2017-07-24

<NickHu> Is this kexec stuff a new/recent discovery or something?
<gchristensen> something about getopt long
<gchristensen> but perhaps you solved that problem while doing something else
<sphalerite[m]> Anyone know a good self-hostable service that's similar to owncloud but not a PHP nightmare to maintain? Maybe something with an existing NixOS module?
<et4te> honestly i wouldn't be averse to donating some BTC or something
<domenkozar> aristid: it was something in staging branch
<copumpkin> or something
<LnL> Sonarpulse: copumpkin: let me know if I can help build something
<LnL> joepie91: oh, well something like that should work :)
<copumpkin> there isn't something I can do inside the expression to make it work?
<copumpkin> is there something I can do from inside the expression?
<niksnut> pierron: were you also in utrecht for a while? or stratego user day or something
<copumpkin> ixxie: I think I missed something :)
<ixxie> copumpkin: I think its suppose to mean something like "how long are security updates rolled out for a stable version", or no?
<pie_> ixxie, you could just cite the authors of the first nixos paper or something
<nh2> it just claims that it can't fetch the package, even though it's there on the other machine. And it works when the command is run from a different machine. I'm wondering if it has something to do with the ssh-agent socket shown in the strace
<gchristensen> this feeds in to my organizational bug report: a private copy of nixpkgs on github or something, which should come with a private bug tracker for security issues, set up in Hydra as a private job. this way we can pre-build and pre-test security issues without having them leak publicly, but when they do leak they'll already be in the cache
<NickHu> It seems without passing in the empty set at the end the expression doesn't evaluate or something
<NickHu> Does something like this work? http://lpaste.net/357165
<NickHu> domenkozar: Something like callCabal2nixWithSubpath?
<LnL> GlennS: it's fine to amend changes, unless it's something nontrivial then it can be useful to see what changed
<mudri> manveru: so just gem 'something' from a repository should work fine?H
<manveru> mudri: gem 'something', path: '../something'
<mudri> Is bundlerEnv a recent addition to nixpkgs or something?
<Infinisil> rycee[m]: I may have overengineered it a bit, but considering htop crashes when something is even slightly wrong I'd consider it ok :)
<lukego> The crowd-pleaser would be to do something web-based, with d3.js etc, but personally I'd rather put my hand in a blender :)
<lukego> Could be so :). Here is my rationalization though. The tools should be much more general than only for Snabb - something you can extend to any application, like Emacs. So Lua will not be a prereq for other people. Then that using Lua for GUI seems very low level e.g. I will be taking a Tk/GTK/Qt/etc binding and doing everything from scratch.
<dhess> lukego: Why Pharo rather than something Lua-based, since you presume the user can already write Lua for the Snabb side?
<dhess> maybe not. Anyway, nixos/modules/hardware/cpu/intel-microcode.nix does something similar to what I want.
<dhess> or something like that
<bachp> I'm thinking about implementing something similar to coreos, where each machine get a lock from e.g. etcd or cobsul and then reboots. This way only one machine at the time would reboot.
<LnL> bachp: maybe, diff the kernel version in /run/current-system with uname or something
<et4te> i've been trying to get Hydra to build something, everytime it tries to build gcc.. Is that normal? I've tried setting NIX_PATH and binaryCaches but nothing seems to change it
<rycee[m]> Infinisil: About the htop fields. It's not possible to do exactly what you want (as far as I know). But you could set the type to something like `listOf (enum (attrNames htopFields))` and that should check the values, including giving nice error messages...
<johnramsden> clever: I was getting a stream of 'no version information available' messages. I don't know if I need to set something up for my graphics card, but ethminer https://github.com/ethereum-mining/ethminer didn't mention anything.
<pie__> woo finally got something

2017-07-23

<Zer000> so I just got a bluetooth dongle with the hopes that I can get my phone to play music through my computer, so it thinks the computer is a headset or something. But I quickly realized I have no idea what I'm doing, and besides hardware.bluetooth.enable I'm out of my element.
<manveru> i don't think there's a better way than RUBYLIB for now if you want something working fast
<manveru> and that's just not something nix does well
<Infinisil> clever: I'm writing something down, need to know: have you had the error "timed out waiting for the VM to connect" on your local machine once? I can't remember
<schoppenhauer> manveru: on the other hand … isn't it possible to define a gem in a tarball or something?
<deltasquared> clever: the compiler puts it there, it's something the C standard says
<deltasquared> your compiler must be doing some kind of escape analysis or something...
<Infinisil> clever: So you arrive at something which works pretty much like IPFS
<clever> even something as simple as write this string to a file, depends on which platform you run it on
<Infinisil> clever: Isn't something like this done with all mallocs?
<bhipple[m]> Is there a way to do something like:
<mpickering> I just ran nix-shell and it died whilst downloading something
<mpickering> meh I can't even get bazel to print out the commands it is running, time to work on something else
<clever> so the perl code just starts it on-demand, the first time something is done to it
<Infinisil> clever: Are you on RAM/Disk capacity limit or something?
<srhb> If the database doesn't run full again or something.
* hodapp throws something at Alan Kay
<hodapp> dash: so they're clearly just doing it because it's popular. or something.
<hodapp> I'd settle for something like C++... just... without most of what makes it C++
<deltasquared> I'd probably settle for something like C++ with an extra semantic pass on top to add rust's lifetime checking rules in addition
<deltasquared> this is starting to sound like something I wouldn't want anywhere near any of my systems
<deltasquared> hodapp: maybe it's compiling something to be used in the build process itself, like the kernel does
<deltasquared> clever: would it not be possible to simply include a patch as part of the "standard" openCV packaging for nix here? a patch to disable the hashing (and let something else do it)
<Infinisil> deltasquared: I'm not too experienced with openGL, but aren't thread-locals just something GPU's have?
<hodapp> it's not clear to me whether this just never worked right in OpenCV 3.2 (I never used xfeatures2d until now and need a feature only in 3.3.0-rc), or whether this was due to me getting some repo revs wrong or something
<hodapp> so is the fact that CMake is even trying to download an indication that something went wrong there?
<joepie91> ToxicFrog: definitely not all. something something name bucketing
<zarel> deltasquared, clever: I read that it runs a dropbear instance and everything about the host keys but I don't know how the stage 1 behaves when something like cryptsetup-askpass is waiting for an input somewhere that is not the SSH shell I open
<gchristensen> mpickering: you're trying to do this inside a nix-shell or something?
<joepie91> deltasquared: well, 'frameworks' are something you'll want to just avoid in general in perpetuity
<LnL> rvolosatovs: I do something like this https://gist.github.com/LnL7/e645b9075933417e7fd8f93207787581
<hodapp> hrmph. if I run this build (opencv updated to 3.3.0-rc) normally, something in it causes xfeatures2d to be disabled. If I step through it in nix-shell (so far, unpackPhase and eval $postUnpack and then cmakeConfigurePhase), xfeatures2d is properly enabled
<Infinisil> hodapp: Hmm, actually the manpage says this argument doesn't apply to nix-shell, something like nix-shell -E 'with import path/to/nixpkgs {}; callPackage path/to/foo { enableBar = true; }' should probably work
<Infinisil> hodapp: Does something like nix-shell -A foo --arg enableBar true work?
<hodapp> I was running "nix-shell -A foo", but now it turns out that I need something more like (foo.override { enableBar = true; }) and am a bit lost as to how
<kiloreux> Sorry if I am interrupting something

2017-07-22

<Infinisil> mpickering: It might be something with having cmake as a build-time dependency, because you wouldn't have that without a cmake project
<Fare> Infinisil, I believe I am. Should I? Should I not? Need I add something to its configuration?
<clever> sed may just silently ignore things if upstream changes something, and then you get odd errors
<cherrybl0ss0m_> I have something like this for configuring the Hydra server https://www.irccloud.com/pastebin/4tek8E5a/
<clever> netflix uses widevine on chrome and something else on firefox
<joepie91> then I get distracted because ~shiny~ and go do something else, and return to the original thing later
<Infinisil> joepie91: I feel like if I'd start learning about Rust or something else that interests me a lot, I'd get obsessed for a while and I'd waste all my time, while I really should be studying for exams
<mpickering> If I am packing something which says "run a script" and it has a shebang at the top with an absolute path, what is the right fix for this?
<mudri> I usually end up with a default.nix, a shell.nix, and a pkg.nix or something.
<mudri> Also, something that has been catching me out: is it true that the default.nix for something I'm developing shouldn't take arguments without default values?
<mudri> clever: aah, yes, I do something vaguely like that for packages, so I can see it working.
<Infinisil> sorcus: You might have done something like building a package file in nixpkgs directly with nix-build <file>, which doesn't work because these files are all functions
<sheenobu> I made a project that is sort of like a DSL for dotfile management. I want to build something similar but for complex projects and microservices
<gchristensen> I wish I had a project to learn haskell on :) books never do it for me unless I already have something in mind.
<silver_hook> clever: I was thinking something similar, yes. And if I eventually get rid of NixOS, I can probably get Mageia to override Grub again.
<mpickering> I want to 1) Know how it works so I can write something similar for another tool 2) Fix the documentation if it is out of date
<butchery> can anyone recommend a little http server with live reloading thats already in nixpkgs? I normally use something from npm but I'd rather not
<ben> ok concretely i want to eval something that has a secret in it that doesnt need to go into the store
<srhb> alunduil: Where you've done something like let YourCouchDB = nixpkgs.pkgs.haskell.packages.${compiler}.CouchDB.overrideAttrs (oldAttrs: { doCheck = false })
<ottidmes> gchristensen: That was in response to cost, but the reason I said it was to stop you from moving the conversation to something irrelevant to me
<alunduil> srhb, sorry, I was thinking something crazy.
<srhb> alunduil: something like let myCouchDb = couchdb.override { doCheck = false; }
<bennofs> I just figured out something really cool how you can debug wrapped programs with gdb: set exec-wrapper
<srhb> Here's something interesting. Running the old spotify placed the widewine thing in ~/.cache/spotify/... and now the new version does not complain about it missing, but still segfaults
<deltasquared> could be worse... at least there's no ARM TrustZone code or something
<joepie91> deltasquared: in this case, I mean it as "people believing they have above-average knowledgee on something they are clueless about"
<srhb> Unfortunately, something has changed, so the build needs revising, and I'm not sure how to identify what
<tilpner> Fare - If you do want nix-env, you would put something like import path/to/your/local/nixpkgs {} or import <nixpkgs> {} (with NIX_PATH set accordingly) into ~/.nix-defexpr/nixpkgs.nix and delete other entries in that directory as needed

2017-07-21

<ottidmes> yegortimoshenko: so in the installPackage of your new package do something like "cp ${pkgs.target-package}/bin/target-bin $out/bin"
<celph_> ottidmes: the problem with that is some users don't know what they did to cause an issue. there was something I read at some point about how it's a good idea to have a general error catcher thing that catches a crash and logs the program state at the time of the crash. I forget ideas about how to implement such a thing.
<celph> joepie91: when I was searching around for resources I found some documentation on how/why the wiki was killed and people admiring the Arch wiki a lot (which is something I also really like). Maybe as I figure things out I'll try to help with the documentation
<ToxicFrog> I assume that this is something going wrong with the wacky chroot tricks that NixOS does to get Steam to run at all, but I definitely do not have the time or the inclination to debug that
<celph> Something I was having trouble with last night: In order to build a package for Nix using dependencies in nixpkgs, do I need to clone nixpkgs?
<dualinity> you always have to fix something on arch is how it feels
<joepie91> so if you have eg. $out/share/themes containing something, that something will be symlinked from /run/current-system/sw/share/themes automatically
<ottidmes> joepie91: That could still be caused by a wrong sitting position, chair, keyboard, keyboard placing, not having enough breaks, I would really try to experiment, it is basically your body telling you strain certain parts too much, so something has to change
<joepie91> (and I'd recommend that anybody considering TypeScript would instead spend some time to get acquainted with JS *properly*, or if different tradeoffs are required, to consider something like Flow instead)
<ottidmes> deltasquared: My desktop right now is zooming, bzzzzzzzzz... half of the time due to a half working GPU fan... But not something playing some Metal cannot fix ;)
<ottidmes> Filystyn: Yeah, I assume it is something external, and udisk is often (always?) used for handling that, and polkit to determine privilages (probably your issue)
<ottidmes> Filystyn: Do you have something like udisk / gvfs on your system?
<simpson> ottidmes: When I said "theory of computation", I was hoping for something like lambda calculus, pi calculus, CoC, JVM-style, CVM-style, Horn clauses, Post correspondence, Wang tiles, etc.
<mbrock> that's something I love about Nix too, how it's based on this really great purely functional core, but then it's also a huge blob of shell scripts
<joepie91> ottidmes: yep, familiar with the points mentioned there. it's not a big issue for me, because I have a tendency to just build the necessary tooling myself if something is missing :P
<ottidmes> I find it somewhat funny that I recently started to use C for something else then my Pebble watchface, while if you start out, Rust would be way better, but I have yet to write something considerable in Rust
<deltasquared> well you could always change something that won't effect behaviour, like a config option or something
<srhb> cwre: Patches not long, rebuilding the kernel might take 10 minutes or something
<ottidmes> In technology age does not really mean a lot, only that you are likely to have more knowledge to fallback on, but other that, things are moving so fast that most things are not that old, so someone just starting good be just as much as an expert on something as someone who has a lot more working experience in general (absolute in years)
<srhb> pi3r: You'll want to do something like haskell = super.haskell // { packages = super.haskell.packages // { ghc821 = super.haskell.packages.ghc821.override { ...
<ottidmes> dshin: I generally tend to just Google for it with: nixpkgs pkg-name, which then gives me the service (if there is any) and the package, one benefit is that you have to be less exact, but the downside is that it might miss something, I have not run into this problem, with Github, yes, plenty of time it missed something, with Google search, not that I can remember
<NickHu> cwre you might need to set the acpi_backlight kernel parameter to something else
<srhb> I don't know if that's the driver going bonkers or something else.
<srhb> cwre: Right, so something like cat /sys/class/backlight/*/brightness
<deltasquared> must be logind doing it's pause thing or something...
<cwre> Konsole has to be doing something stupid because it's being symlinked to oblivions
<NickHu> I'm almost certain it's something to do with your graphics drivers
<NickHu> yeah something like that
<cinimod> But also I seem to have to ask questions to which the answer is something like
<fresheyeball> is there a nix way to do something like
<dalaing> joehh: normally you run nix-build from outside of the nix-shell (unless I've missed something about your use case)
<catern> gchristensen: sure sure but you could still conceivably merge them together with a bind mount or something
<dtzWill> "wait are you sure you don't want to like... process some post-install hooks or something for 10 minutes?"
<ottidmes> cwre: If you just want things to work out of the box, and you do not care about them, as long as they work, you should be sticking to something like MacOS, Windows, or Ubuntu or a similar distribution
<ottidmes> cwre: Because even if you fix the cursor, something similarly might popup
<ottidmes> cwre: BTW I think you still would need to set nixpkgs.config, but you could probably use something like: nixpkgs.config = import /etc/nix/nixpkgs-config.nix;
<ottidmes> cwre: Exactly, what Infinisil said, I have something like packageOverrides = pkgs: { premium-xcursor-theme = pkgs.callPackage ./premium-xcursor-theme { }; }
<Infinisil> cwre: Typically you do something like packageOverrides = pkgs: { hello = pkgs.hello.override... ; }
<ottidmes> cwre: e.g in your systemPackages something like this: (pkgs.callPackage /path/to/premium-xcursor-theme.nix { }) (I would add it to packageOverrides first in your nixpkgs config and then just use pkgs.whatever-name-you-gave-it)
<ottidmes> cwre: In the worst case, if something really is messed up, you could always use a basic config (just enough to get your system to be useful) and then install your actual config again, this has helped me in the past

2017-07-20

<celph> or something
<celph> I think I'm misunderstanding something about how program execution works
<Infinisil> Ahh, the endless battle of how annoyed you are of something vs the effort necessary to fix it
<sphalerite[m]> copumpkin: I'd really like that too, I think we should add a matchGlob builtin (or something, I'm sure there's a better name for it) which wraps the relevant libc function and makes implementing gitignore trivial
<clever> sphalerite[m]: heh, sounds like something i would do with not-os
<boomshroom> clever: Or I should be able to create it. ( I am the one and only user of my local machine. If there's something I can't do locally, I'd be very upset. )
<boomshroom> I've spent some time trying to get Nix installed on my university account. Is there some way that I could make the process easier using something like NixOps?
<deltasquared> lvm, too. couldn't really expect the kernel to understand something that complex, as it were
<deltasquared> the entire world shalt bow to the functional package manager something something mumble...
<boomshroom> deltasquared: Side-effect-less systems only cause issues when something assumes it's a side-effect-full system.
<queiw> it's marked as broken and it's an older version. I just upgraded NixOS but apparently broke something
<boomshroom> deltasquared: Nix has a makeWrapper or something similar to set environment variables. Can it do the same for CWD?
<deltasquared> srhb: I've just spotted something rather inefficient in how the nix libc searches the paths
<srhb> deltasquared: (or something like that)
<srhb> deltasquared: Did you mean something like objdump?
<deltasquared> srhb: hmm, so which part of the manual should I stare at... I suspect it has something to do with nix expressions
<deltasquared> something to do with the dep tree?
<srhb> So either you have Arch and a virtual machine with NixOS running on arch, or you have Nix on Arch, (or something is going to go very wrong in a little while)
<cwre> Do I need to do something special with Nix?
<LnL> yeah there must be something special with the response that causes curl to abort
<LnL> FRidh: unless I did something wrong nix-build https://github.com/NixOS/nixpkgs/archive/staging.tar.gz -A python.pkgs.pyparsing
<disasm> gchristensen: yeah, something wrong with either how gmail is displaying the message or how mailvelope is handling it in javascript, annoying...
<disasm> gchristensen: it's either something messed up with gmail or mailvelope, message is fine, I downloaded source and it verified just fine, I'll continue debugging...

2017-07-19

<LnL> yeah, that's what I was talking about not something like docker
<xplat|work> (thin containers are basically containers that only contain the nix environment and mount (sometimes just parts of) the host's store, something you can really only do properly with nix)
<clever> there was something worse in the pam 2 factor auth
<simpson> JameySharp[m]: I know the feeling WRT tedium of recompiling. In general we want to avoid even theoretically-transparent caching since it always seems to be impure in practice. ISTR that clever had something for this.
<srhb> simukis: Probably something similar to hardware.pulseaudio.systemWide then :)
<krav_> when upgrading to nixos-unsable, should I change system.stateVersion to something else?
<joko> disasm: hopefully, it was something simpler for pt, an issue from upstream. If I compile this rev https://github.com/monochromegane/the_platinum_searcher/commit/c7d8eec66dca50773e6b4ee7dfdad2174860b9b1, it works
<copumpkin> pierron, gchristensen: so here's an example of the thing I'm talking about. Exhibitor (a zookeeper management service) has what's effectively an (implicit) sum type for its config. If this were Haskell, it would be something like `data Mode = S3 {- a bunch of s3-specific options -} | Zookeeper {- a bunch of ZK-specific options -} | File {- a bunch of filesystem options -} | None`. This results in a giant list of assertions that are a bit of
<ben> it's like having a shadow cabinet or something
<clever> yeah, it mounted something over /etc/passwd
<LnL> dtzWill: I think the build couldn't report back to the daemon and since it restarted the new instance didn't know about the build something
<dtzWill> happens for me all the time, think it's some mix between large jobsets, jobsets that build on one another, and jobs that have dependencies that are explicit jobs in other jobsets. I got lost in that code before sorting anything, could all be due to something else which is why I thought I'd ask
<simpson> postFixup = ''mv $out/bin/internal $out/bin/ct''; # something like this maybe
<copumpkin> nix-prefetch-url --unpack is what you need for fetchFromGitHub or something

2017-07-18

<boomshroom> I was thinking of doing something similar with Minecraft.
<clever> i think it might be in "-k kernel" or something like that
<clever> spinus: something came up, cant type much now
<spinus> bennofs, yeah I thought it's that, but looked strange in this sentence for me, I understood something like "there is way X, and brb" (which is way Y) :)
<copumpkin> or something like that
<schoppenhauer> tilpner: because intero somehow needs this … or something
<tilpner> It should output something else before and after the attempt
<gchristensen> I mean more sketchy :D like replacing /nix/store/hash-package with something else
<bennofs> copumpkin: oh.. so something like: foo = { variant = A; bar = ...; } and foo = { variant = B; qux = ... } ?
<copumpkin> so I've been noodling about something
<S0rin> I should have done something wrong because I thought the override was only temporary
<cinimod> Seems to be something to do with cautious-file whatever that is
<cinimod> Is this something in the default.nix file?
<srhb> NickSeagull: Perhaps I missed something.
<Zer000> hang on there must be something i'm doing wrong because a minimal example works

2017-07-17

<jtojnar> yegortimoshenko: seems to have something to do with update walker https://github.com/NixOS/nixpkgs/commit/07d9de713a1789b90f4153ca82afafc3a8ea1b87#commitcomment-17379503
<clever> so it was only working because something you installed was leaking in
<tnks> the docs say something about build-time versus run-time.
<clever> alex-v: if something was ran as root in the past, the ownership may be mixed
<bennofs> but the PR is quite old + has no comments. so we should do something about it
<gchristensen> or something else?
<sphalerite[m]> Or generally if I'm using a USB stick and not running software off it or something
<clever> in theory, if the guest writes the correct headers to the file, qemu can start to interpret it as a copy-on-write image of something else, and read an attacker controlled path
<clever> bennofs: something else i have been thinking of, is taking the hash in the <hash>.narinfo, and undoing the base32
<dtzWill> clever: oooo I was thinking about something similar myself
<dtzWill> bennofs: I believe those errors should be retried with exp backoff until you get a different response code or retry count is hit; hopefully this is better documented in AWS docs or something
<bennofs> etu: to install python with nix-env, I use something like: nix-env -iE '_: with (import <nixpkgs> {}); python3.withPackages (ps: with ps; [ requests ipython lxml beautifulsoup4 nclib dns ])'
<hodapp> friend worked around some of these years ago, and reported that with the support contracts they were under, they'd come in on a Monday and see that the machine had something fail at 3 AM on Saturday and someone from IBM already showed up to replace the part while they were out because the machine had just phoned home
<S0rin> Like that if I want to test something I can change it in the local copy of nixpkgs and use it directly in the config file
<gchristensen> but last I checked we needed them for something
<S0rin> It seems that the override is not working on the 32 bit part so maybe I need to change something in the configration
<ertes-w> i notice that python packages are inconsistently named, usually ridiculously outdated, etc., which basically just hints at the lack of something like cabal2nix/haskellPackages for python… is there any solution in sight, or are python packages just too non-uniform?
<makefu> someone here on the channel wrote about the experience of waiting 24hours or something for optimize to finish
<makefu> apparently it does this sometimes but other times the issue cannot be solved with this. thought i've read something like this on the ml
<Infinisil> Well I just need to know if it outputs something for you
<Infinisil> Yeah I got something similar
<adisbladis[m]> scott: If you go to "System settings -> Look And Feel" do you see something along the lines "The shared library was not found"
<adisbladis[m]> Anyone else running KDE who can confirm something for me?
<ottidmes> Infinisil: Agreed, I am not fully set on that yet, I wanted it to be something short 2-3 chars, and it contains more then just my config files, so it has to be generic enough

2017-07-16

<Infinisil> boomshroom: Indeed, and if something's missing you can mostly do it yourself without much effort
<gchristensen> schoppenhauer: there might be something in the issue tracker
<Infinisil> Wizek__: Maybe something like `nix-shell -p ghc --run 'runhaskell ./test.hs'` could work for you
<dtzWill> boomshroom: do the RISCV guys have a fork/branch somewhere with the patches applied? Seems like something they probably have...
<boomshroom> dtzWill: LLVM is useful, but it's so complex that it's difficult to add a new backend. Does anyone know if there exists a simple retargetable compiler? I don't care about optimizations, because those add complexity and right know I just want to get something running.
<boomshroom> I guess it would basically just be like the LiveCD/DVD/USB, but within an already installed *nix system. I don't care much about it, just something that idly crossed my mind.
<Infinisil> I'm using latest master, maybe that has something to do with it?
<sciguy> the scientific community is longer for something that makes "re-doing" the same experiments easy, even for those who arent coders
<bennofs> S0rin: theoretically, it should also be possible to do something like: let pkgsStable = import <nixpkgs> {}; pkgsUnstable = import <nixpkgs-unstable> {}; in ... if you setup your NIX_PATH correctly
<cherrybl0ss0m_> It feels like there is something wrong happening well before the chroot part of the image building process?
<Filystyn> I won't drop with this tor problem it's definatly something I miss in configuration.nix
<bennofs> Eisfreak7: if you want to do it declaratively in your configuration.nix, i think you need to use something like this: https://github.com/bennofs/etc-nixos/blob/master/conf/default.nix#L86-L96
<srhb> Something's weird...
<ij> gchristensen, Should I have done something more other than adding it to contents of the buildImage?
<joepie91> it's something I've been putting off for a while and I should probably just get to it :P
<joehh> looks like i'll have to right something up if/when I succeed :)
<gchristensen> you asked for a specific example of where I'm having trouble. can I turn that around and ask for a specific example of something I've written, and how it might be better?
<ixxie> is there something NixOS specific going on possibly?
<boomshroom> Something about how primary monitors are assumed to be to the left, while mine is to the right.
<celph> boomshroom: I don't feel to strongly about my DM so I may also just switch to something less barebones
<boomshroom> In that case it should be as simple as `services.xserver.displayManager.slim.enable = true; services.xserver.windowManager.i3.enable = true;` and then reboot. Assuming nixpkgs isn't doing something stupid, then SLiM should be able to find i3 just fine and your i3/config should be able to set the xrandr settings.
<boomshroom> celph: I'm guessing you're trying to access something called "singleton" on line 112 of your configuration.nix?
<boomshroom> error: attempt to call something which is not a function but a set, at /nix/store/5jc75g6f5zw7b1x9cd83ng57mh428hcs-nixos-17.09pre110213.01c3847b9c/nixos/pkgs/stdenv/generic/default.nix:348:52
<cherrybl0ss0m_> boomshroom: this might be the old way but you can do something like this `(terraform.overrideAttrs (oldAttrs: oldAttrs // { doCheck = false; }))`. I have that particular thing as an entry in the `buildIntputs` in a project shell.nix somewhere.
<celph_> Here's something pretty weird: error: syntax error, unexpected $undefined, expecting '}', at /etc/nixos/configuration.nix:117:34
<ottidmes> Infinisil: That was something I kept running into with Agda, when your proofs got complicated enough, it became way too slow
<ottidmes> Infinisil: I think I proofed that one, or something similar, as well a few years back, but then in Coq. Can't say I am fan
<Infinisil> ottidmes: The naturals can have pretty bad runtime when it's not optimized.. Try evaluating something like 7! using Nats in the interactive terminal
<Infinisi1> radvendii: I believe Haskell has often that problem: because of lazyness, you can never be sure when something actually gets executed, because it may only be needed much later, really hard to predict in a bigger program. And it may even be the case that a big calculation gets done twice, because it just calculates it lazily, without allocating memory for it
<ottidmes> Haha, that is insane indeed, something about critical mass

2017-07-15

<spinus> any upgrade? maybe new driver or something?
<Infinisil> spinus: Something like 05ac:9227?
<Infinisil> Maybe it has something to do with "hotplug", of which I have no clue
<spinus> maybe that way you can find "something" which will lead you to the source
<simpson> catern: I am trying to introduce you to something new, then; do you think that that's air that you're breathing? Why do you think that your *observations* reflect the actual *nature* of reality?
<simpson> catern: So, what I'm saying is: What if numbers as they really exist aren't something we perceive, but instead we only perceive their shadows? What if the ring of integers, and their ideals the primes, are actually just quirks of the way that our minds work? There's no compelling reason for this not to be the case other than that it's always surprising to be reminded that humans don't directly observe reality.
<schoppenhauer> is there something like "rubyWithPackages"?
<Infinisil> I have ubuntu installed, should i test something?
<wizwizwizwiz> or just something sufficient
<srhb> wizwizwizwiz: I'm just grasping though, but that's something easy to try out
<gchristensen> schoppenhauer: so now the problem is the builder can't access "/home/christoph/tmp/nixops/chat" because it doesn't have arbitrary access to your filesystem, you're passing a string, you should pass it something of the datatype path.
<Sonarpulse> or you mean something else?
<johnw> maybe something else is wrong...
<johnw> for a regular file with fetchurl, I would untar ${mySrc} or something
<schoppenhauer> hi. is it possible to define a (local) package with a git repo that is not public? like, specifying an SSH-Key or something?
<clever> until somebody upstream changes something
<jophish> I have this closed source binary (modelsim-ae) and it (obviously) doesn't have an RPATH which reads the nix store, so I've wrapped it with something setting LD_LIBRARY_PATH (I'm not patching the RPATH because it takes hours to install, and I want a quicker iteration time). If I put nix's glibc in the library path it just prints garbage and dies
<simukis> clever: another related question I had recently is why stage1 a shell spaghetti, instead of, you know, just using systemd target or something
<srhb> Maybe I don't have python properly in.. scope or something
<socksy> but i think i must be doing something wrong because it doesn't seem to give me a real FHS env, and besides, it's sim linking to /nix/store which is mounted read only
<ToxicFrog> I have this horrible feeling that I'm going to end up with two versions of KDE installed or something though.
<LnL> for my build machine I have a .bashrc symlink to something IIRC
<gchristensen> instantepiphany: well I was hoping we make that silly mistake, vs. something harder to fix :P
<MoreTea> It's more a random thought, than something grand plan
<pie_> ok it *is* a possibility that i did something dumb with my global state at some point wrt installing bash for example
<pie_> i mean this is something people would expect to work by default
<pie_> and now it doesnt work (i probably fixed it with something funky)
<TimePath> I remember oflag=seek_bytes fixed something, that may have been my build script though
<ixxie> sphalerite[m]: something to think about.... thanks :)
<TimePath> you can bump bs to something higher if it works
<clever> and if you add something to systemPackages that has a $out/etc/xdg/autostart, the same thing will work
<LnL> nixy: you could do something like pkgs."foo" to allow arbitrary attribute names
<Infinisil> Something like that yes
<Infinisil> Some security protocol was using something like Enc(Hash(Enc(...), ..., ...), years later it was discovered that this extra nesting doesn't do anything and could have been left out. They just used it to be on the safe side, no proof
<Infinisil> simpson: I'd just start hacking something together first
<TimePath> tried cross compiling something that depended on perl to arm once, didn't go well
<gchristensen> I made something similar to this for converting installer tests in to installation docs
<Infinisil> Well before I even wanna start discussing docstrings again, I'll first try to build something that would do that and then come back and show it to get opinions

2017-07-14

<LnL> either way something like that should work
<sphalerite[m]> Main reasons I got it are A, so I have something that will run libreboot just because that's cool, and B, for taking to the US. I don't want to take my normal driver laptop with me
<ToxicFrog> Shouldn't be too hard to have an "include nonfree y/n" toggle and put a little ⚠ next to them or something.
<joepie91> ToxicFrog: if you hate CSS, that means you've learned something about styling :)
<pbogdan> how is trust of a binary cache established? is it a case of "I trust this entity" or is there something in nix that would enforce that contents of said cache can't be tampered with?
<fetchurl> Why is it that Nix when a different hash has been found than expected and you correct this, that it will download the file again, while it should still have the previously downloaded file available? It should be able to check for the already downloaded in the store, right? Am I misunderstanding something?
<dash> gchristensen: exarkun did something like that
<schoppenhauer> srhb: seems like it needs root access via ssh or something
<Eisfreak7> Hm yes I was talking about runtime. I was thinking something like firejail integration: I can enable firejail globally / on a per application basis an nix generates "firejail wrappers" that automatically call the binaries with firejail and an appropriate config
<copumpkin> something like that :P
<copumpkin> niksnut: I'm starting to think a more useful primitive is something like "deepScopedImport" might be easier to implement (with caching) and cover 90% of use cases (and all of mine)
<joepie91> something more recent than what's on the wiki
<joepie91> this is not something you can infer from a single post, or a single person's response, or any one single bit of data. but the moment somebody starts riling other people up with little to no constructive value, that's the first warning sign that there's a situation you need to deal with. that doesn't mean you need to immediately resort to bans, but you can't just wave it away as "just ignore it" either
<joepie91> nh2: so yes, the first attempt or even attempts (plural) should be to communicate with a problematic user and try to get them to change their behaviour towards something more constructive. but if somebody is actively hostile towards that, and is producing nothing but noise, then at some point the chances run out.
<joepie91> nh2: and getting discouraged by hostile behaviour is a long-term process; it's not something you can pin down on any one particular comment.
<__Sander__> this guy is not secretely terry a. davis or something? :D