're at fosdem right now
<pie_[bnc]> khm
<samueldr> I would be at fosdem, if it wasn't so far away
<qyliss> I get the sense from afar that Guix is probably a pretty small core of dedicated people
<drakonis> despite being gnu affiliated, it does the pragmatic route and just disallows nonfree code in the main repository
<samueldr> pragmatic? isn't that idealistic?
<qyliss> drakonis: as opposed to...?
<pie_[bnc]> im not gonna derail on that
<samueldr> >> pragmatic: dealing with things sensibly and realistically in a way that is based on practical rather than theoretical considerations.
<samueldr> (I'm not questioning, it's more about english as a second language user here)
<drakonis> i'm also a english as second language user here
<drakonis> they reached a middle ground
<drakonis> debian has main, contrib and nonfree, the average gnu distro only has main
<drakonis> guix has main and contrib
<samueldr> pragmatic I think would be allowing unfree things because it's useful for some cases, like firmwares, but still marking them, like nixpkgs does, right?
<drakonis> nonfree exists outside of official channels
<drakonis> its easy to find but not endorsed, not unlike fedora and rpm fusion
<pie_[bnc]> a tagging system makes me fantasize about stuff that goes on on sites like danbooru
<pie_[bnc]> all the tags
<drakonis> that's also me but no danbooru
<drakonis> i'd probably just axe most of the existing metadata and do some wild goals
<pie_[bnc]> low_quality_package needs_metadata split_stuff id
<pie_[bnc]> idk
<drakonis> also guix has declarative channels, its so sweet.
<drakonis> things that should be learned from, grafting
<drakonis> it allows making small changes to the store, so it allows avoiding mass rebuilds
<drakonis> great for security updates
<gchristensen> say more about what exactly it does and what you'd use it for?
<pie_[bnc]> might be better to just link the docs
<drakonis> yes
<drakonis> their docs are also vastly better
<drakonis> oh man this is so clean.
<pie_[bnc]> tbf there seems to be something about small teams of dedicated people
<drakonis> plus they have all of guile documented
<drakonis> it solved rolling release by providing a good mechanism to pin packages
<drakonis> so, grafts are useful for updating only pieces of packages
<pie_[bnc]> we have alllll the packages but they somehow seem to be doing the architectural stuff that matters
<drakonis> that's the thing
<drakonis> and they have a nix importer
<qyliss> pie_[bnc]: what are you thinking of, specifically?
<pie_[bnc]> not to be depressing or anything, thats just what i feal like from genreal vibes
<pie_[bnc]> my feelings are not a valid indicator of the state of nixos
<drakonis> so allllll the packages doesn't matter as much
<pie_[bnc]> whelp. :P
<gchristensen> we have a pretty incredible project, and I think everybody involved in Nixpkgs should be extremely proud
<pie_[bnc]> gchristensen: i agree though
<drakonis> guix wouldnt do all these amazing things without nix
<gchristensen> I look at Guix as doing More Good
<gchristensen> and Guix doing More Good means NixOS can do More Good too -- and none of this detracts or makes either Less Good
<pie_[bnc]> as drakonissaid, round two but trying to fix mistakes
<pie_[bnc]> but like, how can a big project like nix do a round two without forking i guess?
<pie_[bnc]> s/forking/rewrite/
<qyliss> I'm wondering whether we're just sitting here thinking that the grass is greener
<pie_[bnc]> qyliss: it crossed my mind
<pie_[bnc]> qyliss: but im not entirely convinced
<gchristensen> lots of green grass :P
<qyliss> pie_[bnc]: well, what is it about Guix that makes you think they're doing architecture better?
<gchristensen> pie_[bnc]: one step at a time
<simpson> qyliss: Exactly!
<drakonis> you should peek at their docs
<drakonis> they're quite good at that
<qyliss> I've looked at their docs
<qyliss> I look at their packages quite a lot to see how they do things sometimes if I'm struggling
<qyliss> Although that's never been all that useful
<drakonis> this whole conversation is a "how can we make nix better by learning from guix"
<pie_[bnc]> im just jealous x)
<drakonis> situation
<gchristensen> it is okay to feel a bit of jealousy. don't /just/ be jealous. effect change!
<qyliss> pie_[bnc]: of what?
<drakonis> that's what's happening at least
<drakonis> a little push comes to shove and now we have people getting into developing prototype for package features
<drakonis> a prototype
<qyliss> I keep asking you this not because I'm saying there's not anything to be jealous of, but because you've so far not identified anything
<drakonis> can i answer that?
orivej has quit [Ping timeout: 260 seconds]
<qyliss> you can, but I'm still interested in pie_[bnc]'s answer
<drakonis> cool.
<drakonis> i'll let him do it first
<drakonis> they
<pie_[bnc]> I dont have a proper answer. It just looks like theyre doing good stuff.
<simpson> qyliss: I have something I'm jelly of; I think that $(guix challenge) is interesting and useful.
<qyliss> Oh yeah it looks great
<drakonis> as for me, i'm jealous of how elegant their user interface is
<drakonis> it is easy to extend and comes with all kinds of tools that's available as separately maintained tooling
<qyliss> I'm jealous of their commitment to freedom
<drakonis> that too
<qyliss> But that's sadly not a technical thing we could even implement
<drakonis> i'm jealous of their ability to actually offer piece meal rebuilds instead of always doing mass rebuilds when touching core packages
<qyliss> That makes me nervous
<drakonis> its optional
<qyliss> oh that's cool
<samueldr> so you can somehow *insert term here* to replace a part of the system impurely?
<drakonis> their commitment to reproducibility and bootstrapping things
<qyliss> samueldr: graft
<drakonis> it doesnt do it impurely
<samueldr> neat
<qyliss> Yeah the bootstrap stuff is cool
<drakonis> they recalculate the hashes and all that
<qyliss> I really wish I understood how Nixpkgs' bootstrap worked
<samueldr> still impure in the sense that the dependents are not rebuilt, right?
<qyliss> yes
<samueldr> that would help working on those library issues in bit rooty things like Qt or GTK
<drakonis> g-expressions are also really nice
<drakonis> they can define the build steps to build any artifact
jared-w has joined #nixos-dev
<pie_[bnc]> they do use mes :D https://www.gnu.org/software/mes/
<drakonis> they created mes!
<qyliss> I'd love it if we could use Mes
<pie_[bnc]> drakonis: :O
<qyliss> I'm also jealous of their website
<drakonis> infact the author of mes is in here
<drakonis> the person that designed their website is pretty cool
<drakonis> they also did guile's website
<niksnut> at the reproducible builds summit in 2017 I hacked with Jan a bit on a mes-based stdenv bootstrap
<drakonis> jan is also doing gash and gash-coreutils
<drakonis> its impressive
<drakonis> that's nice!
ixxie has joined #nixos-dev
<drakonis> ah, something else i'm jealous about, guix actually has fully realized profiles
<drakonis> i can have profiles for a variety of packages and enable them as i desire
m1cr0man has joined #nixos-dev
<gchristensen> oh?
<drakonis> profiles in guix are declarative
<gchristensen> neat
<gchristensen> oh cool, they took our manifest.nix and just went ahead and made it user-editable
<drakonis> and you can have multiple profiles AND stack them
<niksnut> I initially wanted something like that for the 'nix profile' command, but then decided against it
<niksnut> since we already have 'nix-env --set' / 'nix build --profile'
<drakonis> yeah but isn't nix supposed to eventually replace the old tooling?
<drakonis> supercede
<niksnut> nix build --profile
<drakonis> ah i see.
<niksnut> which you can give a declarative spec of your profile
<drakonis> cool.
Synthetica has quit [Quit: Connection closed for inactivity]
<drakonis> speaking of which, whatever happened to nix app?
<drakonis> looked like you had fairly lofty goals for it
<drakonis> 'nix app' that is
<gchristensen> tbh I'd like to see something built inside nixpkgs as a PoC
<gchristensen> proving its viability and then merging in to nix
<drakonis> a experimental sandbox should be in order for this
<niksnut> PoCs never get merged
<gchristensen> fair
<pie_[bnc]> cough
<pie_[bnc]> ok im crashing i should just go to bed
<pie_[bnc]> gnight folks
<drakonis> gnight
<infinisil> gchristensen: How do you imagine "as a POC"?
<infinisil> I can't think of anything
<pie_[bnc]> im sleeping i swear, put a tool in for a release
<pie_[bnc]> if its good, deprecate the old stuff, repeat :p
<gchristensen> like a collection of library functions and executables
noonien has quit [Quit: Connection closed for inactivity]
ixxie has quit [Ping timeout: 258 seconds]
<drakonis> these few hours got rid of the elephant in the room, i was feeling a bit uncomfortable to be here without using nix
<gchristensen> I have long suspected that might be the case :)
rajivr___ has quit [Quit: Connection closed for inactivity]
<drakonis> i'm glad all is fine.
<simpson> By pig-and-chicken logic, I'm a little irritated at how the conversation started, but it seems to have been a worthwhile conversation.
<simpson> (Did not realize who was not a pig!)
ixxie has joined #nixos-dev
<drakonis> i have gained perspective from this
<drakonis> simpson: i didn't intend to demean anyone's work when i sparked the conversation
<drakonis> at any point
<simpson> drakonis: And I don't think that you did. Nonetheless, https://en.wikipedia.org/wiki/The_Chicken_and_the_Pig
<drakonis> ah, so that's what the expression means, cool.
<drakonis> its a interesting fable.
ixxie has quit [Ping timeout: 240 seconds]
<drakonis> i suppose i might be the chicken in here.
<simpson> Exactly. Not a bad thing, but a thing.
<drakonis> i can still go ahead with the project i have in mind, nothing prevents me from being in both camps
<infinisil> When I close somebody else's PR, can they reopen it?
joko has joined #nixos-dev
<drakonis> they can, yes.
<drakonis> you have to lock it
<infinisil> I see, well I *want* them to be able to do that :)
<drakonis> fair enough
<drakonis> https://github.com/nixpkgs can someone reclaim this org?
<drakonis> at least to prevent someone else from squatting on it?
<gchristensen> I *think* someone friendly holds it. checkin.
<drakonis> simpson: i've been relapsing into using nix over the years because nobody else in the traditional distros can provide what nix does, i might find some little issue with nix but i always come back.
<drakonis> the community is great.
<simpson> Sure, good times.
<drakonis> there's been some, hm, "personal growth", so i've been trying to be more committed to things
<drakonis> gchristensen: what gave it away?
<drakonis> was it all the uncomfortable shuffling around namedropping guix?
teozkr has joined #nixos-dev
ixxie has joined #nixos-dev
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos-dev
<infinisil> Here's a script for collecting all open PRs: https://paste.infinisil.com/Ov89v-JHrI
<infinisil> You can do `script > prs` then `firefox $(cat prs | shuf | head -1)` to open a random one then
joko has quit [Quit: WeeChat 2.6]
<infinisil> (don't forget to set your auth token)
<gchristensen> infinisil: "$(echo "$QUERY")" -> "$QUERY", no?
<infinisil> Oh heh yes, left-over from "$(cat query)"
<gchristensen> nice, recursion
zarel has quit [Quit: ZNC 1.7.4 - https://znc.in]
zarel has joined #nixos-dev
<infinisil> Hm what to do with open r-ryantm PRs that nobody cares about
<{^_^}> #69719 (by r-ryantm, 15 weeks ago, open): openwsman: 2.6.9 -> 2.6.11
<gchristensen> I'd ping deepfire again
<infinisil> Hm did that
<infinisil> I'm also going to be a bit more liberal with closing PR's without any activity for a while
<yorick> infinisil: hub pr list?!
* infinisil tries that
orivej has joined #nixos-dev
<qyliss> I think we should merge r-ryantm PRs that build, if nobody responds
* infinisil while hoping it doesn't work
<qyliss> that's the best way to get people to care
<infinisil> yorick: It's been running for like a minute now
<yorick> qyliss: chaotic neutral
<yorick> infinisil: yeah it has to fetch like 30 pages
<infinisil> Ohh yeah now it's done
<infinisil> yorick: Well the script i wrote is much faster anyways, only takes ~10 seconds :)
<infinisil> Compared to `hub pr list -f '%U'` which takes ~67 seconds here
<infinisil> My effort shall not be in vain!
<infinisil> I should integrate getting a random PR into nixbot
<infinisil> ,random-pr
<gchristensen> btw you might be able to set a per_page limit or something to 1,000
<infinisil> gchristensen: 100 is the max unfortunately
<gchristensen> huh
<gchristensen> right, v3 is the same
<infinisil> I was kind of hoping not having to do paging with graphql
__monty__ has quit [Quit: leaving]
rajivr___ has joined #nixos-dev
lassulus has quit [Ping timeout: 265 seconds]
lovesegfault has joined #nixos-dev
lassulus has joined #nixos-dev