asuryawanshi has quit [(Ping timeout: 260 seconds)]
<hyper_ch> you don't have a scanned in pdf with different languages that ocr can be tested on? preferreably non-roman letters?
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
NightTrain has joined #nixos
<hyper_ch> and tomorrow I need to test if I can properly run my win8.1 vm with kvm/qemu instead of vbox
<samueldr> hyper_ch: throwing this in, I've been using libvirtd with virt-manager successfully lately with nixos, and intend to replace virtualbox for all my uses with that
<infinisil> hyper_ch: go to wikipedia, select an article (preferably scientific), switch language to the one you want to test, search for that articles title with pdf on google
<hyper_ch> well, I already converted the vbox vdi to qcow2
<samueldr> (it manages the VMs, can be remotely accessed, uses qemu-kvm)
<hyper_ch> infinisil: well, in order to do ocr you need image pdfs ;)
<hyper_ch> most stuff online is text pdf
<infinisil> hyper_ch: oh, well then just make an image of those pdfs
<hyper_ch> samueldr: I use kvm/qemu with virt-d to manage the vms on the server
<hyper_ch> infinisil: that requires some advanced hacker skillz ;)
<hyper_ch> (actually, I think OCRmyPDF can do that as well)
<infinisil> hyper_ch: screenshot/camera
<hyper_ch> as said, advanced hacker skillz
<infinisil> ':)
<hyper_ch> "If --force-ocr is issued, then all pages will be rasterized to images, discarding any hidden OCR text, and rasterizing any printable text."
<hyper_ch> sounds like text pdfs will be converted to images and then it will be ocred again
<infinisil> yeah probably
erictapen has quit [(Ping timeout: 248 seconds)]
phreedom has quit [(Quit: No Ping reply in 180 seconds.)]
ts__ has quit [(Quit: Page closed)]
phreedom has joined #nixos
ryantm has joined #nixos
* ryantm Is there an automated list of all the nixpkgs packages that need updating?
<samueldr> yes!
* samueldr forages through notes
<hyper_ch> off to bed
<hyper_ch> cya
<infinisil> hyper_ch: see ya!
<samueldr> it's not from nixos, but assuming it's up to date, still useful
* ryantm samueldr: Thanks!
infinisil_erc has joined #nixos
ilyaigpetrov has joined #nixos
<gchristensen> ryantm: are you manually prefixng your messages with /me ?
<infinisil> haha didn't even notice that
* dhess` laughs
infinisil_erc has quit [(Quit: Not feeling comfortable in emacs :P)]
<samueldr> who could?
* samueldr ducks
<infinisil> samueldr: are you referring to my quite message?
<samueldr> yes
<infinisil> heh, well I'm getting more and more into emacs, it's getting better
<samueldr> yeah, I would have said the same for anything, I'm not really into flame wars
<samueldr> your phrasing made it too tempting
deepfire has quit [(Remote host closed the connection)]
<infinisil> well I was a vim user before (and still use it), and I'm using evil-mode in emacs
<infinisil> Whatever floats ones boat ;)
<gchristensen> org-mode got me to switch to emacs
<gchristensen> oh, also, #emacs' self-deprecation
<infinisil> org-mode seems nice, but I can't quite use it that well just yet
<infinisil> gchristensen: What's that?
<samueldr> btw, gchristensen, did you get the link I sent you?
agjacome has quit [(Ping timeout: 248 seconds)]
<gchristensen> if you ask if emacs is better or worse than X, they'll do ,salespitch and fsbot will reply "We aren't gonna lie. Emacs sucks. Some of us tolerate it, but we can't tell you if YOU'LL be able to. Try it and make up your own mind."
<gchristensen> in this way, they cut off any potential superiority thing or flame war at its knees
<gchristensen> samueldr: I did!
<manveru> infinisil: you still using tilix?
<gchristensen> though I haven't saved it yet ...
<samueldr> n/p, it's not going away
<samueldr> when the 2TB of that server fills up, it may
<gchristensen> haha cool, thank you :D
<samueldr> the trend says it won't go away soon
nixos1022 has quit [(Ping timeout: 248 seconds)]
<samueldr> do you have other ideas/needs that may need to be tracked statistically?
<manveru> i just got started with spacemacs a few days ago :)
<infinisil> manveru: I don't even know what tilix is, you must be mistaken me with somebody else ;)
<manveru> hmm
<samueldr> I want to run a task periodically to make similar graphs
<gchristensen> samueldr: you should chat up globin
<manveru> infinisil: https://github.com/NixOS/nixpkgs/pull/28394 mentioned you, so i thought you knew something about it :)
<manveru> i've got a segfault with it, it's a terminal emulator
<gchristensen> samueldr: if you need a place to host a cron, I can run it on nix.gsc.io
mahalel_ has quit [(Ping timeout: 255 seconds)]
<dhess`> Is there anything like a "nix-lint"?
<samueldr> thanks, I have my own underused server where it will happen
<gchristensen> dhess`: maintainers/scripts/nixpkgs-lint.nix
<gchristensen> but I'm not sure it is any goood?
<dhess`> gchristensen: ahh cool, I'll check it out.
<infinisil> manveru: Oh god, that's due to a kinda blunder i did, where i refactored like 2000 files to use http://my.url.com instead of "http://my.url.com" and it got merged in a day without proper thought
<manveru> ah :)
markus1199 has joined #nixos
ma27 has quit [(Ping timeout: 240 seconds)]
<manveru> welp, any ideas how to debug segfaults?
<samueldr> infinisil: the last one touching a file owns any issues with it :)
<infinisil> manveru: turn on debug symbols and stuff
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] grahamc pushed 2 new commits to unstable-aarch64: https://git.io/vFj3D
<NixOS_GitHub> nixpkgs/unstable-aarch64 704ad56 Graham Christensen: fixup kernelTarget
<NixOS_GitHub> nixpkgs/unstable-aarch64 23d6ccf Graham Christensen: Refactor the netboot setup and support aarch64-linux
NixOS_GitHub has left #nixos []
<manveru> i'll try latest version first
<infinisil> samueldr: Heh, I'm just glad mention-bot doesn't work anymore :P
<manveru> it's just D... so probably will take a while to build :|
markus1189 has quit [(Ping timeout: 268 seconds)]
erictapen has joined #nixos
ylwghst has joined #nixos
erictapen has quit [(Ping timeout: 250 seconds)]
<ylwghst> Anyone here's using Lightdm gtk ?
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] andir opened pull request #32057: ipv6calc: 0.99.2 -> 1.0.0 (also fixes gcc-7 compilation) (master...ipv6calc-gcc-7) https://git.io/vFj3j
NixOS_GitHub has left #nixos []
<dhess`> gchristensen: have you build uboots for any aarch64 systems?
<dhess`> s/build/built/
<gchristensen> I don't think so
wolfcub has quit [(Remote host closed the connection)]
<gchristensen> but I'm seriously an aarch64 noob
<dhess`> I think you do all right :)
<dhess`> anyway np
<dhess`> I want to update uboot on my armv7l BeagleBones, but I'm not exactly sure how that's done with NixOS
<samueldr> dhess`: as of right now, the best location for nixos on arm documentation is the wiki
<dhess`> seems like I just need to build the derivation and then copy the files myself
<dhess`> samueldr: yeah, I know of it and use it frequently, but it doesn't cover how to update uboot
<dhess`> not last I checked, anyway.
<samueldr> Dezgeg updates the article regularly
tgunb has quit [(Ping timeout: 248 seconds)]
dan_b has quit [(Ping timeout: 248 seconds)]
<samueldr> the uboot should be buildable with nix, and then using the `dd` invocation under Beaglebone Black, updated
<samueldr> but yeah, I don't think there's anything that automatically updates the bootloader binary
<dhess`> yeah that's what I figure. It's a bit odd that it doesn't happen automatically a la grub
<dhess`> maybe it's too platorm-specific.
<samueldr> well, right now it is
<samueldr> raspberry pis will use a file in the fat fs (iirc) for uboot
<samueldr> and every platform will have their own uboot
<samueldr> the system once booted does not exactly know (I think) which system it is
<samueldr> it can be discovered, but that's another thing that's not one-size-fits-all
<gchristensen> dhess`: mostly I haven't done much of anything but support and benefit from Dezgeg's efforts, honestly :)
<samueldr> though, it sure woule become a welcome addition!
<dhess`> samueldr: once I figure it out for BB, I will try to put something together.
<dhess`> I also want to make it possible to build DTBs and DTBOs from Nix
<samueldr> hmmmmm, I don't remember the hostname of my armv7l board
<samueldr> that's annoying
<Dezgeg> yea, a module could be written,, but care needs to be taken that the bootloaders don't trash any filesystems
<dhess`> oh hi Dezgeg :)
<sphalerite> Oh jeez, the JDK bootstraps are hosted on dropbox
<samueldr> ah, got it
<dhess`> Dezgeg: sure but that's true of grub as well.
<sphalerite> There also doesn't seem to be any info on how they're produced
<samueldr> on efi, much less of an issue
<samueldr> a bit like how on raspberry pis it's a file on a FS
<Dezgeg> grub does many checks
<Dezgeg> yes rpi is safe/easy but for allwinners you write to random sectors
<dhess`> Dezgeg: have you documented anywhere how to update uboot on BB for NixOS?
<sphalerite> shlevy: you seem to have made openjdk bootstrap tarballs at some point, could you enlighten me as to how?
<Dezgeg> wiki has it iirc
<dhess`> hmm I think that just refers you to Debian. I could be wrong, it's been awhile.
<dhess`> oh no it is there!
<dhess`> nice!
<dhess`> well, not the part where you build the uboot derivation in Nix, but that should be relatively easy.
<manveru> btw, how do you guys handle nix-shell with emacs
asuryawanshi has joined #nixos
<dhess`> manveru: do you mean getting Emacs to call it ?
<manveru> yeah, for compilation and such
<manveru> i often rely on languages that need commands from arbitrary node_modules and such :|
<dhess`> specifically, nix-sandbox
<dhess`> then you can hook that into your modes as needed
<manveru> ah
<manveru> interesting...
<samueldr> dhess`: possibly nix-build $NIXPKGS -A ubootBeagleboneBlack
<dhess`> samueldr: cool, thanks
<samueldr> (my armv7l board is currently working)
<samueldr> (but it seems to be doing the right thing)
ssmike has joined #nixos
<samueldr> I really need a faster sd card in there
<dhess`> ugh me too
asuryawanshi has quit [(Ping timeout: 248 seconds)]
ylwghst has quit [(Quit: Lost terminal)]
<manveru> dhess`: looks like that won't work nicely with spacemacs
<dhess`> manveru: ahh too bad.
<manveru> spacemacs has a nixos layer, but it's not particularly useful for nix-shell
<manveru> can probably copy it over though
dhess` is now known as dhess
<infinisil> hmm spacemacs..
phreedom has quit [(Ping timeout: 240 seconds)]
fragamus has joined #nixos
<infinisil> ohhh there needs to be an LSP backend for nix
dan_b has joined #nixos
<manveru> lisp?
Neo---- has quit [(Ping timeout: 268 seconds)]
<manveru> ah
phreedom has joined #nixos
<infinisil> so once you have an LSP frontend, such as in vim/emacs/atom, you can use any backend such as one for rust/haskell/python
<infinisil> which gets you autocompletion, syntax-highlighting, refactoring support, etc
<dhess> none of the Emacs modes I use are LSP-capable. It sounds great but seems a ways off yet.
<infinisil> dhess: There is this I just found recently: https://github.com/haskell/haskell-ide-engine
<dhess> infinisil: that looks promising but again, I don't know of an Emacs Haskell mode (or hook) that can use it yet.
<infinisil> dhess: Just found this: https://github.com/emacs-lsp/lsp-mode which mentions https://github.com/emacs-lsp/lsp-haskell which uses haskell-ide-engine :)
<dhess> it does seem like it could be the One True IDE API, eventually
<infinisil> Indeed
<infinisil> It is rather new though
<dhess> "This package is still under development, and is not recommended for daily use."
<infinisil> Yeah it might take a while, but I'm glad there's development happening in that direction :)
<dhess> me too, don't get me wrong.
<dhess> it's just not there yet.
<dhess> The LSP stuff is interesting enough that it makes VS.Code tempting... but it's not quite *that* interesting :)
oida has quit [(Ping timeout: 240 seconds)]
<infinisil> dhess: I assume you have a decent working setup for your development languages then?
<samueldr> dhess: after nix-build, I have the u-boot.img file under the result/ symlink
<samueldr> (with an MLO file for bealbone black)
<dhess> infinisil: yes
<dhess> samueldr: awesome, thanks for trying that out. I will give it a shot soon, got this BB test board sitting here waiting to go
<samueldr> I really need to upstream the u-boot build for my board
<dhess> samueldr: I'm going to update uboot to the latest release while I'm at it. It recently got the ability to load DTBOs before boot
oida has joined #nixos
NightTrain has quit [(Quit: Lost terminal)]
<dhess> samueldr: which board?
<samueldr> an orange pi, checking the actual version
<samueldr> Orange Pi PC
<samueldr> not that great, but really cheap
<dhess> cool, are you doing anything interesting with it? GPIO etc?
<samueldr> I wanted one for the armv7l port, before I found out aarch64 was more supported
<samueldr> bought it especially for playing around with nixos
<samueldr> like the pine64 I still haven't got around to setup
<dhess> samueldr: I think it really depends on your hardware. All of the armv7l stuff I've tried so far has worked great. aarch64 is a mixed bag.
<samueldr> that one will be fun to use, it can use an emmc chip (which I bought)
<samueldr> raspi3 works fine with the stock aarch64
<samueldr> and the pine64 I bought it especially to further help withe aarch64 port
<samueldr> the A64-LTS
<samueldr> (and hopefully, get a pinebook)
<dhess> The pine64 looks interesting. Does it have any closed-source bits? I can't recall.
<samueldr> yes
<samueldr> about the same as most allwinner
<samueldr> upstream u-boot could be good enough if no gpu is needed, not sure yet
<samueldr> (upstream will be needed for the way nixos boots)
<samueldr> (the allwinner fork of u-boot is just ancient)
<samueldr> otherwise, there's a surprisingly good community with good enough documentation (through code)
<samueldr> I was surprised to see that there are *build instructions* with the recommended distribution images
acarrico has quit [(Ping timeout: 248 seconds)]
<samueldr> too often those cheaper or lesser used boards, you have to use the mystery meat builds and hope for the best
<dhess> Yeah that's why I've stuck with the BeagleBone on the armv7l side.
<clever> samueldr: ive used the nixpkgs u-boot on the banana pi r1
<dhess> it has an excellent kernel fork for the stuff that isn't in upstream yet.
<clever> samueldr: but my PSU strugles to keep it stable, and it has yet to get into a linux kernel
<dhess> several, actually.
<samueldr> right now, rpi3 is probably the best for nixos, since it's mostly supported with aarch64
<samueldr> even does modesetting properly
<samueldr> with the upstream kernel!
<dhess> I really like my Jetson TX1 with NixOS aarch64
<dhess> I'm not using the GPU though, no idea whether that works.
<clever> i had gone with the bpi1 because it includes a 5 port gigabit switch
<samueldr> that price though
<clever> i was planning on using it as a router
<dhess> sure, but I'm using it for a build server, so it's worth it in time saved on builds alone :)
<samueldr> yes, definitely
<samueldr> I was half-tempted
<dhess> also I think I saw somewhere the other day that it's down to $200
<clever> there is only one critical flaw with the one i got, the switch IC defaults into all ports linked mode
<samueldr> but instead sank money into a workstation (used) with a 6 cores xeon and 64GB of ram
<clever> so the wan port is essentialy just a 5th lan port, until the OS boots
<clever> zero firewall
<dhess> samueldr: I wish cross-compiling with Nix was less of a mystery. I might only need the ARM build server for Haskell stuff if I could cross-compile everything else.
<dhess> I think there are some PRs in flight to clean it up quite a bit.
* samueldr forages through notes
<clever> dhess: have you seen the 96 core arm servers packet.net has?
<dhess> clever: yeah, they're amazing!
<dhess> clever: they are also incredibly expensive@!
<samueldr> older discussion with clever
<dhess> I could buy an additional TX1 a month for what they're charging, I believe.
<samueldr> still haven't got around to test that
<dhess> samueldr: I'm going to wait for these PRs to merge, then have another go. It sounds like the story will be much better soon-ish.
<samueldr> I think I'm trying to replicate the setup next
<samueldr> I mean, it's been month that I said I was going to do it
dan_b has quit [(Ping timeout: 248 seconds)]
<clever> i recently fixed my armv8 build slave
ris has quit [(Ping timeout: 240 seconds)]
<clever> turns out the new glibc doesnt agree with the nssswitch.conf of rasbbian
<clever> but the nix sandbox fixes it
NightTrain has joined #nixos
mahalel_ has joined #nixos
ssmike has quit [(Ping timeout: 248 seconds)]
lambdamu_ has joined #nixos
acarrico has joined #nixos
lambdamu has quit [(Ping timeout: 248 seconds)]
<petersjt014[m]> is there any way I can 'refactor' configuration.nix and pull all of the imports and such in?
<petersjt014[m]> or at least cat out that result?
<clever> petersjt014[m]: nix-repl '<nixpkgs/nixos>' and then try to eval config i believe
nuncanada has quit [(Ping timeout: 240 seconds)]
<petersjt014[m]> did you mean nix-shell? nix-repl doesnt seem to be on the installer iso
<clever> nix-env -iA nixos.nix-repl
<clever> also, it needs a different command if you want to view the configuration.nix of a different install
<petersjt014[m]> okay, got the repl installed
<clever> petersjt014[m]: what exactly do you want to do?
<petersjt014[m]> im booted into the graphical installer iso and I want to see what the experssion looks like with all the imports imported
<petersjt014[m]> because somewhere in there are some drivers that I nned
<petersjt014[m]> I know where to find some of the imports, but there's a lot to dig through and getting it all in one place+refactoring it would make it more visible I'm thinking
mbrgm has quit [(Ping timeout: 250 seconds)]
* petersjt014[m] sent a long message: petersjt014[m]_2017-11-26_02:36:24.txt <https://matrix.org/_matrix/media/v1/download/matrix.org/KVimFhwfKglDaGxeISpwhOKv>
<petersjt014[m]> and others in the same-ish directory
<petersjt014[m]> somewhere in there is the reason wpa_supplicant doesn't work on my main install
<petersjt014[m]> and a few other things
mbrgm has joined #nixos
astronav_ has joined #nixos
Supersonic112 has quit [(Disconnected by services)]
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
ryantm has quit [(Quit: Connection closed for inactivity)]
<samueldr> oh \o/, got my xeon to build arm stuff using the trickery I linked earlier
<dhess> samueldr: ooooh
<dhess> samueldr: got a one- or two-liner?
<samueldr> not 100% sure it works, I'll have to wait for the build to come back
<dhess> it doesn't use qemu or something like that, does it?
<samueldr> it *does some work*, but what it does... I'm not sure if it's valid
<samueldr> qemu-user
<samueldr> much faster than full-blown qemu
<dhess> hmm. I suspect my Jetsons will be faster.
<samueldr> possibly
<samueldr> I'm curious what we could try to benchmark the build speed
<samueldr> I mean, it's a valid thing to know
<samueldr> there's only one thing that bugs me right now: I'm not sure it's actually armv7l, it may be working as armv6l and I don't know if it's important
<samueldr> but it's started the kernel build I asked for
seanparsons has quit [(Ping timeout: 248 seconds)]
seanparsons has joined #nixos
rogue_koder has joined #nixos
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixpkgs-17.09-darwin advanced to https://github.com/NixOS/nixpkgs/commit/bd80745771 (from 18 hours ago, history: https://channels.nix.gsc.io/nixpkgs-17.09-darwin)
<nix-gsc-io`bot> Channel nixpkgs-unstable advanced to https://github.com/NixOS/nixpkgs/commit/45a85eaceb (from 20 hours ago, history: https://channels.nix.gsc.io/nixpkgs-unstable)
nix-gsc-io`bot has quit [(Client Quit)]
dieggsy has joined #nixos
mahalel_ has quit [(Ping timeout: 240 seconds)]
amfl has quit [(Ping timeout: 250 seconds)]
mahalel_ has joined #nixos
amfl has joined #nixos
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
dieggsy has quit [(Quit: ERC (IRC client for Emacs 27.0.50))]
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixos-unstable-small advanced to https://github.com/NixOS/nixpkgs/commit/3385c85fb5 (from 11 hours ago, history: https://channels.nix.gsc.io/nixos-unstable-small)
nix-gsc-io`bot has quit [(Client Quit)]
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixos-17.09-small advanced to https://github.com/NixOS/nixpkgs/commit/b03e6f9d35 (from 4 hours ago, history: https://channels.nix.gsc.io/nixos-17.09-small)
nix-gsc-io`bot has quit [(Client Quit)]
<gchristensen> wow!
sibi_ has quit [(Quit: Connection closed for inactivity)]
<samueldr> channels were stuck?
<infinisil> whowie
mahalel_ has quit [(Ping timeout: 258 seconds)]
acarrico has quit [(Ping timeout: 260 seconds)]
jtojnar has quit [(Ping timeout: 240 seconds)]
dieggsy has joined #nixos
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixos-unstable advanced to https://github.com/NixOS/nixpkgs/commit/2f1a818d00 (from 21 hours ago, history: https://channels.nix.gsc.io/nixos-unstable)
nix-gsc-io`bot has quit [(Client Quit)]
amfl has quit [(Read error: Connection reset by peer)]
amfl has joined #nixos
ryantm has joined #nixos
jtojnar has joined #nixos
* ryantm @gchristensen no... maybe something with irccloud.
asymmetric1 has quit [(Ping timeout: 268 seconds)]
astronav_ has quit [(Remote host closed the connection)]
* ryantm testing /me
<ryantm> not /me
iyzsong has joined #nixos
<samueldr> (what it looks like to me)
<ryantm> It seems that doing a deliberate /me message fixed irccloud's state
<samueldr> looks like it
* ryantm Sorry about that.
<ryantm> oh
<ryantm> I figured it out! It's typing ctrl-enter instead of enter
mizu_no_oto has joined #nixos
jtojnar has quit [(Remote host closed the connection)]
johnw_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] adisbladis closed pull request #32043: vaapi-intel: fix missing depency to get X11 backend (master...fix-missing-x11-in-vaapi-intel) https://git.io/vFjI8
NixOS_GitHub has left #nixos []
slack1256 has joined #nixos
<infinisil> test
jtojnar has joined #nixos
babyflakes has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
mizu_no_oto has joined #nixos
slack1256 has quit [(Remote host closed the connection)]
fragamus has joined #nixos
schoppenhauer has quit [(Ping timeout: 248 seconds)]
schoppenhauer has joined #nixos
jtojnar has quit [(Quit: jtojnar)]
hiratara has quit [(Ping timeout: 240 seconds)]
dieggsy has quit [(Remote host closed the connection)]
jtojnar has joined #nixos
hiratara has joined #nixos
jsgrant_ has quit [(Ping timeout: 240 seconds)]
aw has quit [(Remote host closed the connection)]
spacefrogg has quit [(Remote host closed the connection)]
jtojnar has quit [(Quit: jtojnar)]
aw has joined #nixos
spacefrogg has joined #nixos
mkoenig has quit [(Ping timeout: 240 seconds)]
asuryawanshi has joined #nixos
jsgrant_ has joined #nixos
mkoenig has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
asuryawanshi has quit [(Ping timeout: 260 seconds)]
jtojnar has joined #nixos
* pcarrier oh really?
<pcarrier> good to know, thanks, ryantm.
jtojnar has quit [(Remote host closed the connection)]
jtojnar has joined #nixos
asuryawanshi has joined #nixos
MP2E has quit [(Remote host closed the connection)]
jtojnar has quit [(Remote host closed the connection)]
jtojnar has joined #nixos
asuryawanshi has quit [(Ping timeout: 240 seconds)]
johnw_ has quit [(Quit: Textual IRC Client: www.textualapp.com)]
<petersjt014[m]> I have successfully booted an install and logged in
<petersjt014[m]> (smol applause)
<petersjt014[m]> thanx to some people here (forgot who, sorry) otherwise I wouldn't have internets on it rn
jb55 has joined #nixos
<petersjt014[m]> although if anyone can point me to an example setting of media keys that'd be nice because I am unable to turn down my brightness and my battery is about to kick it
<petersjt014[m]> (I have other devices w/riot, but this is just a bit less tedious)
asuryawanshi has joined #nixos
asuryawanshi has quit [(Ping timeout: 268 seconds)]
NightTrain has quit [(Quit: Lost terminal)]
asuryawanshi has joined #nixos
mrkgnao has joined #nixos
asuryawanshi has quit [(Ping timeout: 240 seconds)]
asuryawanshi has joined #nixos
jb55 has quit [(Ping timeout: 240 seconds)]
asuryawanshi has quit [(Ping timeout: 240 seconds)]
jmc_fr has joined #nixos
jb55 has joined #nixos
Wizek_ has quit [(Ping timeout: 240 seconds)]
rogue_koder has quit [(Ping timeout: 248 seconds)]
jsgrant_ has quit [(Ping timeout: 248 seconds)]
justan0theruser has quit [(Ping timeout: 252 seconds)]
jsgrant has joined #nixos
jsgrant has quit [(Remote host closed the connection)]
jsgrant has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] volth opened pull request #32061: varnish: 5.1.3 -> 5.2.1 (master...patch-72) https://git.io/vFj4u
NixOS_GitHub has left #nixos []
jsgrant has quit [(Ping timeout: 248 seconds)]
asuryawanshi has joined #nixos
ryantm has quit [(Quit: Connection closed for inactivity)]
avn has quit [(Read error: Connection reset by peer)]
avn has joined #nixos
justan0theruser has joined #nixos
pie__ has quit [(Ping timeout: 248 seconds)]
asuryawanshi has quit [(Ping timeout: 260 seconds)]
fuzzy_id has joined #nixos
<fuzzy_id> i want to write a nixos test for a service i am developing
<fuzzy_id> hence, when i try to run the test with `nix-build path/to/test.nix` i get:
<fuzzy_id> "The option services.cassandra defined in […] does not exist"
<fuzzy_id> i think i forgot to import the service module
kvz has quit [(Quit: ZNC 1.6.3 - http://znc.in)]
<fuzzy_id> ah, that did the trick!
lukec has quit [(Quit: WeeChat 1.4)]
jmc_fr has quit [(Ping timeout: 240 seconds)]
ma27 has joined #nixos
<hyper_ch> adisbladis: why split it up into seperate things? the problem is if one of the dependencies is missing, it won't work
<hyper_ch> adisbladis: no idea what you mean here Don't pass python versions explicitly when called from pythonPackages.
<hyper_ch> adisbladis: no idea what you mean here It is better to use pythonPackages.callPackage, and list python dependencies in the arguments of the function that returns the derivation.
<hyper_ch> adisbladis: no idea what you mean here Its better to list all dependencies in the function that returns the derivation.
Isorkin has quit [(Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)]
<adisbladis> hyper_ch: Its more clear when you are digging through history and triaging issues if everything is its own commit
<hyper_ch> it's not clear
<adisbladis> hyper_ch: When something is called from pythonPackages with callPackage the correct interpreter version is already set
<hyper_ch> it won't work with python2
<adisbladis> hyper_ch: And rather than using with python3.pkgs its better to explicitly list all dependencies in the function arguments
<hyper_ch> I don't know what you mean by that
<adisbladis> hyper_ch: I'll make you a PR to your fork and walk you through the changes :)
Maxdamantus has quit [(Ping timeout: 248 seconds)]
Maxdamantus has joined #nixos
jb55 has quit [(Ping timeout: 248 seconds)]
FRidh has joined #nixos
jb55 has joined #nixos
sbjorn has joined #nixos
mrkgnao has quit [(Ping timeout: 240 seconds)]
ssmike has joined #nixos
simukis has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] FRidh closed pull request #31555: python3.buildEnv: pick outputs correctly, refactor (master...python-buildenv/dev-nightmare) https://git.io/vFwXN
NixOS_GitHub has left #nixos []
<joepie91> oh nice, nix-info is a thing now
* joepie91 trashes his custom shell alias :)
leat has quit [(Ping timeout: 240 seconds)]
mrkgnao has joined #nixos
leat has joined #nixos
sigmundv has joined #nixos
<joepie91> so I've just set up Hydra and added a user using hydra-create-user and now it's claiming that my password -- the same one I gave to hydra-create-user! - is wrong
<joepie91> any thoughts?
<joepie91> nothing meaningful in journalctl
<pcarrier> has anybody automated nix-push'ing their system and user profile derivations and rsync'ing them over to a remote host used as a cache?
<pcarrier> I could save time not building that from scratch :)
<joepie91> pcarrier: I don't know much about it, but this sounds vaguely like something in the area of nix-copy-closure
<joepie91> or something along those lines
sigmundv has quit [(Ping timeout: 255 seconds)]
<joepie91> urgh, why is Hydra not working :(
<joepie91> ... aha.
<joepie91> it seems the hydra- tools are interacting with an sqlite DB in /var/lib/hydra, not with the postgresql service that the web interface talks to
<pcarrier> joepie91: yeah so I want a binary cache, nix-copy-closure seems to be to populate nix stores
<pcarrier> nix-push is very much what I'm after I'm afraid
<pcarrier> as in, I don't want to trust the remote store, I want signed NARs there.
Harpalus has joined #nixos
<pcarrier> have all my systems push their stuff there regularly, and then use it as a cache everywhere :)
asuryawanshi has joined #nixos
<Harpalus> Good morning everyone. I got a question regarding updating to 17.09. I'm on 17.03 and in 17.09 guile-xcb is marked as broken. When I try upgrading by setting allowBroken=true, the build process fails. Is there any way to solve this? I'd appreciate any hint on what I could do. Thanks in advance!
<joepie91> pcarrier: hold
asuryawanshi has quit [(Remote host closed the connection)]
<pcarrier> joepie91: ha ha I was about to start writing a derivation based on systemd.services to build a systemd timer running a shell script that uses `nmcli --terse device show wl0|grep GENERAL.CONNECTION|cut -d: -f2-` and `sshfs` and `nix-push` in a `minijail0` :)
<pcarrier> OK that's a lie I totally started writing it.
<pcarrier> but I'll hold because that doesn't make me happy at the moment:)
ssmike has quit [(Ping timeout: 248 seconds)]
<pcarrier> that being said I'd like to be able to tweak towards that general direction or have something that do at least most of what I want
<joepie91> there we go
<pcarrier> maybe a prebuilt solution will be even better in which case yay
<joepie91> hydra login works now
<joepie91> seems the problem was that the env vars needed for the commandline tools to work, are set in /etc/profile - and unless you load that manually, I don't think it gets loaded for a new shell unless you relogin?
<pcarrier> oh to use a binary cache it needs to have entry points?
<pcarrier> that's deeply dissatisfying at the moment
<joepie91> pcarrier: to be honest, a timer service is probably the sanest approach for automating this, short of some sort of post-build hook (not sure if this is possible?)
<pcarrier> oh yeah no I don't want to do that on the fly
<joepie91> not sure about the sshfs bit though
<pcarrier> I might not have a fast uplink at that point
<joepie91> right
<pcarrier> I want this to only run on a few privileged networks when I know it's a good idea
<joepie91> hmmm
<pcarrier> and have systemd retry regularly when it fails, but once it succeeds not for another 24 hours or something
<joepie91> pcarrier: idea: dbus-activated service? that responds to dbus connection event for those newtorks
<joepie91> networks*
<pcarrier> so if I disconnect/reconnect often I don't want that to run too often
<joepie91> will probably only work if you're using networkmanage
<joepie91> networkmanager*
<pcarrier> and if I decide not to run yet, when would I choose to run again?
<pcarrier> oh yeah exactly
awi has joined #nixos
<pcarrier> as I was saying, `nmcli --terse device show wl0|grep GENERAL.CONNECTION|cut -d: -f2-`
<joepie91> oh, doh
<pcarrier> that `wl0` device name comes from a udev rule so yay.
<joepie91> but yeah, you can basically just listen to a dbus event then
<joepie91> afaik networkmanager emits such events
<pcarrier> right so what if I don't want to run right now because I ran 12 hours ago
<pcarrier> I don't want to possibly not run again until I reconnect, which could be days later
<pcarrier> I'm pretty picky about system design for stuff like backups :)
<joepie91> pcarrier: the most robust approach would be to just 1) write a shell script that checks the timestamp of last sync and syncs if over X seconds, 2) create a periodic timer service that calls this shell script, and 3) create a thing that listens to a networkmanager connection event for those interfaces
<joepie91> that way you have the highest chance of it actually occurring while connected to the right network
<joepie91> 'last timestamp' can just be /var/lib/mysyncthingie
<joepie91> or whatever
<pcarrier> I mean yeah sure
<joepie91> stateful file that just has the timestamp of last sync
<pcarrier> you just described something perfect here
<pcarrier> but if I just run every 10 minutes until I stop failing then only after 24 hours I'm gold
<pcarrier> it's much simpler and much more predictable and I have a full history in my journald (which I replicate as well)
<pcarrier> I _think_ systemd can do this. I have to check.
<pcarrier> but I think its timers are designed for weird scheduling like this.
<pcarrier> hmmm yeah I think I've convinced myself that this could work well.
<pcarrier> that's the kind of stuff I would never bother to make reproducible or carry around systems before so wasn't worth writing
ixxie has joined #nixos
<pcarrier> but with nix why the hell not.
<joepie91> what you're trying to do almost sounds like an ad-hoc Hydra setup :P
<ixxie> pcarrier: I'm just coming in now but I think that line should be a T-Shirt
<joepie91> probably conceptually simpler though
<joepie91> and yes, +1 :p
<ixxie> pcarrier: "...but with Nix why the hell not!"
<pcarrier> yeah I mean hydra is right in many ways
<pcarrier> if I was building infrastructure with nix I would want hydra FOR SURE
jb55 has quit [(Ping timeout: 248 seconds)]
<pcarrier> but for now it's for my personal machines and those I need to be able to tweak on the fly in arbitrary conditions
<pcarrier> some of which might include in the not-too-far future "I need this shit on my laptop and I'm in New Zealand for the month"
<pcarrier> so infrastructure investments must be very passive, conditional, and require very little infrastructure (eg a host I can ssh into and can act as a binary cache is something I could provision in the "local" EC2 with nixops for really cheap).
<pcarrier> ie a personal, volatile setup, not at all what I'd do for my employer.
zraexy has quit [(Ping timeout: 248 seconds)]
<pcarrier> eg I never want to have to deal with SSL certificates that work for whichever DNS entry my cache has today. that's too hard to port around.
<pcarrier> I'm really a huge fan of nix's flexibility because I can both play low-cost tricks and build solid "businessey" systems. I'm just not sure I'll ever get to use it professionally for anything other than my own workhorses.
orivej has quit [(Ping timeout: 240 seconds)]
fuzzy_id has quit [(Ping timeout: 240 seconds)]
<pcarrier> it would need quite some polish and documentation, and I would have to onboard a few colleagues personally first, before I could go to my team and argue we should build on it as a company. I'm trying the early steps of that process though and might have some free time on my hand to contribute :)
<pcarrier> but clearly Nix is everything I always wanted and it seems impossible to ignore the potential.
<samae> pcarrier: +1
<pcarrier> the way everything I do can turn into recipes and easily they are to reuse and share makes me think a "product" touch on top of it could have wonderful results.
<pcarrier> a private AWS cloud with hydra-based CI and app workers for github repositories by writing documented high level expressions could blow every other solution out of the water
<pcarrier> with nice composition to build docker images and push them to an integrated docker repository, oh my god
<pcarrier> then support deploying the same thing to kubernetes and you can bring a lot of players to the table
<samae> thinking along those lines too. I think it is feasible to have build a user friendly ubuntu-like distro on top of nix, and that would rock, with auto updates that don't break the system and defaults apps that are living coherently right from the config file
<TimePath> you've just described how I use nix
<pcarrier> tat stuff is super hard to build any other way, and each company invents their own ad hoc super limited solution
<pcarrier> then invests ridiculous amounts of time to make it work nearly as well as nix would out of the box
<pcarrier> and they have large teams building and maintaining that stuff
<pcarrier> TimePath: what would it take to bring your solutions to dozens of San Francisco startups, in environments they're familiar with, and with nice high-level abstractions we potentially improve over time to support new customers?
<joepie91> my notes from setting up Hydra, so far: https://gist.github.com/joepie91/c26f01a787af87a96f967219234a8723
<TimePath> pcarrier: all I did was read through the nix repo, figure out it can produce docker images, wrote some kubernetes descriptors in nix expressions and converted them to JSON, and ran the whole thing in CI
<joepie91> samae: ~that's my project~
<joepie91> :p
<TimePath> the building blocks are all there already
<pcarrier> I know
<pcarrier> but it needs some cute packaging on top of be something startups will pay for
<Ralith> is there a configuration.nix option to enable udev rules for yubikeys?
<Ralith> strange that it's not supported by default
<pcarrier> cause their engineers have better things to do than decide for themselves whether they're building the correct solution
<pcarrier> Ralith: yes you can add your own udev rules very easily
<pcarrier> services.udev.extraRules as a string
ixxie has quit [(Ping timeout: 248 seconds)]
<Ralith> pcarrier: I'm aware, just wondered if there was an out of the box thing
<pcarrier> Ralith: which udev rules do you need? I can search for them and/or contribute them back. I have most iterations of yubikeys to test with.
<TimePath> there's a u2fAuth in nixpkgs/nixos/modules/security/pam.nix, which might be relevant
<TimePath> right below is usbAuth
<pcarrier> OK you're trying to do PAM auth with a Yubikey U2F or 4/4C?
hotfuzz_ has joined #nixos
<Ralith> pcarrier: I just got a couple FIDO U2F keys, and I mean I'm perfectly capable of copying the relevant bit of https://github.com/Yubico/libu2f-host/blob/master/70-u2f.rules into my config myself
<Ralith> not currently trying to set up PAM, just enable chrome to talk to it for now
<pcarrier> OK cool. well that's on my checklist for my setup as well so we can be friends
<pcarrier> I'll get on it too :)
hotfuzz has quit [(Ping timeout: 248 seconds)]
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<pcarrier> Ralith: I'll put something together in nixos/modules/services/hardware/yubikey.nix
<Ralith> sweet, thanks!
<samae> joepie91: oh nice!
<joepie91> samae: it's very long-term though :P
<joepie91> a lot still needs to happen to really get it right
<samae> sure!
<pcarrier> nevermind Ralith, it should be in pkgs/development/libraries/libu2f-host
whald has joined #nixos
<samae> joepie91: what's missing?
<pcarrier> wait it does that already Ralith
petsa has joined #nixos
<joepie91> samae: lots of different things in lots of areas; from the underlying NixOS perspective there's a lot of missing documentation and developer tooling, as well as some inconsistencies / missing policy / missing extensibility in nixpkgs
<Ralith> pcarrier: the file exists there, but is there a way to use that to get it into the system configuration?
<joepie91> samae: then on a higher level I'm very unhappy with the current state of desktop environments and window managers, as well as the poor integration between different applications
<Ralith> pcarrier: a .rules file in /nix/store somewhere doesn't really do any good on its own
<pcarrier> yup yup yup
<pcarrier> getting there :)
<joepie91> samae: but even if you ignored that, for it to be truly user-friendly you need a way to represent package and system management to the user in a graphical, understandable way - without giving up on Nix' guarantees
<joepie91> which means that all existing package and system management tooling is 100% useless because it's designed around global systems
iyzsong has quit [(Read error: Connection reset by peer)]
<joepie91> often to a degree that retrofitting it into NixOS would be more work than just rolling your own
iyzsong has joined #nixos
<samae> joepie91: do you actually need that?
<joepie91> and then there's the need to figure out how to represent NixOS' *unique* features in a user-friendly way; that is, things for which there is no prior work in terms of UX, like for example system rollbacks
<joepie91> samae: 'that' referring to?
<samae> "package and system management to the user in a graphical, understandable way"
<joepie91> well, yes
<TimePath> if you want to take over ubuntu, yes
FRidh has quit [(Quit: Konversation terminated!)]
<joepie91> exactly
<joepie91> commandlines are nice for developrs
<joepie91> not so nice for the average user
<TimePath> and sysadmins
<samae> Does an average user install stuff?
<joepie91> ...yes?
<samae> ok…
<samae> I'm so far gone then :P
<joepie91> how else would they come by the software they need?
<TimePath> unless you happen to bundle everything they need..?
hiroshi- has joined #nixos
<samae> With nix you could have bundles
<joepie91> TimePath: that works until your distro leaves the lab :)
<joepie91> and it meets real-world users
<samae> like: the web navigation bundle, the document writing bundle, etc.
<joepie91> "oh, but I need <obscure application for creating knitting patterns>!"
mahalel_ has joined #nixos
<TimePath> joepie91: it is an impossible task :P
<joepie91> samae: user needs are far more varied than that, and even then you *still* need a graphical interface
<TimePath> samae: so which office suite does that include?
<TimePath> all of them?
<joepie91> that somehow layers over Nix without restricting it
<joepie91> which is possible, don't get me wrong
<joepie91> there's just no prior work, no tooling for it
<Ralith> hmm
<samae> TimePath: is it that important in general? Which office suite you want to use?
<Ralith> I enabled that udev rule and switched, and reinserted my device, and chrome is still crapping out
<Ralith> not sure how to verify that it's working
<joepie91> Ralith: are you sure the problem isn't chrome :D? (kidding)
<Ralith> no, I'm not, but I have little else to go on :P
<TimePath> samae: not that important to me, no. For the web one, what browser would that be?
<samae> Ralith: don't you need to also restart udev or something like that?
hiroshi has quit [(Ping timeout: 240 seconds)]
hiroshi- is now known as hiroshi
<joepie91> samae: users tend to care a lot about familiar interfaces
Synthetica has joined #nixos
<samae> sudo udevadm control -R ?
<samae> joepie91: that is true
<Ralith> samae: I kind of assumed that nixos-rebuild switch would do that when I touched the rules
<samae> TimePath: yeah so it all boils down to what kind of users we are talking about
<joepie91> samae: so yeah, realistically, to compete with Ubuntu et al, you need to offer an equivalent 'store' interface
<Ralith> there's definitely no non-root permissions on /dev/hidraw0
<samae> ubuntu didn't care about what browser, it installs a default one
<joepie91> which brings me to another point: the lack of serious package metadata in nixpkgs
<samae> that most user accept
<joepie91> it has a title, description, website, and not much else
<TimePath> samae: and users familiar with chrome go and install that one
<joepie91> can't build a good package manager out of just that
<Ralith> restarting systemd-udevd manually didn't fix that
<joepie91> so that's yet more work to be done for a good UX
<samae> Then, why not first aim for a distro with NO package management
<joepie91> I'm not complaining or anything, to be clear, just wanting to point out how much work still needs to be done in a variety of aspects
<Synthetica> Is there any DNS caching other than from nscd enabled by default?
<joepie91> to get to an Ubuntu-equivalent user-friendly distro
<joepie91> or preferably, beyond that
<samae> something you can have just run, and that works?
<Ralith> the udev rule I added definitely has the right product/vendor IDs
<joepie91> samae: because people don't use an OS that you can't install things on
<TimePath> joepie91: I'd love to use nixos to manage a fleet of computers
<joepie91> TimePath: how do you mean?
<TimePath> joepie91: on employee workstations
<joepie91> in the sense of employee workstations running NixOS?
<samae> joepie91: we clearly don't have the same idea on "people". That'd be a first thing to do, actually identifying the user group
<joepie91> and managing them from a central expression or so
<TimePath> yeah, but managed centrally
<joepie91> right
<TimePath> yup
<joepie91> TimePath: it's something that's definitely possible, but I personally wouldn't consider this as a major target group (yet), because in enterprise environments it tends to be exceedingly difficult to get people to switch *anything*
<TimePath> it's orthogonal to managing a local computer, but would be nice to have a similar graphical experience
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
asuryawanshi has joined #nixos
erictapen has joined #nixos
<joepie91> samae: final target group for what I'm working on: the average user who is currently probably using Windows or maaaybe Ubuntu, who's not technically inclined other than general interaction with a computer when it isn't broken
<joepie91> samae: but initial target group: the not-so-clueless-but-curious users
<joepie91> er
<joepie91> not-so-clueful-but-curious*
<Ralith> pcarrier: you have a yubikey working, right? I switched in the new udev rule which adds the uaccess tag to the correct subsystem/vendor/product, but the generated /dev/hidraw0 is still 0600 root:root
<Ralith> what might I have missed?
<samae> joepie91: and this group needs to install things? I mean, rare things?
<pcarrier> Ralith: I'm working on it :)
<joepie91> which is a good adoption group because they're often willing to try out new things even at the cost of a temporarily nonfunctional system, but they will still be fairly non-technical and complain about things that aren't user-friendly enough
mahalel_ has quit [(Ping timeout: 246 seconds)]
<joepie91> good first adoption group *
<samae> I see
<joepie91> think the kind of person who would have built a geocities site back when that was still a thing, by barely learning HTML
<TimePath> samae: there is a certain class of user that does everything from a web browser, they won't really benefit from using nixos over whatever they are currently
<joepie91> samae: both groups need to install things; for the final group ("average user") it's mostly ready-made software for non-computer hobbies (think model train control software, knitting pattern software, that kind of thing), for the first group ("curious but not clueful user") it'll be largely handy utilities for tweaking their system without needing too much technical knowledge (think teracopy-like software etc.)
<pcarrier> Ralith: I'm building the system configuration then trying it out. `services.yubikey.enable=true` installs the udev rules in /etc/udev/rules.d through environment.etc
<pcarrier> probably should have started from nixpkgs-channels unstable to save time for experiments, the build might take a while.
asuryawanshi has quit [(Ping timeout: 260 seconds)]
<Ralith> that is cool, but the udev rules don't actually seem to be working for me
<Ralith> I wonder if there's something else
<pcarrier> yeah nevermind I'll rebase my branch.
<joepie91> Ralith: you're sure your yubikey is Linux-compatible?
<Ralith> (yeah experimenting atop nixos-unstable is a huge timesaver)
<joepie91> I vaguely recall there being non-Linux-compatible models
<Ralith> I am not aware of any such
<joepie91> at some point in time
<Ralith> regardless, this one is clearly advertised as compatible
<joepie91> right, okay
<joepie91> samae: btw, another item on the pile of work; to reach the final target group, you'd ideally want some way for users to provide feedback from *within the OS itself*, eg. filing support issues when software doesn't work or crashes, and so on
<joepie91> not only does that require having such reporting software in the first place, it also requires actual human support
<joepie91> a big problem with "Linux for the masses" is that a lot of software *could* theoretically run on Linux if you run it through eg. WINE, but a lot of software is too obscure to be supported out of the box and requires special handling, or even native Linux software may need weird deps that distros tend to note have
<joepie91> not have*
<joepie91> these are all problems that need solving if you want to build a distro that can actually take over the 'average user' demographic
simukis has quit [(Quit: simukis)]
simukis has joined #nixos
dhess has quit [(Remote host closed the connection)]
<joepie91> and the problem is that even being "as good as Windows" isn't enough - that's not a reason for people to expend the effort to switch
<joepie91> you need to actually beat it in UX
<joepie91> so yeah, hence "very long-term"
dan_b has joined #nixos
simukis has quit [(Client Quit)]
simukis has joined #nixos
Isorkin has joined #nixos
<TimePath> joepie91: yes, many people don't think about cost of switching
<TimePath> you have to be better than the cost of switching on top of the thing you're trying to replace
<Ralith> pcarrier: googling around is giving me the impression that uaccess is broken in udev.extraRules
<Ralith> all the nixos examples have hardcoded mode+group stuff bodged in
<Ralith> pcarrier: appending `RUN{builtin}+="uaccess"` to the rule seems to fix the permissions but I don't understand why it's necessary and I doubt it's the correct fix
<Ralith> perhaps local rules run after uaccess for some reason
<Ralith> incidentally /etc/udev/rules.d appears to be a symlink to some udev specific magic in the store so installing via environment.etc might not be as intended
<Ralith> ooh
<pcarrier> yeah I figured that stuff out
<pcarrier> my system builds, lemme see what happens
<pcarrier> if I need to create a group & co I'll be happy to
<Ralith> I think putting libu2f-host in udev.packages might be the proper magic
<Ralith> testing
<hyper_ch> adisbladis: still here?
<Ralith> yes, that appears to do it!
<vaibhavsagar> Who do I talk to about getting https://github.com/NixOS/nixpkgs/pull/31395 merged?
<vaibhavsagar> It's been a while now :)
<adisbladis> hyper_ch: Yep
<Ralith> (and chrome is happy with that as well)
<avn> Folks, anyone use mopidy on nixos?
<hyper_ch> so, setup server to do now ocr from my original nix expressions.... will get some food and afterwards I can try to redo that pr
<pcarrier> Ralith: PR coming
tmaekawa has joined #nixos
<Ralith> pcarrier: using the udev.packages method?
<pcarrier> yup, packaged neatly hopefully
<Ralith> \o/
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] adisbladis opened pull request #32068: sudo: 1.8.20p2 -> 1.8.21p2 (master...sudo-1_8_21p2) https://git.io/vFjgn
NixOS_GitHub has left #nixos []
<pcarrier> services.u2f.enable = true and yipee.
<Ralith> looks very tidy!
<pcarrier> can't wait to hear if it can be made prettier
thc202 has joined #nixos
<Ralith> PR title looks/commit message looks out of date
<pcarrier> it was automatically filled by github
<pcarrier> not to say you're wrong, maybe I did something wrong?
<Ralith> from the commit message, which looks out of date
<Ralith> :p
dan_b has quit [(Ping timeout: 255 seconds)]
<Ralith> your commit message says "nixos/modules/services/hardware/yubikey" but the commit contains "nixos/modules/services/hardware/u2f.nix"
tmaekawa has quit [(Quit: tmaekawa)]
<pcarrier> oooh I see what you're saying sorry
<Ralith> np, just a quibble
<pcarrier> no no thanks for noticing
ixxie has joined #nixos
<awi> Hi, cannot access UEFI/Bios menu anymore. The boot menu disapeared as well. System boots normally. Anyone here that can help?
<pcarrier> awi: what happens when you boot?
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] adisbladis opened pull request #32070: virt-manager: 1.4.2 -> 1.4.3 (master...virt-manager-1_4_3) https://git.io/vFjgd
NixOS_GitHub has left #nixos []
<pcarrier> awi: if it doesn't find anything, use a livecd, mount your root in /mnt and your /boot in /mnt/boot, bind-mount /mnt/sys, /mnt/dev, /mnt/proc, and run bootctl --install in a /mnt chroot
<pcarrier> awi: first try to go into your EFI interface and create a new boot entry
<awi> @pcarrier: Black-Screen until Nixos starts up. I'm missing the usual "Thinkpad Startup Screen" as well.
<pcarrier> oh it goes straight to the first entry? where's your nixos config? is there grub and/or systemd-boot?
<pcarrier> you can check the timeout in their config, maybe? if it's 0 it might jump to the first entry without delays?
<pcarrier> awi: sorry I didn't read "boots normally" :D
<awi> boot.loader.systemd-boot.enable = true; boot.loader.efi.canTouchEfiVariables = false; boot.loader.timeout = 8;
dmi3y has joined #nixos
<pcarrier> and that's reflected in /boot/loader/loader.conf ?
<awi> but yes, it goes to the first entry - if I do not press any keys during startup.
<pcarrier> it should have something like "timeout 8\ndefault nixos-generation-XX'
<pcarrier> you're saying you're not getting the boot menu anymore though. can you be more specific?
<pcarrier> it still waits for 8 seconds but doesn't display anything?
<awi> yes, actually timeout and default generation is in.
<awi> yes, does not display anything.
<pcarrier> eg if during those 8 seconds you press "<up><up><enter>", it boots into a previous generation?
<awi> hmm, just removed any old generation...
<pcarrier> could it be that you updated your EFI and the graphical primitives are broken here? did you start plugging a different display?
<pcarrier> sorry, trying to go for the obvious first.
<pcarrier> it's often something stupid for me.
<pcarrier> but it kinda sounds like your EFI doesn't know how to output to your display anymore which seems pretty far removed from a nix problem
<pcarrier> I could, of course, be completely wrong.
<awi> no, I have not done any EFI update. Have not tried with a different display. Have the impression that display is powered but black.
<pcarrier> and if you type the traditional key to get into the EFI menu during the boot sequence nothing happens?
<pcarrier> if so maybe it's time to talk to your hardware vendor
<pcarrier> or browse /sys/firmware/efi/efivars? :)
<pcarrier> you can start with `efibootmgr` which can tell you _some_ very limited stuff
<pcarrier> then upgrade to `efivar -l`
<pcarrier> some EFIs come with a lot of options that could theoretically go wrong
<pcarrier> not sure how you'd have gotten into this situation in the first place if your hardware isn't faulty though.
<sphalerite> that doesn't sound like something that the nixos installer can cause
<pcarrier> nope
<awi> hmm, the single thing I once did was a "ls | grep -v <current-generation-name> | xargs rm" in /boot/loader/entries . Besides of that I always used Nixos tools to manage system.
asuryawanshi has joined #nixos
<pcarrier> awi: so there's no reason to think the entries are there, this is a EFI/hardware problem best I can tell
<pcarrier> I meant no reason to think the entries aren't there or causing any problem, sorry
cmacrae has joined #nixos
<awi> ok, I'll try to "systemctl reboot --firmware-setup", hope I'll get there and try upgrading efi/bios.
<cmacrae> o/ Yo! Is there a means of ensuring a Gnome3 session (launched from GDM) runs in Wayland instead of X11? I've set `services.xserver.displayManager.gdm.wayland = true;` - but it seems I'm still being launched into an X11 session
awi has quit [(Remote host closed the connection)]
patrl has joined #nixos
asuryawanshi has quit [(Ping timeout: 248 seconds)]
patrl has quit [(Client Quit)]
mrkgnao has quit [(Ping timeout: 255 seconds)]
dmi3y has quit [(Quit: dmi3y)]
dmi3y has joined #nixos
ris has joined #nixos
cmacrae has quit [(Remote host closed the connection)]
Maxdamantus has quit [(Ping timeout: 255 seconds)]
dywedir has joined #nixos
Maxdamantus has joined #nixos
ixxie has quit [(Ping timeout: 246 seconds)]
asuryawanshi has joined #nixos
ma27 has quit [(Ping timeout: 250 seconds)]
mahalel_ has joined #nixos
asuryawanshi has quit [(Ping timeout: 240 seconds)]
mrkgnao has joined #nixos
justan0theruser has quit [(Ping timeout: 255 seconds)]
dmi3y has quit [(Quit: dmi3y)]
LexLeoGryfon has joined #nixos
<pcarrier> does anybody have an example of introducing nixos modules through overlays?
<pcarrier> I want to derive a tool to switch between display configurations from my system configuration
<adisbladis> pcarrier: Sadly overlays cant provide modules
<pcarrier> :(
FRidh has joined #nixos
ambro718 has joined #nixos
<makefu> pcarrier: but what you can do is simply import module descriptions
<avn> pcarrier: custom modules in repo with configs works well
Myrl-saki has quit [(Read error: Connection reset by peer)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #32071: libu2f-host: 1.1.1 -> 1.1.4 (master...libu2f-host-1_1_4) https://git.io/vFj28
NixOS_GitHub has left #nixos []
<pcarrier> nice, thanks makefu
<pcarrier> maybe slightly crazier question: I'd like a derivation that writes a shell script that calls nix-build on a derivation
<pcarrier> how can I make the derivation available at runtime for the script?
euniarte has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vFjwk
<NixOS_GitHub> nixpkgs/master 167e9fb adisbladis: xonsh: 0.5.12 -> 0.6.0
<NixOS_GitHub> nixpkgs/master cea1246 Jörg Thalheim: Merge pull request #32067 from adisbladis/xonsh-0_6_0...
NixOS_GitHub has left #nixos []
<Isorkin> ow to configure zsh plugin - antigen - http://pastebin.ru/Y6RmN6fx - not normal work config. Error - -antigen-cache-generate:zcompile:54: can't write zwc file: /nix/store/...-antigen-2.2.1/share/antigen/antigen.zsh.zwc
justan0theruser has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] MikePlayle opened pull request #32073: libamqpcpp: init at 2.7.4 (master...add-libamqpcpp) https://git.io/vFjwl
NixOS_GitHub has left #nixos []
ma27 has joined #nixos
<FRidh> pcarrier: are you referring to "import from derivation"? Or do you explicitly want to call `nix-build` during a build?
<pcarrier> I want to call `nix-build` at runtime in the script
<pcarrier> it's meant to have side-effects
<pcarrier> maybe I'm doing this wrong? :)
<pcarrier> I want it to have access to the closures I want to backup, and only those, to run nix-push to an sshfs mount
<pcarrier> I also want it to have a secret key without having a create a dedicated user and provision a home dir & co
<pcarrier> seems like a decent approach, not sure if it makes sense?
tgunb has joined #nixos
<pcarrier> hmmm it's a bad idea to include it in the closure I'm backing up.
<FRidh> in the shell script you could include ${nix}/bin/nix-build . For the Nix expression, you could copy the whole expression into the store, say next to this shell, or you refer to a path outside of the store. After you nix-build the script, you could run the script and it would run nix-build
<pcarrier> > For the Nix expression, you could copy the whole expression into the store, say next to this shell
<pcarrier> yes, how? :)
<FRidh> you can include a file or a folder with ./file or ./folder
<pcarrier> but how do I get the expression accessible as a path?
<FRidh> the variable is a path
<pcarrier> oh right. jesus.
<pcarrier> I keep forgetting how amazingly all this works.
<pcarrier> just gotta accept everything is always easy.
petsa has quit [(Ping timeout: 248 seconds)]
fendor has joined #nixos
euniarte has quit [(Remote host closed the connection)]
<hyper_ch> FRidh: got some time?
<gchristensen> samueldr: they weren't stuck, but it is cool so many are passing all at oncce
fendor has quit [(Ping timeout: 240 seconds)]
euniarte has joined #nixos
petsa has joined #nixos
awi has joined #nixos
asymmetric1 has joined #nixos
<awi> @pcarrier: Thanks for your help. I managed to get the menu back again, so I could change boot order and upgrade firmware.
<awi> The "Enter Setup" screen appeared when a I booted, after a failed resume from suspend to disk.
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] Izorkin closed pull request #28544: Add ModSecurity Connector (master...master) https://git.io/v5q9u
NixOS_GitHub has left #nixos []
awi has quit [(Quit: Leaving)]
rihards has joined #nixos
<Isorkin> How to fix PR https://github.com/NixOS/nixpkgs/pull/28544 Need change branch Izorkin/nixpkgs:master to Izorkin/nixpkgs:master Izorkin/nixpkgs:ModSecurity-beta-Nginx-Connector ?
<gchristensen> Isorkin: like this? https://github.com/NixOS/nixpkgs/pull/28545
<Isorkin> yes
<gchristensen> that is fine, just keep the new PR
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] grahamc pushed 1 new commit to unstable-aarch64: https://git.io/vFjof
<NixOS_GitHub> nixpkgs/unstable-aarch64 0ea9e57 Graham Christensen: Patch over aarchc64 support for now
NixOS_GitHub has left #nixos []
<hyper_ch> adisbladis: so, I think I have addressed all issues/remarks by you and FRidh, how to update the pr?
sibi_ has joined #nixos
b has joined #nixos
nuncanada has joined #nixos
truh has joined #nixos
b has quit [(Quit: Lost terminal)]
<viric> how can I get with nix-shell an environment for an attrpath of nixpkgs AND autoconf, automake, ...?
euniarte has quit [(Remote host closed the connection)]
euniarte has joined #nixos
<adisbladis> hyper_ch: The PR is updated if you push to the source branch. So it's already updated.
<hyper_ch> adisbladis: ok :)
<adisbladis> I would still like it split up into one commit per package
<hyper_ch> it's way too late for that
Wizek_ has joined #nixos
ylwghst has joined #nixos
erictapen has quit [(Ping timeout: 240 seconds)]
erictapen has joined #nixos
<Harpalus> Hi! I got a question regarding updating to 17.09. I'm on 17.03 and in 17.09 guile-xcb is marked as broken. When I try upgrading by setting allowBroken=true, the build process fails (as expected). Is there anything I can do to upgrade? I'd appreciate any hint. Thanks in advance!
civodul has joined #nixos
<gchristensen> looking, Harpalus
iyzsong has quit [(Ping timeout: 240 seconds)]
<hyper_ch> samueldr: awake?
ylwghst has quit [(Quit: Lost terminal)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] Izorkin opened pull request #32075: nginx: add ModSecurity Nginx Connector (master...ModSecurity-beta) https://git.io/vFj6S
NixOS_GitHub has left #nixos []
Maxdamantus has quit [(Ping timeout: 258 seconds)]
<Isorkin> gchristensen: that PR https://github.com/NixOS/nixpkgs/pull/32075/files normal?
Maxdamantus has joined #nixos
<gchristensen> Harpalus: I think it should be fixable, can you open a github issue about it, note in the issue it builds fine on unstable, and tag @vyp?
<adisbladis> hyper_ch: I fixed up your commits
ylwghst has joined #nixos
<hyper_ch> adisbladis: not sure what you did
thblt has joined #nixos
<adisbladis> hyper_ch: Look at the commit history of your PR now https://github.com/NixOS/nixpkgs/pull/32053/commits
asuryawanshi has joined #nixos
<hyper_ch> but wouldn't changes be needed to my in my fork?
<hyper_ch> s/to my/to bed made/
<hyper_ch> s/bed/be/
<adisbladis> hyper_ch: I did
<hyper_ch> why do you have write access to my fork?
dan_b has joined #nixos
astronav_ has joined #nixos
<Harpalus> gchristensen: I'll open an issue; hadn't so far because I couldn't imagine being the only one depending on guile-xcb. (although, tbh, I have no clue what package is pulling that in…)
<gchristensen> Harpalus: the list of "failed to build" should reveal it
asuryawanshi has quit [(Ping timeout: 240 seconds)]
<adisbladis> hyper_ch: Check the "Allow edits from maintainers" checkbox on your pr
<Harpalus> gchristensen: is that part of the output of nixos-rebuild? if so, I'll retry, but iirc it just claims guile-xcb to have failed
<hyper_ch> adisbladis: ah... well, git... it's complicated
<gchristensen> Harpalus: yeah, retry, it should say ...
erictapen has quit [(Ping timeout: 248 seconds)]
astronav_ is now known as astronavt_
<hyper_ch> adisbladis: thanks for the work
mahalel_ has quit [(Ping timeout: 255 seconds)]
dan_b has quit [(Ping timeout: 268 seconds)]
ylwghst has quit [(Ping timeout: 248 seconds)]
erictapen has joined #nixos
<pcarrier> @makefu how do you generate/maintain README.md?
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] sjau opened pull request #32076: Adding myself (master...patch-2) https://git.io/vFjPf
NixOS_GitHub has left #nixos []
<makefu> pcarrier: you mean of nixos/nixpkgs? is it generated? according to the history it seems like it is a shared effort
<pcarrier> @makefu no https://github.com/musnix/musnix
<pcarrier> it looks formatted a lot like `nixos-config` does
<makefu> ah, i just linked to it, i am not the owner of the project :D
<pcarrier> oh sorry :P
<makefu> but again this looks like a manual effort
<pcarrier> yeah wow
<hyper_ch> adisbladis: I'm giving up... I don't wanna go through 100s of hoops
erictapen has quit [(Ping timeout: 240 seconds)]
<adisbladis> hyper_ch: I can help you if you want to
<hyper_ch> no need, it works for me
<adisbladis> Its just one tiny thing left :>
<hyper_ch> that's ok
<Isorkin> How to configure zsh plugin - antigen - http://pastebin.ru/Y6RmN6fx - not normal work config. Error - -antigen-cache-generate:zcompile:54: can't write zwc file: /nix/store/...-antigen-2.2.1/share/antigen/antigen.zsh.zwc
rihards has quit [(Quit: rihards)]
acarrico has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] sjau closed pull request #32053: OCRmyPDF: v.5.4.3 (master...master) https://git.io/vFjO8
NixOS_GitHub has left #nixos []
bennofs has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] sjau closed pull request #32076: Adding myself (master...patch-2) https://git.io/vFjPf
NixOS_GitHub has left #nixos []
petsa has quit [(Ping timeout: 248 seconds)]
ShalokShalom has joined #nixos
orivej has joined #nixos
<mudri> Isorkin: I'll check my config. If I remember correctly, I don't use Nix for it.
phreedom has quit [(Quit: No Ping reply in 180 seconds.)]
greymalkin has quit [(Ping timeout: 240 seconds)]
phreedom has joined #nixos
<mudri> Isorkin: Yeah, I just have the antigen repository cloned and sourced from my ~/.zshrc.
<mudri> I assume I just followed their instructions.
<mudri> Not saying that's the best way, but at least it works for me.
astronavt_ is now known as astronavt
Acou_Bass has quit [(Ping timeout: 248 seconds)]
bbsl has joined #nixos
dmi3y has joined #nixos
<Isorkin> mudri: from the home folder so it works too, not work with /etc/zshrc (
bbsl has left #nixos []
<pcarrier> can I iterate over a list of files in a nix derivation?
ris has quit [(Ping timeout: 264 seconds)]
goibhniu has joined #nixos
nuncanada has quit [(Ping timeout: 258 seconds)]
tgunb has quit [(Ping timeout: 268 seconds)]
rogue_koder has joined #nixos
<pcarrier> oh builtins.readDir :)
<chris|> quick question: if I deploy a machine using nixops that requires compiling from source on a target arch that is different from the client arch, will that be a problem?
<gchristensen> yes, you'll need to setup remote builders
<Wizek> What can cause this error? `$ nix-env -i terminator # error: syntax error, unexpected $undefined, expecting IND_STR or DOLLAR_CURLY or IND_STRING_CLOSE, at /nix/store/v8qp0cmicpbzradiin7g9pamypc29d9q-nixpkgs-18.03pre121253.2f1a818d00/nixpkgs/pkgs/applications/misc/hubstaff/default.nix:30:77`
goibhniu has quit [(Ping timeout: 250 seconds)]
<chris|> gchristensen: is there any documentation on how to do that? Can't find a reference in the manual
ryantm has joined #nixos
<gchristensen> Wizek: what version of nix?
<Wizek> $ nix-env --version # nix-env (Nix) 1.11.2
<gchristensen> ahh you should probably update
<Wizek> gchristensen: ^
<Wizek> okay, looking into how
<gchristensen> chris|: can you run nix-info?
<chris|> gchristensen: system: "x86_64-linux", multi-user?: yes, version: nix-env (Nix) 1.11.15, channels(root): "nixos-17.09.2182.7f6f0c49f0", channels(chris): "", nixpkgs: /nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs
fragamus has joined #nixos
<gchristensen> cool, nixos then?
tgunb has joined #nixos
<Wizek> gchristensen: Google doesn't want to tell me how I can upgrade nix. Do you know how?
<gchristensen> Wizek: what are you running nix on?
<vaibhavsagar> my usual approach of pinging periodically doesn't seem to be working
<Wizek> gchristensen: ubuntu based distro
<vaibhavsagar> Wizek: nix-env -u nix?
<Wizek> vaibhavsagar: giving that a try
<gchristensen> ^ that should do, I've gotta go :)
<Wizek> gchristensen: okay, thanks
<gchristensen> chris|: note sshUser must be a trusted-user according to nix. most people use root. gotta go :) good luck
<Wizek> vaibhavsagar: btw, I have a long-running nix-build in the background, do you think it is safe for me to upgrade nix in parallel?
<Wizek> vaibhavsagar: it should be because of the hashes, no?
bitchecker has quit [(Quit: bye!)]
<chris|> gchristensen: thanks :)
Maxdaman1us has joined #nixos
Maxdamantus has quit [(Ping timeout: 240 seconds)]
bitchecker has joined #nixos
arximboldi has joined #nixos
ylwghst has joined #nixos
jmc_fr has joined #nixos
<vaibhavsagar> Wizek: I would expect it to work
asuryawanshi has joined #nixos
<Wizek> vaibhavsagar: huh, I gave it a try, and get the same error :/
<Wizek> error: syntax error, unexpected $undefined, expecting IND_STR or DOLLAR_CURLY or IND_STRING_CLOSE, at /nix/store/v8qp0cmicpbzradiin7g9pamypc29d9q-nixpkgs-18.03pre121253.2f1a818d00/nixpkgs/pkgs/applications/misc/hubstaff/default.nix:30:77
rihards has joined #nixos
bitchecker has quit [(Ping timeout: 248 seconds)]
bitchecker has joined #nixos
<Lisanna> Man, state sure does suck
ylwghst has quit [(Quit: Lost terminal)]
<vaibhavsagar> wtf
<Lisanna> Lots of nixos services have all sorts of state which effectively makes it not reproducible at all ):
<vaibhavsagar> Lisanna: what are you talking about?
<infinisil> Lisanna: Agreed
<vaibhavsagar> and what do you mean by reproducible? I think your definition might be stricter than mine
<Lisanna> vaibhavsagar: stuff like databases and "private" data like key / password files
<vaibhavsagar> so the goal is to be able to put the same lines in configuration.nix on a different machine and have the state be bit-for-bit identical?
<Lisanna> you can't even get that on the same machine :)
<vaibhavsagar> put everything in a git repository
jb55 has joined #nixos
<Lisanna> <_<
<vaibhavsagar> ;)
<vaibhavsagar> but seriously, git is pretty good in terms of reproducibility
<Lisanna> edit configuration.nix to have text A --> result A, edit it again to have text B --> result B, revert it back to text A --> result C
<vaibhavsagar> I wish we used it more as a database
<vaibhavsagar> now that is strange
<pcarrier> so I'm running into "/usr/bin/env: bad interpreter: No such file or directory"
<pcarrier> and it kinda blows my mind. details @ https://gist.github.com/pcarrier/249f7e2e54770188f22316e6a08a2d89
<pcarrier> I'm been at this for a good 30 minutes.
<Lisanna> vaibhavsagar: not really. running services causes the filesystem to become altered, and that doesn't get undone because you might lose data
<vaibhavsagar> but how does the altered filesystem become part of the derivation?
Harpalus has quit [(Quit: Leaving)]
<Lisanna> vaibhavsagar: it doesn't. the services still use the filesystem though. the filesystem changes, so the results of running a service can change.
astronavt has quit [(Remote host closed the connection)]
<vaibhavsagar> ah, I thought you meant the actual result symlink points to a different place
<Lisanna> ...no
<vaibhavsagar> that would be extremely odd
<pcarrier> and "/usr/bin/env bash" works great in "nix-shell /nix/store/2pq02823q72hmjmhkq46cxlkja831h0w-monorepo-env-dep.sh.drv"
Harpalus has joined #nixos
<sphalerite> aaah I love zfs. My backups are so much faster now
<pcarrier> sphalerite: yeah it's really slick. what do you use to automate your backups?
<sphalerite> nothing yet. Still so much better than before, when I had a script which used borgbackup
<vaibhavsagar> pcarrier: just to be absolutely sure, what is the output of `ls -l /usr/bin/env`?
<FRidh> pcarrier: bash is unavailable during build; add bashInteractive
<pcarrier> vaibhavsagar: if I add it after PATH: $PATH, I see /nix/store/izkijl7w7qkjdfr09qk43v1n3bnbdnvm-stdenv/setup
<pcarrier> sorry I meant "ls: cannot access '/usr/bin/env': No such file or directory"
<FRidh> to buildInputs, or patch the source to hardcode a store path
<pcarrier> well /usr/bin/env doesn't seem to be in the build env!?
pie__ has joined #nixos
bitchecker has quit [(Excess Flood)]
ma27 has quit [(Ping timeout: 250 seconds)]
bitchecker has joined #nixos
<pcarrier> what the.
ylwghst has joined #nixos
ma27 has joined #nixos
bitchecker has quit [(Remote host closed the connection)]
<Wizek> vaibhavsagar: any ideas if even nix-env -u nix fails with that error?
jb55 has quit [(Ping timeout: 240 seconds)]
<pcarrier> mmmkay thanks FRidh
bitchecker has joined #nixos
<pcarrier> so I need to first copy the script then patch its shebang
<pcarrier> slightly annoyed here.
<vaibhavsagar> Wizek: try doing a fresh install of Nix: `curl https://nixos.org/nix/install | sh`
<ylwghst> How could I change cursor theme within display-manager eg. lightdm?
cmacrae has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] peterhoeg pushed 3 new commits to master: https://git.io/vFjys
<NixOS_GitHub> nixpkgs/master 50ad43b Peter Hoeg: libqmatrixclient: init at 0.1
<NixOS_GitHub> nixpkgs/master c5717d0 Peter Hoeg: tensor: documentation regd libqmatrixclient
<NixOS_GitHub> nixpkgs/master 388c732 Peter Hoeg: quaternion: git -> 0.0.4
NixOS_GitHub has left #nixos []
jb55 has joined #nixos
<cmacrae> Should I expect to get a Wayland session in Gnome3 by using `displayManager.gdm.wayland = true;` ?
<FRidh> pcarrier: that reminds me. It would be useful to have a nix function for patchShebangs that simply copies the derivation, but patches its shebangs.
<pcarrier> FRidh: agreed.
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] MP2E pushed 1 new commit to master: https://git.io/vFjyR
<NixOS_GitHub> nixpkgs/master 2e29a1e Cray Elliott: wineUnstable: 2.19 -> 2.21
NixOS_GitHub has left #nixos []
<sphalerite> pcarrier: you could also explicitly call the interpreter on it, depending on the context
<ylwghst> cmacrae: nope, this only runs GDM as wayland
<cmacrae> ylwghst: Ah! Okay, thanks
<pcarrier> oh right those derivations need internet access
ma27 has quit [(Quit: WeeChat 1.9.1)]
<neonfuz> does anyone here run with their own git checkout of nixpkgs?
<cmacrae> Looking through GH issues, I guess Wayland support for Gnome is still being worked on
<pcarrier> I realize that's wrong on paper, but in this context it's really appropriate
<neonfuz> and can I do that with nixos? (like, to update the system packages of nixos too)
<pcarrier> my goal is to not have to rewrite that stuff. can I somehow let it resolve?
<vaibhavsagar> neonfuz: we do essentially this at work
<neonfuz> nice
<Wizek> vaibhavsagar: giving that a try
<ylwghst> cmacrae: you cantry run gnome with gnome-shell --wayland but it propably doesn't work on nixos yet
<cmacrae> ylwghst: Alright, thanks!
<Wizek> vaibhavsagar: seems to have worked, thanks!
ylwghst has quit [(Quit: Lost terminal)]
cmacrae has quit [(Remote host closed the connection)]
erictapen has joined #nixos
<pcarrier> it can't possibly be that weird to want my builds to access a hostname ?
<sphalerite> pcarrier: you can make it a fixed-output derivation
astronavt has joined #nixos
<sphalerite> in which case you need to specify the hash of the output
<pcarrier> sphalerite: and then it will have /etc/resolv.conf?!
<sphalerite> yeah. It's how fetchurl and friends are implemented
<pcarrier> what if I don't want to maintain those hashes? :)
<pcarrier> I'm not packaging stuff like a distro would here. I want to leverage nix to wrap around scripts that I'm confident have fixed outputs despite connecting to external services to grab resources, because those resources are immutable-as-in-I'm-willing-to-trust-that-they-are
ylwghst has joined #nixos
jb55 has quit [(Ping timeout: 240 seconds)]
<vaibhavsagar> Wizek: glad to hear it :)
justan0theruser has quit [(Ping timeout: 240 seconds)]
<FRidh> pcarrier: you may use __noChroot and build-use-sandbox "relaxed"
Acou_Bass has joined #nixos
arximboldi has quit [(Quit: arximboldi)]
<pcarrier> FRidh: I'd like to keep the chroot if at all possible
<pcarrier> FRidh: it'd be nice if it just worked where I run "nix-shell" in the repo
<pcarrier> FRidh: sorry if I'm asking for too much
<pcarrier> am I trying to break abstractions in a completely unacceptable ways here?
<sphalerite> you are trying to break purity, which is the guarantee that makes nix so wonderful
<sphalerite> it's not a use case that's well supported by nix
<pcarrier> https://github.com/NixOS/nix/issues/520#issuecomment-98072186 is the best summary of what this is to me
<sphalerite> there is stuff like builtins.fetchurl though
<pcarrier> yeah I mean my goal was to start encapsulating something that tries to do something a bit nix-ish
<pcarrier> without having to rewrite on top of nix, by abstracting it away
<pcarrier> but it doesn't look like I can "just trust it" as something I would only ever use in a nix-shell anyway
<pcarrier> where "side effects OMG" is kind of the point.
<pcarrier> (of how this environment is used)
<pcarrier> but it's not like it would be able to infect anybody else through the store right?
<pcarrier> I'm sort of unclear about how there isn't an escape trap I can take here
arximboldi has joined #nixos
<pcarrier> seems hardly possible at all unless I'm missing something
zraexy has joined #nixos
<FRidh> pcarrier: you could have a look at nixUnstable (1.12) https://github.com/NixOS/nix/commit/647291cd6c7559f68d49a5cdd907c2fd580790b1
euniarte has quit [(Remote host closed the connection)]
bitchecker has quit [(Remote host closed the connection)]
bitchecker has joined #nixos
JosW has joined #nixos
<hyper_ch> so, I think I can remove vbox from my setup.... ported the win81 machine to kvm, virtio works fine for disk and interface, auto-resizing the virt-viewer windows works also... need to test some usb stuff tomorrow (smart card reader and label printer)
<pcarrier> FRidh: ok so I'm not 100% sure I understand the restrictions so would need to play with it to get a good feel for what it can and can't do. so I wanted to use this now I'd have to switch to nix unstable. can I still benefit from the public binary cache?
<FRidh> pcarrier: yes
<pcarrier> interesting.
<FRidh> note hydra runs nixUnstable
<pcarrier> how come? is everything built multiple times, of is the nix major version not an input to the hashes?
<pcarrier> seems like the latter would create a lot of backward-compatibility concerns right?
<clever> the version of nix you build something with shouldnt impact its hash
<FRidh> Indeed, it does not depend on Nix version. All though in practice there may be different impurities (e.g references to /tmp vs /build that remained)
<pcarrier> OK interesting, thanks.
<pcarrier> well this is nixos, what's the risk anyway.
euniarte has joined #nixos
bennofs1 has joined #nixos
<sphalerite> probably the worst that could happen is a corrupted nix database. Which is quite a pain
<clever> 2nd worst would be a corrupt bootloader config
nwspk has quit [(Quit: Quit: *.banana *.split)]
andi- has joined #nixos
Baughn has quit [(Ping timeout: 240 seconds)]
<sphalerite> but who knows, it might lead to the world ending. The butterfly effect's a crazy thing.
bennofs has quit [(Ping timeout: 240 seconds)]
Harpalus has quit [(Remote host closed the connection)]
Harpalus has joined #nixos
taktoa has joined #nixos
<infinisil> sphalerite: if we used zfs we could just rollback [ the world ]
jmc_fr` has joined #nixos
<sphalerite> :)
Baughn has joined #nixos
jmc_fr has quit [(Read error: Connection reset by peer)]
<clever> ah, snapshots on /nix would preserve nix.db, and keep it in-sync with the rest of the store
<pcarrier> sphalerite: yeah so I have lots of zfs snapshots
<clever> but i always turn snapshotting off there, because then garbage collect doesnt work
<sphalerite> :D
<pcarrier> sphalerite: and a liveUSB with zfs support
<clever> pcarrier: or my rescue boot config
<clever> pcarrier: if you add this to your imports, it will put a ~300mb live environment into /boot, and in the grub menu
<clever> so as long as /boot survives, you can boot a full nixos up
<pcarrier> clever: OK this is really cool and I now truly regret how small my /boot is
<clever> the only potential problem i can see, is that the rescue env is always rebuild from the same nixpkgs as the host
<pcarrier> clever: unfortunately it is my understanding that it won't move to zfs
<clever> so if there is some fatal kernel problem in the linux nixpkgs provides, both die
<clever> yeah, grub has some trouble reading zfs
<pcarrier> clever: can't you access the nix generation and use that to create a new env everytime? :)
nwspk has joined #nixos
argc_ has joined #nixos
<clever> yeah, but then it will cost 300mb of /boot for every generation
<pcarrier> clever: next time I can be bothered I'm making a 5GB /boot on my laptop.
<clever> i usually make it 512mb or 1gig
jmc_fr`` has joined #nixos
<MichaelRaskin> clever: caring is for losers. I am moving to the idea that a 200MiB compressed initramfs that requires 2GiB to boot is actually OK
<argc_> is anyone successfully using sway (nixos 17.09, sway 0.15 from unstable)? Everything installs fine and I can run it from a tty, but it doesn't seem to register keypresses of any kind (although mouse movement works). Anyone seen similar behavior?
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<pcarrier> MichaelRaskin: why not use squashfs instead?
<pcarrier> do you really want your full system in RAM?
<MichaelRaskin> No, it is actually initramfs.
jmc_fr` has quit [(Read error: Connection reset by peer)]
<pcarrier> so loaded into RAM by the bootloader right?
<MichaelRaskin> I just cannot decide if bothering with static builds and musl and busybox is worth it.
<MichaelRaskin> And why bother choosing what firmware to put where.
ylwghst has quit [(Quit: Lost terminal)]
<MichaelRaskin> Why not just let eudev and LVM2 etc. do their job.
<sphalerite> argc_: worked for me last time I tried, a couple weeks ago or so
<pcarrier> I'm still confused. you want to dedicate 2GB of RAM to an initramfs decompressed by your loader at boot time and use that system for?
<MichaelRaskin> Actually, I could afford that.
<MichaelRaskin> But the point is, the initramfs will be eventually cleaned up anyway.
<pcarrier> yeah sorry I'm looking for the benefits here.
<pcarrier> you want to PXE-boot everything maybe?
<MichaelRaskin> I do not use mainline NixOS, because systemd makes some minor features impossible.
<clever> pcarrier: a full PXE netboot server in a nixos module
<clever> complete with NAT so a laptop could serve netboot over ethernet, and share its wifi
<sphalerite> MichaelRaskin: what are they?
<argc_> sphalerite: did you do anything other than launch "sway" from a tty after placing a default config file in ~/.config/sway/?
fragamus has joined #nixos
<MichaelRaskin> Basically, anything that requires vt management by non-PID1 processes (even UID0)
<sphalerite> argc_: as far as I remember, no.
<argc_> sphalerite: okay, thanks, i'll keep poking at it. thanks. maybe i'll try a newer unstable or something.
<pcarrier> clever: what does it serve how?
<argc_> (right now i'm pinned exactly to the ref that upgraded sway to 0.15 and overlay it atop my 17.09 config)
<pcarrier> I guess I'm asking what's in netbootIpxeScript I guess
Neo-- has joined #nixos
<sphalerite> Hah, I have different welcome messages on tty1, 2 and 3
<sphalerite> Each from a different system generation
<clever> pcarrier: it shares the internet access of the wifi to the wired users
<clever> pcarrier: so you can plug a machine into the laptop's ethernet port, and get both internet and netboot
<pcarrier> ok so that's the cool part, not what it serves over PXE being driven by nix or something?
<clever> the cool part is just being able to configure all of that with a single file
<clever> most distro's dont make it so simple
<pcarrier> oh yes nix is great.
<manveru> anyone here using elm-mode?
<hyper_ch> (or you can split up the config in as many parts as you want for specific things... all up to you)
<manveru> just made a little PR so you can use it with nix-sandbox: https://github.com/jcollard/elm-mode/pull/130
<pcarrier> so how mergeable would an sftp binary cache sound?
argc_ has quit [(Quit: Leaving)]
<pcarrier> my intuition is that it's a great idea :P
<hyper_ch> what would that be good for?
erictapen has quit [(Ping timeout: 240 seconds)]
<pcarrier> being able to read and write binary stores over SSH without the host(s) supporting nix?
<pcarrier> so for example being able to use rsync.net as a writeable binary cache.
<hyper_ch> well, I fail to see what it's good for but then, I'm just a simple user
<pcarrier> so a binary cache is a great way to deduplicate work across your fleet
<hyper_ch> (I don't have a fleet)
<pcarrier> OK so if you wanted to garbage collect your machine's nix store aggressively
<pcarrier> but wanted to avoid ever rebuilding
<pcarrier> you could set up an s3 bucket and use that as a writeable binary cache
<pcarrier> and s3 storage scales much better than laptop storage.
<hyper_ch> well, I don't see a need for that... and I wouldn't use amazon anyway
<pcarrier> exactly. you would use your own server
<hyper_ch> but if you think that would be useful for you, go ahead
<pcarrier> and you'd want your nix systems to be able to push to it and pull from it
<pcarrier> ssh seems nice to get authentication and co
<clever> nix supports S3 as a store
<pcarrier> but then you need to set a contract on your ssh binary cache: it has to have nix-daemon or nix-store (ssh:// vs ssh-ng://)
<pcarrier> whereas I'd like to be able to use whatever box I have around with storage, including eg rsync.net or insert-random-NAS-manufacturer
bennofs has joined #nixos
<pcarrier> clever: yeah I know that's why I used it as an example :)
<pcarrier> clever: unfortunately right now it seems to be either S3 or nix-on-the-other-side for all the writeable binary cache stores available
<pcarrier> clever: another interesting one could be http(s) with POST for writes
<pcarrier> clever: adding that as an option in the query parms.
<FRidh> pcarrier: nix copy -f . hello --to file:///path/to/cache also supports --to ssh://server
<pcarrier> FRidh: yeah, and that will spawn a nix helper on the server
<pcarrier> right?
<pcarrier> FRidh: btw what's "nix copy" sorry? do I not know of a utility call nix?
<FRidh> pcarrier: that's nix 1.12
<FRidh> new cli
<pcarrier> oh LOL
<pcarrier> OMG OMG OMG
bbsl has joined #nixos
arximboldi has quit [(Quit: arximboldi)]
thblt has quit [(Ping timeout: 240 seconds)]
<pcarrier> FRidh: yeah so that's still using nix on the remote right?
bbsl has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vFj5I
<NixOS_GitHub> nixpkgs/master 3fc7f92 Tim Steinbach: linux-copperhead: 4.14.1.a -> 4.14.2.a
NixOS_GitHub has left #nixos []
devrando1guy has joined #nixos
<FRidh> guess so
<pcarrier> well or I should be looking at another branch than master
dmi3y has quit [(Quit: dmi3y)]
Mateon2 has joined #nixos
Mateon1 has quit [(Ping timeout: 258 seconds)]
Mateon2 is now known as Mateon1
<FRidh> I suppose you could first push it to another folder locally, just to create the cache/nar files, and then use rsync or so to put it elsewhere
fendor has joined #nixos
arximboldi has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] c0bw3b pushed 4 new commits to master: https://git.io/vFj5S
<NixOS_GitHub> nixpkgs/master 9083624 Renaud: lib.licenses: add Open Software License...
<NixOS_GitHub> nixpkgs/master 11c92ba Renaud: libsmbios: 2.3.2 -> 2.3.3
<NixOS_GitHub> nixpkgs/master 203a59e c0bw3b: restya-board: license is OSL 3.0
NixOS_GitHub has left #nixos []
euniarte has quit [(Remote host closed the connection)]
alhariel has joined #nixos
alhariel has quit [(Max SendQ exceeded)]
alhariel has joined #nixos
<pcarrier> FRidh: yeah so that was my original idea, either using rsync or even sshfs
<pcarrier> FRidh: but quite frankly neither feels too right
Sonarpulse has quit [(Ping timeout: 255 seconds)]
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<sphalerite> pcarrier: yeah I use sshfs for that personally but it's really not ideal
<pcarrier> I still run into "curl: (6) Could not resolve host: github.com" with __impure = true; in system: "x86_64-linux", multi-user?: yes, version: nix-env (Nix) 1.12pre5732_fd10f6f2, channels(pcarrier): "", channels(root): "nixos-18.03pre120728.9ce33dd8fb", nixpkgs: /home/repos/nixpkgs
astronavt has quit [(Remote host closed the connection)]
<clever> pcarrier: are you defining an output hash and a output hash algo?
<pcarrier> sphalerite: right? wouldn't you like a sftp solution? curl supports sftp and nix already uses curl.
<pcarrier> clever: well OK so we're back to not having my problem solved at all here :D
<sphalerite> pcarrier: well sshfs does use SFTP AFAIU
<pcarrier> sphalerite: but like, you wouldn't have to use fuse and mountpoints anymore :)
<sphalerite> yeah that would be nice
phreedom has quit [(Quit: No Ping reply in 180 seconds.)]
endformationage has joined #nixos
phreedom has joined #nixos
<pcarrier> so OK, right now if I resolve by talking UDP to 8.8.8.8 and then pass the IP to curl, my recipe works
<pcarrier> the reason it doesn't is because I need curl to find something that works in /etc/resolv.conf right?
<pcarrier> ls
orivej has quit [(Ping timeout: 255 seconds)]
<clever> pcarrier: can you gist your nix expression?
<Lisanna> In my top-level NixOS configuration.nix, I'd like to extend a NixOS service to add an option. Specifically, I want to extend services.dockerRegistry to add an option "enableGarbageCollection". Is there a complete example of doing something like this anywhere? I can only find incomplete examples.
<clever> pcarrier: thats not a fixed-output derivation, so nix wont setup networking
<pcarrier> yeah so what's the point of __impure sorry?
<clever> ive never used that flag
ryantm has quit [(Quit: Connection closed for inactivity)]
tgunb has quit [(Ping timeout: 248 seconds)]
justbeingglad has joined #nixos
justbeingglad has left #nixos []
<Lisanna> is it as simple as just doing config.services.dockerRegistry.enableGarbageCollection = mkOption {...}?
<pcarrier> yeah so I'm confused. I can have arbitrary inputs from the local filesystem very easily
<clever> pcarrier: foo = ./path/to/thing;
<pcarrier> but not from the network in any convenient way
<pcarrier> unless I know what the other side will say
<clever> pcarrier: at eval time, nix will hash that input, and if the input changes, it will trigger a rebuild
<pcarrier> right so every single evaluation makes absolutely sure that nothing changes, assuming the compilers we're running are deterministic
tgunb has joined #nixos
<pcarrier> OK so the thing is, in my mind those scripts are compiling a complex expression that does have its set of inputs, and is "deterministicish" in its output. it's just that the transformation needs network.
jmc_fr`` has quit [(Ping timeout: 248 seconds)]
<clever> what is the network doing to transform it?
<pcarrier> it's turning something that's maintained as "I want this HTTPS URL extracted this way and set those environment variables and [...]" into "the env has those files on it"
<clever> pcarrier: do you know what the https url is before the build begins?
<pcarrier> which in turn lets me expose the tooling dependencies of my monorepo into shareable-cache-that's-exposed-read-only
<pcarrier> I do. but I don't want to maintain a bunch of hashes
<pcarrier> I don't want to understand exactly the logic my script is using to picking outputs
<clever> then you might want a small util that pre-fetches everything and gets the hashes, then inputs those into nix
<pcarrier> yeah I mean you're offering a solution that kind-of involves rewriting the content of those scripts to be integrated in nix
<pcarrier> and then add maintenance overhead so people who want to update the version of X now have to run that tool before they commit
<pcarrier> or they will break stuff for nix users without knowing
<Lisanna> clever: I do this for my nix setup, I have a bash script which I can run that generates .nix files containing an attrset of hashes
<pcarrier> I was trying to leverage nix for my monorepo environment without completely rewriting.
devrando1guy has quit [(Quit: Lost terminal)]
<sphalerite> nix is just quite opinionated about that sort of stuff. You could always try to get your team to use nix ;)
dywedir has quit [(Remote host closed the connection)]
<Ralith> that's what I did
<Ralith> worked great!
<samueldr> the qemu-user arm trickery worked, compiled the kernel in ~4h instead of ~17h native on my armv7l board
<clever> samueldr: nice
<samueldr> a native (x86) build for a kernel takes around 25 minutes
<sphalerite> huh
<sphalerite> kernel built in 6.5h on my ARM chromebook
<sphalerite> WIth a single job (although it has 4 cores)
<samueldr> your arm chromebook hopefully is way faster than the I/O starved cheap single-board computer I used
<sphalerite> That is, the full configuration in nixpkgs
<samueldr> (I think that I/O killes the build speed)
<sphalerite> ah ok
<sphalerite> yeah that seems likely to me
ambro718 has quit [(Ping timeout: 268 seconds)]
fragamus has joined #nixos
<sphalerite> so I've been making a dump of the root device on the chromebook, copying to my main laptop, using `ssh root@green dd if=/dev/mmcblk2 bs=1M > chromebook-mmcblk2-dump`… Connection was interrupted halfway through, is there a good way to continue the transfer without having to start over?
<sphalerite> Something like rsync, but I'm not sure it works on block devices…
<Lisanna> sphalerite: doesn't dd have an offset parameter?
<sphalerite> Lisanna: I'm scared of trying that because I've managed to clobber perfectly good data with it so often
<clever> samueldr: i prefer ddrescue for things like that, but it still needs an FS
<Lisanna> sphalerite: it's called seek
<clever> sphalerite: ddrescue stores a log of things its done, and can resume
<samueldr> sphalerite: ^
<sphalerite> clever: can it go through SSH as well?
<samueldr> possibly a mix of skip, and outputting to a new file, then merging them... but verifying the integrity would be a must
<sphalerite> clever: I don't want a filesystem, I want just the contents of the block device
<sphalerite> especially for the partition table, but all the other stuff on it as well
<clever> sphalerite: maybe over sshfs it can work
<sphalerite> ah, screw it, I'll just try with the seek and skip stuff and if it fails I'll just do it from the beginning overnight
<Lisanna> sphalerite: to use seek in your expression, you'd need to use "of" instead of ">"
<sphalerite> Lisanna: I'd need to add another dd in fact.
<samueldr> wouldn't skip, and ">" to a second file be enough?
<sphalerite> `ssh root@green dd if=/dev/mmcblk2 bs=1M skip=8k | dd of=chromebook-mmcblk2 seek=8k conv=notrunc`
<sphalerite> I believe
<Lisanna> sphalerite: that looks right, not sure why you added the second dd though
<samueldr> then, when both files are complete, cat file1 file2 > complete_file
<sphalerite> Lisanna: because one of them runs remotely and the other remotely.
<Lisanna> oh, okay
<samueldr> sphalerite: see skip_bytes and seek_bytes in the man page, just in case
<Lisanna> the lack of quoting in that scares me though, you should add some :)
<sphalerite> Where?
<Lisanna> ssh root@green "dd if=/dev/mmcblk2 bs=1M skip=8k" | dd of=chromebook-mmcblk2 seek=8k conv=notrunc
<sphalerite> ooh right. Meh, there aren't any variables involved
<Lisanna> It's not variables so much as where does the SSH part of the command start and where does it end
<samueldr> bash won't have trouble with it, but the reader might
<Lisanna> I don't like trusting implicit parsing like that
<sphalerite> oooh right
<Lisanna> it's like using || and && in the same C expression without parens, it's bad practice
<samueldr> though, for having written the start of posix sh compliant parser, there isn't anything implicit in the initial command
<sphalerite> idk it seems fairly obvious to me
<Lisanna> samueldr: likewise, there's nothing implicit about the C expression "foo(a) && bar(a) || bazz(a)"
<samueldr> sure
<sphalerite> Lisanna: but if that's what your concern is, I think `{ ssh root@green dd if=/dev/mmcblk2 bs=1M skip=8k ; } | dd of=chromebook-mmcblk2 seek=8k conv=notrunc` is a better way to write it, no?
slack1256 has joined #nixos
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
arximboldi has quit [(Quit: arximboldi)]
<Ralith> Linus: I think the questionable bit is what ssh considers an argument to itself vs. to the command
<Lisanna> sphalerite: well, that's bad for a different reason :)
<samueldr> a subshell too could be usedm possibly
<samueldr> (ssh ...) | dd ...
<sphalerite> Ralith: yeah but in this case it's up to the shell, not up to SSH, how much SSH gets
mrkgnao has quit [(Ping timeout: 240 seconds)]
arximboldi has joined #nixos
<sphalerite> samueldr: but is there any good reason to do so?
<samueldr> no
fragamus has joined #nixos
<samueldr> but it's another way to represent the command to make it easier to comprehend to the reader
<samueldr> in all cases, the shell will decide what's given as arguments, when not quoted, the program will receive all arguments
<Lisanna> be careful with subshells and things like that, error handling starts breaking down
<samueldr> in the case of ssh, it will receive X number of arguments, that it all passes through the other side
<samueldr> only issue is that whitespace won't be preserved, if they were important (most times they aren't)
<sphalerite> samueldr: and into a shell on the other side.
<samueldr> yes, that's "the other side", which I was eliding
<Ralith> Linus: ssh gets it all but what does it consider to be part of the remote command vs. a flag to process locally?
rogue_koder has quit [(Remote host closed the connection)]
phreedom has quit [(Ping timeout: 248 seconds)]
<samueldr> with quotes, it will still be passed to the other side, but ssh will see only one argument, which will have all whitespace preserved
rogue_koder has joined #nixos
astronavt has joined #nixos
<sphalerite> Oooh right that too
<samueldr> (what happens on the other side is implementation dependent)
<samueldr> (ssh to a windows host would do something different!)
<samueldr> (but generally, it's shelled out)
tmaekawa has joined #nixos
<samueldr> if I understand the man page correctly
<samueldr> anything passed AFTER the hostname is a command
tmaekawa has quit [(Client Quit)]
<samueldr> I misunderstood it!
tmaekawa has joined #nixos
phreedom has joined #nixos
patrl has joined #nixos
patrl has quit [(Client Quit)]
orivej has joined #nixos
<schoppenhauer> hi.
<samueldr> hi!
<schoppenhauer> the boost package in nixos does not appear to have a pkgconfig definition. how is the preferred way to compile projects inside a shell that contains libboost?
<bennofs> schoppenhauer: doesn't boost have its own config program? something like boost-config? i think it used to be that way, not sure if it changed though...
arximboldi has quit [(Quit: arximboldi)]
tmaekawa has quit [(Quit: tmaekawa)]
ixxie has joined #nixos
bennofs1 has quit [(Ping timeout: 240 seconds)]
<schoppenhauer> bennofs: there are no programs called boost-something
jb55 has joined #nixos
mahalel_ has joined #nixos
justan0theruser has joined #nixos
pie__ has quit [(Remote host closed the connection)]
pie_ has joined #nixos
<sphalerite> there is a CMake module for it but idk if that's all there is
pie__ has joined #nixos
pie__ has quit [(Remote host closed the connection)]
Alchemical has joined #nixos
rogue_koder has quit [(Ping timeout: 240 seconds)]
rogue_koder has joined #nixos
ssmike has joined #nixos
erictapen has joined #nixos
srdqty has joined #nixos
<johnw> has anyone written a FUSE filesystem to represent the attribute tree of <nixpkgs>?
<johnw> i'd like to browse attributes using a file browser
arximboldi has joined #nixos
ssmike has quit [(Remote host closed the connection)]
erictapen has quit [(Ping timeout: 276 seconds)]
ssmike has joined #nixos
Mic92 has quit [(Quit: WeeChat 1.9.1)]
<neonfuz> anyone use uim on nixos?
ssmike has quit [(Client Quit)]
ssmike has joined #nixos
ylwghst-nix has joined #nixos
<ylwghst-nix> What's up :-)
<neonfuz> johnw: that'd be neat
<neonfuz> or 9p
<LnL> that's an interesting idea
<johnw> ok, if no one has done it, it shouldn't be hard to write in Haskell fairly quickly, based on hnix
<jeaye> dooo it!
<neonfuz> (9p would be neat)
<johnw> it would let me find packages using... "find"
<johnw> or find all attributes containing a value
<johnw> since the contents would all be post evaluation, not pre
<ylwghst-nix> I get this while nixos-install: changing mode of /nix/store/*-update-users-groups.pl to 10444: Read-only file-system
<johnw> neonfuz: what is 9p?
mistnim has joined #nixos
<neonfuz> it's what plan9 used to mount filesystems. It's simple to implement and is network transparent. Nixos actually uses it for some things, like serving the nix-store to guest VMs so I think the nixos kernel would have support by default.
<neonfuz> they probably have a haskell library for implementing 9p?
<neonfuz> and there's alos a 9p / fuse adapter
<neonfuz> s/os/so/
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] adisbladis pushed 2 new commits to master: https://git.io/vFjjp
<NixOS_GitHub> nixpkgs/master 5f5d8e6 Lancelot SIX: gnupg: 2.2.2 -> 2.2.3...
<NixOS_GitHub> nixpkgs/master 21d8f25 adisbladis: Merge pull request #31901 from lsix/update_gnupg_2_2_3...
NixOS_GitHub has left #nixos []
jb55 has quit [(Ping timeout: 240 seconds)]
ixxie has quit [(Quit: Lost terminal)]
Ross has quit [(Ping timeout: 260 seconds)]
Ross has joined #nixos
ixxie has joined #nixos
Arcaelyx_ has joined #nixos
<bennofs> johnw: what should be the contents of the files in that system? derivations?
<johnw> bennofs: for example, if I mount <nixpkgs>, then "ls" would show all the top-level attributes (8242 of them on my system right now)
<johnw> I could then "cd haskellPackages"
arximboldi has quit [(Ping timeout: 276 seconds)]
<johnw> and "ls" to see all the available haskellPackages
<johnw> I could: cat haskellPackages/lens/version
ylwghst-nix has quit [(Remote host closed the connection)]
<bennofs> ah so the contents should be the attributes of the derivation
<johnw> the main advantage here being that both the hierarchy, and the contents of its leaves, would become accessible to standard unix searching and navigation tools
Arcaelyx has quit [(Ping timeout: 250 seconds)]
<johnw> which the only reason I use nix-repl
<johnw> there's no better way (currently) to "discover" which haskell packages are actually available on my OS
<bennofs> you could even use nix-index to add a haskellPackages/lens/files (if the pkg is build by hydra) :D
<bennofs> (which would contain the file tree of the package if it was installed, but without contents)
<bennofs> but for such a file system to work smoothly, we probably need to optimize nix-instantiate quite a bit
<bennofs> listing all attributes in nixpkgs takes quite a while, even on a SSD
<johnw> that's where hnix would come in
<johnw> where we can easily do lazy parsing and evaluation
<johnw> and caching/memoization
bennofs1 has joined #nixos
<neonfuz> an example of a program that I know of that uses 9p is plan9port's program 'fontsrv'
<johnw> what is 9p, and how does it relate to what we've been discussing?
<neonfuz> usage is like 'fontsrv -m ./fonts/' or something, and it hosts your system fonts as a filesystem and converts the fonts to plan9 format on the fly
<neonfuz> it's comparable to fuse but network transparent
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] Izorkin opened pull request #32081: antigen: 2.2.1 -> 2.2.2 (master...shells-antigen) https://git.io/vbevL
NixOS_GitHub has left #nixos []
<neonfuz> and it's just a suggestion, I think it would be neat
<neonfuz> and nixos already uses it in other parts of the os so it would fit in
<johnw> oh, I see
MrNickMyNick has joined #nixos
<johnw> the reason I reached for FUSE is that it's also well established on Mac
<neonfuz> so like with that fontserv program an example path would be "./fonts/Helvetica/20a/font" to get the plan9 format, size 20, anti-aliased Helvetica font
jmc_fr`` has joined #nixos
miko__ has joined #nixos
<neonfuz> ah, yeah I suppose if it's not just for nixos then yeah. There's always 9pfuse too
<MrNickMyNick> neonfuz: mounting the 9p files that plumber and fontsrv expose is mostly broken.
drakonis has joined #nixos
mistnim has quit [(Remote host closed the connection)]
<neonfuz> lol really, it works fine for me
<neonfuz> what's the problem?
<pcarrier> so it might depend on your mount options as well, and maybe newer releases trying to use newer versions of the protocol / supporting more feature?
slack1256 has quit [(Ping timeout: 260 seconds)]
<MrNickMyNick> the linux in-kernel 9p module and the fuse modules don't use quite the same protocol
<neonfuz> isn't 9p2000 the latest version?
<neonfuz> oh, hmm
<pcarrier> neonfuz: it's the biggest, for Linux
<pcarrier> so the linux kernel module has mount options
<pcarrier> to turn stuff on and off
<neonfuz> what do you mean the biggest
nallar has joined #nixos
Ross has quit [(Ping timeout: 260 seconds)]
nallar is now known as Ross
<MrNickMyNick> look up harveyos's docs on what to use for mounting and stuff. they disscuss it there.
<neonfuz> k
<pcarrier> oh OK sorry so there's 9p2000.u and 9p2000.L as well
<pcarrier> so like a crazy larger surface because Unix
<pcarrier> but if you want symlinks and hardlinks and ACLs and pipes and stuff, that's more API surface
rtjure has quit [(Ping timeout: 248 seconds)]
tempeh has joined #nixos
<pcarrier> good old 9P is a crazy smaller protocol to have to implement as a library.
<neonfuz> I mean in the case of doing something like mounting the nix options to a filesystem, you wouldn't really need any of the extensions
<neonfuz> like symlinks
<pcarrier> and the message packing and opscode and structure descriptions are always kept as simple a possible in all versions, it's really just about scope
rtjure has joined #nixos
<pcarrier> oh yeah absolutely. sorry I don't know what's out there for languages that would interop nicely with nix
bbsl has joined #nixos
<johnw> neonfuz: interesting, there's a 9p2000 binding for Haskell too
srdqty has quit [(Quit: WeeChat 1.9.1)]
<johnw> neonfuz: worth a look
<neonfuz> yeah that's probably what you'd want
<pcarrier> honestly otherwise it's just really easy to write
<pcarrier> like seriously the protocol part is really really tiny
<neonfuz> I should finish learning haskell and learn hnix
srdqty has joined #nixos
<neonfuz> I read through LYAH until like, monads I think
<johnw> it should be easy to support either 9p or FUSE, the underlying abstraction isn't the hard part as pcarrier suggests
<pcarrier> then you have to implement the logic of having file descriptors and doing operations on them, keeping tables of network IDs, etc.
<neonfuz> yeah
<pcarrier> I mean no 9p000 is crazy small and elegant
<pcarrier> FUSE is a different story
<johnw> ok, then 9p2000 is the better first target
<johnw> you've sold me with small and elegant
<pcarrier> :P
<johnw> neonfuz: is that what you use? ^
<pcarrier> it's about as small and simple as it can be. and you can implement it over virtio DMA.
<johnw> hmm.. 9pfs in all-packages depends on fuse
<johnw> and fuse declares itself 'linux only' and depends on utillinux
<neonfuz> I use nixos
<neonfuz> so no
<neonfuz> but maybe that's good? idk
<neonfuz> I've never used it
<johnw> i'd like the fuse build in nixpkgs not to require linux
<neonfuz> well yeah
<MrNickMyNick> johnw: you can fork the kernel 9p module and modify it to work with plan9 9p and get rid of the dependencies and most performance issues like that.
Harpalus has quit [(Quit: Leaving)]
<neonfuz> if that mac9p works well with the haskell 9p2000 bindings you're good
srdqty has quit [(Quit: WeeChat 1.9.1)]
<johnw> neonfuz: ok
johnsonav has joined #nixos
<bbsl> anyone use nixos at work here? Im thinking of installing it on my workmachine, but I have reservations about dealing with other peoples python-virtualenv and sbt projects at work
<neonfuz> hmm, if you're starting from a computer that doesn't have 9p or fuse installed already, 9p might be smaller
<neonfuz> bbsl: iirc there are some people that use it at work on a work wide scale, but I don't know about individuals that choose it on their own
asuryawanshi has quit [(Remote host closed the connection)]
<neonfuz> I mean I do but I'm just a freelancer and I'm currently unemployed lol
<pcarrier> yeah so mac9p only supports 9p2000
<pcarrier> the original
<pcarrier> or at least that was the case last I checked
<neonfuz> so that's probably a good thing in this specific case
<bbsl> neonfuz: yea this is my issue :) I can make it work for all my projects at home but I havent realy experimentet with doing things the "non nix way", as in working on outer peoples projects without migrating it to nix
<bbsl> sorry for the horribad spelling :)
<pcarrier> sorry, 9p2000 or 9p2000.u depending on mount flags. not 9p2000.L though. https://github.com/benavento/mac9p/blob/master/kext/vfsops.c#L161-L167
JosW has quit [(Quit: Konversation terminated!)]
<pcarrier> so yeah R/O 9p over immutable data structures should be ridiculously easy. the state transitions and internal state are very clear
<MrNickMyNick> are you guys talking about implementing the protocol library from scratch? cause there dozens of implementation in C and Go in the wild...
bbsl has quit [(Remote host closed the connection)]
hotfuzz has joined #nixos
<sphalerite> Is there a way to control where data will be swapped to preferentially? I have zram and an external hard disk set up for swap, and I only want to use the external hard disk when zram is being used to its full capacity
nixos1022 has joined #nixos
Acou_Bass has quit [(Ping timeout: 248 seconds)]
hotfuzz_ has quit [(Ping timeout: 240 seconds)]
orivej has quit [(Read error: Connection reset by peer)]
<sphalerite> MrNickMyNick: cool thanks. Hm, there doesn't seem to be a nixos option for it
_ris has joined #nixos
ixxie has quit [(Quit: Lost terminal)]
babyflakes has quit [(Quit: Connection closed for inactivity)]
MinceR has quit [(Ping timeout: 255 seconds)]
orivej has joined #nixos
<sphalerite> Can I use a different compression algorithm for a binary cache, when for instance IO/space is more abundant than CPU time and I'd prefer gz over xz?
<sphalerite> MrNickMyNick: huh, must be a new addition since 17.09 then
<MrNickMyNick> :D
MinceR has joined #nixos
<sphalerite> Oh no, I just can't read
<sphalerite> x)
<MrNickMyNick> :P
jmc_fr`` has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] Izorkin opened pull request #32083: zsh-command-time: init at rev-2111361 (master...shells-zsh-command-time) https://git.io/vbeJ5
NixOS_GitHub has left #nixos []
whald has quit [(Quit: Leaving)]
<MrNickMyNick> anyhow, if someone can backport this https://github.com/NixOS/nixpkgs/commit/0317d5d083a7d58c75bd6df2acb31dde44efcb0c#diff-b905f64fc74a09e764fa1fe8ae6106ed I'm sure someone will be happy other than me. bye!
MrNickMyNick has quit [()]
Acou_Bass has joined #nixos
<judson> What did I do wrong to have all my man pages go away?
asymmetric2 has joined #nixos
wajsel has joined #nixos
asymmetric1 has quit [(Ping timeout: 268 seconds)]
ylwghst has joined #nixos
devrandomguy has quit [(Quit: leaving)]
ilyaigpetrov has quit [(Quit: Connection closed for inactivity)]
wajsel has quit [(Quit: Leaving)]
wajsel has joined #nixos
justan0theruser has quit [(Ping timeout: 240 seconds)]
dieggsy has joined #nixos
<sphalerite> Anyone know how I can copy the secondary GPT header and partition table on a block device into the primary one?
<samueldr> (if you didn't find that already via google)
<sphalerite> samueldr: gdisk just says there's no GPT present
asuryawanshi has joined #nixos
<sphalerite> Aaah managed to find it using the advanced menu
<sphalerite> Weird that it doesn't mention that the backup table exists...
<samueldr> maybe when there is no trace of gpt, it doesn't assume it's a backup table?
<Lisanna> I tried to add a new option to an existing NixOS module, and it isn't working :/ garbageCollection.nix contains "{config, lib, pkgs, ...}: let cfg = config.services.dockerRegistry; in { options.services.dockerRegistry.garbageCollection = {...}; config = mkIf ...; }" and I import it with "imports = [ ./hardware-configuration.nix ./garbageCollection.nix ];" Doing just this produces no errors, but I get an error that the option doesn't exist
<Lisanna> if I actually try to use it.
<Lisanna> am I missing a step?
<samueldr> Lisanna: can you gist/pastebin the whole garbageCollection.nix file?
asuryawanshi has quit [(Ping timeout: 248 seconds)]
Lisanna2 has joined #nixos
ylwghst has quit [(Quit: Lost terminal)]
laggy_wifi_ has joined #nixos
rodgort` has quit [(Quit: Leaving)]
<Lisanna> I have no idea if that works or not, but the first thing I'm trying to do is to just get it to actually recognize services.dockerRegistry.garbageCollection as an option I can set in my NixOS config
laggy_wifi_ has quit [(Client Quit)]
slack1256 has joined #nixos
<samueldr> hmmm
<samueldr> it seems to work for me
<samueldr> used: `services.dockerRegistry.garbageCollection.enable = true;`
<Lisanna> Okay, new information: it works in my main nixos configuration, but not in a container's nixos configuration
<Lisanna> and the container is declarative
<samueldr> did you import the garbageCollection.nix in the container? (I don't know how adding custom options works with containers, yet)
<Lisanna> No, I didn't try that... was hoping it would just get inherited somehow
<nixos1022> Is there any attempt to generate configuration for fcitx (or any other IM)?
<Lisanna> I'm sure that would work now that I think about it, but is there a better way to do this, so I don't have to import this everywhere?
ma27 has joined #nixos
<Lisanna> other than running my own nixpkgs
<manveru> anyone here using elm? :)
ylwghst has joined #nixos
<gchristensen> I use it a teensy tiny manveru
rodgort has joined #nixos
<manveru> wonder if anyone of you can try it...
<Lisanna> containers = { registry = { config = import ./registry.nix; autoStart = true; }; };
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vbek7
<NixOS_GitHub> nixpkgs/master 3fb57fe Peter Romfeld: jazzy: init at 0.9.0
<NixOS_GitHub> nixpkgs/master ab48481 Jörg Thalheim: Merge pull request #31073 from peterromfeldhk/jazzy...
NixOS_GitHub has left #nixos []
ylwghst has quit [(Client Quit)]
<manveru> gchristensen: basically, in a project with shell.nix and this config, you should be able to use elm-mode now without having to start emacs in the nix-shell
rodgort has quit [(Client Quit)]
<gchristensen> ooohhh
<gchristensen> neat!
<manveru> i'd like to add that to other modes too... but it's unfortunate that there seems to be no central place i can inject this...
Thra11 has joined #nixos
FRidh has quit [(Quit: Konversation terminated!)]
hariel has joined #nixos
rodgort has joined #nixos
alhariel has quit [(Ping timeout: 248 seconds)]
Thra11_ has joined #nixos
Thra11 has quit [(Ping timeout: 240 seconds)]
azdle has quit [(Remote host closed the connection)]
azdle has joined #nixos
johnw has quit [(Quit: ZNC - http://znc.in)]
<samueldr> hmmmm
<samueldr> I have an issue where I can't nixos-rebuild after playing around with sandbox paths
<samueldr> (happily, it's all done with nixos-rebuild test)
hiratara has quit [(Ping timeout: 258 seconds)]
<samueldr> (but I'm too lazy to simply reboot, let's work hader for my laziness)
<samueldr> any idea on how I could cheat my way around "error: while setting up the build environment: getting attributes of path ‘/proc/sys/fs/binfmt_misc/’: Too many levels of symbolic links
<samueldr> nix.conf file is readonly, from being in the store
dieggsy has quit [(Remote host closed the connection)]
<neonfuz> any way to run random mono .exe files?
<neonfuz> like, a package or an environment maybe
ma27 has quit [(Quit: WeeChat 1.9.1)]
<neonfuz> or just install mono?
hiratara has joined #nixos
ma27 has joined #nixos
ylwghst has joined #nixos
<Thra11_> Having updated nixos-unstable, gcc 6.4.0 fails to build, with "configure: error: C preprocessor "/lib/cpp" fails sanity
<Thra11_> check"
<Thra11_> should it really expect to find it at /lib/cpp?
<Thra11_> I would have thought it would be either in the nix store or in the tmp build directory
ylwghst has quit [(Quit: Lost terminal)]
<samueldr> ah, I was able to fix it by using a previous configuration.nix that already was built, it didn't need to rebuild, so didn't need the sandbox
ma27 has quit [(Ping timeout: 276 seconds)]
ylwghst has joined #nixos
<ylwghst> Does anyone here using lightdm gtk greeter?
simukis has quit [(Ping timeout: 248 seconds)]
<samueldr> ylwghst: might as well share the issue, it might not be specific to lightdm gtk greeter
<samueldr> (but I'm personally using sddm)
fendor has quit [(Quit: Leaving)]
<ylwghst> samueldr: well are you able to configure your cursor theme?
<samueldr> for the user, but haven't done it through nix
Synthetica has quit [(Quit: Connection closed for inactivity)]
<ylwghst> samueldr: i mean within display-manager
<samueldr> no, haven't tried, but I'm actually curious and will try
<samueldr> this feels like something that's not dependent on the DM
<ylwghst> samueldr: sure it's not but it could be configured throug DM config too
<samueldr> what have you tried?
<ylwghst> samueldr: I actually wasn't able to pass icon theme files
<ylwghst> samueldr: I have tried to configure lighdtm dm gtk greeter by lightdm.gtk.greeter.extraConfig = 'icon-theme-name: Theme;''; This actually doesn't work. I think it just looks for icon theme in /usr/share/.../...
<samueldr> (I'm booting up my test computer)
<samueldr> humour me, try adding : environment.pathsToLink = [ "/share" ];
<samueldr> I know *some* software will load themes only when linked
<Lisanna> Hey samueldr , did anything else look wrong in my garbageCollection.nix file? For some reason the systemd service isn't showing up at all (systemctl status docker-registry-garbage-collection returns that it could not be found)
<ylwghst> samueldr: @pbogdan have done this but itsn't in upstream yet https://github.com/pbogdan/nixpkgs/commit/dfe9a1522928ea02df36712b953b171a8bbb055f
<Lisanna> not sure how much you know about nixos modules
<samueldr> Lisanna: only checked the presence of the systemd service in the FS, and I'm not using docker, sorry I can't help
<Lisanna> okay
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] NeQuissimus pushed 1 new commit to release-17.09: https://git.io/vbetu
<NixOS_GitHub> nixpkgs/release-17.09 3c93a19 Tim Steinbach: sbt: 1.0.3 -> 1.0.4...
NixOS_GitHub has left #nixos []
vod has joined #nixos
<vod> Hello everyone!
<ylwghst> samueldr: brb
ylwghst has quit [(Quit: Lost terminal)]
ylwghst has joined #nixos
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
<ylwghst> that has no effect
<ylwghst> actually we just need to set environemnt variable XCURSOR_PATH = "path/to/nix/package/icon/theme";
<samueldr> (I'm setting up lightdm gtk right now and I'll try soon after)
<ylwghst> yep
<ylwghst> and propably --set XDG_CONFIG_HOME "${theme}/share" so it can read the cfg
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] NeQuissimus closed pull request #32068: sudo: 1.8.20p2 -> 1.8.21p2 (master...sudo-1_8_21p2) https://git.io/vFjgn
NixOS_GitHub has left #nixos []
miko__ has quit [(Ping timeout: 250 seconds)]
<ylwghst> I do think it can reach the config too.. If I configure other options then just cursor-theme it has no effect on lightdm
<ylwghst> it can't reach
hiratara has quit [(Quit: ZNC - http://znc.in)]
<vod> could someone point me in the direction to search: I need to add a file to "/etc/kbd/consolefonts", but this already is a link to pkgs.kbd, if I overrideAttr it triggers a rebuild of all kbd deps, how should i go about just adding a couple of files to path?
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] NeQuissimus pushed 2 new commits to master: https://git.io/vbeqT
<NixOS_GitHub> nixpkgs/master 5c936f1 name: antigen: 2.2.1 -> 2.2.2
<NixOS_GitHub> nixpkgs/master ec8731e Tim Steinbach: Merge pull request #32081 from Izorkin/shells-antigen...
NixOS_GitHub has left #nixos []
<slack1256> qt programs depend on the qt runtime version value and thus need to be updated when it changes on the base system right?
<slack1256> that is a kind of hard to deal with impurity, right?
hiratara has joined #nixos
<samueldr> were you using `icon-theme-name: Theme;`? the syntax seems to be `icon-theme-name = Theme`
rihards has quit [(Quit: rihards)]
<ylwghst> samueldr: this defines only only icon-theme
<ylwghst> samueldr: not cursor
<samueldr> yes, but I wanted to know if your syntax was right or not
<ylwghst> cursor-theme-name = Paper;
<ylwghst> samueldr: icon-theme-name can be currently set using nix
<ylwghst> like this theme.name = "Greybird";
<ylwghst> theme.package = "${pkgs.greybird}";
nixos1022 has quit [(Remote host closed the connection)]
<gchristensen> slack1256: yes indeed, however if all those programs are in systemPackages on nixos, it is an easy issue to resolve
<ylwghst> this is improved version which handles even cursor-theme
bennofs1 has quit [(Quit: WeeChat 1.9.1)]
<ylwghst> vod: overrideAttr overrides original derivation and builds from sources with changes you specify
<vod> ylwghst: what should I use instead?
<ylwghst> samueldr: error: The unique option `environment.variables.XCURSOR_PATH' is defined multiple times, in `/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs/nixos/modules/programs/environment.nix' and `/etc/nixos/configuration.nix'.
<samueldr> ylwghst: are you using pbogdan's fixed version?
<samueldr> it seems to patch ligthdm-gtk-greeter
<ylwghst> samueldr: i'm not
<ylwghst> samueldr: i want to stay up with stable channel
<samueldr> okay, wanted to be sure I was testing the same thing you were testing
Thra11_ has quit [(Quit: IRC for Sailfish 0.9)]
<ylwghst> samueldr: but look how he achieved it
<ylwghst> i get the error if i set
<ylwghst> environment.variables = { XCURSOR_PATH = "${pkgs.paper-icon-theme}/share/icons/"; };
<ylwghst> vod: what do you want to add?
<vod> ylwghst: I just want to add a few *.psf.gz into "/etc/kbd/consolefonts/" dir
<vod> ylwghst: without triggering a rebuild of "kbd" package
<vod> currently "/etc/kbd/consolefonts" is a link to a dir in nix/store/kbd-version
<ylwghst> vod: you should use overrideAttrs
<vod> ylwghst: but it tries to rebuild everything that depends on kbd package.... any way around this?
<vod> it shouldn't require a rebuild of a system to just add a few console fonts, shoud it?
<ylwghst> vod: it just rebuilding the kbd package after the rebuild it will remain in /nix/store/ and then it will not rebuild again until you change the overrideattrs ssomehow again
<samueldr> ylwghst: afaict, pbogdan's configuration for lightdm will need patches for lightdm...
<samueldr> ... and it looks like the PR has been closed this summer: https://github.com/NixOS/nixpkgs/pull/25974
<ylwghst> samueldr: yep
<vod> ylwghst: 6+ hours to add a font on my machine...
<ylwghst> vod: this propably will note help you but try to look here https://nixos.org/nixos/options.html#environment.etc
<vod> ylwghst: thanks!
<vod> ylwghst: problem is the dir is a link to store, not the files inside, so when i tried to create a file in env.etc it permission denied me
<ylwghst> vod: thats how nix works
<vod> ylwghst: is there a way to override the result of an already built derivation? ie, tell nix to take a store path and create another one with additional links in it?
<ylwghst> vod: i have no idea at all
<ylwghst> samueldr: i actually wasn't able to change the cursor theme within all avalable dm's
<ylwghst> samueldr: XCURSOR_PATH = [ "$HOME/.icons" ];
<ylwghst> is already ^
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<samueldr> I was able to replicate these results: https://github.com/NixOS/nixpkgs/pull/25974#issuecomment-305997110
jb55 has joined #nixos
<samueldr> it needed to change the lightdm gtk wrapper
<samueldr> I'm now checking if it's doable without changing the lightdm derivation
<ylwghst> samueldr: oh, i haven't tried this, not bad at all
<wajsel> Hi anyone having experience with setting up a elixir phoenix environment in a shell.nix?
<ylwghst> samueldr: the DM doesn't uses these env variables or do?
<ylwghst> samueldr: i mean it can't reach my $HOME/.icons if i'm not logged as "myuser" yet
<samueldr> these variables aren't used for services, they are used for login shells, afaik, which affects the logged-in session, when logging-in
<ylwghst> samueldr: thats whay they pass theme packages path in DM derivations
<ylwghst> samueldr: yep, thanks
bennofs has quit [(Quit: WeeChat 1.9.1)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej closed pull request #32003: cloudfoundry-cli: init at 6.32.0 (master...ris-cf-cli) https://git.io/vFpKS
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej pushed 1 new commit to release-17.09: https://git.io/vbeYl
<NixOS_GitHub> nixpkgs/release-17.09 0990eea Robert Scott: cloudfoundry-cli: init at 6.32.0...
NixOS_GitHub has left #nixos []
fragamus has joined #nixos
<ylwghst> samueldr: What exactly this does https://github.com/NixOS/nixpkgs/pull/25974#issuecomment-305997110 ?
<ylwghst> I cant understand it..
<ylwghst> I see
<samueldr> there are two parts
<ylwghst> it creates $out/share/icons/default/index.theme
<samueldr> 1) it creates a theme called "default", which is ... yes, what you just said
<ylwghst> then this uses LightDM
<samueldr> and 2) (3 in his notes) patches lightdm to use the environment variables in the wrapper
<samueldr> I'm trying to see where lightdm loses environment variables from the systemd unit, as setting them in the systemd unit does not seem to work
<samueldr> it might be necessary to wrap the wrapper or something else
ssmike has quit [(Ping timeout: 248 seconds)]
<samueldr> I'll look into it a bit later, I'll eat and do random stuff
bfrog has joined #nixos
asuryawanshi has joined #nixos
<ylwghst> samueldr: It's unlcear to me why he just closed the PR
nuncanada has joined #nixos
asuryawanshi has quit [(Ping timeout: 268 seconds)]
thc202 has quit [(Ping timeout: 258 seconds)]
<sphalerite> https://lugn.sphalerite.org/build-times.sh gets build times for derivations with fairly poor accuracy
<sphalerite> Someone else might find it useful though :)
<sphalerite> Or, even better, know of a better way to find out how long a derivation took to build
<neonfuz> some chinese characters are still not showing up :/
<neonfuz> this is really damn annoying
tgunb has quit [(Remote host closed the connection)]
statusbot has joined #nixos
vod has quit [(Remote host closed the connection)]
jb55 has quit [(Ping timeout: 255 seconds)]
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
ylwghst has quit [(Quit: Lost terminal)]
ylwghst has joined #nixos
nuncanada has quit [(Quit: Leaving)]
nuncanada has joined #nixos
acarrico has quit [(Ping timeout: 252 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] xaverdh opened pull request #32090: oneko: set correct meta.platforms (master...master) https://git.io/vbe3z
NixOS_GitHub has left #nixos []
ylwghst has quit [(Quit: Lost terminal)]
ylwghst has joined #nixos
<Lisanna> which package provides su?
<Lisanna> or, rather, the su wrapper that I guess NixOS uses?