myguidingstar has joined #nixos
sigmundv has quit [(Ping timeout: 260 seconds)]
simukis__ has quit [(Ping timeout: 258 seconds)]
shanemikel_ has quit [(Quit: Konversation terminated!)]
systemfault has joined #nixos
<hodapp> what's the method for dropping to the correct environment for some build, but not actually running the phases of that build?
<c74d> hodapp: nix-shell
ambro718 has quit [(Ping timeout: 260 seconds)]
<Mateon1> copumpkin: Can we talk here rather than the Github issues?
<copumpkin> sure
<copumpkin> hi!
<hodapp> ah, right... now I'm just trying to figure out how to run the early phases
<hodapp> e.g. fetching from 'src'
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<Mateon1> copumpkin: I'm not 100% sure on how the Nix build system works, as I'm relatively new to Nix, however I believe the part that's failing is "fixed-output"
<Mateon1> Let me get the source
<copumpkin> oh okay
<c74d> hodapp: I thiiink it's just running `fetchPhase` in the shell
<hodapp> c74d: no such command seems to exist
<c74d> mh
<hodapp> $src isn't set to anything either
<copumpkin> Mateon1: oh, I see there's a fetchgx
<Mateon1> copumpkin: https://github.com/NixOS/nixpkgs/blob/master/pkgs/applications/networking/ipfs/default.nix#L11L14 - this is the "gxdeps" part of the IPFS build, it has a sha256 attached, which I assume means it's fixed-output and allowed to talk to the network
Avogadro has quit [(Ping timeout: 260 seconds)]
<copumpkin> Mateon1: yeah, then never mind, I should've checked the source :)
<Mateon1> Then, any idea on what the issue could be?
[0x4A6F] has quit [(Ping timeout: 260 seconds)]
<hodapp> unpackPhase won't work because $src/$srcs isn't set... not sure what to do
<copumpkin> if I were to guess
[0x4A6F] has joined #nixos
<copumpkin> Mateon1: compare to fetchgx source which says SSL_CERT_FILE = "${cacert}/etc/ssl/certs/ca-bundle.crt";
<copumpkin> just needs to be updated to say NIX_SSL_CERT_FILE
shanemikel_ has joined #nixos
<copumpkin> 99% sure that'll fix it immediately
<Mateon1> Ah, that's probably it. I wonder why it didn't fail on my machine, then
<hodapp> well.. hmm. I did "nix-shell -p python35Packages.python-mapnik" while what I'm actually looking at is pkgs/development/libraries/mapnik/default.nix, a dependency thereof
<copumpkin> Mateon1: not sure
<hodapp> because I'm trying to get at the scons build, which is responsible for producing the packages in python35Packages.python-mapnik (among other things)
mizu_no_oto has joined #nixos
kampfschlaefer has quit [(Ping timeout: 240 seconds)]
davidak has quit [(Quit: Leaving.)]
stepcut has quit [(Remote host closed the connection)]
<hodapp> can anyone think offhand of a module which produces both Python 2.x and 3.x bindings?
<hodapp> mostly, I'm looking for an example to refer to of how it passes appropriate options based on "which" Python
shanemikel has quit [(Quit: Leaving)]
<clever> hodapp: i think most of them go into top-level/python-packages.nix, and then nixpkgs will call the package twice, once with python2, and once with python3
<clever> hodapp: then each version shows up under python3Packages and python2Packages
<hodapp> clever: oh, hmm, yeah, I guess I knew that... but was looking for a way that the build could tell apart if it was being built with Python 2.x or Python 3.x
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<hodapp> this particular build needs python2 for Scons, but then Scons assumes it's building for python2 even if it's python35Packages.python-mapnik that is being built
<hodapp> I see that python-mapnik simply includes mapnik as a dependency, so I'd have to figure out how to inject in python3 somehow... or something?
aneeshusa has quit [(Quit: WeeChat 1.7)]
Wizek_ has quit [(Ping timeout: 260 seconds)]
ryanartecona has quit [(Quit: ryanartecona)]
elninja44 has quit [(Remote host closed the connection)]
justan0theruser has joined #nixos
mizu_no_oto has joined #nixos
justanotheruser has quit [(Ping timeout: 260 seconds)]
<dejanr> hey is there a way to query package store location?
<dejanr> i wanna point to some file from package
systemfault has quit [(Quit: Bye!)]
shanemikel_ has quit [(Quit: Konversation terminated!)]
<hodapp> okay, I am looking right at https://github.com/NixOS/nixpkgs/blob/master/pkgs/development/libraries/mapnik/default.nix#L26 yet if I run "nix-shell ~/nixpkgs -A mapnik" and try to run configurePhase it says "no configure script, doing nothing"
<clever> dejanr: the storepath after its built, or its nix expression path?
<dejanr> clever: the store path after its installed
<clever> hodapp: typing 'configurePhase' into the shell will run the default configurePhase bash function, not eval the $configurePhase variable
<clever> hodapp: nixpkgs is weird like that
<hodapp> ugh
<hodapp> is this documented clearly anywhere?
<dejanr> clever: so i need store path from outside of nix, is there a nix command to get that?
<clever> [clever@amd-nixos:~]$ nix-instantiate --eval -E 'with import <nixpkgs>{}; "${hello}"'
<clever> "/nix/store/bdjyhh70npndlq3rzmggh4f2dzdsj4xy-hello-2.10"
<hodapp> clever: that doesn't really explain it any better
<clever> hodapp: line 870 is using some weird bash syntax to either eval $configurePhase or run configurePhase
<clever> hodapp: you could just run the entire genericBuild function to get all of it handled for you
<hodapp> I didn't want to run the entire build, I wanted to look at things partway through to figure out what options needed to be changed for scons
<clever> hodapp: then you probably just eval "$configurePhase"
ryanartecona has joined #nixos
<ericnoan> i wish there was a wiki so this kind of knowledge could be congregated in one place
<ericnoan> i feel the only way to learn nix is by reading other modules and copy paste
<hodapp> uhm, yes, this is all stuff that is seemingly documented nowhere and I've gotten inconsistent answers
<hodapp> but I'm not sure how else I'm to go about debugging shit that is broken
<MP2E> reading existing modules and nix code of other people is pretty much how i've learned so far
<hodapp> the documentation is usually great, then I run into corners like this where nothing is really explained
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<ericnoan> there is also useful to read the issue tracker and this website has valuable info http://sandervanderburg.blogspot.com/2013/12/using-nix-while-doing-development.html
filterfish has joined #nixos
mbrgm has quit [(Ping timeout: 268 seconds)]
mbrgm has joined #nixos
<dejanr> clever: thanks
agjacome has quit [(Quit: leaving)]
stepcut has joined #nixos
takle_ has quit [(Remote host closed the connection)]
byteflame has joined #nixos
Supersonic112 has quit [(Disconnected by services)]
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
<alphor> how do you guys do directory management with nix? one shot idempotent services?
<grahamc[m]> Can you describe more of what you're trying to do, alphor?
byteflame has quit [(Ping timeout: 260 seconds)]
takle has joined #nixos
<alphor> sorry, should've clarified nixOS: if you've got a service expecting a certain directory structure (and appropriate permissions), my current solution is to have a wantedBy unit that does it for the service. just wondering if that's canon.
myguidingstar has quit [(Read error: Connection reset by peer)]
sigmundv has joined #nixos
nicknovitski has quit [(Ping timeout: 260 seconds)]
takle_ has joined #nixos
<alphor> ie mkdir -p /dir/needed && chgrp -R /dir/needed group && chmod ...
takle has quit [(Ping timeout: 258 seconds)]
mizu_no_oto has joined #nixos
takle_ has quit [(Ping timeout: 255 seconds)]
marusich has joined #nixos
infinisil has joined #nixos
<copumpkin> alphor: systemd tmpfiles has a NixOS module
<copumpkin> your module can just use systemd.tmpfiles.rules
<copumpkin> and make directories appear
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
wanttolovenix has joined #nixos
<clever> alphor: ive always used a prestart script
athan has quit [(Ping timeout: 240 seconds)]
<wanttolovenix> Hi! I'm setting up my first ever nixos installation with LUKS, partitioning with GParted. Just want to ask if this partitioning is correct: Table type=gpt, /dev/sda1 = 200MiB FAT32, /dev/sda2 = rest of disk ext4 ? Many thanks for reading my question
athan has joined #nixos
<wanttolovenix> The fat32 partition I aim to put on /boot, and have NixOS put GRUB on thre while /dev/sda2 would be the LUKS-encrypted partition where the system would live
phreedom has quit [(Ping timeout: 245 seconds)]
justan0theruser has quit [(Ping timeout: 240 seconds)]
justanotheruser has joined #nixos
joneshf-laptop has quit [(Remote host closed the connection)]
clever has quit [(Ping timeout: 240 seconds)]
nicknovitski has joined #nixos
takle has joined #nixos
kristoffer has quit [(Remote host closed the connection)]
takle has quit [(Ping timeout: 255 seconds)]
shanemikel has joined #nixos
matthewbauer has joined #nixos
<shanemikel> is there a qmake2nix script?
infinisil has quit [(Quit: Leaving)]
<shanemikel> I'm just developing something, not packaging it.. But trying to Make my project (after running qmake), I get some strange linker messages: https://clbin.com/DvH8J
<shanemikel> I've got it to successfully build with a derivation mostly copied from the "qtpass" package, but I'd like to be able to work with qt tools w/o wrestling too much with Nix for the time being
<shanemikel> Also, does anybody know how to get the Examples and Demos installed with Qtcreator?
<NixOS_GitHub> [nixpkgs] 0BobTheJanitor closed pull request #23547: libqmlbind: init at 0.2.0 (master...master) https://git.io/vy8I5
clever has joined #nixos
takle has joined #nixos
darlan has joined #nixos
ryanartecona has quit [(Quit: ryanartecona)]
hexagoxel has quit [(Ping timeout: 258 seconds)]
takle has quit [(Ping timeout: 252 seconds)]
matthewbauer has quit [(Remote host closed the connection)]
clever has quit [(Ping timeout: 268 seconds)]
hexagoxel has joined #nixos
takle has joined #nixos
clever has joined #nixos
takle has quit [(Ping timeout: 245 seconds)]
hamishmack has quit [(Ping timeout: 255 seconds)]
<NixOS_GitHub> [nixpkgs] copumpkin opened pull request #24964: [WIP] make-disk-image: change to be less VM-centric (copumpkin/nixos-install-refactor...make-disk-image-refactor) https://git.io/vSdmF
preyalone has joined #nixos
<preyalone> why does Nix complain nix-env: not found ?
<NixOS_GitHub> [nixpkgs] copumpkin closed pull request #21943: [WIP, don't merge] make-disk-image: change to be less VM-centric (master...images-with-less-vm) https://git.io/vMKHP
mizu_no_oto has joined #nixos
<preyalone> example Dockerfile triggering nix-env: not found https://github.com/NixOS/nix/issues/555#issuecomment-294397181
taktoa has quit [(Remote host closed the connection)]
takle has joined #nixos
sigmundv has quit [(Ping timeout: 258 seconds)]
ryanartecona has joined #nixos
pie_ has joined #nixos
takle has quit [(Ping timeout: 258 seconds)]
wanttolovenix has quit [(Remote host closed the connection)]
proteusguy has quit [(Remote host closed the connection)]
taktoa has joined #nixos
byteflame has joined #nixos
pie_ has quit [(Ping timeout: 258 seconds)]
byteflame has quit [(Ping timeout: 258 seconds)]
darlan has quit [(Remote host closed the connection)]
moet has joined #nixos
justbeingglad has joined #nixos
justbeingglad has left #nixos []
takle has joined #nixos
<shanemikel> ericnoan: thanks, I got it worked out..
shanemikel has quit [(Quit: Konversation terminated!)]
takle has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
<unlmtd[m]> has anyone made nix work with musl?
<unlmtd[m]> do you think that nixpkgs could eventually serve different kernels/Clibs in a single branch?
<unlmtd[m]> might become a bit bloaded I guess
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<simpson> unlmtd[m]: I think that there are small
<simpson> Er. I think that there are small musl efforts here and there, but I don't think that there's any single unified musl effort.
<unlmtd[m]> iv tried on void musl a while back but it was just a failure
takle has quit [(Ping timeout: 240 seconds)]
systemfault has joined #nixos
systemfault has quit [(Max SendQ exceeded)]
systemfault has joined #nixos
systemfault has quit [(Client Quit)]
moet has quit [(Ping timeout: 260 seconds)]
Zer000 has quit [(Quit: Leaving)]
moet has joined #nixos
erigoaj has joined #nixos
takle has joined #nixos
<erigoaj> Hi, I'm fiddling with LUKS and have set boot.initrd.luks.devices[0].device to be "/dev/disk/by-uuid/7a...". Turns out when rebooting that was the wrong path, as that UUID is to the already luks-opened partition. My question is: How can I know what path below /dev/... to set as `device` inside of boot.initrd.luks.devices?
takle has quit [(Ping timeout: 258 seconds)]
<erigoaj> Or should I set `device` to point to /dev/sda2, which is the partition that is LUKS-encrypted?
takle has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
<erigoaj> Giving up and installing Ubuntu instead... Feels bad but 2 days into trying to get it running I just can't stand it no more
nslqqq has quit [(Ping timeout: 260 seconds)]
ryanartecona has quit [(Quit: ryanartecona)]
johnsonav has quit [(Ping timeout: 255 seconds)]
<clever> erigoaj: what does blkid say when ran on /dev/sda2 ?
<clever> erigoaj: that may show the uuid of the encrypted version of the device
alx741 has quit [(Quit: alx741)]
nslqqq has joined #nixos
<erigoaj> clever: Thanks for the input. It returns `/dev/sda2: UUID="1f82dd32-c077...." TYPE="crypto_LUKS" ... ...`. Is that whay should be put inside the (first and only) entry in boot.initrd.luks.devices ?
<erigoaj> Wiped the disk and re-trying Nixos installation once again right now...
<clever> erigoaj: yeah, that sounds like the uuid you wanted to use
<erigoaj> I see... Thank you very much for the help
arkad has joined #nixos
arkad is now known as ob7dev
taktoa has quit [(Ping timeout: 260 seconds)]
taktoa has joined #nixos
<ob7dev> anyone know how to boot nixos from ISO?
preyalone has quit [(Quit: Connection closed for inactivity)]
<erigoaj> ob7dev: I ussume you want to make a USB stick to boot from? Plug it into a computer where you have the iso, then run `sudo dd bs=4m if=~/Downloads/nixos-graphical-17.03.945.5acb454e2a-x86_64-linux.iso of=/dev/disk2s1` (replacing the place of the nixos iso and of= with where your device is)
Itkovian has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
taktoa has quit [(Ping timeout: 255 seconds)]
taktoa has joined #nixos
takle has joined #nixos
byteflame has joined #nixos
shadow-x has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
byteflame has quit [(Ping timeout: 240 seconds)]
moet has quit [(Quit: leaving)]
<erigoaj> clever: I did the formatting and installation once again, and it worked as you said. Thank you for pointing me in the right direction.
<clever> yep :)
<ob7dev> @preyalone - yes I usually d dump isos for this... but what I meant is how to directly boot from without disk dumping it... specifically how do I boot it from a multiboot usb as one of the available iso installers
<ob7dev> ...see I have a USB that contains many ISO installers... an iso for arch linux, debian, etc...
<erigoaj> Have right now a freshly booted system, LUKS asking me for the password to decrypt the disk. Turns out at this stage, input from my keyboard are not appearing. It worked well in the (graphical) live USB-drive I had, but now now. It is a Logicool K400 Plus (a shit keyboard, but what I have at hand). It is wireless, with a USB dongle that you plug into the computer. Is this fixable somehow?
<ob7dev> but when I load the nixos iso to this usb, it fails to get to stage 2.... as it cannot find root device ( /dev/root/)
<ob7dev> @erigoaj - anyway u can get a wired keyboard pluged in?
<erigoaj> ob7dev: have no such - bought this one assuming it would (from the perspective of my computer) be a normal USB keyboard, as it has a USB dongle
<clever> erigoaj: https://nixos.org/nixos/options.html#boot.initrd.ker
MP2E has quit [(Quit: gotta go bbl)]
<ob7dev> ^
<clever> erigoaj: using this option, you can force what kernel modules it will load within the initrd
moet has joined #nixos
<erigoaj> clever: ob7dev: thank you. Assuming you mean "boot.initrd.availableKernelModules"? Is there some way to find out what value to put in that list? The doc gives only this example: [ "sata_nv" "ext3" ]
moet has quit [(Client Quit)]
<clever> erigoaj: mostly by looking at lsmod when things are working, and making an educated guess
<erigoaj> Ok... I think I'll just install Ubuntu instead right now. Thank you good folks for answering my questions so far. I hope to come back to Nix at some point in the future
<clever> erigoaj: you can always go the non-luks route, that one should have less problems
<clever> currently, i only use luks on portable machines
<erigoaj> Sadly I kind of require hardware encryption for this computer, so can't do that. But cheers for hanging out here on IRC and helping a Nix noob like me
mounty has quit [(Ping timeout: 252 seconds)]
mounty has joined #nixos
zeus_ has quit [(Remote host closed the connection)]
zeus_ has joined #nixos
zeus_ has quit [(Ping timeout: 260 seconds)]
hamishmack has joined #nixos
proteusguy has joined #nixos
takle has joined #nixos
calvertvl has quit [(Ping timeout: 260 seconds)]
ebzzry has quit [(Ping timeout: 258 seconds)]
takle has quit [(Ping timeout: 260 seconds)]
nicknovitski has quit [(Ping timeout: 255 seconds)]
shadow-x has quit [(Quit: Leaving)]
CripperZ has joined #nixos
mounty has quit [(Ping timeout: 240 seconds)]
mounty has joined #nixos
takle has joined #nixos
ob7dev has quit [(Quit: WeeChat 1.5)]
takle has quit [(Ping timeout: 255 seconds)]
mk-fg has joined #nixos
mk-fg has quit [(Changing host)]
mk-fg has joined #nixos
william__ has joined #nixos
ixxie has joined #nixos
wkennington has quit [(Read error: Connection reset by peer)]
thc202 has joined #nixos
william__ has quit [(Client Quit)]
takle has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
marsel has joined #nixos
cripperz_ has joined #nixos
CripperZ has quit [(Ping timeout: 260 seconds)]
cripperz__ has joined #nixos
cfricke has joined #nixos
cripperz_ has quit [(Ping timeout: 240 seconds)]
cripperz_ has joined #nixos
byteflame has joined #nixos
cripperz__ has quit [(Ping timeout: 240 seconds)]
MoreTea has joined #nixos
byteflame has quit [(Ping timeout: 240 seconds)]
nslqqq has quit [(Ping timeout: 240 seconds)]
systemfault has joined #nixos
systemfault has quit [(Max SendQ exceeded)]
Filystyn has joined #nixos
Filystyn has quit [(Changing host)]
Filystyn has joined #nixos
systemfault has joined #nixos
nslqqq has joined #nixos
mudri has joined #nixos
cripperz_ has quit [(Remote host closed the connection)]
cripperz_ has joined #nixos
cripperz_ has quit [(Remote host closed the connection)]
cripperz_ has joined #nixos
simukis__ has joined #nixos
cfricke has quit [(Quit: WeeChat 1.7)]
<sphalerite> I want to try out zfs, but it seems that the installer doesn't include the zfs kernel module. Is adding boot.extraModulePackages = with config.boot.kernelPackages; [ zfs ]; then nixos-rebuild switch the way to go?
<sphalerite> update, that doesn't actually work. But it also turns out I'm using an old installer image (17.03 beta) so I'll try using an up-to-date one first.
<kahiru> sphalerite: boot.supportedFilesystems = [ "zfs" ];
<kahiru> and then nixos-rebuild switch
cripperz_ has quit [(Remote host closed the connection)]
cripperz_ has joined #nixos
<sphalerite> kahiru: great, thanks
<kahiru> ywc, let me know how it goes
<avn> sphalerite: do you plan have / on zfs, and do you plan to boot from it?
cripperz_ has quit [(Remote host closed the connection)]
thaega has joined #nixos
cripperz_ has joined #nixos
<sphalerite> avn: yes
whenov has joined #nixos
<avn> sphalerite: so don't use dedup, use optimize store for /nix/store. And copyKernels = true; /* grub unable to read kernels from /nix/store on zfs, too much hardlinks */
whenov has left #nixos []
<clever> i prefer to keep an ext4 /boot
<clever> and also i always make a seperate zfs filesystem for /, /home, and /nix
<avn> clever: can be sane option, yes
m3tti has joined #nixos
takle has joined #nixos
<sphalerite> ah ok
erigoaj has quit [(Remote host closed the connection)]
<sphalerite> this is just a throwaway experiment, but good to know for the future where I do intend to probably use zfs for actual machines :)
<clever> i usualy enable snapshots for / and /home, but disable snapshots for /nix
<clever> and compression, it varies depending on disk size and cpu speed
<clever> i also try to make the pool name align to the hostname
<avn> clever: lz4 work for everything, except maildir (gzip-9 there)
<clever> so pools dont collide if i ever need to rescue 1 hdd in another machine
<avn> clever: also sane idea, but I affraqid to rename my tank ;)
<clever> ive never tried to rename them
<clever> but as long as you dont use tank again elsewhere, you should be fine
<avn> Anyway always be ready to quest like "I need boot media, with zfs unstable and recent kernel, crap how to do it faaaast"
<clever> i also have netboot configured in my lan
<clever> with a full nixos install env in the initrd
<clever> so if i just flip on netboot in any machine, it drops me to the same login prompt the iso gives
<avn> clever: anyway you need netboot with proper kernel and zfs module ;)
takle has quit [(Ping timeout: 252 seconds)]
<avn> I was too insane to do `zpool upgrade` on 0.7.0rc ;)
<clever> yeah
<clever> but i should be able to upgrade that netboot with just a nix-build
<clever> if i remember where i left it, lol
<clever> clever@c2d /var/www/c2d $ ls -ltrh
<clever> lrwxrwxrwx 1 clever clever 51 Dec 7 07:37 netboot -> /nix/store/w46h600glmq9m2v5cjpc2q06jd61winq-netboot
<clever> and there is a netboot.nix in there that contains exactly 1 derivation
<avn> clever: would be nice, if we promote netboot to mainline nixos ;)
<clever> avn: but there is no way to customize the configuration.nix
<clever> avn: https://gist.github.com/cleverca22/75e3db4dabeff05b743e9be77a2341b9#file-netboot-nix-L1 this replicates that code, and loads a custom configuration.nix
<sphalerite> well, it booted!
<avn> I should think about this, may be I invent how to make "builtin" rescue netboot to host configuration ;)
<sphalerite> that worked pretty well.
<sphalerite> I love how quick it is to install nixos
<clever> avn: that was also in my plans
<clever> avn: i just took the netboot kernel+initrd, and ran kexec on them
<clever> avn: and in theory, you can also drop those into /boot and add a grub option
<sphalerite> also, there should totally be a netboot nixops backend.
<clever> sphalerite: also, do you see the justdoit in my last gist?
<sphalerite> I saw a justdoit in some gist of yours
<sphalerite> which I adapted and added to an installer's root bashrc :D
<clever> this one creates a bash script in the systemPackages of the netboot
<sphalerite> yeah, saw that
kampfschlaefer has joined #nixos
<clever> so when it boots, you can simply type in 'justdoit' and it will do the entire install
<NixOS_GitHub> [nixpkgs] joachifm pushed 1 new commit to master: https://git.io/vSdlN
<NixOS_GitHub> nixpkgs/master f029fc8 Taahir Ahmed: csscompressor: init at 0.9.4 (#24946)
<clever> avn: oh, and i was thinking of using the kexec in nixops also
<clever> avn: needs less control over the network then full netboot
takle has joined #nixos
<avn> clever: I am not so familiar with nixops sadly
<NixOS_GitHub> [nixpkgs] joachifm closed pull request #24945: htmlmin: init at 0.1.10 (master...htmlmin) https://git.io/vS5zd
<clever> avn: ive read its source and helped a few people with it, but have yet to actually use it
<sphalerite> clever: what would be really neat is maintaining a hostname map on the server so that they don't all end up with the hostname autoinstall
<sphalerite> I did a terrible hack along those lines when I was last messing about with it, but it's really not ideal
<clever> sphalerite: i think it would be safer for it to register itself as booted in some central server, and then wait for a hostname, sshpubkey, and the order to self-nuke
<clever> so it doesnt nuke every machine you accidentaly boot it on
<sphalerite> yeah
<sphalerite> there hasn't been much risk of that so far though because I've just been connecting my laptop straight to the target machines
<sphalerite> there's also the issue of many of the machines having bad hard drives :( so I have it run badblocks first, but that takes ages
takle has quit [(Ping timeout: 260 seconds)]
systemfault has quit [(Quit: Bye!)]
ambro718 has joined #nixos
cripperz_ has quit [(Remote host closed the connection)]
<clever> ah, and then hard-code a map of ether->hostname into the image, via ethers
marsel has quit [(Ping timeout: 255 seconds)]
<sphalerite> yep
FRidh has joined #nixos
cripperz_ has joined #nixos
ebzzry has joined #nixos
<FRidh> how to ignore errors when evaluating?
<sphalerite> Could also serve it up via HTTP or something, that would avoid a bunch of rebuilding
<FRidh> $ nix-instantiate --eval -E 'with import <nixpkgs> {}; builtins.filter lib.isDerivation (builtins.attrValues python3.pkgs)'
<FRidh> error: CDDB-1.4 not supported for interpreter python3.5m
cripperz_ has quit [(Client Quit)]
phreedom has joined #nixos
<FRidh> for context, I am after the names of all the packages and prefer not to use nix-env
frheajgre has joined #nixos
<clever> sphalerite: if you are using the boot.php from my original gist, you could also modify the netboot.ipxe it chainloads, and embed that into the php script
<clever> sphalerite: and then pass a hostname as a kernel parameter
<clever> sphalerite: then it can just read /proc/cmdline to find arguments
<sphalerite> I'm not, but that's neat too
<sphalerite> then it could fetch its configuration at http://host/hostname.nix :D
<NixOS_GitHub> [nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vSd8D
<NixOS_GitHub> nixpkgs/master 2d1115b Peter Hoeg: sensu: 0.28.0 -> 0.29.0
<clever> yeah, once you throw some http into the mix, it can just use a remote database
takle has joined #nixos
filterfish_ has joined #nixos
filterfish has quit [(Remote host closed the connection)]
<frheajgre> Hi! I'm on my very first nixos installation. Trying to use nix-shell which seems very cool. I'm trying to use docker-compose via `nix-shell -p docker_compose`. Works great! But when in the shell, running `docker-compose up` gives "ERROR: Couldn't connect to Docker daemon at http+docker://localunixsocket - is it running?". I suppose I somehow need to start the service defined at
<frheajgre> https://github.com/NixOS/nixpkgs/blob/master/nixos/modules/virtualisation/docker.nix#L107, but how does one do that? Not finding anything on this in the manual
<clever> frheajgre: https://nixos.org/nixos/options.html#docker.enable this option will need to be set in /etc/nixos/configuration.nix
takle has quit [(Ping timeout: 252 seconds)]
<sphalerite> FRidh: builtins.map (d: if d.value ? name then d.value.name else null) (builtins.filter (v: v.success && lib.isDerivation v.value) (map builtins.tryEval (builtins.attrValues python3.pkgs)))
frheajgre has quit [(Remote host closed the connection)]
<sphalerite> key being tryEval
takle has joined #nixos
<sphalerite> tryEval needs documented...
<FRidh> sphalerite: I just happened to saw it in old release notes, but without actual documentation indeed.
<FRidh> thank you
<FRidh> funny though that it still complains about broken packages
<FRidh> I would expect it would ignore those as well.
dejanr has quit [(Quit: WeeChat 1.7)]
marsel has joined #nixos
<Isorkin> Install php with libressl not support?
<NixOS_GitHub> [nix] lheckemann opened pull request #1337: Manual: document tryEval (master...doc-tryEval) https://git.io/vSd47
<sphalerite> FRidh: ^ :)
<sphalerite> I like the PR number as well
Dezgeg_ has quit [(Ping timeout: 255 seconds)]
<FRidh> heh
hotfuzz has joined #nixos
byteflame has joined #nixos
hotfuzz_ has quit [(Ping timeout: 252 seconds)]
FrozenCow has joined #nixos
FrozenCow has quit [(Client Quit)]
byteflame has quit [(Ping timeout: 240 seconds)]
<MichaelRaskin> tryEval is a hack that I have written long long ago for a specific release.nix-related task that everyone would love to kill but it is still used by Hydra, so it lives.
<sphalerite> MichaelRaskin: is there an alternative for use cases like FRidh's?
<MichaelRaskin> Well, maybe his use case is close enough to what Hydra needs…
<MoreTea> MichaelRaskin, I found it quite useful to try to extract information from every package's metadata without coding in lots of checks
<MichaelRaskin> In some sense, listing the packages is the original use case
<MichaelRaskin> That we need to do better and nobody knows how
pie_ has joined #nixos
pie_ has quit [(Ping timeout: 260 seconds)]
schoppenhauer has joined #nixos
<MoreTea> Using a typed language might help
pie_ has joined #nixos
<MichaelRaskin> Not sure: a bottom value propagation logic doesn't need typing but needs an overview of everyhting, so nobody has tried yet
<MichaelRaskin> How typing would help. though?
jmc_fr has joined #nixos
<jmc_fr> Hello - I'm a newbie in Nixos, comming from debian, but loving functionnal languages - I'd like to know if NixOS is OK for an old laptop (ACER 5820TZG)
<MichaelRaskin> HDD size, RAM size?
khebbie has joined #nixos
khebbie has quit [(Client Quit)]
<jmc_fr> HDD = 200 Gb ; RAM = 2 Gb
m3tti has quit [(Ping timeout: 260 seconds)]
<jmc_fr> Pentium P6100
<sphalerite> Should be, I think. I've installed and used (briefly) nixos on a laptop with similar specs.
cfricke has joined #nixos
<avn> jmc_fr: should work, at least I have succesful install on thinkpad t400 with 100g hdd and 4g ram
khebbie has joined #nixos
khebbie has quit [(Client Quit)]
khebbie has joined #nixos
<MoreTea> MichaelRaskin, I meant that for this specific issue, typing would prevent me from getting into trouble with inconsistencies in data structures in pkg.*.meta
byteflame has joined #nixos
khebbie has quit [(Client Quit)]
<ktosiek> what's the current way of installing Nix as root?
<ktosiek> ...on Ubuntu
<ktosiek> it seems .debs are gone for nix>1.11
khebbie has joined #nixos
khebbie has quit [(Client Quit)]
mk` has joined #nixos
byteflame has quit [(Ping timeout: 260 seconds)]
khebbie has joined #nixos
<jmc_fr> MichaelRaskin, is it a reasonnable way to try a live CD ?
<MichaelRaskin> Well, installing anything inside LiveCD would eat RAM, but whatever is already there should work reasonably
<unlmtd[m]> could nix give problems to deoplete (vim auto-completion plugin). I just tried it with the deoplete-go installed, opened a go file, typed a few letters and got a bunch of errors
<MichaelRaskin> 5B
<MichaelRaskin> Maybe you just lack some auxillary tools?
sibi has joined #nixos
<MichaelRaskin> I have never used deoplete, though
<unlmtd[m]> sorry youre right
<unlmtd[m]> ahah, I need gocode
* unlmtd[m] slaps his forehead
mk` has quit [(Ping timeout: 260 seconds)]
khebbie has quit [(Quit: leaving)]
ixxie has quit [(Ping timeout: 245 seconds)]
khebbie has joined #nixos
arjen-jonathan has joined #nixos
Jookia2 has joined #nixos
<Jookia2> o/ Is GCC 6.3.0 broken on ARM for anyone else? It fails to build since it can't find crt1.o or crtio.o
khebbie has quit [(Quit: leaving)]
plumps has joined #nixos
khebbie has joined #nixos
khebbie has quit [(Client Quit)]
khebbie has joined #nixos
khebbie has quit [(Client Quit)]
khebbie has joined #nixos
khebbie has quit [(Client Quit)]
khebbie has joined #nixos
mudri has quit [(Ping timeout: 258 seconds)]
<NixOS_GitHub> [nixpkgs] NeQuissimus pushed 1 new commit to release-17.03: https://git.io/vSdgz
<NixOS_GitHub> nixpkgs/release-17.03 3069f72 Ambroz Bizjak: ntpd: Allow additional syscalls in seccomp filter....
<NixOS_GitHub> [nixpkgs] bjornfor pushed 1 new commit to release-16.09: https://git.io/vSdga
<NixOS_GitHub> nixpkgs/release-16.09 ab88d3f Bjørn Forsman: nixos/lighttpd: set $HOME for gitweb sub-service...
<NixOS_GitHub> [nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vSdgM
<NixOS_GitHub> nixpkgs/master 5a7b029 Tim Steinbach: linux: 4.11-rc6 -> 4.11-rc7
Jackneill has joined #nixos
jmeredith has quit [(Quit: Connection closed for inactivity)]
derjohn_mob has joined #nixos
<NixOS_GitHub> [nixpkgs] bjornfor pushed 1 new commit to release-16.09: https://git.io/vSd2k
<NixOS_GitHub> nixpkgs/release-16.09 764f69e obadz: keepass: 2.34 -> 2.35...
Avogadro has joined #nixos
<copumpkin> niksnut: any objections to merging the latest version of the installer PR? I think all specific concerns have been addressed and enough people have chimed in to suggest that they share my problems with the current approach. I've also opened another PR that uses this one in make-disk-image.nix
<copumpkin> I realize you still don't like it much but I think as I make more incremental changes it'll become more palatable :)
mudri has joined #nixos
khebbie_ has joined #nixos
khebbie_ has quit [(Client Quit)]
khebbie has quit [(Ping timeout: 240 seconds)]
khebbie has joined #nixos
<grahamc[m]> Jookia2, what version of NixOS?
<Jookia2> grahamc[m]: release-17.03
<grahamc[m]> What type of ARM?
ixxie has joined #nixos
<Jookia2> armv7
<grahamc[m]> I know there was a patch around armv8 on 17.03 in the past few days. Not sure about armv7 though...
<Jookia2> what was the patch?
<dtz> sphalerite: wow that is incredible haha, thanks for sharing
agjacome has joined #nixos
<NixOS_GitHub> [nixpkgs] bjornfor pushed 1 new commit to release-17.03: https://git.io/vSdaf
<NixOS_GitHub> nixpkgs/release-17.03 1d45682 Bjørn Forsman: nixos/lighttpd: set $HOME for gitweb sub-service...
<Jookia2> grahamc[m]: Yeah... It might be the same bug
contrapumpkin has quit [(Quit: My MacBook Pro has gone to sleep. ZZZzzz…)]
jmc_fr has quit [(Ping timeout: 255 seconds)]
boothead has quit [(Ping timeout: 240 seconds)]
<Jookia2> grahamc[m]: I'll tell you in a few days once the rebuild happens. :)
<Jookia2> my rebuild, rather
chakerbenhamed has joined #nixos
khebbie has quit [(Ping timeout: 260 seconds)]
<gchristensen> sounds good / painful, Jookia2 :)
<Jookia2> Yeah. Night! :)
<NixOS_GitHub> [nixpkgs] bjornfor pushed 1 new commit to release-17.03: https://git.io/vSdaS
<NixOS_GitHub> nixpkgs/release-17.03 cba47bc Peter Hoeg: spotify: 1.0.52.717 -> 1.0.53.758...
Avogadro has quit [(Ping timeout: 260 seconds)]
sigmundv has joined #nixos
<gchristensen> niksnut: ping again re type 2a
samotarnik has joined #nixos
drasich has joined #nixos
proteusguy has quit [(Ping timeout: 252 seconds)]
byteflame has joined #nixos
<ronny> i presume its unlikely that kde5/gnome get fixed within the next 1-2 weeks? as things are i can neither invest pesonal nor work time into nix, and the repeating breakages make it impossible to justify using nixos
<copumpkin> MichaelRaskin: have you opened an issue for your overlayfs on /etc idea? I really like it
<copumpkin> MichaelRaskin: or does this fall under the "I don't really care about NixOS" category for you?
<copumpkin> MichaelRaskin: I do think that getting the NixOS folks to figure out the details might save you some work even if you don't use it
<MichaelRaskin> Haven't opened it yet, because I tried to understand where overlayfs is going
<copumpkin> ah
<MichaelRaskin> I mean, I have figured out the details for my use case (I can live with theoretically-overwriteable static parts)
<MichaelRaskin> But I am not sure if it is good enough for NixOS
<copumpkin> user permissions still guard that on both NixOS and overlayfs, right?
<MichaelRaskin> That's true
<copumpkin> like large swaths of /etc are just symlinks to /etc/static right now
<samotarnik> hey everyone, nixos newbie here. when i do `nixos-rebuild switch`, i get back an error "error: file ‘nixpkgs/nixos’ was not found in the Nix search path". is there a way to recreate the symlinks in /.../nix/profiles/... ?
<MichaelRaskin> They look like real files in my case
<copumpkin> and those are just as overwritable (barring user permissions) as overlayfs today
<copumpkin> yup
contrapumpkin has joined #nixos
<copumpkin> I like that
<MichaelRaskin> Hm, good point
<MichaelRaskin> samotarnik: what does $NIX_PATH contain, what does nix-channel --list say (as root, I would guess)?
Avogadro has joined #nixos
<MichaelRaskin> copumpkin: OK, I take your point about current symlink situation
<MichaelRaskin> Then I can write-up my current state as is, you are right
<copumpkin> I'd love that, yeah
byteflame has quit [(Ping timeout: 255 seconds)]
<MichaelRaskin> (Also there is a _really_ nasty trap if you make an easily avoidable mistake…)
<copumpkin> that's also very valuable :)
<MichaelRaskin> Did you know that nix-store --gc doesn't stop at bind mount boundary inside nix-store?
<samotarnik> MichaelRaskin: (both as root) printenv NIX_PATH -> "nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs:nixos-config=/etc/nixos/configuration.nix:/nix/var/nix/profiles/per-user/root/channels"; nix-channel is nixpkgs-unstable
<peterhoeg> gchristensen: ref DM (or whatever those are called in IRC lingo), you've got a write-up regd security fixes?
<MichaelRaskin> I think you need a nixos-named channel
byteflame has joined #nixos
<samotarnik> MichaelRaskin: sorry, my bad, the exact output is "nixos https://nixos.org/channels/nixpkgs-unstable"
<MichaelRaskin> OK, risky but should work
<MichaelRaskin> What is in /nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs ?
<peterhoeg> gchristensen: excellent, thanks. Do you mind if I incorporate that as part of a bigger policy/procedure doc for submissions as a proper RFC ?
<gchristensen> nope
<gchristensen> here let me put a license on it
<gchristensen> peterhoeg: I added a license that says you can do whatever you want
<samotarnik> MichaelRaskin: that's the issue "ls -la /nix/var/nix/profiles/per-user/root/channels/nixos" has only nixos in it (a symlink to specific file inside my /nix/store). also, the "/.../root/channels" itself is a link to a generation that (i think) was gc-ed
<MichaelRaskin> nix-channel --update
<samotarnik> did that
<samotarnik> didn't help
<copumpkin> MichaelRaskin: mind highlighting me on the /etc thing when you write it up? I'm curious to see it
<MichaelRaskin> Maybe just switch channel to nixos-unstable?
<peterhoeg> Cool, I'll share it with you when it starts getting into shape. The idea is to make it "required reading" for new contributors - it basically covers a lot of the questions I had when starting out.
<MichaelRaskin> copumpkin: OK
<MichaelRaskin> copumpkin: I remember NixOS spending seconds walking /etc and looking for now-obsolete /etc/static links to remove. Is it still true?
<gchristensen> probably :)
<copumpkin> MichaelRaskin: I don't recall, sorry. I do think there's something about user permissions in there
<copumpkin> which is ugly and I'd love to do better if there is a better way
<copumpkin> user ownership, that is
<copumpkin> Ideally we'd get to the point where we can abolish /etc for all static config, but we're not there yet :)
<MichaelRaskin> I just want to make a claim about what NixOS does and do not want to build a heavy enough NixOS configuration for all those things to be obvious
<copumpkin> probably never will be
<copumpkin> but I can dream
<MichaelRaskin> copumpkin: /etc/hosts does make some sense, after all
<copumpkin> yeah
<copumpkin> a couple of them do, most of them don't need to be there
<MichaelRaskin> I want some things like /etc/X11 to be there _optionally_
<copumpkin> it calls find(
<MichaelRaskin> Because it saves three meaningless steps for every rebuild when debugging
<gchristensen> (NixOS Confessions) I like putting i3's config in /etc because then you can reload the i3 config without restarting the whole display manager, and it is a global state I actually like. (embarrassed)
<copumpkin> and then applies user ownership based on files in there and adjusts other things
* copumpkin banishes gchristensen from the anti-global-state club
<MichaelRaskin> Shouldn't i3 config be per-user anyway?
<samotarnik> MichaelRaskin: thanks, switching to nixos-unstable did something, gonna switch to cable now to see if it upgrades
<gchristensen> MichaelRaskin: I also want nix to manage it
<MichaelRaskin> But my current idea of proper state management is becoming a complicated mess of lifetime management based on cost:benefits of updating each specific part faster/less often
<MichaelRaskin> gchristensen: I have a script that nix-builds some stuff for my $HOME and just puts it with -o
<gchristensen> neat
<MichaelRaskin> But for a lot of stuff I do not care much as long as it ends up being a symlink into one of the main Monotone checkouts
<copumpkin> I think your overlayfs plan could also make it fully atomic
<MichaelRaskin> I mean, I want history and ease of reverts etc., right? Any VCS gives that, not everything should go through Nix
<copumpkin> \
<MichaelRaskin> Yes, my global fream is to make _FS_ state switch fully atomic
<gchristensen> MichaelRaskin: I made this module for other stuff in my home: https://gist.github.com/grahamc/1efe78f25434b44ea1758ef911cc899e
<MichaelRaskin> As in, current-system link changes when everything else is ready, and _everything_ magically points into correct places immediately
ryanartecona has joined #nixos
<MichaelRaskin> Also /bin/sh and /usr/bin/env
byteflame has quit [(Ping timeout: 268 seconds)]
<MichaelRaskin> I guess I don't have enough distinct Nix-running systems
samotarnik has quit [(Ping timeout: 260 seconds)]
<MichaelRaskin> So Monotone is OK for me
<MichaelRaskin> «somewhat more atomic»
<MichaelRaskin> OK…
<MichaelRaskin> By the way, why current-system link lives under /run ?
<gchristensen> "I guess I don't have enough distinct Nix-running systems" what is this, practicality? :)
<MichaelRaskin> I mean, for recovery etc it would make a ton of sense to boot from USB and see /var/current-system
calvertvl has joined #nixos
<copumpkin> /run!
<copumpkin> :)
<MichaelRaskin> copumpkin: sorry?
<copumpkin> the symlink is in /run/current-system
sigmundv has quit [(Ping timeout: 260 seconds)]
<MichaelRaskin> I know
<MichaelRaskin> That's what I am asking about
<MichaelRaskin> Because /run is non-persistent
<copumpkin> oh
<copumpkin> I see
<copumpkin> sorry :)
<MichaelRaskin> And the switch was deliberate but I never understood the rationale
ebzzry has quit [(Ping timeout: 240 seconds)]
LnL has quit [(Ping timeout: 252 seconds)]
hyphon81 has joined #nixos
<hyphon81> Hello. How do we use nested "if" in nix expression or can't it?
<gchristensen> use more ('s and )'s
ryantrinkle has quit [(Ping timeout: 258 seconds)]
LnL has joined #nixos
<joko> Hello, on my Hydra server I am getting "Output limit exceeded", any idea what does that mean and how to fix it?
<hyphon81> gchristensen: Thanks! My problem seems to be solved!
<MichaelRaskin> copumpkin: in reality, my solution is slightly more fragile than the current one, because Just Opening For Writing As Root results in «RO FS» error right now and can succeed with overlayfs
dridus has joined #nixos
byteflame has joined #nixos
ixxie has quit [(Ping timeout: 260 seconds)]
<NixOS_GitHub> [nixpkgs] NeQuissimus opened pull request #24966: rocker: init at 1.3.0 (master...rocker_1_3_0) https://git.io/vSdoV
<MichaelRaskin> gchristensen: copumpkin: well, not as short as I thought it would be… (the /etc/ management writeup; I tagged both of you)
jmeredith has joined #nixos
myguidingstar has joined #nixos
<copumpkin> thanks, looks good!
<copumpkin> what were the issues with aufs, out of curiosity?
<copumpkin> oh I guess it's not maintained
<copumpkin> "After several attempts, the author has given it up.[6]"
<MichaelRaskin> Well, to start with the obvious and undeniable one, it is maintained as much as the time permits
<MichaelRaskin> Giving up is about upstreaming
<MichaelRaskin> There is a 4.10 friendly release now, but you need to check when updating the kernel
<myguidingstar> hi all, I am using nixos-17.03 (hence NixOps 1.5). I am learning nixops and tried the GCE example in NixOps manual
<myguidingstar> when deploying, this error happens:
<myguidingstar> proxy....> setting up /etc...
<myguidingstar> proxy....> rm: cannot remove '/var/setuid-wrappers': Device or resource busy
<MichaelRaskin> copumpkin: gchristensen said aufs is complicated enough that stability under non-trivial operation sequences is not well-tested enough
<myguidingstar> proxy....> setting up tmpfiles
<copumpkin> boo
<myguidingstar> proxy....> error: unable to activate new configuration
<myguidingstar> (proxy is one of the GCE machines. The same error appears in other machines, too)
<myguidingstar> I tried re-run "nixops deploy ..." several times but the errors are still there
<myguidingstar> any idea what to do now?
ryanartecona has quit [(Quit: ryanartecona)]
agjacome has quit [(Remote host closed the connection)]
<NixOS_GitHub> [nixpkgs] copumpkin closed pull request #23026: Refactor nixos-install to separate out filesystem build logic (master...nixos-install-wip) https://git.io/vDbUq
<obadz> copumpkin: congrats :)
<copumpkin> obadz: thanks!
<sphalerite> any reason to use aufs and not overlayfs?
<copumpkin> sphalerite: MichaelRaskin was proposing to use overlayfs
ryanartecona has joined #nixos
<MichaelRaskin> I do list some reasons to prefer aufs
<MichaelRaskin> Basically, it doesn't have one annoying limitation of overlayfs
<NixOS_GitHub> [nixpkgs] copumpkin deleted copumpkin/nixos-install-refactor at d990aa7: https://git.io/vSd6s
<copumpkin> could always add functionality to it :)
<MichaelRaskin> Looking at the politics around overlayfs/unionmount merge…
<MichaelRaskin> Basically, aufs was not merged because unionmount looked more promising, but it never got completely ready for merging and then overlayfs was merged because it is simpler
plumps has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<copumpkin> sure, but adding a small feature to it might not be too hard now
<myguidingstar> anyone can help with the above?
<copumpkin> :P
<MichaelRaskin> copumpkin: a proper version of the feature breaks its internal data structure layout
cfricke has quit [(Quit: WeeChat 1.7)]
agjacome has joined #nixos
mguentner has joined #nixos
eacameron has joined #nixos
<sphalerite> what is the feature?
<MichaelRaskin> Union FS mount where there is an RW branch and an RO branch, but RO branch takes precedence (so it can shadow preexisting files in RW branch, and you cannot write to a file present in the RO branch)
joneshf-laptop has joined #nixos
MoreTea has quit [(Ping timeout: 240 seconds)]
<sphalerite> hyphon81: you can. if cond1 then if cond2 value1 else value2 else value3
<sphalerite> whoops, didn't realise I was in the history
<sphalerite> aah right
<NixOS_GitHub> [nixpkgs] gilligan opened pull request #24969: broadcom-sta: fix linux-4.8 patch sha (master...fix-broadcom-hash) https://git.io/vSdPf
ixxie has joined #nixos
mkoenig has joined #nixos
iyzsong has joined #nixos
<sphalerite> so my X server just crashed
ryantrinkle has joined #nixos
<sphalerite> at least I think it did. All the applications I had open died with "Resource temporarily unavailable"
<sphalerite> there's nothing in the journal about it crashing though
<sphalerite> (oh, and my screen went black)
<sphalerite> Where would it make sense to go hunting for more information?
erasmas has joined #nixos
takle has quit [(Remote host closed the connection)]
drasich has quit [(Ping timeout: 255 seconds)]
mk` has joined #nixos
derjohn_mob has quit [(Ping timeout: 252 seconds)]
takle has joined #nixos
takle has quit [(Remote host closed the connection)]
<gchristensen> MichaelRaskin: I was using aufs in fairly extreme environments, and this was about 2 years ago. it is probably
<gchristensen> MichaelRaskin: I was using aufs in fairly extreme environments, and this was about 2 years ago. it is probably better now, and probably fine under less extreme use cases
estewei has joined #nixos
<MichaelRaskin> Well, hopefully, but then you also have this version-tracking question
<NixOS_GitHub> [nixpkgs] romildo opened pull request #24970: vivaldi: 1.6.689.34-1 -> 1.8.770.56-1 (master...upd.vivaldi) https://git.io/vSdXA
<MichaelRaskin> For my own system, that question is actually more relevant
roconnor has joined #nixos
bennofs has joined #nixos
marusich has quit [(Quit: Leaving)]
bennofs1 has joined #nixos
bennofs has quit [(Ping timeout: 245 seconds)]
takle has joined #nixos
<gchristensen> man, ext4 has unfairly lost favor IMO
<gchristensen> it is still an excellent filesystem, even in the face of more modern ones
<roconnor> what am I supposed to be using instead of ext4?
<gchristensen> btrfs and zfs have gained a lot of fans, and it isn't undeserving of some fans, but ext4 is (still) a fantastic FS
takle has quit [(Ping timeout: 252 seconds)]
<dash> xfs of course
takle has joined #nixos
ryanartecona has quit [(Quit: ryanartecona)]
ryanartecona has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
<Sonarpulse> Profpatsch: there are intentionally multiple fixpoints, for now
<Sonarpulse> with my listy thing :)
<Sonarpulse> stdenv is weird, but I suspect overriding it in a bootstrapping stage should work
nicknovitski has joined #nixos
<MichaelRaskin> «Google's QUIC protocol leaks the exact length of sensitive information when transmitted over the Internet»
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #24927: rustup: init at 1.2.0 (master...rustup) https://git.io/vS7Mq
<gchristensen> great :|
<gchristensen> TLS & HTTP are pretty good. not perfect, but pretty good
<MichaelRaskin> What about people having had some time to find the basic pitfalls there and even to fix most of them
<gchristensen> yep
<MichaelRaskin> The recommended fix is to switch QUIC off in all the tools that allow to do that, and block QUIC at the firewall level until a fix is deployed
<gchristensen> MichaelRaskin: can you link me to the issue?
takle has joined #nixos
<gchristensen> nice branding
<MichaelRaskin> It is a fresh link at lobste.rs
<gchristensen> this is a funny sentence " This could allow an eavesdropper to learn the exact length of someone's password when signing into a website like Google's E-mail system called G-mail."
<ixxie> lol
<NixOS_GitHub> [nixpkgs] ttuegel pushed 1 new commit to master: https://git.io/vSdHs
<NixOS_GitHub> nixpkgs/master 4f5391e Thomas Tuegel: dropbox: 23.4.18 -> 23.4.19...
<NixOS_GitHub> [nixpkgs] ttuegel pushed 1 new commit to release-17.03: https://git.io/vSdHc
<NixOS_GitHub> nixpkgs/release-17.03 24ce0db Thomas Tuegel: dropbox: 23.4.18 -> 23.4.19...
chakerbenhamed has quit [(Ping timeout: 240 seconds)]
obadz has quit [(Ping timeout: 258 seconds)]
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #24969: broadcom-sta: fix linux-4.8 patch sha (master...fix-broadcom-hash) https://git.io/vSdPf
takle has quit [(Ping timeout: 260 seconds)]
takle has joined #nixos
reb has joined #nixos
<reb> How can I install pandoc-citeproc on nixos please?
<gchristensen> `nix-shell -p haskellPackages.pandoc-citeproc` seems to do it
<reb> Thanks!
<ixxie> the package I am trying to make gives a homeless shelter error; yesterday someone said this has to do with missing $HOME directory, but what I don't understand is where do I need to create one
<MichaelRaskin> Do you know what it wants to do with $HOME?
<ixxie> as far as I know nothing
<ixxie> there is no $HOME mentioned anywhere in my code
<MichaelRaskin> Maybe add set -x somewhere to see what commands are executed
obadz has joined #nixos
chakerbenhamed has joined #nixos
<LnL> ixxie: you can also set it to /tmp/foo, and see what ends up in there
ryanartecona has quit [(Quit: ryanartecona)]
<ixxie> it seems its trying to cd into nothing
<MichaelRaskin> You had «cd $in», have you fixed that?
<ixxie> I guess my problem is I don't know how to set $in
<MichaelRaskin> in=$PWD
<MichaelRaskin> (whenever you are in the directory where you want to return later)
<MichaelRaskin> pushd/popd could also work
<ixxie> `whenever` :P this is another problem: I am trying to fill in attributes in stdenv.mkDerivation but I don't know how it executes what and when
<gchristensen> copumpkin: how can I get back the not-clang stdenv?
<MichaelRaskin> Why do you even want to come back to the source directory
<MichaelRaskin> Can't you copy everything in one go?
nicknovitski has quit [(Ping timeout: 255 seconds)]
<ixxie> thats what I tried to do really
<ixxie> but apparantly its not in the right directory
derjohn_mob has joined #nixos
<ixxie> perhaps it is because my package file is with the code it uses
JagaJaga has joined #nixos
<MichaelRaskin> Why are you trying to do cd $in at all?
<ixxie> most packages fetch from somewhere else
<MichaelRaskin> In reality that doesn't matter
<ixxie> MichaelRaskin: because the installPhase cannot find directories that are in my project
<MichaelRaskin> What is your current code version?
<ixxie> 0.0.1
<MichaelRaskin> Why do you change directory in the build phase??
<MichaelRaskin> Grrr
<MichaelRaskin> I mean a link
<ixxie> hold on I will push
<MichaelRaskin> Whenever your code starts copying anything, you should just copy everything you ever want to copy
<MichaelRaskin> This doesn't have any cd anymore
<MichaelRaskin> OK, and what is the output when it fails?
aminechikhaoui has joined #nixos
<MichaelRaskin> Looks like a success
<ixxie> oops
<ixxie> set -x managed to hide that from me :P
<ixxie> thanks
DutchWolfie has joined #nixos
DutchWolfie has quit [(Changing host)]
DutchWolfie has joined #nixos
nicknovitski has joined #nixos
<ixxie> MichaelRaskin: so whatever I put in bin should be directly executible in the shell right?
<MichaelRaskin> Yes
JagaJaga has quit [(Ping timeout: 255 seconds)]
iyzsong has quit [(Ping timeout: 255 seconds)]
<copumpkin> gchristensen: hmm? On Darwin?
<gchristensen> copumpkin: yeah, isn't that a thing that changed?
ilyaigpetrov has joined #nixos
<copumpkin> It changed to LLVM 4. It's been using LLVM for years now
<copumpkin> You can use a previous LLVM stdenv if you're having trouble but it adds a long time to compile the other version of LLVM unless Hydra already has it
<gchristensen> right, makes sense
<copumpkin> See rustc right now for an example using LLVM 3.9 I think
<reb> The haskellPackages.pandoc-citeproc works well. Thanks. How did you know its name when it doesnt show up in nix-env -qaP '.*pandoc.*'?
<LnL> you can use llvmPackages_37.stdenv to get the old one IIRC
<ixxie> MichaelRaskin: maybe I am misunderstanding... I just put a simple script that echoes a test, and its not accesible when type it's name into nix-shell
<MichaelRaskin> Well, nix-shell is another set of troubles
<MichaelRaskin> It only loads _dependencies_ by default
<MichaelRaskin> Not the package itself
<ixxie> so I should go for a normal install... one moment
Dezgeg_ has joined #nixos
zeus_ has joined #nixos
zeus_ has quit [(Remote host closed the connection)]
zeus_ has joined #nixos
<ixxie> nix-env -f ./mypackage.nix -i doesn't find stdenv (cannot auto-call func with argument without default value)
<MichaelRaskin> Well, I guess you can remove pkgs argument and add default for stdenv as (import <nixpkgs> {}).stdenv
<maurer> Hm? Are we considering switching the default stdenv to be clang-based, or has it just always been that way on osx?
<MichaelRaskin> On OSX many things want clang, so clang is default there; Linux-specific stuff expects gcc quite often
arjen-jonathan has quit [(Ping timeout: 260 seconds)]
<ixxie> Cheers MichaelRaskin, it works!
<ixxie> It would be nice to know WHY :P my gut says if I used it as a module imported into configuration.nix I wouldn't have this problem
<MichaelRaskin> Why what? Why it needs stdenv to be either passed or have a default?
<MichaelRaskin> Because where could it get a value?
<ixxie> I guess what I am asking is, why do you see many packages on nixpkgs with no defaults working?
<ixxie> even though I am not explictly passing values when I install them
<aminechikhaoui> aszlig: any chance you can look at https://github.com/NixOS/nixops/issues/648 , I found you and niksnut are the ones that touched most of that part in the git blame so I think it would be very helpful if you can provide your input ;)
<MichaelRaskin> Because you install them via attributes in all-packages.nix where callPackage _does_ pass them the values?
<ixxie> and nix-env -i pkgname ?
<ixxie> (I don't use all-packages.nix - is that a Nix-on-other-distros thing?)
<MichaelRaskin> It scans the attributes inside all-packages.nix to check which of them has the correct name (this is slow), but ends up installing via all-packages.nix
<MichaelRaskin> It is the repo structure
<MichaelRaskin> If you use checkout, you _do_ use it
<MichaelRaskin> And with the channel — probably too
<ixxie> checkout?
<MichaelRaskin> Have you read the Nix manual?
<joko> Anyone to help me with "Output limit exceeded" on Hydra? I have set max_output_size to 4 GB and still nothing :(
<ixxie> yes, not front to back but yeah
<MichaelRaskin> Primary nix model is about working with Nix expressions probably in some repository and building from source
<MichaelRaskin> NixOS installs channels which just means you do not manage the expressions or a repository checkout, but it doesn't change the actual evaluation
<schoppenhauer> hi.
<schoppenhauer> is there a reason why the palemoon package builds palemoon itself rather than using the binary tarball and applying patchelf
ambro718 has quit [(Ping timeout: 260 seconds)]
<MichaelRaskin> Because building from source is the default approach of Nix, I guess
<MichaelRaskin> You could add palemoon-bin as an option
reb has left #nixos []
<NixOS_GitHub> [nixpkgs] nicknovitski opened pull request #24971: elm-format: symlink latest version to `bin/elm-format` (master...just-elm-format) https://git.io/vSdpa
<ixxie> MichaelRaskin: aha I see checkout is just a subversion term I am unfamiliar with :) thanks for clearing that up
<MichaelRaskin> git actually also uses this word
ryanartecona has joined #nixos
<MichaelRaskin> (and saner DVCSes use this word too)
erictapen has joined #nixos
byteflame has quit [(Ping timeout: 245 seconds)]
sigmundv has joined #nixos
<NixOS_GitHub> [nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vSdjW
<NixOS_GitHub> nixpkgs/master b47e022 Nick Novitski: elm-format: symlink latest version to `bin/elm-format` (#24971)...
mk` has quit [(Ping timeout: 240 seconds)]
Jackneill has quit [(Remote host closed the connection)]
sigmundv has quit [(Ping timeout: 240 seconds)]
davidak has joined #nixos
pie_ has quit [(Changing host)]
pie_ has joined #nixos
digitus has joined #nixos
drasich has joined #nixos
sibi has quit [(Quit: Connection closed for inactivity)]
endformationage has joined #nixos
<NixOS_GitHub> [nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vSFTl
<NixOS_GitHub> nixpkgs/master da9a147 Andrew Miloradovsky: dale: init at 20170416...
<NixOS_GitHub> nixpkgs/master e8f2046 Jörg Thalheim: Merge pull request #24941 from amiloradovsky/dale...
ambro718 has joined #nixos
vandenoever has joined #nixos
<copumpkin> schoppenhauer: patching binaries is error-prone and hard to track
<copumpkin> schoppenhauer: so probably 99.x% of nixpkgs is built from source
<copumpkin> also gives us a ton of flexibility and control over e.g., hardening flags and patches to handle our non-FHS paths
<MichaelRaskin> We have a few tens of closed-source packages, so 99.x% may be a bit optimistic
<schoppenhauer> copumpkin: the problem is that I have a bug on palemoon, and people there want me to try their binary package, and stuff. it is always the same with software … it is always "the other project" that does the error.
<copumpkin> yeah, that can be tricky to deal with
<copumpkin> sometimes we do introduce bugs, too!
<copumpkin> usually the goal is to apply as few nix-specific customizations as possible in nixpkgs
<copumpkin> and get upstream to fix their stuff as much as possible
<MichaelRaskin> schoppenhauer: you could try to mimic firefox-bin
<copumpkin> so we don't need to patch
<MichaelRaskin> I think palemoon-bin will get accepted as an optional thing, just not as a replacement of palemoon built from source
ilyaigpetrov has quit [()]
<Ralith> I notice that electron also isn't built from source, presumably because it's hellishly complicated and slow to do so
byteflame has joined #nixos
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #24385: fsharp: init at 4.1.7 (master...master) https://git.io/vStOp
<copumpkin> hasn't stopped us in the past, Ralith :)
<copumpkin> but I guess past a certain point we might not bother
<Ralith> hydra time isn't free after all
byteflame has quit [(Ping timeout: 258 seconds)]
chakerbenhamed has quit [(Ping timeout: 240 seconds)]
arjen-jonathan has joined #nixos
<copumpkin> orly
<Ralith> I mean, I assume
<Ralith> someone's paying for the hardware and upkeep
<gchristensen> it is true
<gchristensen> but generally we don't worry too much about how much it takes to buildstuff
<copumpkin> nope
<copumpkin> I refuse to believe it
<schoppenhauer> MichaelRaskin: I actually thought of that. But first I try some more debugging
erictapen has quit [(Ping timeout: 260 seconds)]
<schoppenhauer> in any case, I will probably add a pull request for some more optional flags.
<schoppenhauer> soometimes soon
MP2E has joined #nixos
mudri has quit [(Ping timeout: 252 seconds)]
cfricke has joined #nixos
arjen-jonathan has quit [(Quit: WeeChat 1.7)]
byteflame has joined #nixos
mojjo has joined #nixos
<mojjo> hi! anyone experienced with npm2nix? I get 'registry error parsing json' when trying to build a generated expression...
<schoppenhauer> is it enough to add autoconf213 to nativeBuildInputs, to make sure it is actually built with that version of autoconf?
ertesx has joined #nixos
<Mic92> mojjo: I remember the nix expression in nixpkgs/pkgs/development/node-packages/generate.sh to work fine some days ago
ryanartecona has quit [(Quit: ryanartecona)]
sellout- has joined #nixos
aminechikhaoui has quit [(Ping timeout: 240 seconds)]
ertes has quit [(Ping timeout: 255 seconds)]
ertesx is now known as ertes
mojjo has quit [(Ping timeout: 258 seconds)]
<dash> I'm trying to package some code that uses CMake's 'externalproject_add' to pull in a dependency. I'm not a cmake expert, is there a convenient way to replace that with a nix-provided dependency instead?
ixxie has quit [(Quit: Lost terminal)]
marsel has quit [(Ping timeout: 240 seconds)]
ryantm has quit [(Quit: Connection closed for inactivity)]
Filystyn has quit [(Quit: Konversation terminated!)]
<NixOS_GitHub> [nixpkgs] joachifm pushed 4 new commits to release-17.03: https://git.io/vSFsb
<NixOS_GitHub> nixpkgs/release-17.03 376bda6 ndowens: logcheck: 1.3.17 -> 1.3.18...
<NixOS_GitHub> nixpkgs/release-17.03 d55d1ff Joachim Fasting: logcheck: point homepage to alioth.debian.org...
<NixOS_GitHub> nixpkgs/release-17.03 058b899 Tobias Pflug: broadcom-sta: fix linux-4.8 patch sha...
cfricke has quit [(Quit: WeeChat 1.7)]
DutchWolfie has quit [(Quit: Konversation terminated!)]
cfricke has joined #nixos
mudri has joined #nixos
alx741 has joined #nixos
erictapen has joined #nixos
JagaJaga has joined #nixos
plumps has joined #nixos
aneeshusa has joined #nixos
<ToxicFrog> Argh
Itkovian has joined #nixos
<ToxicFrog> Ever since updating to 17.03, the permissions on my LetsEncrypt certs keep getting reset, which breaks everything
hotfuzz_ has joined #nixos
Itkovian has quit [(Client Quit)]
hotfuzz has quit [(Ping timeout: 255 seconds)]
drasich has quit [(Ping timeout: 240 seconds)]
marsel has joined #nixos
myguidingstar has quit [(Ping timeout: 240 seconds)]
<mbirkis> is it possible to disable the firewall only for a specific wireless network (i am using networkmanager)=
<mbirkis> ?
dejanr has joined #nixos
JagaJaga has quit [(Ping timeout: 260 seconds)]
Avogadro has quit [(Read error: Connection reset by peer)]
drasich has joined #nixos
AllanEspinosa has joined #nixos
<Ralith> ToxicFrog: the acme stuff has a "group" option; that should be all you need to give access to things that need it
<ToxicFrog> Ralith: yeah, and I set that option, and it worked fine until I switched to 17.03 last week
<ToxicFrog> And now they keep resetting from nginx:mail to nginx:nginx
<vandenoever> konsole is built with cmake, yet i do not get cmake in the path when i do this: nix-shell '<nixpkgs>' --attr konsole
<Ralith> haven't had any issues on my server yet
Itkovian has joined #nixos
agjacome has quit [(Quit: leaving)]
byteflame has quit [(Ping timeout: 240 seconds)]
<hodapp> for a cmake builder, must one add 'make install' manually in some form?
<hodapp> ...wait a minute. nevermind. I was running 'nix-build', not 'nix-env'.
pareidolia has joined #nixos
<NixOS_GitHub> [nixpkgs] joachifm pushed 1 new commit to master: https://git.io/vSF4y
<NixOS_GitHub> nixpkgs/master aeaa79d Joachim Fasting: ratpoison: 1.4.8 -> 1.4.9...
jarrodwb has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
ryanartecona has joined #nixos
<jarrodwb> Hi. I am pretty new to Nix but I am plugging away at it. Currently I am trying to make a derivation that uses ammonite-repl as a builder. The problem is that ammonite-repl/bin/amm is a shell script that use /usr/bin/sh and also java. How do I provide these things to the derivation?
<gchristensen> can I use a custom path to the config.nix in a nix-shell invocation?
<hodapp> https://github.com/NixOS/nixpkgs/pull/24901 huh, the CI build seemed to fail just because of some transient network issue
<hodapp> not sure what to do about that
lush has joined #nixos
<jarrodwb> Sure.
<jarrodwb> What is the best way to paste preformated code into here?
<MichaelRaskin> At what level? I think you could set NIXPKGS_CONFIG, and inside default.nix/shell.nix you could reimport nixpkgs and pass config argument
<jarrodwb> with (import <nixpkgs> {}); derivation { name = "amm-test"; builder = "${ammonite-repl}/bin/amm"; args = [ ./s.sc ]; system = builtins.currentSystem; }
<lush> hey! is "texlive.combined.scheme-full" the usual texlive environment for LaTeX?
<jarrodwb> agg. Sorry about that.
<jarrodwb> That was supposed to be my nix file. After that I would do nix-build amm.nix
<lush> any latex user here?
<hodapp> lush: that's what I use, I believe
<lush> hodapp: alright ty :-)
mudri has quit [(Ping timeout: 245 seconds)]
<MichaelRaskin> I have a list of stuff to add to medium
cfricke has quit [(Quit: WeeChat 1.7)]
Itkovian has joined #nixos
<NixOS_GitHub> [nixpkgs] Hodapp87 opened pull request #24972: osm2pgsql: init at 0.92.1-unstable (master...osm2pgsql) https://git.io/vSF0w
<hodapp> yay, I did another PR
dridus has quit [(Quit: znc? wassat?)]
<hodapp> I sort of wish I didn't have to set up an entire Postgresql database just to convert OpenStreetMap XML to a form that tools can read, but whatever
jarrodwb has quit [()]
<joko> Is it possible to delete manually a path in /nix/store?
<MichaelRaskin> Stronger than nix-store --delete /nix/store/path ?
<joko> MichaelRaskin: I think so, "nix-store --delete" mentions that it can't delete the file, it's still alive
erictapen has quit [(Ping timeout: 245 seconds)]
<MichaelRaskin> Well, you can manually delete a path, but it breaks various correctness invariants
<MichaelRaskin> Why don't you find corresponding roots with -q --roots and kill them?
erictapen has joined #nixos
<joko> Well, I know the roots, it's from a Hydra job... It's actually strange, I get output limit exceeded although the builder hosts have built the package. So now I am trying to remove it manually in case it starts working
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #24970: vivaldi: 1.6.689.34-1 -> 1.8.770.56-1 (master...upd.vivaldi) https://git.io/vSdXA
wak-work has quit [(Ping timeout: 255 seconds)]
<hodapp> hm, if I need to fix something in a PR, should I just push another commit to the branch? or squash & rebase and do a new PR?
<MichaelRaskin> You can even force-push
<hodapp> oh, that works?
<MichaelRaskin> Yes
<joko> hodapp: commit --ammend and push --force
byteflame has joined #nixos
aminechikhaoui has joined #nixos
<NixOS_GitHub> [nixpkgs] 7c6f434c closed pull request #24972: osm2pgsql: init at 0.92.1-unstable (master...osm2pgsql) https://git.io/vSF0w
bennofs1 has quit [(Ping timeout: 252 seconds)]
tv has quit [(Read error: Connection reset by peer)]
tv has joined #nixos
AllanEspinosa has quit [(Ping timeout: 252 seconds)]
<NixOS_GitHub> [nixpkgs] the-kenny pushed 1 new commit to master: https://git.io/vSF2D
<NixOS_GitHub> nixpkgs/master 383706f jraygauthier: digikam: add video support through libqtav (#24866)...
AllanEspinosa has joined #nixos
drasich has quit [(Ping timeout: 240 seconds)]
wak-work has joined #nixos
mudri has joined #nixos
seanz has joined #nixos
zeus_ has quit [(Read error: Connection reset by peer)]
zeus_ has joined #nixos
dridus has joined #nixos
lush has quit [(Quit: Konversation terminated!)]
<NixOS_GitHub> [nixpkgs] Ericson2314 opened pull request #24974: Introduce `mapNullable` into lib and use it in a few places (master...mapNullable) https://git.io/vSF6J
mudri` has joined #nixos
marsel has quit [(Ping timeout: 258 seconds)]
mudri has quit [(Ping timeout: 260 seconds)]
takle has quit [(Remote host closed the connection)]
<mbirkis> is it possible to disable the firewall only for a specific wireless network based on ssid (i am using networkmanager)?
mudri` is now known as mudri
<joko> mbirkis: check when your machine is on that network if shutting down the firewall service is ok and if so, tell networkmanager to shutdown the fw service when on that wireless network
<joko> It seems that the "output limit exceeded
takle has joined #nixos
<joko> problem is gone... I had to uninstall all the Hydra-related stuff, rm hydra and postgres and delete the old cache >_<
takle_ has joined #nixos
takle has quit [(Read error: Connection reset by peer)]
danharaj has joined #nixos
myguidingstar has joined #nixos
<joko> All that for setting max_output_size in hydra.conf after the initial run...
_c_ has left #nixos ["-"]
<NixOS_GitHub> [nixpkgs] Ericson2314 closed pull request #24974: Introduce `mapNullable` into lib and use it in a few places (master...mapNullable) https://git.io/vSF6J
<joko> Damn it, I was too optimistic, it still fails with that error... I'm giving up for tonight, bb, ppl
wak-work has quit [(Ping timeout: 245 seconds)]
<jeremejevs> Hi, my system suddenly semi-bricked =/ Made some innocent change (just added a font, I think), quit my window manager, then my display manager failed to start, so I rebooted, and then got greeted with the following: "Error loading \efi\nixos\<hash>-linux-4.9.22-bzImage.efi: Not Found" I think I've done garbage collection since the last reboot, that's
<jeremejevs> probably the cause. Loading recent generations doesn't help, only the very first few ones work (without a window manager and such). And problem is, the old generation I got to work is 16.09, but I've been on unstable for a few days now. Doing a "nixos-rebuild switch" in that working generation results in "Read-only file system:
<jeremejevs> '/boot/loader/entries/nixos-generation-<number>.conf.tml'". And the number changes between reboots. Running "nixos-rebuild switch" again then complains about something different: "Read-only file system: '/boot/efi/nixos/<hash>-linux-4.4.52-bzImage.efi'". What do I do? :(
wak-work has joined #nixos
FRidh has quit [(Quit: Konversation terminated!)]
ryanartecona has quit [(Quit: ryanartecona)]
mbirkis has quit [(Quit: WeeChat 1.4)]
danharaj has quit [(Ping timeout: 268 seconds)]
mbirkis has joined #nixos
<NixOS_GitHub> [nixpkgs] Ericson2314 closed pull request #24610: Platform normalization (master...platform-normalization) https://git.io/vSBhY
plumps has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<jeremejevs> Should I maybe run "nixos-install" or something? Would there be any risk of losing my home dir?
AllanEspinosa has quit [(Ping timeout: 260 seconds)]
* gchristensen is over the moon
* copumpkin pulls gchristensen back down
takle has joined #nixos
<gchristensen> I just got the final thumbs up to use Nix at work
<copumpkin> yay!!
<NixOS_GitHub> [nixpkgs] lheckemann opened pull request #24975: [WIP] inkscape: add potrace (master...inkscape-potrace) https://git.io/vSFyM
takle_ has quit [(Read error: Connection reset by peer)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<MichaelRaskin> jeremejevs: does dmesg say anything about errors on /boot?
Itkovian has joined #nixos
simukis__ has quit [(Ping timeout: 240 seconds)]
<jeremejevs> MichaelRaskin: I only see a single error: "FAT-fs (nvme0n1p1): error, fat_free_clusters: deleting FAT entry beyond EOF", and then "Filesystem has been set read-only". I assume this is about /boot.
AllanEspinosa has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<NixOS_GitHub> [nixpkgs] hamishmack opened pull request #24976: defaultIconTheme: Enable darwin support (master...darwin-icons) https://git.io/vSF9F
zeus_ has quit [(Read error: Connection reset by peer)]
zeus_ has joined #nixos
hiratara has quit [(Ping timeout: 258 seconds)]
AllanEspinosa has quit [(Ping timeout: 240 seconds)]
hiratara has joined #nixos
<MichaelRaskin> Well, that explains why it is read-only…
<MichaelRaskin> Basically, you need to dosfsck it in the unmounted state then mount it back
byteflame has quit [(Ping timeout: 255 seconds)]
<MichaelRaskin> (may be easier to do it with some reboots, and maybe with an external boot media)
drasich has joined #nixos
<jeremejevs> MichaelRaskin: Thanks, will try this tomorrow. What could cause this? Garbage collection?
AllanEspinosa has joined #nixos
aminechikhaoui has quit [(Ping timeout: 260 seconds)]
<MichaelRaskin> Definitely no, GC doesn't touch outside /nix/store
<MichaelRaskin> Well, outside /nix
<MichaelRaskin> A generic FS error, dunno, maybe unclean shutdown
Guest60 has joined #nixos
LnL has quit [(Ping timeout: 260 seconds)]
kampfschlaefer has quit [(Ping timeout: 260 seconds)]
seanz has quit [(Quit: Leaving.)]
<NixOS_GitHub> [nixpkgs] dsprenkels opened pull request #24977: mpDris2: 0.6 -> 0.7 (master...master) https://git.io/vSFdb
LnL has joined #nixos
AllanEspinosa has quit [(Ping timeout: 240 seconds)]
Guest60 has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
steven__ has joined #nixos
justanotheruser has quit [(Ping timeout: 240 seconds)]
<Mic92> fat32 on nvram, sounds kind of weird.
hiratara has quit [(Quit: ZNC - http://znc.in)]
<Mic92> jeremejevs: if you only used /boot for nixos, it is probably easier to re-format it and use nixos-install to fix it
<MichaelRaskin> Why? Main storage in the system is nvram, and UEFI requires FAT32
justanotheruser has joined #nixos
hiratara has joined #nixos
<Mic92> of course
<Mic92> There is no way around it
<Mic92> But think in what time fat32 was developed and what we have today.
AllanEspinosa has joined #nixos
steven__ has quit [(Ping timeout: 240 seconds)]
Guest60 has joined #nixos
mkoenig has quit [(Remote host closed the connection)]
<jeremejevs> Mic92: Yes, just NixOS. Will do that.
myguidingstar has quit [(Remote host closed the connection)]
<jeremejevs> By the way, my last working generation was 99 (and there's also an empty file for the 100th). Coincidence? :D
<Mic92> jeremejevs: what is the filesystem of your system partition?
byteflame has joined #nixos
<jeremejevs> ext4
<jeremejevs> Mic92: Why?
vandenoever has quit [(Quit: ䷴៚)]
justin_ has joined #nixos
<Mic92> jeremejevs: looks like you had an outage, while updating. you probably want to run `nix-store --verify --repair --check-contents` afterwards to ensure consistency
<Mic92> empty files happen on lost updates.
<Mic92> metadata was updated, but not the content
erictapen has quit [(Ping timeout: 255 seconds)]
byteflame has quit [(Ping timeout: 240 seconds)]
<jeremejevs> An outage? As in a network outage? Highly unlikely. But either way, why wouldn't a previous generation load then? In fact, out of 95 generations, only the first 5 load, then a half gets kernel panic, and the other half prints what the newest generation printed.
<jeremejevs> *out of 99
<clever> jeremejevs: power outage
thc202 has quit [(Ping timeout: 240 seconds)]
<jeremejevs> clever: Ah, no, the last switch went smoothly. I then simply quit i3, observed the display manager failing to load (don't know for what reason), and then hard-rebooted. There was nothing going, I wasn't even logged in.
<Mic92> probably the filesystem was not synced
<clever> a hard-reboot will also corrupt the disk
<NixOS_GitHub> [nixpkgs] zraexy opened pull request #24978: grub module: fix efiInstallAsRemovable description (master...zraexy-grub) https://git.io/vSFp0
<Mic92> zfs seems to be more imune regarding that
<jeremejevs> Doesn't ext4 deal with that?
<Mic92> depends on your mount flags
<clever> zfs guards against it by making all data and metadata immutable
<clever> so it never overwrites data, it creates a new copy of the data with the changes
<clever> and new metadata
<clever> and i think it changes the metadata in every dir in the tree up to the root folder, which is in a circular log
<Mic92> jeremejevs: next time, you could use magic sysreq to let your os sync the filesystem before reboot: https://wiki.archlinux.de/title/SysRQ
<Dezgeg> even with ext4 you have to explicitly mount with some unsafe mount options to make it non-power-fail-safe
<Dezgeg> a bigger problem is that the nix package manager itself isn't power-fail-safe
<clever> Dezgeg: though there is a sync call in nixos-rebuild
<clever> Dezgeg: at times, i have noticed it hanging, because it had 100mb's in the write buffer, and i just killed power, lol
<Dezgeg> well, it doesn't help if you kill the box during that sync()
<clever> yeah
<Mic92> Dezgeg: what makes it unsafe?
<clever> SD cards have a low write speed, and linux buffers too much
<clever> so when the sync does happen, it hangs for unusualy long periods
<clever> and now your paying for the fake speed the write-cache gave
Wizek_ has joined #nixos
<clever> and i incorrectly thought it was crashed
digitus has quit [(Quit: digitus)]
<Dezgeg> nix doesn't call fsync() (recursively) for built/fetched store paths before marking them valid in the sqlite database
<jeremejevs> Nix doesn't run anything in the background, after a switch, right?
markus1189 has joined #nixos
markus1199 has joined #nixos
taeradan_ has quit [(Quit: No Ping reply in 180 seconds.)]
<Dezgeg> it doesn't but the operating system will write out data asynchronously without an fsync()
<Dezgeg> anyway, what nix does is not related to /boot corruption, just got side-tracked
<taktoa> ZFS is the nix of filesystems :^)
<clever> :D
<taktoa> (in that it's based on persistent/immutable data structures)
<calvertvl> the only issue I'm aware of for ZFS is you can't have swap-on-zvol to hibernate :/
taeradan has joined #nixos
markus1209 has quit [(Ping timeout: 260 seconds)]
markus1219 has quit [(Ping timeout: 240 seconds)]
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #24978: grub module: fix efiInstallAsRemovable description (master...zraexy-grub) https://git.io/vSFp0
<clever> calvertvl: for non-luks systems, i would just use a normal swap partition
<Mic92> Is swap on zvol a good idea?
<clever> Mic92: ive heard that it has problems locking up
<calvertvl> clever: yeah, but I didn't know that when I set this box up, so I don't have one
<clever> calvertvl: ah yeah
<Mic92> Maybe the situtation improved a little, because zfsonlinux has a new memory allocation method
<Mic92> It now uses Slab-based allocator, which fits better into the linux architecture
<NixOS_GitHub> [nixpkgs] Mic92 pushed 1 new commit to release-17.03: https://git.io/vSFhd
<NixOS_GitHub> nixpkgs/release-17.03 452f64c zraexy: grub module: fix efiInstallAsRemovable description...
mudri has quit [(Quit: WeeChat 1.7)]
<NixOS_GitHub> [nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vSFjk
<NixOS_GitHub> nixpkgs/master 457bc9b Linus Heckemann: inkscape: add potrace...
<NixOS_GitHub> nixpkgs/master ae49647 Jörg Thalheim: Merge pull request #24975 from lheckemann/inkscape-potrace...
elninja44 has joined #nixos
<NixOS_GitHub> [nixpkgs] Mic92 pushed 2 new commits to staging: https://git.io/vSFj5
<NixOS_GitHub> nixpkgs/staging d6b42b4 Maksim Bronsky: perl: 5.22.2 -> 5.22.3
<NixOS_GitHub> nixpkgs/staging f78c032 Maksim Bronsky: perl: init 5.24.1 and set as default perl
filterfish_ has quit [(Ping timeout: 258 seconds)]
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #24521: instead-launcher: init at 0.6.3 (master...instead-launcher) https://git.io/vSWrs
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #24608: Fix consul version being "unknown-unknown". Fixes #24606. (release-17.03...fix-24606-consul-version-unknown-17.03) https://git.io/vSB9S
<NixOS_GitHub> [nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vSbvu
<NixOS_GitHub> nixpkgs/master 6b7c5ba Jörg Thalheim: display-manager: fix argument handling of sddm...
<NixOS_GitHub> [nixpkgs] Mic92 closed pull request #23264: display-manager: fix argument handling of sddm (master...sddm) https://git.io/vyq2H
<jeremejevs> Mic92 clever Dezgeg Going off right now, but just wanted to report: inserted my thumb stick with 16.09 on it, updated nix-channel to nixos-unstable (on the stick), ran nixos-rebuild switch, then ran "mkfs.fat" on my boot volume, then mounted in /mnt/boot, along with my system volume in /mnt, and ran nixos-install (looked all good). Then rebooted, and the old
<jeremejevs> generation list is still there (ending in 102 now, but that didn't increment after I went through the procedure again). The 102nd generation doesn't boot, gets stuck on running start job for some device for 1.5m, and then fails. Not sure what to do...
erasmas has quit [(Quit: leaving)]
<clever> jeremejevs: you need to update /mnt/etc/nixos/hardware-configuration.nix with nixos-generate-config --root /mnt
<clever> jeremejevs: otherwise, the resulting nixos will try to mount the old (and missing) boot filesystem
<clever> jeremejevs: thats one common issue when /boot is getting formated
<jeremejevs> clever: I see, thanks. I wanted to run it, but feared that it might overwrite my configuration.nix (which isn't backed up yet). Will it?
<clever> no
<clever> it only overwrites hardware-configuration.nix
<NixOS_GitHub> [nixpkgs] Mic92 pushed 1 new commit to release-17.03: https://git.io/vSbvQ
<NixOS_GitHub> nixpkgs/release-17.03 40a5498 Jörg Thalheim: display-manager: fix argument handling of sddm...
<jeremejevs> Okay, will give this a quick go.
Guest60 has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
byteflame has joined #nixos
drasich has quit [(Ping timeout: 258 seconds)]
<clever> jeremejevs: ah, and that start job its stuck on, that might be trying to find the old /boot
Jookia2 has quit [(Quit: Jookia2)]
ijophst has joined #nixos
justin_ has quit [(Remote host closed the connection)]
<ijophst> Is this a typo? https://github.com/NixOS/nixpkgs/blob/release-17.03/nixos/modules/system/boot/initrd-ssh.nix#L35 I don't even have any /bin/ash on my (fresh) nixos install
justin_ has joined #nixos
<clever> ijophst: that path will only be valid inside the initrd
<clever> ijophst: everything in the extrautils/bin path (which includes all of busybox) gets symlinked to /bin
byteflame has quit [(Ping timeout: 252 seconds)]
<ijophst> Wow, that's neat! I've never seen the shell from the shebang `#! @shell@`, is that a template? Are the @-symbols in this script of special meaning?
<clever> thats handled by the substitute function in setup.sh
filterfish_ has joined #nixos
justin_ has quit [(Ping timeout: 255 seconds)]
filterfish_ has quit [(Remote host closed the connection)]