<clever>
mfiano: nixos auto-detects the .fstype of the filesystem configured for / and /nix
<mfiano>
That could be the problem?
<clever>
and the problem is before grub even found its own config
<clever>
my only guess is that grub cant make sense of your btrfs, and you need an ext4 /boot partition
ashkitten1 has quit [Remote host closed the connection]
<mfiano>
Heh that's strange that it works on 2 other linux distributions. The only thing I did was format the root partition and mounted nix/store under that
<clever>
mfiano: the subvolumes could be making it more complex
<mfiano>
I don't see how, considering they are the same, minus the nix-store sub-subvolume, but that's post boot
<pie__>
my plasmashell or something died again and i keep forgetting the executable names to restart this stuff so now i cant switch out of irc but i can run programs
<pie__>
can anyone tell me how to get my window manager back?
Myrl-sak1 has joined #nixos
ashkitten has joined #nixos
<pie__>
(stuff is still runnig i just cant switch, and i dont want to kill everything)
Myrl-saki has quit [Ping timeout: 256 seconds]
<pie__>
anyone awake? xD
jtojnar has quit [Remote host closed the connection]
jtojnar has joined #nixos
<MichaelRaskin>
I am awake, but my advice would be to got to console ant install icewm or something
<MichaelRaskin>
(Something very lightweight by capable to switch)
<mfiano>
I'll try again some other time. Thanks for the help
<pie__>
would that work? everything except the window switcher and taskbar seem to be runnig
<MichaelRaskin>
I though KDE WM was KWin but that's before Plasma (maybe)
<MichaelRaskin>
Not sure, but could?
<clever>
pie__: in theory, you can run any window manager, and it should just take over and give you window switching
<MichaelRaskin>
IceWM is small; if the WM place is not free for taking IceWM will safely refuse
<pie__>
oh i think kwin was the thing i always forget
<pie__>
hm its not fixing the problem this time around though
<mfiano>
sda6 is boot. sda7 is the btrfs container. sda8 is swap
mk-fg has joined #nixos
mk-fg has quit [Changing host]
mk-fg has joined #nixos
<clever>
mfiano: sda1 is way too big
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] volth opened pull request #35510: jdk: use oraclejdk on aarch64 (master...jdk-on-aarch64) https://git.io/vAwEQ
NixOS_GitHub has left #nixos [#nixos]
<clever>
mfiano: also, what does `mount` say?
<mfiano>
clever: That was created by Windows install, and any attempt to make it smaller results in me not able to boot. Additionally, any other Linux distribution I tried has no issue there.
<disasm>
samueldr: but nix 1.12 is still the default, it just adds nixStable2 instead of nixUnstable as the name of 2.0, hence why that change is much less complex than the other PR and is already merged.
<clever>
mfiano: what does `mount` say?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] deepfire opened pull request #35511: GHC 8.4 configuration: more package fixes (master...ghc-8.4) https://git.io/vAwEb
<clever>
mfiano: the bios boot partition on sda1 only needs to be about 2mb, and it must never be mounted or formated with an FS
<clever>
mfiano: also, is the CSM enabled in the bios?, your configured for legacy booting, not efi
<mfiano>
clever: I understand that but this was auto-created and any attempt to resize it prevents me from dual booting. I'm a game developer and need to test my code on Windows too
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] volth opened pull request #35512: i2p: force openjdk (master...patch-98) https://git.io/vAwuv
NixOS_GitHub has left #nixos [#nixos]
<mfiano>
Windows installer did that for me
<disasm>
mfiano: are you booting efi or bios with grub?
<mfiano>
CSM is enabled and set to legacy
mk-fg has quit [Ping timeout: 245 seconds]
<clever>
mfiano: dual-booting can be restored, once nixos is working, with the right nixos config for grub
<mfiano>
This motherboard is pretty horrible and it can't even detect a usb drive at boot without CSM on/set to legacy
<mfiano>
I'll nuke the whole partition table, create a 2mb unformatted bios at head, and the rest as a btrfs container. I'll reinstall Windows on another drive at some point when I need it
<mfiano>
Anything else you suggest?
<clever>
mfiano: that should work
<mfiano>
Also I'm very curious as to why other distributions don't have any problems
<clever>
mfiano: my primary desktop is a 2 drive raid for nixos, and a 3rd drive for windows
<clever>
mfiano: its very puzzling how just shrinking the bios boot partition breaks booting, because re-running nixos-install should fix it, and make it only boot nixos
mk-fg has joined #nixos
mk-fg has quit [Changing host]
mk-fg has joined #nixos
<mfiano>
What partition type should the 2mb partition be? I'm using cfdisk
<disasm>
mfiano: if your system supports EFI, why not just have a 1 GB EFI vfat partition mounted as /boot? Then you don't need to worry about grub being able to talk to the btrfs partition.
<mfiano>
disasm: It does support it, but it doesn't work
<disasm>
oh, lol :)
<mfiano>
I cannot boot from a usb drive with EFI
<clever>
mfiano: one sec
<mfiano>
It just doesn't even detect hardware present, even in the boot menu
<clever>
mfiano: if you install nixos with boot.loader.grub.efiInstallAsRemovable = true; and the vfat mounted to /boot, it should become bootable
<mfiano>
Hmm?
<clever>
mfiano: that tells grub to configure the sata drive as-if it was a removable disk
<clever>
and most bios should then boot it automatically without the special efi vars config
<mfiano>
I have a blank partition table right now. I'm asking if the partition table needs to have type partition type byte set
<clever>
depends on if you want MBR or GPT, and if you want legacy or EFI
<mfiano>
GPT and legacy considering I can't even boot EFI
<mfiano>
the install media that is
mk-fg has quit [Ping timeout: 245 seconds]
<clever>
mfiano: boot.loader.grub.efiInstallAsRemovable lets you do an EFI install when not booted from EFI
<clever>
mfiano: which might fix your EFI issues
<mfiano>
I'll try that
<mfiano>
I have actually never did an EFI linux install before, and I suspect this will be a lot of research and troubleshooting. I need this ready for work on Monday
<mfiano>
So I'm a bit hisitant
<mfiano>
hesitant rather
<clever>
mfiano: we could just do legacy on gpt then, a 2mb bios boot partition, and the rest for btrfs and it should work
<mfiano>
I'm wondering now if I should have swap after biosboot or end of disk
<MichaelRaskin>
I would actually create a 1GiB /boot with FAT anyway
<mfiano>
Why, and why FAT?
<clever>
mfiano: so you can upgrade to efi later on
<MichaelRaskin>
FAT because if you ever want to try migrating to UEFI, you can do a quick attempt and roll back if it doesn't go well
<mfiano>
ok so 2mb bios boot, 1gb FAT /boot, 8gb swap, and rest btrfs sound okay?
<MichaelRaskin>
Also because GRUB2 not reading FAT32 correctly would be big news, and GRUB2 having a limited support for some subvolumes on BtrFS would be a Tuesday.
<clever>
mfiano: yeah, and another optional thing you could try, is leave some empty space after that for windows
dan_b has quit [Ping timeout: 260 seconds]
<mfiano>
Ok
TonyTheLion has quit [Ping timeout: 260 seconds]
<mfiano>
How can I install gparted on the kde live media?
<clever>
mfiano: nix-env -iA nixos.gparted
<mfiano>
I want to set the correct partition byte flags
<mfiano>
Thanks
mounty has joined #nixos
<simpson>
MichaelRaskin: "For you, the day btrfs lost your data was the most important day of your life. But for btrfs, it was Tuesday."
<mfiano>
Well I've lost my fair share of data before, which is why I have periodic cron jobs that backup to 3 remote servers of mine.
<clever>
simpson: i once tried to resize an lvm volume, while an xfs one was mounted, something borked, and it overwrite a decent chunk of an ext3 volume in lvm
<MichaelRaskin>
mfiano: note that I mentioned not irrevocable loss of data, but a problem with reading fancy-featured BtrFS variants by minimalistic GRUB2 implementation
<clever>
mfiano: ive got zfs with redundancy configured, and snapshots
<MichaelRaskin>
Frankly, at some point I was in a situation where BtrFS losing my /nix was just another Tuesday for _me_.
<mfiano>
I use borgbackup hourly for data deduplication so i can roll back to any hour with minimal space usage
<mfiano>
across 3 remote machines
<simpson>
clever: Yikes. I've had LVM incidents. It is a sharp and multifaceted footgun.
<MichaelRaskin>
I just have nightly backups. Hourly is too advanced for me.
Arcaelyx_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mfiano>
btrfs will just be a local redundancy measure, so I'm not scared, and haven't lost data with it yet
<clever>
simpson: my only guess, is that xfs looked up the mappings, then bypassed lvm, and got confused when lvm began shifting blocks around
cmacrae has joined #nixos
<MichaelRaskin>
I think some of my dead BtrFS partitions were not really dead, I could mount them read only, I just couldn't fix them to be r/w and not fail after two operations.
<MichaelRaskin>
As it was /nix/store, I didn't care enough to investigate.
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vAwzc
<NixOS_GitHub>
nixpkgs/master 3d8a664 Peter Hoeg: neomutt: 20171215 -> 20180223
NixOS_GitHub has left #nixos [#nixos]
<MichaelRaskin>
But I would say that BtrFS advanced a lot w.r.t. stability, and it definitely did have a space for advancement.
cmacrae has quit [Ping timeout: 245 seconds]
<mfiano>
Anyway, I've had my eye on nixos for a couple years. Installed it in a VM once or twice and I really like it. I've been using Arch Linux since 2001 when it was released, and I dislike how much it has gone downhill in the last year or 2.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to master: https://git.io/vAwzg
<NixOS_GitHub>
nixpkgs/master 7d7b7a0 Shea Levy: libgpgperror: Add derivation for gen-posix-lock-obj to help in cross-compilation
NixOS_GitHub has left #nixos [#nixos]
<mfiano>
For example, the latest binary nvidia driver is not picked up by the latest chromium, and thus everything is rasterized on the cpu, and video decoding as well. even moving my mouse cursor is laggy. Nearly 20 years without any issues whatsoever on the same install, and then about 10 things pile up. That's why I'm switching
kelleyNif has joined #nixos
kelleyNif has quit [Client Quit]
<clever>
mfiano: one of the major strenghts of nixos, is that once you know a rev of nixpkgs that works good, you can always go back to that, no mater how bad things get, at the cost of downgrading to the versions within that rev
<MP2E>
well unfortunately we're not immune to regressions either. But you can always rollback
acarrico has quit [Ping timeout: 252 seconds]
<mfiano>
All software is horrible and regressions and new bugs are a way of life. I hear that. But yes, nixos let's me roll back.
<MichaelRaskin>
The real strength is that you can rollback, then still install that one package you really need — from a fresh version
<mfiano>
Arch Linux has just got completely unusable for me after almost 2 decades of use, and it's quite sad
<mfiano>
But there are better now.
<MichaelRaskin>
When I learned of NixOS, it was an instaswitch (ten years ago)
<MichaelRaskin>
The fact that I was trying to get something similar by manipulating unionfs in a A/B LFS system did play a role…
<samueldr>
not long after actually looking at nix, and nixos, I HAD to switch, didn't regret it
<mfiano>
MichaelRaskin: Curious, being a Lisp guy, what made you choose NixOS over GuixSD?
<MichaelRaskin>
GuixSD didn't exist back then?
<mfiano>
Fair point.
<MichaelRaskin>
And then, the extra extra purity of deblobbing Linux kernel of all blobs for add-on processors — I never got englightened enough to be able to keep track of hardware on that level
<mfiano>
For me, it's mostly staying away from the GPL
<clever>
MichaelRaskin: and what is guix going to do with microcode blobs and meltdown?
<MichaelRaskin>
Also, I am not actually that pure a Lisp person — I have written shell scripts — complicated shell scripts — before writing anything large in Lisps
<MichaelRaskin>
clever: good question, but I think the main protection against Meltdown is still KPTI
<clever>
MichaelRaskin: and spectre
<clever>
MichaelRaskin: i think there might have also been something about KPTI ruins performance, but the microcode can fix it faster
<MichaelRaskin>
Spectre is just a mess, really.
goibhniu has quit [Ping timeout: 260 seconds]
<samueldr>
MichaelRaskin: nothing wrong with shell scripts :)
<clever>
spectre-meltdown-checker is also in nixpkgs, so you can test your boxes
<MichaelRaskin>
clever: I think on some generations KPTI can be implemented as-is. And on some generations speedup comes from tuning caches for KPTI
coot has joined #nixos
<MP2E>
is there a tool I can run to figure out why an executable is not running? The steam package seems to be broken for me but nobody else can reproduce the issue I have... it just gives me "~/.local/share/Steam/steam.sh: line 755: ~/.local/share/Steam/ubuntu12_32/steam: No such file or directory"
<gchristensen>
anyone want to buy a town with me?
<clever>
MP2E: are you using the hardened linux kernel
<MichaelRaskin>
Basically, for _build scripts_ around shell-based build systems I actually prefer shell to Guile.
<MP2E>
I don't think so, my kernel is just "linux = pkgs.linuxPackages_4_14.kernel;"
<MP2E>
(with linuxPackages overriden as well)
<clever>
MP2E: if you override boot.kernelPackages, it will set the kernel automatically
<clever>
MP2E: if you run ls -lh ~/.local/share/Steam/ubuntu12_32/steam, what does it say?
<mfiano>
how big should my nix store be for a graphical desktop?
<MP2E>
it shows this :
<clever>
mfiano: for my laptop, its using 62gig, and my desktop, 46gig
<MP2E>
-rwx------ 1 cray nogroup 4.4M Nov 17 2015 /home/cray/.local/share/Steam/ubuntu12_32/steam*
<MichaelRaskin>
mfiano: I like GPL for many kinds of things, AGPL not so much because I don't understand it.
<MichaelRaskin>
For store size, remember a simple fact: a glibc upgrade means you need to store both all the old stuff and all the new stuff at once.
<MP2E>
"/home/cray/.local/share/Steam/ubuntu12_32/steam: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux.so.2, for GNU/Linux 3.2.32, BuildID[sha1]=12e429cc13f7db94a201d1ecf83bd7bdff5b2457, not stripped"
mk-fg has joined #nixos
mk-fg has quit [Changing host]
mk-fg has joined #nixos
<mfiano>
I've read horror stories about 40gb being too small to build some things. I just want to be safe but have as much unallocated space at the end of my disk as I can get away with
<clever>
MP2E: what does this output? cat /proc/config.gz | gunzip | grep IA32
<MichaelRaskin>
These 40GiBs are probably needed in /tmp
<MP2E>
"CONFIG_IA32_EMULATION=y" and "CONFIG_IA32_AOUT=m"
<MP2E>
2 lines
mbrgm has quit [Ping timeout: 240 seconds]
<clever>
MP2E: yeah, 32bit support is on, are you using the steam chroot stuff from nixpkgs?
mbrgm has joined #nixos
<MP2E>
yep! just installed it with 'nix-env -f ~/nixpkgs -iA steam'. with ~/nixpkgs being a git checkout of nixpkgs that was last pulled a week ago
<MP2E>
I have also tried deleting ~/.local/share/Steam and all .steam* related files
<MP2E>
and ran it again
<MP2E>
it installs and then fails with the same error
<clever>
MP2E: one min
<MP2E>
stranger still, my laptop with a nixos config that is basically the same as this PC, but with a few changes for hardware, runs it just fine
<MP2E>
alright no problem
mk-fg has quit [Ping timeout: 252 seconds]
<clever>
MP2E: try you try it again from nixpkgs f607771d0f5 ?
<MP2E>
sure thing
sigmundv has joined #nixos
asuryawanshi has joined #nixos
<mfiano>
so to clarify, that's FAT16 and not FAT32 or exFAT for the 1GiB /boot partition?
<clever>
mfiano: thats what my EFI system partition looks like when inspected
<mfiano>
as far as I know, vfat is used to denote both FAT16 and FAT32, just using a _f 32 switch during the creation process in the case of the latter.
<mfiano>
-F 32 rather
mk-fg has quit [Changing host]
mk-fg has joined #nixos
asuryawanshi has quit [Ping timeout: 245 seconds]
<MichaelRaskin>
Yes
mk-fg has quit [Ping timeout: 256 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 3204061 Ryan Mulligan: assh: 2.6.0 -> 2.7.0...
<NixOS_GitHub>
nixpkgs/master 529ea1f Michael Raskin: Merge pull request #35504 from ryantm/auto-update/assh-2.6.0-to-2.7.0...
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vAwgE
NixOS_GitHub has left #nixos [#nixos]
gorre has joined #nixos
<gorre>
Hello! Is this channel only for NixOS (like the operating system) realted talks...or also for the Nix package manager (in general)?!
<MichaelRaskin>
Nix+*, really.
<samueldr>
with a list of (package) outputs, assuming I can generate a correct nix list, is there a way to then get a list of paths for those?
<samueldr>
I don't know how I can "just run nix"
<MichaelRaskin>
map (x: x.outPath) ?
<samueldr>
how do I run that nix code?
<samueldr>
that may be my main issue :)
<samueldr>
I'm still learning the nix language
<MichaelRaskin>
What you have?
d4g_ has joined #nixos
Supersonic has quit [Disconnected by services]
<samueldr>
okay, let me rephrase, given nix code, how can I run it? nix-env seems to be to manipulate, like install/uninstall
<gchristensen>
gorre: Nix, Nixpkgs, NixOS the whole Nix ecosystem :)
<gchristensen>
gorre: if your question is super off topic, we'll suggest a different place
Supersonic112 has joined #nixos
<gorre>
Thanks! So, OK...I'm running Ubuntu 16.04 LTS and I installed Docker via `nix-env --install --prebuilt-only docker-17.12.0-ce`. I'm under the assumption that after running that command, that's all I needed to have Docker installed...but I'm always getting: Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?
<samueldr>
"instantiate store derivations from Nix expressions" I thought it meant that was used to "create stuff in the store" :/
<samueldr>
thanks MichaelRaskin!
Supersonic112 is now known as Supersonic
<MichaelRaskin>
It is not even about language…
<MichaelRaskin>
It is about model of evaluation
<clever>
gorre: that command only installs the docker programs, it doesnt configure any docker services
<clever>
gorre: if you want docker to work outside of nixos, you have to manually run the daemon as root, and do all the normal config stuff
jophish___ has joined #nixos
<MichaelRaskin>
There are Nix expressions. They can be evaluated to something, that might include derivations (objects representing stuff that needs building). nix-instantiate primarily writes the derivations to store
<jophish___>
Hi all
<clever>
gorre: but if your on nixos, its as simple as services.docker.enable = true; and your done
d4g has quit [Ping timeout: 252 seconds]
acarrico has joined #nixos
<MichaelRaskin>
But nix-instantiate doesn't launch builds, so that's where --eval-only should be added naturally
<MichaelRaskin>
And I guess Nix 2.0 «nix» command has a new split of responsibilities among subcommands
<jophish___>
oops, ignore that
<samueldr>
MichaelRaskin: it may have been me assuming from the synopsis, because I was swearing it must be it that I saw used for such tasks
<samueldr>
that'll teach me to use `man`
<samueldr>
:)
<mfiano>
clever: I'm not sure if this partitioned correctly. I'm seeing flags I didn't set after a reboot
<gorre>
@clever So, basically, I have to install more stuff...I'm used to the "normal" install
<clever>
mfiano: its possible that your bios is doing funky things like messing with the partition table to "fix" it for you
<MichaelRaskin>
samueldr: maybe this part actually requires the full manual
<samueldr>
well, I'm always referring to the three manuals, but yeah, definitely should actually *read* the nix manual now that I know how to understand it :)
<mfiano>
clever: like i made the first partition unformatted 2mb, and parted is showing it's type as "grub2 core.img". Also, the 1GiB boot partition has the flag "msftdata", and my btrfs-formatted partition is now of an unknown type
Guanin has joined #nixos
Guanin has quit [Remote host closed the connection]
<clever>
mfiano: it sounds like gparted just looked inside sda1, and found the core.img within
<mfiano>
My guess is that the formatting operations were queued but didn't run for some reason. Probably due to the layout changing and kernel not detecting them.
<mfiano>
So it was just reading the flag from the previous table
<clever>
mfiano: id recomend fully closing gparted once your done, then re-opening it to confirm things stuck, and closing again
<clever>
mfiano: then formating things
<clever>
mfiano: ive also seen a friend try to repartition with both fdisk and gparted open at once, and they where overwriting eachothers tables
markus1189 has quit [Ping timeout: 248 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy pushed 2 new commits to master: https://git.io/vAw2Z
<NixOS_GitHub>
nixpkgs/master 87270fb Shea Levy: gnu-efi: Only build on efi-enabled systems
NixOS_GitHub has left #nixos [#nixos]
<mfiano>
I did that actually
mrkgnao has joined #nixos
pkill9 has quit [Ping timeout: 252 seconds]
<mfiano>
It seems I can't even remove the flag. Setting another flag is okay, but removing that other flag just puts msftdata back
<mfiano>
Maybe this is needed for FAT32?
<gchristensen>
shlevy: you are just killing it
<clever>
mfiano: i think thats a result of the uuid type code that was set when creating
cmacrae has quit [Ping timeout: 245 seconds]
<MichaelRaskin>
On GPT you are expected to have _some_ data type, and 0700 is the default
<shlevy>
gchristensen: Slowly but surely getting to systemd :)
<gchristensen>
[ "x86" "arm" ] <- does arm imply aarch64?
<shlevy>
gchristensen: Don't think so, but gnu-efi is currently only building on those two right now
<gchristensen>
I see
<shlevy>
Probably straightforward to get it on aarch64
<gchristensen>
well... I don't know the details, but I'm pretty sure I use create an efi aarch64 image
* gchristensen
places a double-dose of i-have-no-idea-what-im-doing.dog
gorre has quit [Quit: Leaving]
<mfiano>
clever: Ok I'll try this. Is there anything special I have to do to install with EFI using that 1GiB boot partition?
<clever>
mfiano: if your not booted under efi, you have to set the install as removable flag i mentioned earlier
<mfiano>
clever: and that's it other than mounting that fat32 under /mnt/boot as before?
<clever>
mfiano: and run nixos-generate-config --root /mnt, so it knows about that fat32
<chocopuff>
gchristensen: I'm reading Dr. Dolstra's thesis, maybe it was sphalerite_ that referred me to it, he was the guy that laid most of the foundation?
<gchristensen>
chocopuff: yep!
<chocopuff>
I'm assuming, he made it *for* his thesis
<chocopuff>
so nix arose from academic research?
<gchristensen>
yep
<chocopuff>
oh man, that's cool to learn
<gchristensen>
yeah?
<chocopuff>
I feel like this is pretty revolutionary research
<gchristensen>
(Eelco, Dr. Dolstra) is still very involved with Nix and the ecosystem
jooni has quit [Quit: Bye]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar opened pull request #35519: tree-wide: autorename gnome packages to use dashes (master...gnome-normalisation) https://git.io/vAw2R
NixOS_GitHub has left #nixos [#nixos]
MP2E has quit [Remote host closed the connection]
<mfiano>
clever: When is the 1-2mb first partition needed? I noticed your EFI is the first partition on your nvme...
<clever>
mfiano: the bios boot partition is only used for legacy booting
<mfiano>
I see
<clever>
mfiano: with legacy on MBR, grub puts a 512 byte stub into sector 0
<clever>
mfiano: and then it puts core.img between sector 0 and partition 1, in 'unused' space
plakband has quit [Quit: Lost terminal]
<mfiano>
and it wouldn't cause problems with UEFI if i upgrade?
<clever>
mfiano: but gpt uses more then the 1st sector
<clever>
and they are against ugly hacks of using "unused" space
<clever>
so you just make a proper partition to hold that data
<clever>
efi just ignores the bios boot partition
<clever>
instead, it loads the efi system partition, and boots the .efi file from that
<Guanin>
Hi all, I am currently trying to setup a centralized user management (so that I can reuse system users all over different services) and plan to use PAM. As far as I understood from earlier questions and my own research, the unix pam library uses a setuid helper binary to check passwords, but deliberately fails when it is invoked as a non-root user to authenticate a different user. Would you either patch the helper binary (just one
<Guanin>
line needs to be change) or would you chmod the helper binary and /etc/shadow (because group permissions are not dropped, setgid would help here)?
<clever>
but typically, the bios can only boot an efi image configured in the efi vars
<Guanin>
sorry, that became longer than expected
<mfiano>
clever: THanks, I'm going to try using the option you suggested now
MP2E has joined #nixos
<clever>
Guanin: here is a custom pam client i had worked on a few years ago
<clever>
Guanin: i think this file is also involved somehow
<clever>
Guanin: oh, correction to what i said above, hsdm is running as root when it does these calls, and will drop root to the chosen user once it confirms things
<Guanin>
so you do not run into any problems, correct?
<clever>
yeah, that program didnt have any pam problems
nuncanada has quit [Quit: Leaving]
<clever>
other then pam not handling stray signals well
<clever>
haskell uses a timer signal to implement lightweight threads
<clever>
and if pam gets interrupted with any signal,returning an error code that says to try again, pam instead treats it as a hard failure
<clever>
Guanin: have you considered just having normal pam setup everywhere, but using a nixos config file in git to keep them in sync?
plakband has joined #nixos
<Guanin>
what do you mean with normal pam setup?
mk-fg has quit [Ping timeout: 245 seconds]
<clever>
Guanin: ignoring the normal centralized user stuff in pam, and just using git to sync the nixos config that generates users
<Guanin>
Is it possible that we are talking about slightly different things? I want to use the users in /etc/passwd to be able to login on a gogs/gitlab instance or to access directories with http basic authentication, using the same username/password in all places
<clever>
Guanin: ah, yeah, thats a bit different
<Guanin>
So I do not have the problem that I want to feed multiple hosts with the same user
<clever>
Guanin: i was thinking more about things like not even having users in /etc/passwd, and looking them up over the network
<Guanin>
Ok, that would be solvable by putting PAM modules into the config
<Guanin>
e.g. pam_ldap
<clever>
ah, ldap, thats what i was trying to think of
<Guanin>
(I know that ldap _might_ also be a solution for my problem, but it provides way too many features and I failed all five times trying to set it up)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to master: https://git.io/vAw2x
<Guanin>
So, I know that unix_chkpwd drops the privileges for a good reason: to prevent brute force attacks from other users on the same system
<Guanin>
and I know that my problem is probably solvable by either removing that privilege drop, or by chmodding and chowning /etc/shadows and unix_chkpwd
<Guanin>
But to be honest, I'm too new to nixOS to know how to do any of the two approaches
jophish___ has quit [Remote host closed the connection]
<Guanin>
And I do not know which of the two ways would be the cleaner one
mrkgnao has quit [Ping timeout: 260 seconds]
<mfiano>
clever: Heh I got an error for efi installation
<gchristensen>
anyone know how to use go2nix with vendored dependencies?
mrkgnao has joined #nixos
<mfiano>
Ignore that. I messed up the boot partition
<samueldr>
nice, now that I'm not using nix-env on each of 4k pkgs, I can, with parallelism, get all narinfo downloaded from the nix-locate query in 17s :D
<gchristensen>
:o
mk-fg has joined #nixos
mk-fg has quit [Changing host]
mk-fg has joined #nixos
<gchristensen>
!tofu
<{^_^}>
To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected.
<samueldr>
someone needed a specific amount of zeroes
* gchristensen
shifty
<joepie91>
hehe
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master f4a0c57 Franz Pletz: Merge pull request #35514 from rnhmjoj/mitm...
<NixOS_GitHub>
[nixpkgs] fpletz pushed 3 new commits to master: https://git.io/vAwaC
<NixOS_GitHub>
nixpkgs/master 35e8aef rnhmjoj: mitmproxy: remove old patch
<samueldr>
beforehand, it took me 45 minutes to get my first estimate of the total size, because it was: 2 nix-env call per package, with one download per peckage (narinfo), without any parallelism
<samueldr>
45 minutes to 17s is great improvement imho
<gchristensen>
shlevy: can you revert your gnu-efi thing? it broke master
<gchristensen>
shlevy: PS may I recommend a PR so you don't break master? :)
* BlessJah
wonders if nixos would make a good router (VM)
<mfiano>
Well, installation was a success, but I can't boot. But not NixOS's fault this time.
<samueldr>
at least two users here are actively using nixos as their routers
<clever>
BlessJah: i run it on my router, and we have #nixos-on-your-router
<samueldr>
but not as VM
<shlevy>
gchristensen: One sec...
<gchristensen>
BlessJah, samueldr: there is a whole channel for it, #nixos-on-your-router
<samueldr>
gchristensen: I know, I'm idling there :D
<shlevy>
gchristensen: I ran a tarball eval I thought
<mfiano>
clever: Did an EFI install with the option you recommended, rebooted, switched firmware to not use CSM, and now my firmware does not see any boot devices. I can't even change the order in the UEFI. However, it boots fine in legacy mode.
<mfiano>
Something is strange with my motherboard, not letting me boot at all in EFI mode
<gchristensen>
shlevy: this won't catch all the ways master can be broken, but it covers several
<shlevy>
Thanks
<gchristensen>
shlevy: but is there a reason to push right to master?
<clever>
mfiano: that reminds me
<shlevy>
No, I just didn't realize ofborog was ready
<clever>
mfiano: i heard somebody mention in here a week ago, about a bios that could only boot windows
<samueldr>
mfiano: which motherboard is it?
<clever>
mfiano: it was just hard-coded to the path windows uses
<dpc>
ln: failed to create symbolic link '/nix/store/hnr5cvavqcd12iiclcpqy2vvwz2pry7d-neovim-ruby-env/bin/bundle': File exists
<dpc>
builder for ‘/nix/store/lv4khjybbjhh4dgzbjg2xk3bgrgiw4za-neovim-ruby-env.drv’ failed with exit code 1
plakband has quit [Quit: leaving]
<gchristensen>
shlevy: well it doesn't gate, but it is ready in that it has been doing reliable PR checking for 4mo now
<samueldr>
oh, try copying grubx64.efi to /EFI/boot/bootx64.efi in the ESP (so at /boot/EFI/boot/bootx64.efi)
<dpc>
What do I do? :(
<shlevy>
Ah OK. I'll start using PRs then. But we should gate ;)
<mfiano>
samueldr: ASRock Z87 Fatal1ty Killer
<samueldr>
/EFI/boot/bootx64.efi is the default boot program for EFI
<gchristensen>
I agree :) thanks!
<samueldr>
I'm a bit "forced" on a crappy EFI to put refind there
<samueldr>
(cheap cherry trail atom tablet that loves to lose all EFI config)
<mfiano>
samueldr: I'm an EFI newbie. I don't know what you mean by ESP
<samueldr>
and then, if bootx64.efi doesn't work, try what clever linked
* samueldr
shudders
<mfiano>
My problem is that when I switch off CSM, I have no boot devices. The internal drives nor any flash drive plugged in cannot be seen in the firmware menu or its boot selection menu.
<clever>
samueldr: yeah, lol
<mfiano>
Does that make sense to anyone?
<clever>
mfiano: try copying the grub.efi file to EFI\Microsoft\Boot\bootmgfw.efi within the efi partition
<samueldr>
the fact the external drives cannot be seen is weird, but sometimes it's *yet another option*
<clever>
mfiano: you will probably also want to change the \'s to /'s
<mfiano>
Ok I have to reboot into something usable first. brb
Drakonis has quit [Read error: Connection reset by peer]
<shlevy>
gchristensen: That command worked fine... what am I supposed to see in the output?
<gchristensen>
if all goes well, nothing, but in the out-paths result file a list of all outputs for four archs
<shlevy>
gchristensen: Did you check with my latest commit?
<clever>
mfiano: the above copy can be done without nixos-install, just mount it and copy the file, then umount
<samueldr>
it saddens me how the open implementation (tianocore) can be made to work flawlessly on non-EFI hardware (chromebooks) while vendors ship hapazard broken implementations still
<samueldr>
mfiano: then make a mental note: this specific motherboard, `ASRock Z87 Fatal1ty Killer` only boots efi programs at the "microsoft" location
<clever>
samueldr: ive messed with the EFI within virtualbox, and it has a usable UI for editing the boot entries!!
<samueldr>
clever: do you think it would be wise to *also* copy the grub program there?
<clever>
samueldr: as in, i can edit the efibootmgr type stuff, from the firmware
<samueldr>
clever: IIRC it's the same implementation I'm using on my chromebook that's used in virtualbox
<clever>
samueldr: id say just copy it everywhere until it starts working, lol
<mfiano>
ha
<samueldr>
clever: sorry, I meant, for the usb drive for the nixos installer
<clever>
samueldr: ah, maybe
<mfiano>
Ok now I just have to fix my config. Its booting into some unusable wm with a small terminal, when i told it i3
<samueldr>
mfiano: progress is good :D
<clever>
mfiano: ctrl+alt+f1 to get a text console
<clever>
mfiano: then you can edit the config and nixos-rebuild switch followed by systemctl restart display-manager
<mfiano>
I know but what wm does it use hmm
<clever>
samueldr: i need to poke around some more and get efi working under qemu
<clever>
Feb 24 22:57:32 basilisk display-manager[1052]: (EE) VESA(0): V_BIOS address 0x0 out of range
<mfiano>
I am using the dummy 'auto' dm, telling it to start i3 directly. it works in legacy mode
<clever>
Feb 24 22:57:32 basilisk display-manager[1052]: (EE) Screen(s) found, but none have a usable configuration.
<clever>
mfiano: i think the EFI is blocking the video drivers
<mfiano>
So now what is EFI doing wrong? :/
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy opened pull request #35522: Only build libseccomp on supported systems (master...seccomp-unsupported) https://git.io/vAwVy
<samueldr>
though, I'll probably try looking at how impossible it would be to port to either my cheap atom tablet or my main laptop
<samueldr>
now that I have the hardware to unbrick
<clever>
samueldr: this motherboard also has a non-standard double-male usb cable, that connects a special usb port on the mobo, to another pc running special software
<samueldr>
(though, first I'll do a dry run of backing up and recovering the bios)
<clever>
samueldr: then using USB HID packets, it can change the bios settings, simulate the power/reset buttons, and monitor the temps and fan speeds
<samueldr>
clever: is it a non-standard double-male USB or a standard double-male USB? recently learnt that it *is* standard
<clever>
samueldr: even if the motherboard is "off"
<samueldr>
though there's no negotiation with that standard cable, or something akin to that
chocopuff has joined #nixos
<clever>
samueldr: ive been told that its violating the standards to make such a cable
<clever>
samueldr: you have to hit a special button on the motherboard, to switch it into device mode on that port
<samueldr>
I've read that all my life
<mfiano>
this motherboard is horrible. id rather have fun configuring nix than being tortured with a locked down firmware anymore
<clever>
its the same port that handles reflashing itself
<samueldr>
but I bought one for chromebit shenanigans
<samueldr>
and was wondering why it seemed to be sold
<clever>
samueldr: oh, and the stick firmware on this motherboard can take bitmap screenshots of itself, and save them to usb
<clever>
stock*
<samueldr>
I can't find the source of my assertions though
<samueldr>
but while it's unusual and won't do anything by default, USB-A to USB-A male to male is standard
schoppenhauer has quit [Ping timeout: 252 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to master: https://git.io/vAwwG
<NixOS_GitHub>
nixpkgs/master 3c57e77 Shea Levy: Only build libseccomp on supported systems
<NixOS_GitHub>
[nixpkgs] shlevy closed pull request #35522: Only build libseccomp on supported systems (master...seccomp-unsupported) https://git.io/vAwVy
NixOS_GitHub has left #nixos [#nixos]
alex`` has joined #nixos
<clever>
samueldr: ah
<samueldr>
though, while searching, I found out that USB *3* has A-male-to-male in the spec, maybe that's what they were thinking about
<clever>
mfiano: if your using the slim display manager, i think it was f1 or f3 to cycle between the options, and its white on light blue, so its hard to read
<mfiano>
I'm not
<clever>
and auto-login skips that
<mfiano>
I'm using the dummy "auto" dm
<clever>
mfiano: there is an xterm desktop manager that may be enabled by default
<jophish_>
Looks as though I should update the daemon
<clever>
jophish_: ahh, that explains things more
<clever>
its the daemon trying to wrap up the stdout of curl, and ship it over the nix protocol
sigmundv has quit [Ping timeout: 245 seconds]
<mfiano>
thank you. I can auto boot to i3 now
<jophish_>
clever: updating the daemon to 2.0 fixed everything :)
izdmar has joined #nixos
<clever>
jophish_: i should start spreading that to all of my machines then
chrios has joined #nixos
<Li[m]>
> The problem is that we cannot use it in any way, because it's unnamed... joke!
<Li[m]>
From the nix pills #5. Did anyone laugh?
izdmar has quit [Remote host closed the connection]
abbe has joined #nixos
<abbe>
Hi!
<abbe>
I'm a new user of NixOS. I have a question regarding adding new expressions, or modifying existing expressions, e.g. trying to update a package to its latest version.
<clever>
abbe: thats normally done with package overrides
chrios has quit [Ping timeout: 260 seconds]
<abbe>
okay, but how about if I like to make changes, so that I can contribute them upstream (nixpkgs) ?
<clever>
abbe: fork nixpkgs, clone your fork, and then edit the files in that fork
<clever>
abbe: if you run `nix-build -A hello` in the root directory of the fork, it will build a test of the hello world package
<abbe>
oh!
<abbe>
thanks!
<abbe>
that's what I was confused about
asuryawanshi has joined #nixos
endformationage has quit [Quit: WeeChat 1.9.1]
asuryawanshi has quit [Ping timeout: 245 seconds]
<mfiano>
I can't seem to find the right option(s) to turn on the proprietary nvidia driver.
<clever>
mfiano: services.xserver.videoDrivers
asuryawanshi has joined #nixos
<mfiano>
and it will fetch the binary even though it's not open source?
<clever>
mfiano: you can use the keywords from this list to activte it
<clever>
mfiano: it will fail to download because its closed source, and the error explains what you have to add to configuration.nix, to enable unfree software
<clever>
mfiano: oh, and the nvidia drivers may not have trouble with efi
<clever>
it might only be the basic vesa driver that had trouble
<mfiano>
Ha, I'll retry that in a bit then
<mfiano>
So far I am falling in love with nix even if it has a long initial writing of a config step, that's once, and iteratively edited thereafter
<samueldr>
let's say that your motherboard doesn't help :)
jophish_ has quit [Remote host closed the connection]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] samueldr opened pull request #35528: installer: Makes iso image compatible with *wrong* UEFI. (master...feature/bootmgfw.efi) https://git.io/vAwoK
NixOS_GitHub has left #nixos [#nixos]
<samueldr>
let's hope this can be merged, and will be merged
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vAwoX
<NixOS_GitHub>
nixpkgs/master a60774c Peter Hoeg: tome4: 1.4.9 -> 1.5.5
NixOS_GitHub has left #nixos [#nixos]
asuryawanshi has joined #nixos
<clever>
samueldr: ive yoinked the grub efi file from my laptop, and stuck it into the qemu image, to confirm if i'm even using the efi shell right
<samueldr>
and the verdict is?
<clever>
waiting for the 60 second hang
<clever>
it boots
<clever>
just typing fs0: and bootx64 into the shell runs bootx64.efi
<clever>
but without /boot, grub cant find normal.mod, so its stuck in rescue mode
<clever>
i didnt copy enough of it
<samueldr>
by... this day... well... after one sleep cycle, I should have *something* with the manpages, I at least have the full nar archives + restored roots :)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy opened pull request #35529: perl: Work around cross-compilation issue with gcc 7 (master...perl-gcc6-cross) https://git.io/vAwod
<samueldr>
what's left for the first draft is 1) keeping only $out/share/man 2) converting into a format 3) listing :D
chrios has joined #nixos
mrkgnao has quit [Ping timeout: 256 seconds]
ryantm has quit [Ping timeout: 248 seconds]
chrios_ has joined #nixos
<mfiano>
I sort of wish nixos-generate-config was smart enough to copy the mount options. I just realized btrfs isn't even mounted with compression and other options, so I'll have to redo it all after I make my config.
<clever>
it has to query the subvolume, and persist that option correctly, it it wont even mount the right one
chrios has quit [Ping timeout: 245 seconds]
<clever>
mfiano: i prefer zfs, and compression is a setting on the sub-volume, that persists between reboots, so you cant forget it so easily
<mfiano>
Yeah it is only applying the subvol mount option and nothing else to hardware-configuration
<clever>
mfiano: you can also set just one option in configuration.nix, and nixos will merge them
<clever>
fileSystems."/".options = [ "iwantcompression=yes"; ]; for example
<clever>
list based options concat between files
asuryawanshi has joined #nixos
<mfiano>
Anyway, is it considered bad practice to move hardware-configuration to the main config file so you can write your own instead of having it wrongly generate things like that and the frequency scaling governor? The manual says never to edit the hardware-configuration file
<clever>
mfiano: its mostly personal preference
<clever>
mfiano: the only real point i see with hardware-configuration, is to make it easy to re-generate the basic mount stuff with nixos-generate-config
<clever>
but your free to just never use that again, and manage it all yourself
<mfiano>
Right ok
<clever>
and you can just edit hardware-configuration, just be aware that it gets overwritten, if you run nixos-generate-config by mistake
Rusty1_ has quit [Quit: Konversation terminated!]
<mfiano>
Well it looks like NixOS has the same problem that drove me to it from Arch
<mfiano>
I'll have to figure out how to removei t later
<clever>
nix-shell changes dont persist
<mfiano>
chromium uses 100% of all my cores to play videos and to browse certain websites because it is not compatibile with my nvidia gpu or something....but not on all distros: https://i.imgur.com/YwYEckZl.png
<mfiano>
i can't even type into the address bar without it feeling like i'm thrashing ram to disk
<samueldr>
clever: wow no return... that's not useful
<clever>
mfiano: what kind of fps does this get?
<samueldr>
hmmm, mfiano I would have expected such failure to happen with older hardware...
<mfiano>
Oh all other apps work fine. I'm a game developer and I get a perfect GL context elsewhere
<simpson>
clever: Obligatory reminder that glxgears is not a benchmark and that the output of glxinfo is more likely to tell you whether a driver's configured.
<samueldr>
though unhelpful, on a GTX570 it google-chrome seems fine here
<samueldr>
mfiano: have you started with an empty and fresh profile?
<mfiano>
Yes
<clever>
simpson: yeah, and i typicaly expect lower FPS numbers to be better with glxgears
<samueldr>
:/
<clever>
simpson: having 200 fps means you lack v-sync, and the cpu is going nuts rendering a simple thing crazy-fast
<clever>
simpson: but 60 fps means you have v-sync working, and the gpu is probably doing its job
<simpson>
Sure.
<mfiano>
for the record, i get about 200,000fps with glxgears
<mfiano>
(i disabled vertical blank sync)
<clever>
mfiano: definitely sounds like the gpu drivers arent , or that
Guanin has quit [Ping timeout: 240 seconds]
<clever>
mfiano: one min
<mfiano>
This is some incompatibility with newer chromium or nvidia binary...or other. It doesn't occur with ubuntu or mint's stable repo...which tends to hold back a year or 2 for most stuff
<mfiano>
It is a problem on Arch as well. Leads me to believe some bleeding edge incompatability
<clever>
mfiano: try closing chrome fully, then run chromium --disable-gpu
<mfiano>
ok?
asuryawanshi has quit [Ping timeout: 252 seconds]
<clever>
mfiano: does that make any difference?
<mfiano>
Yeah, everything is disabled under gpu, and i'm unable to test my game due to unable to get a webgl context :)
<mfiano>
also super slow
<clever>
mfiano: ok, lets see...
<samueldr>
mfiano: do you have anything for me to test to see if my chromium has accel
<samueldr>
I have one computer with nvidia drivers, and one with intel drivers
<mfiano>
samueldr: yes
<samueldr>
on both, both chromium and google-chrome
<mfiano>
clever, samueldr: Thank you for the help though. I think I'm just going to have to use Firefox until this passes. I only really need chromium once a week for my programming group (Hangouts doesn't support Firefox).
asuryawanshi has joined #nixos
asuryawanshi has quit [Ping timeout: 252 seconds]
<mfiano>
Hmm, my ~/.xinitrc is not being sourced, probably because of the "auto" display-manager. Anyone know how I can have that sourced as usual?
<clever>
there are half a dozen ways of setting env variables in nixos
<clever>
environment.interactiveShellInit just inserts raw strings into /etc/bashrc
<mfiano>
The important bit is them being set before dependency directives needing them
<clever>
environment.shellInit inserts raw strings into /etc/profile
<clever>
programs.bash.shellInit also goes into /etc/profile
<clever>
and environment.loginShellInit
<clever>
environment.variables.A = "1"; deals with escaping and directly handling variables, and environment.extraInit bypasses that and just does raw strings again, both go into a special set-environment script
<mfiano>
You have a lot of nice examples :)
<clever>
and /etc/profile sources set-environment
asuryawanshi has quit [Ping timeout: 240 seconds]
<clever>
i have so many gists that i had to find a dedicated service that deals with finding your own gists, lol
<clever>
it says i have 715 gists....
<Havvy>
clever: Gotta prune 'em once in a while.
<mfiano>
haha yeah. I have over 2000. I learned that the day I found out there is no api to batch delete them, so i just left them there.
<Havvy>
Is there a way to see what the Hydra instance that builds for the nixpkgs unstable channel is currently doing?
<clever>
Havvy: hydra.nixos.org
mrkgnao has joined #nixos
<Havvy>
500 Internal Server Error
<Havvy>
DBIx::Class::Storage::DBI::catch {...} (): DBI Connection failed: DBI connect('dbname=hydra;user=hydra;','',...) failed: could not connect to server: No such file or directory
<mfiano>
I have a lot to learn. I have a very complex .bash_profile/.bashrc system for adding all my vars and appending to path etc. It would be nice to use the nix way
asuryawanshi has joined #nixos
Arcaelyx has joined #nixos
<lopsided98>
Havvy: its broken right now, but hopefully should be fixed soon
simukis has joined #nixos
<mfiano>
clever: stupid question, how do i set multiple of say environment.shellInt variables?
<mfiano>
shellInit rather
<clever>
mfiano: use a ''mult-line string''
<mfiano>
Oh I wasn't aware of that
<mfiano>
Nice
<clever>
both "strings" and ''strings'' accept multiple lines
<clever>
but ''these ones'' will strip out any common spaces between each line
<clever>
so if every line is indented by 4 spaces, the resulting string has no indentation at all
<mfiano>
Not sure I like that though, otherwise the indentation is going to get added
<mfiano>
and I like my nix config to look pretty :)
<clever>
thats the point of ''double single'' strings
asuryawanshi has quit [Ping timeout: 256 seconds]
<clever>
it strips the indent that makes the nix config pretty, keeping the output pretty as well
<mfiano>
Oh just read. Now that is nice
knupfer has joined #nixos
<mfiano>
clever: is environment.shellInit going to expand ''export XDG_CONFIG_HOME="$HOME/.config"'' or have a literal $HOME? I want it to be a literal so it applies to all users
<clever>
mfiano: ${rustc}/bin/rustc will be the path to the binary, even if its not installed
Tobba has quit [Ping timeout: 268 seconds]
robstr has quit [Remote host closed the connection]
robstr has joined #nixos
<clever>
router> restarting systemd...
<clever>
oh god, lol
<clever>
!tofu
<{^_^}>
To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected.
<hyper_ch>
and I always thought tofu is something to eat
<clever>
and the router survived a switch from 16.03 to 17.09!
<mfiano>
clever: still, I get messages when I open a terminal: -bash: rustc: command not found
<clever>
mfiano: then the string i gave wasnt in a nix config file
robstr has quit [Ping timeout: 240 seconds]
<mfiano>
Hmm?
<clever>
mfiano: ${rustc} will expand to the full path of the rustc files, if placed in a nix file
<clever>
mfiano: but anywhere else, its just an unset bash variable
<mfiano>
yes, but I can't assume rustc will be installed, and when it's not I get the error
<clever>
nix will automatically download rustc if you use what i gave above
<mfiano>
error: undefined variable ‘rustc’ at /etc/nixos/configuration.nix:13:34
<clever>
in configuration.nix, it needs to be ${pkgs.rustc}
<clever>
forgot that small detail
<mfiano>
Ah
evhan` is now known as evhan
<clever>
Feb 25 07:48:44 router postgresql-start[8288]: DETAIL: The data directory was initialized by PostgreSQL version 9.4, which is not compatible with this version 9.6.6.
<clever>
also, nix makes it trivial to fix problems like this
<clever>
mfiano: this allows you to access toe nightly, beta, and stable channels
<mfiano>
Yeah, I'll save that for later as I don't even know what an overlay is yet. I was studying Rust but it's not my programming language of choice.
<clever>
Feb 25 08:01:35 router systemd[1]: Started PostgreSQL Server.
<clever>
and the tweak i did above is now in place, my router has postgress again
<sphalerite_>
clever: Dell's firmware, at least the one on the XPS 15 9560, also allows editing boot entries inside the firmware. The UI isn't optimal but it does work, and it saved me a whole lot of effort when I got my motherboard replaced and all the settings were gone.
<clever>
sphalerite_: nice, ive never seen that on real hardware
<sphalerite_>
it's got a graphical file browser for it and everything
<clever>
sphalerite_: i saw similar in virtualbox
<clever>
sphalerite_: also, i just used `nixops deploy` to upgrade a nixos 16.03 (from nixos-unstable) machine to nixos 17.09, and it experienced minimal downtime
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vAwPP
<NixOS_GitHub>
nixpkgs/master f5d145a Peter Hoeg: oath-toolkit: use autoreconfHook
NixOS_GitHub has left #nixos [#nixos]
<clever>
sphalerite_: it wasnt even nixops managed to begin with either!
<sphalerite_>
also, I ran memtester (userspace memtest) on the vocore (that MIPS device I was having trouble with, where running nix-built programs would fail the first 2 or 3 times with various errors — illegal instruction, bus error, segfault…). It was just fine after the first 3 attempts at running it, which crashed
<sphalerite_>
it's still a mystery to me why it would fail 2 or 3 times then consistently work fine
<sphalerite_>
it also doesn't seem to be a caching issue as I thought it might be, since it doesn't matter how far apart in time the first few attempts are
<sphalerite_>
although maybe I should try catting the executable and the libc to /dev/null before running it to rule it out properly
timon37 has joined #nixos
<mfiano>
Interesting. I have the option to enable ssh-agent, and it is indeed running, but it's not remembering keys
<clever>
mfiano: yeah, thats anoying, and why i switched to gpg-agent
<mfiano>
Oh I never used that for ssh keys
<clever>
mfiano: but the nixos option to enable gpg is dumb, and screws up horibly if you ssh into that system
<mfiano>
Is it straightforward to figure out?
<sphalerite_>
mfiano: set the AddKeysToAgent option in your ssh config
<hyper_ch>
sphalerite_: it's not easy "to go to bed"
<mfiano>
and that actually works with the nixos running ssh-agent
<mfiano>
but not without that, and i was hoping to get rid of it :)
<sphalerite_>
mfiano: it would work just fine without cleanup if you used /run/user/$(id -u) instead of ~/.ssh in it :)
<mfiano>
True, when I wrote that it was on a different platform though, and I was hoping to get rid of that hack.
<sphalerite_>
iirc I did it without sourcing its output, and just set SSH_AUTH_SOCK then passed it to ssh-agent via -a
<sphalerite_>
then I switched to gpg-agent because I couldn't use my yubikey with ssh-agent
asuryawanshi has quit [Ping timeout: 256 seconds]
<mfiano>
hmm
<mfiano>
everyone is mentioning gpg-agent being better
<sphalerite_>
I don't think it's better
<sphalerite_>
it just has one feature that I need
<sphalerite_>
I'd still be using ssh-agent if it didn't
<mfiano>
How are you configuring ssh-agent exactly?
<mfiano>
Because it's not remembering anything for me without my hack
<mfiano>
I have AddKeysToAgent for each host in my config
<sphalerite_>
I think I had something like export SSH_AUTH_SOCK=${XDG_RUNTIME_DIR:-/run/user/$(id -u)}/ssh ; ssh-add -l &>/dev/null ; [[ "$?" = 2 ]] && ssh-agent -a "$SSH_AUTH_SOCK"
<mfiano>
Well, better than my hack for sure, assuming that works
<sphalerite_>
just putting `Host *\n\tAddKeysToAgent yes` at the end of your .ssh/config should do it I think?
asuryawanshi has joined #nixos
<larsvm>
has someone had any luck on installing nixos on raspberry pi?
<larsvm>
sphalerite_: i've followed that guide but to no avail
<larsvm>
hangs on "Starting kernel..."
<larsvm>
and the advice regarding that does not seem to be right
<mfiano>
sphalerite_: heh that errors in nix config anyway
<larsvm>
since the guide advices to change extlinux.conf so that it's got console=ttyS0 but it already has
<larsvm>
and it has console=ttyAMA0, too
<sphalerite_>
mfiano: put '' before the ${
<mfiano>
It's actually contained in that
<sphalerite_>
mfiano: it's because of nix's antiquotation syntax coinciding with bash's extended parameter expansion
asuryawanshi has quit [Ping timeout: 245 seconds]
<sphalerite_>
yes, put another '' before the ${ to escape it within the '' quoted string. It's rather confusing, but that's how it works
<sphalerite_>
larsvm: sorry, I don't actually know any more than the existence of that page :(
asuryawanshi has joined #nixos
<mfiano>
sphalerite_: It works great, thanks
<mfiano>
Oh actually
<mfiano>
I get an error when i login
<mfiano>
If I su -, I get:
<mfiano>
bind: No such file or directory
<mfiano>
unix_listener: cannot bind to path: /run/user/0/ssh
<jluttine>
Updated nixos-unstable nixpkgs yesterday. Now sddm doesn't start..
<jluttine>
sddm-helper exited with 11
<sphalerite_>
ah right because it's in the shell init for all users but you're not starting up a full session for root
<mfiano>
What's a fix?
<sphalerite_>
maybe guard the whole thing with if [[ -d ${XDG_RUNTIME_DIR:-/run/user/$(id -u)} ]]
Guest_ has joined #nixos
<larsvm>
do you know who would know?
<larsvm>
meta knowledge
<sphalerite_>
larsvm: maybe Dezgeg
Guest_ has quit [Client Quit]
asuryawanshi has quit [Ping timeout: 252 seconds]
<larsvm>
thanks
<mfiano>
sphalerite_: Thanks!
<sphalerite_>
mfiano: there might be some cases where that doesn't behave the way you want it to though
<mfiano>
Oh?
<mfiano>
hmm
<sphalerite_>
in particular, your agent socket will probably get removed when you log out
<sphalerite_>
which may or may not be what you want
<mfiano>
Define log out
<sphalerite_>
close all sessions (graphical, console and SSH)
<sphalerite_>
you can do loginctl enable-linger to make sure that your runtime dir continues to exist even if you don't have any active sessions
<mfiano>
Only time that ever happens is when I reboot anyhow. This is a single-user system.
asuryawanshi has joined #nixos
<sphalerite_>
I had to do this to make my tmux not break on a remote server since I was putting its socket in /run/user/1000 as well
<mfiano>
and I don't use any display-manager even
<sphalerite_>
right so whenever you close all your console sessions it will delete /run/user/$UID unless you've enabled lingering
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] velovix opened pull request #35531: lxterminal: switch to using GTK3 (master...lxterminal-to-gtk3) https://git.io/vAw1T
NixOS_GitHub has left #nixos [#nixos]
<mfiano>
sphalerite_: I always have X running as my user, which implies always being logged in, but you got me thinking about my remote server with tmux that will eventually be switched over to nix
<sphalerite_>
you can either run tmux with its socket in a different path (which is the default), or enable lingering
<sphalerite_>
it only goes to /run/user/$UID if you set it to using a wrapper or something. Although I think programs.tmux.enable does it too
<sphalerite_>
(programs.tmux.enable should be removed IMHO since it does nothing that needs to be systemwide)
<mfiano>
Ok
<mfiano>
anyone else running on a hidpi display?
<mfiano>
For some reason my mouse cursor is absolutely tiny. Everything else is the right size thanks to Xft.dpi being set, but I can't make out where my cursor is most of the time being this small
asuryawanshi has quit [Ping timeout: 256 seconds]
<sphalerite_>
me!
<mfiano>
Am I missing something? :)
MP2E has joined #nixos
<sphalerite_>
Xresources config for XCursor. Hang on
<sphalerite_>
(this is impure and not really great, there might be a proper nixos option for it by now)
<genesis>
we could reach 7 hundred open PR soon :D
<sphalerite_>
export XCURSOR_PATH=/run/current-system/sw/share/icons in my .xsessionrc, and some breeze package containing the cursor images in my systemPackages
<mfiano>
umm wow
<mfiano>
Why did I never have to do this on Arch? :)
hotfuzz_ has quit [Ping timeout: 240 seconds]
<mfiano>
Xcursor.theme: Vanilla-DMZ\nXcursor.size: 64 doesn't have any effect at all
<sphalerite_>
you also need the XCURSOR_PATH env var set
asuryawanshi has joined #nixos
<sphalerite_>
it doesn't need to be set for everything though, you can test it on just one application
<sphalerite_>
(provided that application supports XCursor)
<mfiano>
yeah i do
<mfiano>
it has no effect :)
<sphalerite_>
did you xrdb -merge <~/.Xresources before starting it?
<mfiano>
yes
<sphalerite_>
and /run/current-system/sw/share/icons/Vanilla-DMZ exists?
<mfiano>
I have that done automatically in services.xserver.displayManager.sessionCommands
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 7 new commits to master: https://git.io/vAw1F
<NixOS_GitHub>
nixpkgs/master 746b36c Kosyrev Serge: ghc841 | http-api-data: SMP fix not on hackage yet
<NixOS_GitHub>
nixpkgs/master 65169aa Kosyrev Serge: ghc841 | uri-bytestring: TH boundaries fix author made no PR yet
NixOS_GitHub has left #nixos [#nixos]
<mfiano>
NoHmm nope
<mfiano>
So what icon is it using then...
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 2 new commits to master: https://git.io/vAw1N
<NixOS_GitHub>
nixpkgs/master 624366d Kosyrev Serge: ghc841: bump to 20180224
<NixOS_GitHub>
nixpkgs/master d1a7cb8 Peter Simons: Merge pull request #35521 from deepfire/ghc84-20180224...
NixOS_GitHub has left #nixos [#nixos]
<sphalerite_>
probably X's builtin ones
<sphalerite_>
if you add vanilla-dmz to systemPackages it should work
asuryawanshi has quit [Ping timeout: 252 seconds]
<sphalerite_>
oh and you did replace the \n with a newline right?
<mfiano>
sphalerite_: yes and no
vidbina has joined #nixos
<mfiano>
the pointer cursor is large now, but the text selection cursor is still very tiny
<mfiano>
Also in chromium, the pointer is small still
viric has quit [Ping timeout: 256 seconds]
<mfiano>
in regular xorg, only the arrow icon has changed
<sphalerite_>
hm
<sphalerite_>
I don't know everything about how XCursor works and docs have been hard to find :/ so I'm not sure I can help you any further. I would like to see it fixed too though
<sphalerite_>
and especially have it configured declaratively, there's so much stateful crud in this setup
<mfiano>
Yeah this is strange.
Guanin has quit [Read error: Connection reset by peer]
Guanin_ has joined #nixos
asuryawanshi has joined #nixos
<mfiano>
Something is completely messed up anyway
<mfiano>
I can't run X apps anymore
<mfiano>
Invalid MIT-MAGIC-COOKIE-1 keyCan't open display :0.0
<mfiano>
such as xrandr
rgc has joined #nixos
<sphalerite_>
I think that's an Xauthority issue
asuryawanshi has quit [Ping timeout: 240 seconds]
<mfiano>
Yeah I do too, but I deleted .Xauthority and rebooted
<sphalerite_>
actually the manual XCURSOR_PATH setting shouldn't be necessary since f80eeb5d854cc5b0ae6f66616c1f29aabf0e0279
<mfiano>
Right now I'm more worried about not being able to execute apps
<sphalerite_>
how are you running X?
<mfiano>
The same as I was when it was working. I have the "auto" display-manager set to run "i3" as "mfiano"
<sphalerite_>
oh, did your xsessionrc exist before adding XCURSOR_PATH to it?
<NixOS_GitHub>
[nixpkgs] griff opened pull request #35536: nixos/dovecot: More sieve options and precompiling of scripts (master...dovecot-precompile) https://git.io/vAwQw
<srhb>
Olgierd: If you're adding a new package that has no maintainers, it's definitely preferable if you add yourself to maintainers so you can get poked if it fails or people need help with it :)
<MichaelRaskin>
Well, if you understand the package itself that you add… Like, what is called «works properly»
<srhb>
manveru: None of them seem to be breaking though. Hmm.
asuryawanshi has quit [Ping timeout: 240 seconds]
<manveru>
where do you have that archive/tar code?
<srhb>
manveru: That's from docker.
<srhb>
manveru: In pkgs/build-support/docker/default.nix: ln -sT ${docker.src}/components/engine/pkg/tarsum src/tarsum
asuryawanshi has joined #nixos
abrar has quit [Quit: WeeChat 2.0]
abrar has joined #nixos
<srhb>
manveru: The solution for now might just be to pin go to 1_9. I was just curious :)
<srhb>
Out of curiosity, is there any kind of automation that I can employ to prevent a commit from reaching stable, without necessarily reverting it in master?
<MichaelRaskin>
You mean, ever reaching stable 18.03?
<srhb>
No, I meant backport :0
<srhb>
I'm guessing no since that's a by-hand process anyway.
<MichaelRaskin>
Put a «Do not backport because X» into the commit message
<MichaelRaskin>
Most things do not ever get backported anyway.
<srhb>
Hmm, I guess I can comment on the commit...
<srhb>
Right.
asuryawanshi has joined #nixos
robstr has joined #nixos
<magnetophon>
Maybe a stupid question, but what is the point of stable, if stuf doesn't get backported?
<srhb>
Security releases are focused mainly on stable. And a lot of things *do* get backported.
<gchristensen>
magnetophon: turning the question over, what is the point of stable if everything is backported?
<magnetophon>
gchristensen: checkmate! :D
kragniz is now known as kgz
<magnetophon>
good point
<gchristensen>
wow! I should play chess, that was easier than I expected
<magnetophon>
srhb, MichaelRaskin: thanks
dan_b has joined #nixos
__monty__ has joined #nixos
<__monty__>
I just ran nix-env -u --leq and it seems like my firefox has been downgraded? I was on 58 installed by nix but now the about menu says it's firefox nightly 57.
<srhb>
Why --leq?
dan_b has quit [Ping timeout: 252 seconds]
<vcunat>
-u by itself won't "update" a package if its version hasn't changed
<vcunat>
Still, dependencies might have been updated, etc.
<srhb>
I see.
<sphalerite_>
nix-env -u sucks anyway :p
<vcunat>
I typically use `--always -uA` with an attrset I defined, but there are many ways.
ryanartecona has joined #nixos
asuryawanshi has joined #nixos
Tobba_ has quit [Read error: Connection reset by peer]
lorne has joined #nixos
<__monty__>
srhb: I have some buildEnv's and they won't update without it.
asuryawanshi has quit [Ping timeout: 245 seconds]
<__monty__>
I don't understand how I could've gotten nightly 57 though, that's not even in nixpkgs.
<srhb>
Then I guess it must have been there before you did anything.
<srhb>
Assuming the nonexistence of magic. :-)
asuryawanshi has joined #nixos
<__monty__>
No, I think it's because it got all the packages from my nixos channel instead of nixpkgs. I changed NIX_PATH to just .nix-defexpr/channels.
<__monty__>
Damn, have to do an entire rebuild : (
<__monty__>
Is there a way to keep my NIX_PATH definition but still specify the order of the channels?
<hyper_ch>
btw, when I use nixos-rebuild swith -I nixpkgs=/path/to/local/repo then I won't need an --upgrade option as it will take that repo and it's checkout?
<srhb>
I don't have that problem. My battery charge lasts for approximately two seconds...
<srhb>
hyper_ch: It will never git pull or anything for you, no.
<srhb>
hyper_ch: I think it might stupidly call nix-channel --update anyway though.
robstr has quit [Remote host closed the connection]
<hyper_ch>
(and I did get used to gigabit way too easily)
robstr has joined #nixos
Itkovian has quit [Client Quit]
<gchristensen>
I was making the connection that 6,680hrs is roughly how long it takes to birth a child
asuryawanshi has quit [Ping timeout: 240 seconds]
<hyper_ch>
vcunat: btw, is there a way to actaully use TO as a full dns server that will request from root dns servers directly and not from ISPs or Google ones?
<vcunat>
hyper_ch: yes, click a checkbox in the Foris UI
<vcunat>
(DNS / disable forwarding)
<hyper_ch>
I see
<hyper_ch>
well, 8.8.8.8 is google tracking you and quad9 is also operated by the London City Police and others.....
<vcunat>
I still consider it a bad decision that it wasn't the default on Omnia.
<hyper_ch>
so I've been pondering to setup my own trusted DNS resolver
<hyper_ch>
but that will only finally work when dns resolving uses tls as well as ISPs can still interfere easily right now
<vcunat>
Well, privacy against ISP is a hard thing.
<vcunat>
Tunnel to a trusted provider is probably the only way.
<hyper_ch>
that's why: running your own dns resolver that uses dns over tls for upsream
<hyper_ch>
there's rfc7858 though
<vcunat>
These don't run on _authoritative_ servers.
robstr has quit [Ping timeout: 252 seconds]
<vcunat>
(aren't standardized for them)
iyzsong-x has joined #nixos
<vcunat>
You may forward over TLS to 9.9.9.9 for example.
mkoenig has joined #nixos
<vcunat>
And BTW, in https handshake, the hostname is in clear text ;-)
<yorick>
can I do overrideDerviation old: {/*fixes*/; meta.broken = false;}?
asuryawanshi has joined #nixos
michas_ has joined #nixos
asuryawanshi has quit [Ping timeout: 245 seconds]
<michas_>
Hi, I just installed nixos and installed the way-cooler and sway package. Both install fine, but both are not able to start, because they cannot find `swrast_dri.so` - shouldn't nix install all dependencies and therefore avoid that kind of error?
<srhb>
yorick: For that you probably want overrideAttrs (old: { ...; meta.broken = false; })
<yorick>
srhb: yeah, found it
<srhb>
:)
asuryawanshi has joined #nixos
ambro718 has quit [Quit: Konversation terminated!]
<michas_>
Does my description make sense? Is anyone able to reproduce that behavior?
<symphorien>
michas_: I have heard that opengl is still impure on NixOS, for technical reasons
katona has quit [Quit: Connection closed for inactivity]
<michas_>
sphalerite_: Thanks at lot. Indeed that fixed it. :)
<sphalerite_>
simpson: that's not the problem, it's just that the option is hidden and as such not in the manual or the options search
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] yegortimoshenko opened pull request #35552: imgpatchtools: init at 0.3 (master...20180225.155732/imgpatchtools) https://git.io/vArfB
NixOS_GitHub has left #nixos [#nixos]
asuryawanshi has quit [Ping timeout: 252 seconds]
<symphorien>
michas_: then you might want to create a page on the wiki about sway or way-cooler: https://nixos.wiki
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar pushed 1 new commit to master: https://git.io/vArfE
<NixOS_GitHub>
nixpkgs/master 20c5476 Ryan Mulligan: babl: 0.1.42 -> 0.1.44...
NixOS_GitHub has left #nixos [#nixos]
<simpson>
sphalerite_: Good point. IIRC configuration.nix used to be auto-populated with commented-out helpers and useful tips which covered this stuff, but maybe that's no longer the case.
<bgamari->
elvishjerricco, to be honest I'm not entirely sure of the semantics of --compiler
<sphalerite_>
simpson: you mean the one generated by nixos-generate-config? It still includes a basic set of examples of what you might want, but those are all non-hidden options as well
asuryawanshi has joined #nixos
<bgamari->
what you have done there looks plausible
<bgamari->
so if it works then LGTM
<elvishjerricco>
bgamari-: Yea it fixes cabal file conditions on GHC version, so cabal2nix seems to support this correctly.
strobelight has quit [Quit: gtg]
<simpson>
sphalerite_: Sure.
vcunat has joined #nixos
<sphalerite_>
oh, the term is actually "internal" as opposed to "hidden". But the effect is the same — it doesn't end up in the manual or the options search. An issue which I will remedy right away :D
<gchristensen>
or: anyone know of a way to do this without the infinite recursion :)
periklis has joined #nixos
<simpson>
Yikes.
jensens has joined #nixos
<elvishjerricco>
gchristensen: The module system is probably strict in `config`, since it differentiates between modules that do `{ foo = bar; }` and `{ config = { foo = bar; }; }`
<elvishjerricco>
Which would mean that it's strict in `cfg.class` in this case, which recurs on being strict on `config`
<gchristensen>
ahh
<gchristensen>
this is why mkIf exists
<LnL>
gchristensen: you can't use config directly
<LnL>
gchristensen: plans."${cfg.class}" could override packet.class but it needs it to evaluate
rauno has joined #nixos
<gchristensen>
ack!
vidbina_ has joined #nixos
asuryawanshi has quit [Remote host closed the connection]
<gchristensen>
thanks elvishjerricco, LnL!
vidbina has quit [Ping timeout: 252 seconds]
<sphalerite_>
michas_: there seem to be modules dedicated to both of the compositors you mentioned though, so you shouldn't need to set hardware.opengl.enable = true; yourself, instead set programs.{sway,way-cooler}.enable = true;
KABA has joined #nixos
<michas_>
sphalerite_: What exactly does it mean to "enable" a program? What is the right kind of documentation for that?
<sphalerite_>
michas_: `man configuration.nix` or the options search at https://nixos.org/nixos/options.html (they contain the same information, although the manpage will contain newly introduced options and stuff while options.html is always for the current stable nixos channel)
<sphalerite_>
michas_: for the most part you'll just want to enable stuff by putting it in systemPackages but some stuff just has extra systemwide dependencies (like the OpenGL setup for wayland compositors) which can't be represented in the package and as such has its own configuration module
<sphalerite_>
same goes for anything that needs setuid, setcap, or systemd services
<symphorien>
michas_: on NixOS installing a program will only make the executables availables in the PATH. it will never have side effects like enabling a systemd service, install logrotate rules and so on. A NixOS option and module is needed for that. So for a window manager, web server etc. installing the program is not enough, you need to swich on nom NixOS option which is usually called "something.programname.enable =
<symphorien>
true"
ryanartecona has quit [Quit: ryanartecona]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar pushed 1 new commit to master: https://git.io/vArUL
<NixOS_GitHub>
nixpkgs/master a31d98f Jan Tojnar: tree-wide: autorename gnome packages to use dashes
NixOS_GitHub has left #nixos [#nixos]
verrchu has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] lheckemann opened pull request #35554: nixos/opengl: unhide and document enable option (master...opengl-unhide) https://git.io/vArUG
NixOS_GitHub has left #nixos [#nixos]
<verrchu>
Hi guys i guess i am kinda stuck with the installation
<KABA>
Any info on the hydra server? When will it be back up?
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master b5c200b volth: jdk: use oraclejdk on aarch64
<verrchu>
and now i am in the system i suppose. But it doesn't look like i installed it. Because if i unplug the usb and reboot my computer it is waiting for a bootable device
<sphalerite_>
verrchu: did you follow the installation instructions in the manual?
<symphorien>
you are in the installer, not in your installed system
<LnL>
jtojnar: should I look at that meson stuff for darwin or do the changes not apply there?
jensens has quit [Remote host closed the connection]
<hyper_ch>
btw, the current hydra hiccups - what are the consequences? Just no prebuilt stuff?
<jtojnar>
LnL: the patch only concerns itself with rpath
<hyper_ch>
hi LnL
jensens has joined #nixos
<LnL>
hi
<jtojnar>
LnL: not even sure what adds the libraries to rpath on Linux, no idea how it works on Darwin
<hyper_ch>
LnL: for my tmux problem you did a git bisect - what did you actually do there?
oahong has joined #nixos
<sphalerite_>
hyper_ch: no channel updates
<hyper_ch>
sphalerite_: hmmmm, it just keeps compiling everything for me now :(
<hyper_ch>
well, on old commits.. it's been compiling stuff for the last 10h or so
<sphalerite_>
hyper_ch: if you're on a channel you should still be getting binaries
<sphalerite_>
oh, then you're probably on a revision that was never the state of a channel
ryanartecona has joined #nixos
<hyper_ch>
sphalerite_: ah ok
<LnL>
jtojnar: the stdenv/cc-wrappers add that, on darwin ld uses the install_name of libraries instead and links against that and we use absolute paths for that instead of @rpath
Tobba has joined #nixos
<michas_>
sphalerite_: Ah, thanks makes sense. However I cannot find programms.sway.enabled in the man page and on options html. - Should they really list all options for all packages? (Or are those also hidden options?)
<LnL>
hyper_ch: sphalerite_: yes, no caches implies that
<sphalerite_>
michas_: are you on nixos-17.09 or nixos-unstable?
raynold has joined #nixos
<KABA>
What determines what nix-env -iA something will use as nixpkgs source? I changed in my configuration.nix the nixpath variable so nixos-rebuild uses that, but nix-env still want to use some channel and I would like to not use -f '<nixpkgs>' all the time
<LnL>
luckily the ofborg infrastructure is separate and now verifies a lot more then before
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sphalerite_>
michas_: the sway module isn't in 17.09, only in unstable and it will be in 18.03
amf has joined #nixos
<michas_>
sphalerite_: good question. :) whatever came with the installer. According to the file name it should be 17.09.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vArTV
<NixOS_GitHub>
nixpkgs/master a035714 Michael Raskin: lispPackages: remove trailing colon from {NIX_LISP_,}LD_LIBRARY_PATH
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 140b163 Michael Raskin: lispPackages: add a few more, escape +
<michas_>
and the web page also refers to 17.09, right?
<sphalerite_>
michas_: yeah you might want to switch to unstable if you want to use wayland stuff, a lot of improvements have been made since 17.09 by the looks of it
<sphalerite_>
yes
<michas_>
can you give me a hint how to switch to unstable? it's the channel thing, right?
<sphalerite_>
Because I found it would become unreachable but was still running
<LnL>
I posed the vm config, no if it's not the default for a isNormalUser
<sphalerite_>
because the /run/user/$UID tmpfs would get unmounted and lost, destroying the socket
<LnL>
the process and it's children disappear
<LnL>
no the socket is in /tmp
<sphalerite_>
but access could be restored by sending USR1 to the tmux daemon
<Mic92>
LnL: you need to enable this manually
<sphalerite_>
huh
<Mic92>
this does not work by default
<sphalerite_>
yeah it still sounds like lingering would fix it. We really need declarative lingering config
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rauno has quit [Ping timeout: 245 seconds]
<hyper_ch>
Mic92: but on every other system I've used I didn't have to do anything
<hyper_ch>
tmux just keeps running
<LnL>
Mic92: nope still dies
i-am-the-slime has joined #nixos
<i-am-the-slime>
Hello. I would like to use fontforge but with a GUI.
<MichaelRaskin>
I think the same problem that systemd broke the very idea of screen/tmux/nohup has already appeared in the past, and some setting has been set…
<MichaelRaskin>
Basically, in any case search through the pkgs/toplevel directory is a good idea
<i-am-the-slime>
I wonder if I'm doing something wrong in general. I always search for the packages, then manually look at the options on github and then go into configuration.nix add the name there and do sudo nixos-rebuild switch
<MichaelRaskin>
Sounds like a valid approach.
rihards has quit [Quit: rihards]
<i-am-the-slime>
Okay. It just feels like I'm doing the computer's work sometimes :D
<MichaelRaskin>
I mean, in general at some point you remember enough out of the Nix and Nixpkgs manuals to use buildEnv and separate expressions for separate activities or something
<MichaelRaskin>
Well, you could try installing nix-index to search for packages.
<i-am-the-slime>
I have nox, and I can search with that. But that doesn't really show me the options.
<i-am-the-slime>
But I'm also fine with this approach right now.
<MichaelRaskin>
In the new and shiny Nix 2.0 (released, not yet in channels) there is nix edit
Arcaelyx has joined #nixos
<MichaelRaskin>
So you can search for attribute name, then open the source.
<MichaelRaskin>
But I think in most cases there is no expectation that you will change the options.
Itkovian has joined #nixos
<MichaelRaskin>
More like pick among fontforge and fontforge-gtk
<samueldr>
so it *technically worked*, only that the webkit-gtk didn't load the files properly
i-am-the-slime has quit [Remote host closed the connection]
ChongLi has quit [Ping timeout: 245 seconds]
<gchristensen>
I wish nixops had a somewhat nicer way to specify where to find nixpkgs
<srhb>
gchristensen: I've been playing around with that. I recently got bitten by restricted mode because I'd been fetching them in my deployments... Still searching for a better model.
<gchristensen>
aye
kitemikaze has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] xeji opened pull request #35564: qemu: add optional virglrenderer and opengl support (master...qemu-virgl) https://git.io/vArLQ
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 2 new commits to master: https://git.io/vArtn
<NixOS_GitHub>
nixpkgs/master 8c45d41 Tim Steinbach: jenkins: 2.89.3 -> 2.89.4
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 0652ed1 Tim Steinbach: sbt: 1.1.0 -> 1.1.1
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vArtW
<NixOS_GitHub>
nixpkgs/master 494d94a Tim Steinbach: jdbc: 5.1.39 -> 5.1.45
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 2 new commits to master: https://git.io/vArtl
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 088f8b8 Tim Steinbach: syslog-ng-incubator: 0.6.0 -> 0.6.2
<NixOS_GitHub>
nixpkgs/master acbb71c Tim Steinbach: syslog-ng: 3.11.1 -> 3.13.2
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vArtB
<NixOS_GitHub>
nixpkgs/master 87658f3 Tim Steinbach: pulseaudio: 11.0 -> 11.1
NixOS_GitHub has left #nixos [#nixos]
chessai has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vArt2
<NixOS_GitHub>
nixpkgs/master 7b2cff9 Tim Steinbach: font-awesome: Add 5.0.6...
NixOS_GitHub has left #nixos [#nixos]
Izorkin_ has quit [Read error: Connection reset by peer]
Izorkin has joined #nixos
michas_ has quit [Quit: Page closed]
Mateon1 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/vArt1
<NixOS_GitHub>
nixpkgs/master fdadd97 Tim Steinbach: keybase: 1.0.40 -> 1.0.42
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vArt9
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 45fcaa6 Tim Steinbach: hicolor-icon-theme: 0.15 -> 0.17
gorre has joined #nixos
joko has joined #nixos
<hyper_ch>
someone's busy
kitemikaze has quit [Remote host closed the connection]
kitemikaze has joined #nixos
<gchristensen>
... lol now _I'm_ getting error: attribute ‘nixosVersion’ missing in nixops, and can't sort it out
<MichaelRaskin>
There is a PR in NixOps
<gchristensen>
oh?
<MichaelRaskin>
Well, there was a cleanup of tagging of NixOS versions in GRUB menu
<MichaelRaskin>
It also renamed the options
<gchristensen>
guh
<MichaelRaskin>
Well, the new putting nixos* into nixos.* makes sense
<MichaelRaskin>
There is a PR in NixOps that updates the interface to NixOS to support both ways
<gchristensen>
great!
<gchristensen>
thank you MichaelRaskin
<MichaelRaskin>
It is stalled, like everything in a Nix* project without twenty committers or so
<MichaelRaskin>
Eelco says that maybe the options should be renamed back or something instead of merging dtz's PR into NixOps…
<gchristensen>
:|
rauno has quit [Remote host closed the connection]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 1 new commit to master: https://git.io/vArqi
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master de149ba Michael Raskin: ripmime: switch of -Werror
knupfer has joined #nixos
cfricke has quit [Quit: WeeChat 2.0]
<MichaelRaskin>
Obviously, he doesn't reject the PR while saying that the rename should be reverted.
ixxie has quit [Quit: Lost terminal]
tmaekawa has joined #nixos
Izorkin_ has joined #nixos
cmacrae has quit [Ping timeout: 268 seconds]
alexteves has joined #nixos
Izorkin has quit [Ping timeout: 245 seconds]
Izorkin_ has quit [Ping timeout: 240 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 5441518 Tim Steinbach: bind: 9.11.2 -> 9.12.0
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vArmW
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] zimbatm opened pull request #35568: nodejs: ship with development headers (master...nodejs-headers) https://git.io/vArmg
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] asymmetric opened pull request #35569: diceware: init at 0.9.3 (master...asymmetric/diceware) https://git.io/vArmo
NixOS_GitHub has left #nixos [#nixos]
<alexteves>
I'm getting a weird error from runghc in a nix-shell under os x (it works in nixos).
<alexteves>
$ runghc Main.hs
<alexteves>
runghc: ramework-path/nix/store/ws9cfrjf7nry1hy1p7612qvdqi87qjwb-apple-framework-PCSC/Library/Frameworks: executeFile: does not exist (No such file or directory)
<alexteves>
"runghc —help" works as expected
<alexteves>
I’m guessing the -framework-path option somewhere is being processed wrong (maybe it’s parsed as -f ramework-path). Any ideas?
chreekat has quit [Ping timeout: 245 seconds]
asuryawanshi has joined #nixos
<LnL>
is this on darwin?
<LnL>
ah, ok
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vArm5
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 3187f3f Tim Steinbach: containerd: 1.0.1 -> 1.0.2
MichaelRaskin has quit [Ping timeout: 260 seconds]
<hyper_ch>
LnL: had a look at that fosdem presentation?
cmacrae has quit [Ping timeout: 248 seconds]
humanoyd has joined #nixos
<sphalerite_>
hyper_ch: I skimmed over it. It seems mostly inferior to nixos :p
betaboon has joined #nixos
robstr has quit [Remote host closed the connection]
<hyper_ch>
sphalerite_: nah, you just need to write according nixos module for it ;)
<betaboon>
hello #nixos, I'm just trying to build a pythonPackage which contains entry_points but the resulting package in nix-store does not contain any py-files. what am i doing wrong ?
<LnL>
kind of skipped it once they started to show a gui
<sphalerite_>
betaboon: that's very hard to divine without any clues as to what your expression looks like :)
chocopuff298 has joined #nixos
<betaboon>
hyper_ch: thanks for the link. i just figure it out. I'm just completly stupid and forgot the __init__.py ...
<hyper_ch>
sphalerite_: you still can't mind-read over the internet?
<sphalerite_>
hyper_ch: after all this time trying, I still can't
<sphalerite_>
they say practice makes perfect though, so I'll keep at it
<hyper_ch>
you can do it. We believe in you!
<betaboon>
sphalerite_: yeah sorry xD turn's out my expressions is perfectly fine and i just forgot the __init__.py in the python package ... should have taken a break at some point Xd
<hyper_ch>
no idea what an __init__.py file is
<LnL>
alexteves: it looks normal at first glance, does it have any special dependencies?
<sphalerite_>
it's a magical file you put into a directory to turn it into a python package
chocopuff has quit [Ping timeout: 256 seconds]
<hyper_ch>
everything uses magic nowadays
<LnL>
python is weird
reinzelmann has joined #nixos
hask_bee_3 has joined #nixos
<hask_bee_3>
Does the binary cache store old binaries too?
<sphalerite_>
yes
<hask_bee_3>
They aren't eventually deleted? (To save storage costs)
<sphalerite_>
not yet, no
<hask_bee_3>
good to hear
<alexteves>
LnL: maybe? it's pointing to a nixpkgs fork, lemme try with the nixos one
<sphalerite_>
I seem to recall hearing that most of the costs are traffic, not storage
<hask_bee_3>
makes sense
<LnL>
alexteves: I mean the Main.hs, a simple hello world worked fine for me
<alexteves>
oh it doesn't even get to read the file, I can literally pass a rubbish name and get the same error
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master ab7a07f Peter Simons: Merge pull request #35321 from ElvishJerricco/fix-haskell-compiler-name-version...
<NixOS_GitHub>
nixpkgs/master a9646b3 Will Fancher: Fix haskellCompilerName version....
<NixOS_GitHub>
[nixpkgs] peti pushed 2 new commits to master: https://git.io/vArOY
<alexteves>
are you on darwin?
<LnL>
what I'm wondering is if that runghc comes from pkgs.ghc or a ghcWithPackages
<LnL>
yep
<hask_bee_3>
Since upgrading to Nix 2, I'm getting this Nix-related error (when doing "stack build" on my Haskell project): dyld: Library not loaded: /nix/store/6a9fr14xc9yr4xwlkyaf4zzqmbh3xngy-gmp-6.1.2/lib/libgmp.10.dylib Referenced from: /Users/haskbee3/.stack/setup-exe-cache/x86_64-osx-nix/Cabal-simple_mPHDZzAJ_2.0.1.0_ghc-8.2.2 Reason: image not found
<hask_bee_3>
Is there any way to grab that particular libgmp manually? Seems like Stack didn't do it for some reason even if it needs it.
<sphalerite_>
based on its store path I'm guessing ghcWithPackages
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jokogr opened pull request #35571: skrooge: 2.10.5 -> 2.11.0 and some changes (master...u/skrooge-2.11.0) https://git.io/vArOG
NixOS_GitHub has left #nixos [#nixos]
humanoyd has quit [Quit: WeeChat 2.0.1]
<LnL>
oh, missed that :)
<clever>
hask_bee_3: that happens when you garbage collect while using stack the "wrong way"
<alexteves>
yeah, it's a ghcWithPackages
<clever>
hask_bee_3: nix has no way to track that those files are in use, and deletes them, so you have to wipe your ~/.stack/setup-exec-cache/
<clever>
hask_bee_3: if things are instead built using nix (stack2nix is an option), then nix knows when things are in use, and wont break it
<hask_bee_3>
clever deleting that dir worked.
<hask_bee_3>
are you saying this is a bug in Stack?
<clever>
hask_bee_3: more that binaries built that way when using nix arent meant to last
<LnL>
alexteves: can you gist the expression, pretty sure it's one of the dependencies that's causing an incorrect flag to be added
null_ has joined #nixos
<null_>
has anyone here installed steam? It needs the 32-bit ligbl for me?
<LnL>
alexteves: if you can't figure out where it's coming from, bisecting the dependencies might help (remove half and see if it works)
betaboon has quit [Quit: WeeChat 1.9.1]
<hask_bee_3>
clever Interesting. Isn't this a bug in Stack then? They're assuming that the packages are gonna stay there. I presume this is a known issue that's on the radar to be fixed?
<clever>
hask_bee_3: it cant really be fixed, its just a side-effect of how nix and nix-shell work
Mateon1 has joined #nixos
hiroshi has quit [Ping timeout: 240 seconds]
hiroshi- is now known as hiroshi
<clever>
hask_bee_3: if you switch over to nix-build, then the problem wont happen
<hask_bee_3>
clever are you proposing I cut Stack out? (I've been thinking about it. It's just that Stack makes it so simple ;) )
<clever>
hask_bee_3: stack2nix lets you use a stack.yaml file with nix
<hask_bee_3>
Interesting. Ok, I'm going to look into this! Thanks.
<hyper_ch>
hi clever, how was the bed?
<clever>
hask_bee_3: good
<clever>
hyper_ch: good, lol
<hask_bee_3>
lol
<elvishjerricco>
clever: hask_bee_3: Stack and Nix have important differences. Mainly, Nix doesn't give you any kind of incremental building, so `stack2nix` is going to require you to recompile the entire project from scratch on any change. Stack's `--nix` should put the proper GC roots in `.stack-work`, so `nix-collect-garbage` won't mess with it. But Stack is really screwy with it's caching under `--nix`, causing it to
<elvishjerricco>
remember old GHC store paths sometimes.
<hask_bee_3>
elvishjerricco i hope this will only be an issue upon major nix upgrades. it happened with me when upgrading to nix 2
<disasm>
well that's interesting... when did chromium start carrying it's own list of trusted certs instead of using the system ca-bundle? Must have been fairly recently.
<elvishjerricco>
hask_bee_3: It shouldn't be Nix updates that cause these problems. It's nixpkgs that would do it
<hyper_ch>
is symantec out of chromium as well?
<null_>
Do I need anything besides the extra kernel modules nvidia_x11 to run bumblebee, I have it enabled and I get [ERROR]Cannot access secondary GPU, secondary X is not active.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 08f8ad2 Nikolay Amiantov: treewide: unquote homepage in my packages...
<NixOS_GitHub>
[nixpkgs] abbradar pushed 57 new commits to master: https://git.io/vAr3i
<NixOS_GitHub>
nixpkgs/master 9813473 Nikolay Amiantov: curaengine: 3.0.3 -> 3.2.1
<NixOS_GitHub>
nixpkgs/master 0a02b55 Nikolay Amiantov: libarcus: 3.0.3 -> 3.2.1
NixOS_GitHub has left #nixos [#nixos]
<null_>
uh maybe i do nvm for a minute I'm bad at searching things
<disasm>
hyper_ch: I see class 1 and 2 G4/G6 authorities under Symantec
<hyper_ch>
hmmm
<hyper_ch>
wasn't all removed?
<hask_bee_3>
elvishjerricco ok. for now, my rule of thumb is: if it fails again (which hopefully will be rare), just delete the ~/.stack/setup-exe-cache dir and try again.
<hask_bee_3>
although i do wonder whether, for very large projects, the garbage collector could kick in consistently and the build always fail
gorre has quit [Quit: Leaving]
<clever>
hask_bee_3: nix will check for files actually in use (in env vars, running, open files)
rardiol1 has quit [Remote host closed the connection]
ryantm has quit [Quit: Lost terminal]
<phry>
heyho,
<phry>
I've got some weird behaviour when adding my own derivation to firefox's "extraNativeMessagingHosts" argument - is someone here right now who could help me? :)
<mfiano>
How can NixOS be MIT?
<phry>
(somehow from that point on, firefox seems to add gcc-wrapper to the build output, but only if I'm using node modules from nixpkgs, but not from the nixos-channel o_O)
cmacrae has joined #nixos
rogue_koder has quit [Quit: Konversation terminated!]
<lejonet>
Mic92: The test case you wanted me to write for the ceph service module, should it just setup the cluster and ensure everything goes nice with that or do you want me to add some cluster interaction in it too? My guess is that no cluster interaction is required as that isn't testing the service module persay, but ceph
null_ has quit [Ping timeout: 260 seconds]
null_ has joined #nixos
<null_>
so yeah, has anyone here set up bumblebee with nvidia successfully?
cmacrae has quit [Ping timeout: 245 seconds]
<null_>
or set up bumblebee for that matter?
<Mic92>
lejonet: It could add a file for example. That way we could ensure, it is not just running but in a failed state.
<null_>
fuck maybe it's working now
<lejonet>
Mic92: whatcha mean "add a file"?
<lejonet>
as in add a file to the cluster?
Acou_Bass has quit [Quit: byeeeeeeeeeeeeeee]
<null_>
yeah I think it's working with services.xserver.videoDrivers = [ "intel" ]; and hardware.bumblebee.connectDisplay = true; my x-server wasn't correctly restarted while testing
<Mic92>
lejonet: or do you need multiple instances to build a cluster?
<lejonet>
Mic92: well, I'm continously checking that the cluster has come up and such thorough the test, ceph-post-file doesn't really give you anything for automated testing atleast
<lejonet>
Mic92: nah, you can do it all on one node, but you don't have to, each daemontype could be on a physically separated machines except ceph-mon and ceph-mgr, you must have a ceph-mgr on each node that has 1 or more ceph-mon
<lejonet>
Mic92: not yet, want to try executing it once before :P
Arcaelyx has quit [Ping timeout: 245 seconds]
ottidmes has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ryantm opened pull request #35579: Auto update/cairo 1.14.10 to 1.14.12 (staging...auto-update/cairo-1.14.10-to-1.14.12) https://git.io/vArGY
NixOS_GitHub has left #nixos [#nixos]
<lejonet>
Mic92: but if we want to, I could write a test with a client, that creates a block device and everything
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ryantm closed pull request #35579: Auto update/cairo 1.14.10 to 1.14.12 (staging...auto-update/cairo-1.14.10-to-1.14.12) https://git.io/vArGY
NixOS_GitHub has left #nixos [#nixos]
<mfiano>
I don't even see a license file on github, but Wikipedia says MIT. What exactly is it?
<sphalerite_>
mfiano: the expressions defining how the software is built are MIT-licenced
<mfiano>
Where is that mentioned please?
null_ has quit [Ping timeout: 260 seconds]
<sphalerite_>
mfiano: COPYING, in the root of nixpkgs
<gchristensen>
another option, of course, is to open a PR fixing it :)
<phry>
okay, as soon as I post my question here I solve it myself.. :) I'll never add $PATH to PATH when creating a script in postInstall again...
<infinisil>
phry: what was the problem?
<disasm>
It seems like something is killing my tmux sessions on my desktop if it's at a login screen. Remote sessions don't die at all if I'm logged in on the desktop.
<phry>
infinisil: I got a lot messages that my output should not rely on gcc-wrapper and stuff - but only on new nixpkgs, which sent me to look at completely unrelated places
FRidh2 has joined #nixos
<phry>
infinisil: in the end, it was just leaking the build-time PATH to an output file -.-
<infinisil>
Ah I see
<amf>
my / device has limited space, and i can't use a symlink for /nix, what are my options?
<LnL>
aristid: it's planned for 18.03 so in a month :D
aborsu has joined #nixos
<mfiano>
Does anyone else experience the issue with chromium that it is NEVER the default browser, and pops open the annoying question every time you start it? I can't figure out how to remove it.
knupfer has quit [Remote host closed the connection]
<dtz>
'afternoon! :)
<sphalerite_>
evening
<hyper_ch>
someone isn't using timeservers
katona has joined #nixos
<LnL>
:)
<joepie91>
just had a bizarre GPU error problem with firefox-bin, switching to firefox somehow fixed it o_O
macslayer has quit [Quit: WeeChat 1.9.1]
macslayer has joined #nixos
sondr3 has joined #nixos
roconnor_ has quit [Ping timeout: 260 seconds]
<mfiano>
What would be a good solution to have my vim configuration be available with both `vim` and `sudo vim` and being able to keep the two in sync in my dotfiles repo?
aardito has joined #nixos
<clever>
mfiano: nixos lets you do that by just not using ~/.vimrc
<clever>
mfiano: this embeds the vim config into the vim binary
Itkovian has joined #nixos
<mfiano>
clever: To be specific nvim
<clever>
mfiano: you can just drop that file into /etc/nixos/ and then add ./vim.nix to the imports list
<mfiano>
and I want to be able to track it in my existing vim dotfiles repository
<clever>
mfiano: oh, and there is also sudoedit
<aardito>
emacs setup question - I'm using exwm with extraPackages and I need to set a variable before some packages are loaded
fragamus has joined #nixos
<aardito>
where can I do that? even at the top of ~/.emacs the packages have already been loaded
<clever>
mfiano: you could also symlink the /root/.vimrc to the other one
<mfiano>
Hmm
noam has joined #nixos
<mfiano>
The problem is I have a several thousand line neovim configuration spanning multiple files, and I want it to be owned by my user in its home directory to be staged in my git dotfiles
<clever>
mfiano: are they mostly in a single directory?
<mfiano>
No
<clever>
mfiano: maybe just make a simple .virmc that does #include /home/clever/.vimrc, and configure that 2nd one to be relative to itself, to find its neighbors?
b has joined #nixos
<mfiano>
neovim convention is to use XDG_DATA_HOME for plugin configuration, scripts, and such, and XDG_CONFIG_HOME for the config file.
<mfiano>
so ~/.config/nvim and ~/.data/nvim on my box
<clever>
mfiano: ah, you could just symlink those 2 nvim directories then, simple
<timokau>
mfiano: If possible you should really use sudoedit instead though
<mfiano>
I never heard of it
<mfiano>
I will look into it to see if it solves my problem
<clever>
mfiano: when you run "sudoedit /root/foo", sudo will copy the file to /tmp, then run $EDITOR without root
<timokau>
mfiano: It basically just copies the file to /tmp and chowns it to your user, then opens an editor with user permissions and finally copies the result back after editing
<clever>
mfiano: and when the editor QUITS, sudo will copy the file back
<clever>
the quit requirement is a bit anoying, but the editor runs as the correct user
<mfiano>
That would break a lot of my nvim functionality I think. I have keybindings to jump to specific files/list specific directories
<timokau>
mfiano: The PWD is preserved if that helps. But yes, it can be inconvenient
<clever>
mfiano: ive heard that emacs also has special plugins, where it will internally call sudo on its own, to gain write access to things, maybe vim has such things?
<timokau>
But otherwise a bug in nvim, any of its plubins or a malicious plugin is a root exploit
<mfiano>
You would use an ex-command to call an arbitrary vim command with sudo
abert has joined #nixos
<disasm>
clever: that is genius :) never doing sudo nvim again, lol -- do you alias it to something shorter?
<clever>
disasm: i use the vim.nix i linked above, so vim just always has the right config
<clever>
disasm: so i just sudo -i and run vim directly as root
<clever>
~/.vimrc just doesnt exist
<catern>
aardito: recent emacs actually has a separate init file that is run before packages are loaded
<alexteves>
LnL: turns out runghc doesn't know the -framework-path flag, so it was interpreting it as -f <path/to/ghc>
chreekat has quit [Quit: quitting]
<catern>
aardito: so if you compile that recent Emacs, you can use that. I think that's probably the best way?
<clever>
alexteves: what is the difference between ghc and runghc?
<M1k3y>
Hello together, I need some help (new to nix development). I successfully build a package for the prometheus grok exporter. But it requires some data from a git submodule at runtime. How can I provide such dependencies at runtime?
abert has quit [Client Quit]
<disasm>
clever: yeah, same here :) but I like the concept of `sudoedit` better from a security standpoint. I need to research it more, but it seems like that would be a good way to give sudo privs to a user that needs to edit files and not allow them to get a shell or edit a different file than the one they requested as a param to sudoedit to enforce proper logging of root actions!
<clever>
M1k3y: set fetchSubmodules=true; inside the fetchFromGitHub
<catern>
aardito: look for early-init-file
<clever>
disasm: yeah, that can be fun
<alexteves>
I don't know specifics but it supposedly runs a .hs file directly, instead of compiling
<mfiano>
sudoedit is pretty nice though. I'm probably going to have to alias that though :)
<clever>
disasm: and in terms of vim plugins gaining root, root controls them fully, via vim.nix
<mfiano>
I like short commands like `vim`, which is an alias for `nvim` :)
<clever>
mfiano: i aliased vi to vim, even shorter!
<disasm>
I'm of the school of thought of sudo -i is only permissible for changes that could potentially render a system seriously broken and only as a backup option to revert said change.
<mfiano>
ha yeah
coot has joined #nixos
<mfiano>
clever: I just don't like calling vim, vi, especially when nocompat is set :P
<clever>
mfiano: i never bothered learning the difference :P
<clever>
mfiano: i just started with vi over a decade ago, and now its just hard-wired into my brain
<mfiano>
hehe, I'm only ahead by a decade.
<M1k3y>
clever: The submodules are getting fetched at build time, no problem there. But I need all the files from this submodule at runtime since they contain matching patterns for log file analysis.
<disasm>
clever: question for ya... I alias nvim -> vi in my zsh config, but then if I sudo vi I don't have it in root config (because I leave root with bash shell). Is there a good way in nixos to specify a global alias across all shell types for some commands that you could set the alias for vi? Otherwise I get vi: command not found if I run `sudo vi` (but sudoedit solves this as well)
FRidh2 has quit [Quit: Konversation terminated!]
<clever>
M1k3y: ah, then the installPhase has to copy those files somewhere under $out, and it has to be patched to know where to look
<clever>
disasm: aliases are handled by bash, and `sudo vi` doesnt check aliases
<clever>
disasm: so you will want to do something like installing writeScriptBin "vi" ''
<clever>
#!/bin/sh
<clever>
exec vim "$@"
<clever>
''
<lejonet>
Mic92: Hmm, is there some way to ensure that the test VM has a certain size on its disk?
<clever>
that generates a vi script that will land in $PATH, so sudo can find it
<M1k3y>
clever: patching won't be a problem since the path is part of the configuration. Any tips where to look for details about the install phase?
<pie__>
clever, hm...im guessing i need to find the source of that mkderivation and see how it sets src?
<mfiano>
I get the rofi dmenu opening, but making a selection does not copy text to the clipboard unless called from bash directly
<mfiano>
works fine in another distro without nix-shell though
<Mic92>
lejonet: tests/docker-tools.nix
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] svanderburg pushed 1 new commit to master: https://git.io/vArWL
<NixOS_GitHub>
nixpkgs/master ceeed46 Sander van der Burg: DisnixWebService: 0.7 -> 0.8
NixOS_GitHub has left #nixos [#nixos]
<Mic92>
virtualisation.diskSize
<clever>
pie__: yeah
<pie__>
christ xD
<pie__>
this seems a bit unnecessarily hard
<clever>
mfiano: bash sources an interactive init file when opening shells, but if you run things directly from the WM, that doesnt get sourced
<sondr3>
hello, I'm running NixOS on my MBP and it's working really well. I have a Nvidia card but I want set some settings in the generated xorg.conf file for it, is there a way to do it in configuration.nix or should I create a separate conf-file for it?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] svanderburg pushed 1 new commit to master: https://git.io/vArWO
<NixOS_GitHub>
nixpkgs/master 7817d87 Sander van der Burg: disnixos: 0.6.1 -> 0.7
NixOS_GitHub has left #nixos [#nixos]
<mfiano>
clever: How does it work outside of nix-shell then?
<clever>
mfiano: when you open a terminal, the shell reads the interactive init stuff, and then the right env vars are setup
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] lblasc opened pull request #35580: buildLuaPackage: propagate modules paths with a setupHook (master...lua-export-path) https://git.io/vArWn
NixOS_GitHub has left #nixos [#nixos]
<lejonet>
Mic92: sweet, thanks
<mfiano>
clever: I understand, but even after a reboot the window manager binding doesn't run the last line correctly
<mfiano>
It's running everything up to that fine anyway
<clever>
mfiano: if you run strings /proc/PID/environ you can peek at the env vars of any process
<clever>
mfiano: check what the demnu process has for an env, its probably missing something key
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] svanderburg pushed 4 new commits to release-17.09: https://git.io/vArWR
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/release-17.09 430e9a6 Sander van der Burg: disnix: 0.7.2 -> 0.8...
<NixOS_GitHub>
nixpkgs/release-17.09 0827a46 Sander van der Burg: dysnomia: 0.7.1 -> 0.8...
<NixOS_GitHub>
nixpkgs/release-17.09 45ea062 Sander van der Burg: DisnixWebService: 0.7 -> 0.8...
<clever>
mfiano: and compare it to what `env` says in a shell
<mfiano>
I can't really get the pid
<clever>
mfiano: you can also query the parent that was launching dmenu
<clever>
mfiano: since dmenu inherits the parent env
<mfiano>
it's not dmenu
<mfiano>
it's rofi running with the -dmenu flag
<clever>
mfiano: you could also try running `env > /tmp/rofi` via rofi, and then `env > /tmp/bash` from a normal shell, then diff the 2 files
<mfiano>
and that indeed works
<mfiano>
the problem is the invocation of the pass command which attempts to store text using xclip
<phry>
mfiano: have you tried adding --pure and then running it from the shell? maybe you are missing a dependency?
<mfiano>
it's not storing the text
<mfiano>
phry: it works from the shell
14WAABGE8 has joined #nixos
<14WAABGE8>
[nixpkgs] Mic92 closed pull request #35580: buildLuaPackage: propagate modules paths with a setupHook (master...lua-export-path) https://git.io/vArWn
14WAABGE8 has left #nixos [#nixos]
94KAAA39D has joined #nixos
<94KAAA39D>
nixpkgs/master 7b76073 Luka Blaskovic: lua packages: propagate libs with a setupHook...
<phry>
btw.: I just finished writing an overlay to use the passb firefox addon in nix - if anyone is interested in something like that. it's my first attempt at writing a derivation, so I'd be happy about feedback as well.
<phry>
good point. actually, I always use that. no idea why i missed it in the docs xD
<mfiano>
clever: Hmm, I can't figure out your env var gist
<mfiano>
I set some in environment.shellInit but I want to make them available inside nix-shell
cmacrae has joined #nixos
<clever>
mfiano: nix-shell --pure disables all of those things
<mfiano>
I'm not using --pure in my script though
<pie__>
clever, this also seems to work: nix-shell-unstable -v -p 'let y = import <unstable> {}; in y.haskellPackages.idris.overrideAttrs (as: {src = y.fetchgit {url="https://github.com/idris-lang/Idris-dev.git";rev="cc7ef9adaf3e7b8dcec420e368b48a120fe07434";sha256="07nzkpr89apz517w9m1fcmp0azbbjhlv9nmjx9w33ly57mi93zfa";};})'
<pie__>
clever, of course im going to put that in a file lol :P
<sphalerite_>
phry: you don't need to unindent the <<EOF bit all the way, nix strips indentation from '' strings
<clever>
pie__: yep, but now you cant modify the src as easily
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] rycee pushed 1 new commit to master: https://git.io/vArlT
<NixOS_GitHub>
nixpkgs/master 48479b6 Robert Helgesson: josm: 13265 -> 13367
NixOS_GitHub has left #nixos [#nixos]
<sphalerite_>
phry: but only the same amount from each line
<pie__>
clever, i cant?
<phry>
sphalerite_: cool, thx :)
<pie__>
oh i see you did set it to the local dir
<clever>
pie__: you would have to edit the rev and hash after every edit
<Guanin>
clever, we talked this morning about PAM, correct?
<clever>
Guanin: yeah
<clever>
pie__: but if you just unpack it locally, you can use src = ./.; and then freely edit anything
<pie__>
right but is that due to the way i did the override or just because i chose fetchgit?
<mfiano>
actually the env vars are getting set in nix-shell, so I'm not sure how to get this crucial script to work :(
<clever>
pie__: the fetchgit
<sphalerite_>
looks good though :)
<pie__>
i mean how does this compare to overridecabal
<pie__>
oh ok re: fetchgit
<clever>
pie__: ah, overrideAttrs vs overrideCabal, i think the difference is which level it overrides at
<sphalerite_>
phry: also great github profile picture
<clever>
pie__: overrideCabal changes the params passed to generic-builder.nix, so it can change haskell level things
<phry>
sphalerite_: ^^ until someday someone sues me for it ;)
* pie__
scratches head
<pie__>
well ok
<clever>
pie__: generic-builder.nix then calls stdenv.mkDerivation, where overrideAttrs can inject its own changes
<clever>
src is just passed down, so it works on either
<pie__>
aha..
<Guanin>
I got it working with patching PAM, but I'm not really happy with recompiling "everything" right now (oh, and sorry for assuming your time zone). As my Patch only affects one helper binary, is there a different way to make this modification? My patch is this: https://clbin.com/7FFOy
<pie__>
ok i roughly get it i thik
<lejonet>
Mic92: I presume that the tests create a nixos-container, is there anyway to afterwards go into it and poke around?
<pie__>
clever, thanks
<clever>
Guanin: its nearly 6pm, and i have slept since we last spoke
<mfiano>
nix-shell is strange... bash: shopt: hostcomplete: invalid shell option name
<lejonet>
Mic92: so that I can troubleshoot why something doesn't work?
<clever>
mfiano: nix-shell uses a more light-weight build of bash that lacks a lot of options
cmacrae has quit [Ping timeout: 240 seconds]
hiratara has quit [Ping timeout: 245 seconds]
<pie__>
clever, your command throws this: error: attempt to call something which is not a function but a path, at (string):1:99
<mfiano>
clever: maybe that's the issue. even if i do -p bash
<Mic92>
lejonet: there was also a way to get into test, but I forgot how. lassulus knows
<clever>
mfiano: you want bashInteractive
<clever>
pie__: oops, i missed an import before <nixpkgs>
<pie__>
oh oops xD
<mfiano>
well that error is gone, but still saying the password i choose is not in the password store. and the env var for the password store is /home/mfiano/.data/pass which is correct
<lejonet>
Mic92: so I could take the node definition I have in the test and run nixos-rebuild -I nixos-config=./testvm.nix build-vm and get an identically specced VM?
<clever>
Mic92: i think the vm leaves a special script in its build dir, and if you -K the build, you can nix-shell into it or something, and re-launch the vm
<clever>
Mic92: and it detects such re-launching, and drops into a shell, rather then trying to re-test itself
<ottidmes>
When are paths copied to the nix store? I just ran into a bug in my config because I references to my overlays by path instead of by string, causing nix-env to error on '/nix/store/<overlayed_package>': No such file or directory.
<pie__>
clever, youre rightbecause they just switched the parser and that was the dep issue
<clever>
ottidmes: if you "/quote/a/path" it wont be copied, its just a string that gets passed to the program
<phry>
mfiano: you could try setting your PASSWORD_STORE_DIR in environment.variables
<clever>
ottidmes: but if you /dont/quote/it at the nix level, it gets copied into the store, and replaced with a storepath
alex`` has quit [Ping timeout: 240 seconds]
<clever>
ottidmes: and inside a string, you need to "do ${./this}" to make it at the nix level
<phry>
mfiano: but you might need to log out of your desktop manager session for that to work
aborsu has quit [Quit: aborsu]
<ottidmes>
clever: But how about imports, those paths are not copied to the store, even though they are not quoted to strings
<macslayer>
Hmmm, this is an odd issue. I'm on 17.09 NixOS, and for some reason, `cargo` and `rustup` aren't working. I try to do something simple like `cargo --version`, and it says "error: no default toolchain configured." If I try to do something like `rustup default stable`, it says "error: missing key: 'url'"
<clever>
ottidmes: some nix functions like import can just read it without copying it into the store
<clever>
ottidmes: and next time you eval, it will read it again
<ottidmes>
clever: So I by default I should assume all my paths are copied over to the nix store, the only exceptions being special builtins like imports
<clever>
ottidmes: yeah
<clever>
ottidmes: the only special things i can think of right now are toString, import, and readFile
<lejonet>
clever, Mic92: How would I add 2 disks to it? it doesn't accept the emptyDiskImages that it accepted for the makeTest one
<clever>
lejonet: extraDisks is a bash variable, not a nix one
jb55 has joined #nixos
<ottidmes>
clever: Thanks for clearing that up for me! BTW the reason why my private keys in initrd config was no longer working was because I switched to grub, which does not support boot.initrd.secrets, while systemd-boot did, which I used when testing that particular config at the time.
<clever>
lejonet: also, try using --fast, its likely a small bug that is misleading you
<clever>
ottidmes: ah, that sounds like something that should be fixed, grub is far better then systemd-boot
<hask_bee_3>
Is it possible in Nix to run an instance of postgresql (pointing at a specific data dir) in the background? My goal is to run postgresql like that through nix -- and the command specified in a systemd file (for reboot on crash or system start).
<hask_bee_3>
Is Nix meant for this particular use case?
<hask_bee_3>
If this is possible, this would be highly useful for maintaining a bunch of postgresql instances on a single Linux machine.
<lejonet>
clever: would seem so, it made everything like it should
<clever>
hask_bee_3: gchristensen actually did something just like that about 3 days ago
<lejonet>
now to get it to use curses interface instead of SDL
<clever>
gchristensen: do you still have that gist link handy?
<ottidmes>
clever: grub already supports it with its extraInitrd option, so if it supports even another extra initrd, that same trick could probably be applied to allow it to have secrets as well, as was mentioned I believe on a issue about it
<clever>
lejonet: pass -curses to the bash script nix generated, or set virtualisation.graphics = false;
<clever>
ottidmes: nice
<ottidmes>
clever: Why is grub objectively better than systemd-boot?
aristid has joined #nixos
Drakonis has joined #nixos
<clever>
ottidmes: systemd-boot dosnt seem to have very many options, and ive always been anti-systemd :P
<clever>
ottidmes: and i know grub inside and out
<hask_bee_3>
hmm, that is for NixOS. I was talking about just running postgresql through Nix, on any Linux distro.
<gchristensen>
hask_bee_3: it isn't for nixos
<clever>
hask_bee_3: it reuses the nixos startup scripts, in a portable way, so it works on any os
<gchristensen>
hask_bee_3: it uses NixOS's tooling to make the database runner
<gchristensen>
what clever said :)
<lejonet>
clever: thanks :) going to do more tomorrow, atleast I've got a semi-functional testcase :P
<mfiano>
phry: The problem is not that variable, but XDG_DATA_HOME. It works if i hard-code the full path in the script, but not if I use that variable in the script, despite seeing XDG_DATA_HOME in `env` inside nix-shell
<mfiano>
phry: ideas?
M1k3y has quit [Quit: Page closed]
<hask_bee_3>
gchristensen i do not understand how this works. Will I need to read both Nix and NixOS manuals to get this?
<clever>
hask_bee_3: the .runner script isnt documented in any of the manuals
<gchristensen>
hask_bee_3: it is sort of expert-level stuff... it took creativity and ingenuitiy to make it work. maybe a reason to not use it :P
<ottidmes>
clever: lol, well, you could say that systemd-boot does one thing and does it well, but I will try and not start yet another discussion about systemd. I personally also prefer grub, also for me the one I have the most experience with
<hask_bee_3>
gchristensen is it possible to run postgres in a more "standard" way? i was thinking something like nix-shell --run or something
<hask_bee_3>
something basic
<gchristensen>
I don't know how to use postgresql, so I really can't say
<clever>
ottidmes: also, in nixos, there is an option to configure how many generations go into the grub config, so your /boot doesnt get bloated
<clever>
ottidmes: systemd-boot just lacks that option
<ottidmes>
clever: But that is not inherent to either, you could easily implement that for systemd-boot as well, right?
<clever>
ottidmes: yeah, but there are other things its also missing
<clever>
ottidmes: and it would just be less work to fix 1 thing in grub, then 5 things in systemd-boot
<LnL>
hask_bee_3: you can use nix-shell -p postgresql --run 'postgres -D /tmp/foo' to run it in the foreground
<mfiano>
Yeah, `echo $XDG_DATA_HOME` in a nix-shell interpreted script emits nothing at all. but `env` inside nix-shell shows it set. Anyone know what nix-shell is doing?
<clever>
mfiano: what about running env in that same script that did the echo?
<hask_bee_3>
LnL: I guess that would do the trick. And putting that into systemd would automagically turn it into a background task
<mfiano>
clever: Odd, XDG is not set there, only when directly called from the nix-shell prompt
<ottidmes>
clever: Agreed, this is one of those places that competition is mostly just fragmention of effort
<clever>
mfiano: so its getting set at a later point, probably when the shell goes interactive and chooses to source more config files
<clever>
ottidmes: also, grub works with both legacy and efi, while systemd-boot is efi only
<mfiano>
clever: I have XDG vars set in environment.shellInit. also tried environment.interactiveShellInit. Any other places I should try?
<clever>
mfiano: try without --pure?
<mfiano>
That IS without pure
<clever>
mfiano: what is launching nix-shell?
<ottidmes>
mfiano: I have them at extraInit
<mfiano>
clever: me typing nix-shell in bash after completely logging out
<LnL>
hask_bee_3: in that case you'd probably want to generate the unit file to avoid going through nix-shell every time
<clever>
mfiano: and if you run env from the same place?
<mfiano>
I see the vars set
<clever>
mfiano: weird
<mfiano>
The XDG vars are set in bash, in nix-shell invoked from bash, but not in a script that calls env
<clever>
hask_bee_3: the version of postgress may change without warning, and its just slower
<clever>
hask_bee_3: if you generate a .service file with nix, then the version will be locked in until you re-generate it, and the files will remain cached
<disasm>
pre-nixos I really liked gummiboot (now systemd-boot) because it took all the hassle out of grub, but nixos takes the hassle out of grub already, so I'm indifferent now. I have half my systems running systemd-boot and others running grub2.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] aij opened pull request #35584: opensm: init at 3.3.20 (master...opensm) https://git.io/vAr4n
<NixOS_GitHub>
[nixpkgs] vitiral opened pull request #35587: add common gcc rust dependencies (master...rust-depends) https://git.io/vAr4M
NixOS_GitHub has left #nixos [#nixos]
<disasm>
gchristensen: ok, that postgresql systemd hack is pretty genius... If I was going to be working at my current place more than 3 days, I'd consider rewriting my nix puppet module to do that instead of hardcode hash/version in hiera for services I want to run!
<gchristensen>
:D
<aristid>
disasm: 3 days sounds like a pretty short tenure
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] xeji opened pull request #35588: cacert: add output "unbundled" for individual certificates (master...cacert-unbundled) https://git.io/vAr45
NixOS_GitHub has left #nixos [#nixos]
<disasm>
aristid: yeah, lets just say I'm excited to play with nix full time :) And haskell is pretty fun too!
<aristid>
disasm: \o/
asuryawanshi has joined #nixos
<Acou_Bass>
eyy folks, im getting a weird error if i run nix-env -u - says error: cannot auto-call a function that has an argument without a default value (‘config’)
<Acou_Bass>
i dont *think* ive changed anything since that error started appearing
simukis has quit [Ping timeout: 252 seconds]
<disasm>
Acou_Bass: did you install anything from a custom nix file using nix-env at any time?
<Acou_Bass>
not recently (though i do have a couple of things installed as such theyve been there a while)
<Acou_Bass>
i also have the nixpkgs-mozilla firefox overlay... could that be doing it?:P
asuryawanshi has quit [Ping timeout: 252 seconds]
alexteves has joined #nixos
alexteves_ has quit [Ping timeout: 245 seconds]
m3lst4d has quit [Quit: Leaving]
chocopuff298 has quit [Quit: WeeChat 2.0.1]
<clever>
Acou_Bass: what is in your config.nix?
m3lst4d has joined #nixos
m3lst4d_ has joined #nixos
m3lst4d_ has quit [Client Quit]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] dtzWill opened pull request #1906: nix search: tests and fix #1893 and part of #1892 (master...fix/nix-search) https://git.io/vArBZ
NixOS_GitHub has left #nixos [#nixos]
<Acou_Bass>
my user one? as in ~/.config/nixpkgs/config.nix?
alexteves has quit [Ping timeout: 248 seconds]
<disasm>
Acou_Bass: yeah
macslayer has quit [Ping timeout: 245 seconds]
<Acou_Bass>
just allowUnfree = true; and a packageOverrides for mumble that gives it pulse, JACK and speex support
<Acou_Bass>
i tried commenting the whole file out and that didnt seem to sort it either
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to master: https://git.io/vArBu
<mfiano>
clever: I tried setting XDG to a bunch of options under environment.* at the same time, rebuilt, rebooted, and then wrote the results of `env` called from inside the script to separate files when called from bash (vars-bash), and my window manager (vars-wm). The result of vars-wm is missing needed things like XDG. https://gist.github.com/anonymous/73a30c540a7190e429731249af0f4900
<NixOS_GitHub>
nixpkgs/master e3cb6e7 Herwig Hochleitner: chromium: build mojo_platform_bindings before chrome...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] bendlas pushed 2 new commits to master: https://git.io/vArRf
<disasm>
Acou_Bass: only thing I can think of to track it down would be remove 1 at a time, run `nix-env -u` until you get down to the one having issues, then add back everything else and make sure it's isolated to that one, then look at the code for that package. One of the reasons I don't use `nix-env` for package management :)
<Acou_Bass>
xD fair enough I'll have a go
<LnL>
hyper_ch: oh, maybe I mistyped a git bisect bad
<LnL>
thought I did test the surrounding commits tho
<Acou_Bass>
i have a couple of self written ones that i haven't upstreamed to nixpkgs, mostly because they're poorly written and would require work i can't be bothered doing to get them up to scratch so it's probably one of those
<hyper_ch>
that's at least 300 commits earlier
<LnL>
it's a merge so that's possible
smichel17 has joined #nixos
duckwho1 has quit [Quit: WeeChat 2.0.1]
<hyper_ch>
LnL: I'm trying to narrow it all down
<hyper_ch>
too like 14 to compile this one df5b0ce589f
<LnL>
cool, post what you find on the issue
<hyper_ch>
at least now it's quick
<LnL>
me messing up somewhere in the bisect would be a pretty good explanation for the seemingly unrelated diff
duckwho has joined #nixos
mkoenig has joined #nixos
<bhipple[m]>
Anyone know roughly how much performance overhead there is when running a single-use /nix/store in proot?
<bhipple[m]>
For execution of the binaries in the store, that is
elninja44 has joined #nixos
jb55 has quit [Ping timeout: 252 seconds]
duckwho has quit [Client Quit]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] TomSmeets opened pull request #35590: android-studio: Depend on ncurses5 (master...android-studio-ndk-fix) https://git.io/vArRu
NixOS_GitHub has left #nixos [#nixos]
<clever>
bhipple[m]: i think proot has more overhead then nix-user-chroot
<hyper_ch>
LnL: I found the offending commit
<hyper_ch>
sorry, maybe not
<mfiano>
clever: You were right. nvidia driver IS working with EFI. Yay!
<LnL>
don't really see how that could cause problems, except that it's a stdenv rebuild
electrocat has joined #nixos
<hyper_ch>
LnL: it was still a small skip... there's another 2 commits in between that I need to test
<hyper_ch>
LnL: no, it's that commit
kitemikaze has quit [Ping timeout: 245 seconds]
nh2 has quit [Ping timeout: 245 seconds]
<clever>
mfiano: i think the vesa driver is trying to use legacy bios API's to change the resolution, and its using dos-era tech to do it all, lol
<clever>
mfiano: but the nvidia driver knows how to actually control the card, and do things right
<mfiano>
I'm happy now. Nix is the first distro that was able to get grub and gpu to play nice with my crappy mobo
<mfiano>
Now if only I could figure out this damn env var issue with my WM that is wasting all day
lambdamu has joined #nixos
lambdamu has quit [Remote host closed the connection]
<mfiano>
clever: let's start over. I'm dumb and forgot to issue nixos-rebuild. I did that just now, restarted the dm, and it's even worse now.
<mfiano>
It seems my PATH is not being set inside the WM now, because the script won't execute AT ALL
<clever>
id just duplicate them then
<mfiano>
but yeah i see the bash script it generated after fixing that typo
<clever>
mfiano: and you can read it to confirm what went into it
a_f_k_ has quit [Quit: Lost terminal]
<mfiano>
I don't see anything that I added
<dtz>
is there a way to have the nixpkgs manual built as part of your current system configuration? like, a configuration.nix setting? Might be a bit noisy but would be nice to have it around JIC :)
<dtz>
(or the NixOS manual, for that matter)
<clever>
mfiano: can you gist the file?
<clever>
dtz: `man configuration.nix` is already available by default, thats a large chunk of it
<hyper_ch>
LnL: no idea why that commit causes it, but I'm off to bed now
<LnL>
yeah, same
<dtz>
neat
<clever>
mfiano: hmmm, line 35 sources /etc/profile, so options D, B, and E should already work
<hyper_ch>
LnL: you're also testing now that commit and the one before?
<clever>
mfiano: aha, and /etc/profile already sources the file i said to add, so A&C should also work
<LnL>
I will tomorrow
<mfiano>
clever: well with those changes I no longer have my PATH set correctly
<clever>
mfiano: but the changes i said to apply are abnormally missing
<clever>
mfiano: something fishy is going on
<hyper_ch>
(I did generate a list of commits between one commit that still works -> post #3 and one that didn't work -> post #2 and just kept halfing the numbers and testing
<clever>
mfiano: ive cloned the gist, and removed the hardware-configuration.nix line, it is now processing...
<mfiano>
Ok
<clever>
the nix eval passes, these paths will be fetched (183.18 MiB download, 863.57 MiB unpacked):
<clever>
mfiano: when done, i'll have a qemu based VM that contains a complete copy of your machine, minus your $HOME files
<deba5e12>
hey, i've been having some trouble getting suspend to work properly on my thinkpad p51
cmacrae has joined #nixos
<deba5e12>
when i close the lid, *sometimes* suspend will work as expected, but probably 10% of the time or less
<mfiano>
clever: Nice, thanks. I can provide any of those files too if needed
<clever>
mfiano: thats why i prefer doing everything with nixos, then they are just included automatically
<deba5e12>
most of the time, the green power LED starts blinking rapidly and the machine becomes unresponsive, and unable to wake up
<clever>
and you never have to think about them on future installs
<mfiano>
clever: That's what I want to do eventually. I'm sick of managing separate branches for all my machines' dotfiles where they differ
<deba5e12>
has anyone run into this sort of issue? i had void linux installed on this box before this, and had no issues with suspend with that setup, so maybe it's a systemd issue?
<clever>
ln: failed to create symbolic link '/nix/store/hnr5cvavqcd12iiclcpqy2vvwz2pry7d-neovim-ruby-env/bin/bundle': File exists
<clever>
mfiano: minor bug in the neovim build, but i doubt vim is involved, so i can just delete it
<mfiano>
I don't get that
<mfiano>
yeah
<clever>
i'm on a different version of nixpkgs
<samueldr>
clever: that error has been reported a couple times here in the last two days
<clever>
mfiano: so which variable is (not) set in what situation?
<mfiano>
Let me open my gist...
<mfiano>
when i add line 143 of my gist and restart display-manager, line 14 (and probably others) are not in effect
<tertle||eltret>
how muc differencedo you think a freesync monitor would make?
<mfiano>
I would guess all of 10-33
<clever>
mfiano: i added `export FINDME=yes` to that block, and it persists to the xterm that opened
<mfiano>
to which block?
<clever>
environment.shellInit
<deba5e12>
(and the same issue happens if I echo "mem" > /sys/power/state. neither enabling nor disabling powerManagment in configuration.nix has any effect.)
<mfiano>
Then something strange specific to PATH
<samueldr>
clever: isn't it only for shells? so xterm has them, but e.g. rofi or i3 wouldn't
<mfiano>
clever: sec...
<clever>
mfiano: cargo and .bin are in $PATH
<mfiano>
clever: here is my path WITHOUT line 143: /home/mfiano/.cargo/bin:/home/mfiano/.bin:/home/mfiano/bin:/run/wrappers/bin:/etc/per-user-pkgs/mfiano/bin:/home/mfiano/.nix-profile/bin:/nix/var/nix/profiles/default/bin:/run/current-system/sw/bin
<clever>
mfiano: that is the path i get in this VM, in xterm
<mfiano>
clever: here is my $PATH WITH line 143: /home/mfiano/bin:/run/wrappers/bin:/etc/per-user-pkgs/mfiano/bin:/home/mfiano/.nix-profile/bin:/nix/var/nix/profiles/default/bin:/run/current-system/sw/bin