zeus_ has quit [(Read error: Connection reset by peer)]
zeus_ has joined #nixos
requisitebits has joined #nixos
alx741_ has joined #nixos
alx741 has quit [(Ping timeout: 240 seconds)]
athan has joined #nixos
lambdamu_ has joined #nixos
<NickHu>
Is xmonad broken on latest nixos-unstable?
<NickHu>
I get an error about
<NickHu>
Failed to load interface for ‘Data.Default.Instances.Containers’
filterfish has quit [(Ping timeout: 258 seconds)]
lambdamu has quit [(Ping timeout: 264 seconds)]
matthewbauer has quit [(Remote host closed the connection)]
svij2 has quit [(Ping timeout: 255 seconds)]
mbrgm has quit [(Ping timeout: 255 seconds)]
filterfish has joined #nixos
mbrgm has joined #nixos
derjohn_mobi has quit [(Ping timeout: 255 seconds)]
alx741_ has quit [(Quit: alx741_)]
requisitebits has quit [(Quit: leaving)]
alx741 has joined #nixos
seagreen has joined #nixos
michas_ has quit [(Quit: This computer has gone to sleep)]
mattyu has joined #nixos
takle has quit [(Remote host closed the connection)]
matthewbauer has joined #nixos
mattyu has left #nixos []
derjohn_mobi has joined #nixos
Wizek has joined #nixos
Wizek_ has joined #nixos
matthewbauer has quit [(Remote host closed the connection)]
s33se_ has quit [(Ping timeout: 255 seconds)]
s33se has joined #nixos
leat has quit [(Ping timeout: 240 seconds)]
snikkers has joined #nixos
snikkers has quit [(Ping timeout: 260 seconds)]
takle has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
s33se_ has joined #nixos
chakerbenhamed has quit [(Ping timeout: 260 seconds)]
s33se has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
alx741 has quit [(Quit: alx741)]
alx741 has joined #nixos
lambdamu has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
lambdamu_ has quit [(Ping timeout: 255 seconds)]
takle has joined #nixos
acarrico has joined #nixos
jsgrant_om has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
jsgrant_om has left #nixos []
jsgrant_om has joined #nixos
johnramsden has joined #nixos
darlan has quit [(Remote host closed the connection)]
hexagoxel has quit [(Ping timeout: 272 seconds)]
takle has joined #nixos
<johnramsden>
is anyone familiar With the way 'networking.bridges' is supposed to be used to create a bridge using a single interface that can then be used in virtual machine or container? I don't want to use it to bridge two physical containers so I'm not sure how I should go about doing this. In the past I have just created a bridge and then used it as my main interface: https://ramsdenj.com/2016/09/23/containerizing-graphical-applications-on
<johnramsden>
-linux-with-systemd-nspawn.html#host-networking-setup how would I go about doing this on NixOS?
<johnramsden>
*physical interfaces
hexagoxel has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
nh2 has quit [(Quit: Leaving.)]
Elizaaa has quit [(Quit: Connection closed for inactivity)]
darlan has joined #nixos
takle has joined #nixos
matthewbauer has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
johnramsden has quit [(Quit: Leaving)]
takle has quit [(Ping timeout: 255 seconds)]
schoppenhauer has quit [(Ping timeout: 272 seconds)]
schoppenhauer has joined #nixos
takle has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/v9SQG
<NixOS_GitHub>
nixpkgs/master 49d3b43 Peter Hoeg: docker: fix reference to unknown attribute
NixOS_GitHub has left #nixos []
takle has quit [(Ping timeout: 268 seconds)]
johnsonav has joined #nixos
matthewbauer has quit [(Remote host closed the connection)]
matthewbauer has joined #nixos
takle has joined #nixos
filterfish has quit [(Ping timeout: 240 seconds)]
ryantm has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
koserge has quit [(Ping timeout: 255 seconds)]
filterfish has joined #nixos
johnsonav has quit [(Ping timeout: 246 seconds)]
reinzelmann has joined #nixos
takle has joined #nixos
takle has quit [(Ping timeout: 264 seconds)]
johnsonav has joined #nixos
ilyaigpetrov has joined #nixos
takle has joined #nixos
leat has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
justelex_ has joined #nixos
riclima has joined #nixos
matthewbauer has quit [(Ping timeout: 260 seconds)]
revtintin has joined #nixos
eacameron has quit [(Remote host closed the connection)]
grantwu has joined #nixos
<grantwu>
I'm having trouble updating youtube-dl
<grantwu>
It seems like every time I try, I reinstall the version I currently have
justelex_ has quit [(Ping timeout: 272 seconds)]
<dash>
grantwu: OK. What did you do?
<grantwu>
uh... let's see
<grantwu>
nix-env --upgrade youtube-dl seems to return silent
<grantwu>
-e and then -i seems to just reinstall the old version
<grantwu>
nix-env -iA nixpkgs.python35Packages.youtube-dl says replacing old ‘youtube-dl-2017.02.04.1’ and then installing ‘youtube-dl-2017.02.04.1’
<dash>
OK. Did you use nix-channel --update? Is there a more recent version in your channel?
<grantwu>
Hrm, no, I didn't do that
<dash>
OK. That's the only way nix has to find out about new stuff
<grantwu>
oh okay, --upgrade works now
<grantwu>
thanks!
eacameron has joined #nixos
takle has joined #nixos
riclima has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
takle has quit [(Ping timeout: 240 seconds)]
riclima has joined #nixos
Mutter has joined #nixos
Mutter has quit [(Client Quit)]
jacob has joined #nixos
jacob is now known as Guest93604
Mateon3 has joined #nixos
nnnn has joined #nixos
alx741 has quit [(Quit: alx741)]
Mateon1 has quit [(Ping timeout: 255 seconds)]
Mateon3 is now known as Mateon1
<nnnn>
Hey, let's say I'm trying to keep track of a dotfiles repo. How terrible of an idea is it to softlink my configuration.nix file to /etc/nixos/ from a directory in my home folder?
Guest73017 has quit [(Killed (sinisalo.freenode.net (Nickname regained by services)))]
Guest73017 has joined #nixos
Guest73017 is now known as jgertm
jgertm is now known as Guest47780
michas_ has quit [(Quit: This computer has gone to sleep)]
katyucha is now known as Guest68587
Guest68587 has quit [(Killed (verne.freenode.net (Nickname regained by services)))]
katyucha_ is now known as katyucha
katyucha_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] LnL7 opened pull request #25725: python-pytest-xdist: disable tests on darwin (master...python-pytest-xdist) https://git.io/v9Sxm
NixOS_GitHub has left #nixos []
Infinisil has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
Itkovian has joined #nixos
takle has quit [(Ping timeout: 268 seconds)]
freusque has joined #nixos
MichaelRaskin has quit [(Quit: MichaelRaskin)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] michalpalka opened pull request #25726: gitlab service: fix uploading artifacts from gitlab-runner (master...gitlab-runner-fix) https://git.io/v9SxP
takle has quit [(Remote host closed the connection)]
nschoe has joined #nixos
<dhess>
Anyone know why the GHC bootstrap process starts with 7.4.2?
<dhess>
Could it be made to start with, say, 7.8.3?
marsel has quit [(Ping timeout: 246 seconds)]
<NickHu>
Is there a way to revert to a particular nixos-unstable commit?
<NickHu>
(Without lots of recompiling)
<hyper_ch>
hmm, I have firewall.allowPing = true
<hyper_ch>
but is that also applied to ipv6?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] MP2E pushed 1 new commit to master: https://git.io/v99Ja
<NixOS_GitHub>
nixpkgs/master 4f51a1e Cray Elliott: nvidia-x11: add dbus as a dependency to settings...
NixOS_GitHub has left #nixos []
derjohn_mobi has joined #nixos
<dhess>
I guess I'll find out :)
<hyper_ch>
ok, ipv6 is always allowed Whether to respond to incoming ICMPv4 echo requests ("pings"). ICMPv6 pings are always allowed because the larger address space of IPv6 makes network scanning much less effective.
<clever>
but in my case, i forgot to setup a v6 firewall of any kind in my router
<clever>
and i dont have randomized v6 addresses for outgoing stuff
<clever>
so if i connect to you, you get my long-term v6 addr, and full access
takle has joined #nixos
proteus-guy has quit [(Ping timeout: 240 seconds)]
thaega has joined #nixos
proteus-guy has joined #nixos
<sphalerite>
If stuff is broken to the point that magic sysrq doesn't work, the only way to proceed is a hardware reset, right?
<hyper_ch>
clever: sweet, you know ipv6 :)
<clever>
sphalerite: the sysrq may work over ps2 or the serial port, even if it doesnt work over usb
<clever>
sphalerite: but getting a real serial port on a modern machine isnt easy
<sphalerite>
nor ps2, on a laptop :d
<clever>
the internal keyboard is probably ps2
<clever>
so if that doesnt work, yeah, hard-reset is about all thats left
swistak35 has joined #nixos
marsel has joined #nixos
johann__ has quit [(Quit: Leaving.)]
<clever>
sphalerite: but i can see a way to get real serial on a laptop, one min
Infinisil has joined #nixos
<sphalerite>
Hm, in the sysfs it's on the serio bus
<clever>
yep, top left, your wifi card is using it
aloiscochard has joined #nixos
zeus_ has quit [(Remote host closed the connection)]
zeus_ has joined #nixos
<dhess>
so there's nothing special about GHC 7.4.2 being used as the bootstrap version; it's just nobody has bothered to add newer ones yet. https://github.com/NixOS/nixpkgs/issues/19926
<sphalerite>
has there been a big KDE update in 17.03?
orivej has quit [(Ping timeout: 255 seconds)]
<Yaniel>
kde5 -> plasma5, yes
<sphalerite>
no I mean in the past few days
<sphalerite>
I'm rebuilding my system and it's downloading a huge amount of KDE stuff again
Infinisil has quit [(Quit: leaving)]
<Yaniel>
doesn't that happen whenever glibc is updated
<Yaniel>
or some similar low level lib
Itkovian_ has joined #nixos
proteus-guy has quit [(Ping timeout: 240 seconds)]
Itkovian has quit [(Ping timeout: 255 seconds)]
nslqqq has quit [(Ping timeout: 260 seconds)]
proteus-guy has joined #nixos
engblom has quit [(Read error: Connection reset by peer)]
engblom has joined #nixos
lsix has quit [(Quit: WeeChat 1.7.1)]
indi_ has joined #nixos
indi_ has quit [(Remote host closed the connection)]
indi_ has joined #nixos
<sphalerite>
I think it might have been the KDE wrappers
<sphalerite>
My system locks up entirely when I try to start the X server with nvidia drivers enabled. journalctl -b -1 and /var/log don't seem to contain anything relevant. How can I debug this?
luto has joined #nixos
<sphalerite>
The lockup happens even if I've set xserver.videoDrivers to mkForce [ "intel" ]; (and it works fine if I set bumblebee.enable to false, which stops the nvidia driver from being included in the system)
nslqqq has joined #nixos
indi_ has quit [(Ping timeout: 272 seconds)]
snikkers has joined #nixos
snikkers has quit [(Client Quit)]
Guest47780 has quit [(Quit: WeeChat 1.7)]
pie_ has quit [(Ping timeout: 268 seconds)]
nslqqq has quit [(Ping timeout: 240 seconds)]
SplitFire has joined #nixos
<SplitFire>
Hello there!
zeus_ has joined #nixos
pie_ has joined #nixos
<SplitFire>
I need help. I can't reconfigure xsane to be usable by gimp
<goibhniu>
hrm .. how about: find /nix/store/d1v4kksrxkav8qc51n7zl7vs9srlli4i-xsane-0.999/ -iname "*plug*"
<SplitFire>
urns nothing
<SplitFire>
*returns
<clever>
a quick check says d1v4kksrxkav8qc51n7zl7vs9srlli4i isnt in the binary cache, so there is a good chance its the overridden version
<goibhniu>
sorry, I guess it doesn't work
<clever>
checking things on this end more
<SplitFire>
clever, yeah i noticed building of xsane
<sphalerite>
Wow. Even lspci locks up the system when nvidia is around.
<goibhniu>
clever++
<sphalerite>
I have had it with their bullshit.
<Yaniel>
sphalerite: waiting for Vega now? :P
<sphalerite>
Vega?
eacameron has joined #nixos
<Yaniel>
amd's new gpu gen
<sphalerite>
No, fuck AMD too.
<clever>
sphalerite: are you locking it up in xorg or text mode?
koserge has joined #nixos
darlan has quit [(Remote host closed the connection)]
<sphalerite>
What I want is a manufacturer that releases free-software drivers so that if htye break everything someone outside the company can fix it
<sphalerite>
clever: instant lockup when starting the X server or running lspci (from text mode obviously)
<clever>
ah, and no kernel panic msg in text mode?
<sphalerite>
Possibly. I have no way of telling.
<sphalerite>
because the display just freezes.
<simpson>
sphalerite: AMD's better than the other two big vendors about this; when I was doing GPUs, they were the only ones that would actually put PDFs of the GPUs into our hands.
<clever>
sphalerite: and ssh is also dead?
<sphalerite>
clever: yes
<sphalerite>
simpson: hasn't that changed though?
<clever>
sphalerite: then a serial console is about the only option, enless you can guess and disable the right driver
<Yaniel>
sphalerite: yeah, nowadays they contribute their stuff to the linux kernel directly
eacameron has quit [(Ping timeout: 240 seconds)]
<sphalerite>
clever: I can get X running by disabling the nvidia driver. Just that then I can't use the discrete graphics at all (nouveau doesn't support the chip, again because of nvidia's proprietary bullshit)
<sphalerite>
Yaniel: oh nice
<Yaniel>
that said, they still have a proprietary firmware blob AFAIK
<Yaniel>
or userspace part
<Yaniel>
not sure where exactly the division goes
<sphalerite>
So like intel
<Yaniel>
but still, using their cards with oss drivers is starting to be a sensible choice
<sphalerite>
The one and only way I've got the GPU working is by starting the X server with nvidia fully disabled, then *switching* to a configuration with the nvidia driver enabled
<clever>
SplitFire: i am seeing some gimp related symbols in the xsane build
<sphalerite>
But heaven forbid I run lspci or start a new X server after that.
<clever>
SplitFire: i am seeing some stuff in "man xsane" saying you need to make a symlink in ~/.gimp-2.0/plug-ins/
<goibhniu>
clever: FWIW, I expect the gimp enabled xsane to create something like share/gimp/plugins/xsane, which can then be symlinked to ~/.gimp-2.8/plug-ins/
<clever>
SplitFire: check the man page and see if it fixes it
<SplitFire>
clever, i saw that but i thought it would be overkill
<clever>
│ │ │ │ │ -binary (e.g., "touch /nix/store/abawa99js846qwwjlqjdqaapsyvxljdh-xsane-0.999/bin/xsane") or delete the plugin cache
<clever>
yeah, some parts of the manual are wrong for nixos
<clever>
but the symlink in $HOME cant hurt and is easy to test
<NickHu>
I was trying to override a haskell package
<gchristensen>
NickHu: I don't know ghc, but it seems all the errors stem from ~/.cabal ?
<sphalerite>
just remove .cabal
<sphalerite>
or rename it rather
<NickHu>
I did
<NickHu>
I guess I need to do some ghc-pkg unregister
<gchristensen>
what does ghc-pkg register do?
<pie_>
does all the thousands of nix symlink stuff thrash ssds
<NickHu>
I think it puts some config stuff under ~/.ghc but not sure
nslqqq has quit [(Ping timeout: 246 seconds)]
<gchristensen>
NickHu: maybe move that one too
<NickHu>
Also can someone help me understand why I need a full attribute path for things defined in ~/.config/nixpkgs/config.nix?
chakerbenhamed has quit [(Ping timeout: 240 seconds)]
<NickHu>
I have two nearly identical machines that I installed nixos on at roughly the same time
<NickHu>
One of them has always been on nixos unstable, and I can do nix-env -i mything fine
<gchristensen>
pie_: symlinks or not, compared to other distros & package managers, nix has higher write volume and disk churn
<NickHu>
The other used to be on 16.09 but then I switched it over to nixos unstable, but I have to do nix-env -iA nixos.mything
<NickHu>
They both give the same output for nix-channel --list and sudo nix-channel --list
<gchristensen>
pie_: and there will be additional read penalty, of course, but I'd expect it to be utterly minimal compared to whatever your application is doing
<NickHu>
And have the exact same config.nix
marsel has joined #nixos
<clever>
NickHu: can you gist the config.nix file?
<clever>
NickHu: and confirm you dont have a ~/.nixpkgs/config.nix
betaboon has joined #nixos
betaboon has quit [(Changing host)]
betaboon has joined #nixos
ent has joined #nixos
nh2 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 1 new commit to master: https://git.io/v99lo
<NixOS_GitHub>
nixpkgs/master ae6f932 Peter Simons: openvpn: update to version 2.4.2 to fix CVE-2017-7478 and CVE-2017-7479
<LnL>
NickHu: I think having stuff in ~/.ghc can cause problems if it was created by an older version of ghc
<gchristensen>
pie_: that doesn't interest me personally very much, and I'm terrible at statistics
<gchristensen>
ent: you're welcome!
<ent>
ooh, that options.html is exactly what I needed :)
<clever>
bbl
<gchristensen>
I also well encourage taking a look at `man configuration.nix`, ent
chakerbenhamed has joined #nixos
<ent>
hmm, I'm usin nix on ubuntu, how do I get that?
<sphalerite>
you don't
vitalii has joined #nixos
<gchristensen>
ohh interesting, so configuration.nix is only for nixos, and the service management options that https://nixos.org/nixos/options.html have are for nixos too
<sphalerite>
Systemwide configuration is nixos-only
magnetophon has joined #nixos
<gchristensen>
there is a way to make a runner script from those options, though
<ent>
ah, but I'm using nixops to create virtualbox machines
<gchristensen>
got it
<ent>
I guess I could of course read the man page on a vm
<gchristensen>
yeah, nixos/options.html is a good choice :)
<magnetophon>
Hi, I'm having trouble getting the clipboard to work in virtualbox. I installed the guest-additions in the VM and rebooted it. What else do I need?
<sphalerite>
Apparently now the GPU has "fallen off the bus". That's a new one!
<Yaniel>
bus factor: way too low
<sphalerite>
indeed
<sphalerite>
At least it's more helpful than "this driver doesn't support this GPU even though the manual says it does"
<Yaniel>
(that's a somewhat novel definition for the term 'bus factor' though)
afics has quit [(Ping timeout: 258 seconds)]
nslqqq has joined #nixos
bennofs has quit [(Ping timeout: 268 seconds)]
proteus-guy has joined #nixos
johann__ has quit [(Quit: Leaving.)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] WilliButz opened pull request #25733: nixpkgs manual: fix typo in ruby section (master...master) https://git.io/v9987
NixOS_GitHub has left #nixos []
<betaboon>
Hello #nixos. I have a question related to preparing a pull-request for nixpkgs: the readme states to "please rebase onto master". I have been working (and regularly rebasing) on a channels/release-17.03 based branch. I'm not sure if I'm doing this right atm. What i did 'checkout master', 'pull --rebase', 'checkout MY-WIP-BRANCH', 'merge --onto master'. could anyone assist me in getting this right ?
<NickHu>
Can I build a haskell package and keep the build dir around to look at it?
<avn>
betaboon: `git checkout my-feature-bracnh && git rebase origin/master && git push -f github my-feature-branch` usually works
WilliButz has quit [(Quit: leaving)]
Willi_Butz is now known as WilliButz
zeus_ has joined #nixos
<betaboon>
avn : the rebase is assuming that the "origin"-remote points to "nixos/nixpkgs", correct ?
<avn>
yes
<avn>
I usually have origin pointing to upstream, and `github` to my own fork
<betaboon>
ah ok, i have "origin" pointing to my fork, and "upstream" to the original
<gchristensen>
betaboon: you don't want to rebase your branch against master
<gchristensen>
(avn^) b/c it'll try to include all the changes from 17.03, too
<gchristensen>
betaboon: how many commits do you have that you want to send in a PR?
<betaboon>
gchristensen, i think my history on that branch is ~50 commits
<gchristensen>
and are all 50 going in to a single PR?
abcrawf has joined #nixos
<betaboon>
i guess so
<avn>
gchristensen: I assumed, that he make PR to master, so he need to rebase to master. If his patch based on other branch, it need to be rebased there
<gchristensen>
betaboon: hmm usually PRs to master have approximately 1 commit. what are your changes?
zeus_ has quit [(Ping timeout: 272 seconds)]
<gchristensen>
you've done a lot of work :D
<betaboon>
ehm, ye :P
<gchristensen>
avn: rebase won't do what he needs though, it'll move over every commit to 17.03 as well
<betaboon>
first time contributing to nixos
<gchristensen>
betaboon: ok, well I'll try to make this easy :)
<avn>
betaboon: you can also squash some WIP commits I believe
<gchristensen>
I would recommend using cherry-pick to copy the commits from this branch, to a new branch based off of master
<betaboon>
gchristensen, thought about that aswell, seems moreclean.
<betaboon>
should i squash the WIP-commits (all related to the actual implementation)
<gchristensen>
absolutely
DutchWolfie has quit [(Quit: Konversation terminated!)]
<betaboon>
if so, how to go on about the dependency commits? i wouldn't like to squash those, as i feel that dirties history
mudri has quit [(Ping timeout: 240 seconds)]
<gchristensen>
each new package requires its own commit
<gchristensen>
but all those taiga module commits should be squashed
<betaboon>
will do that and report back
<gchristensen>
great
<betaboon>
i have one more question right now tho: for the nodepackage i added to node-packages.json -> when i run ./generate.sh the package name generated into node-packages-v4/v6.nix includes the github url etc, why is that? how can i get it to just use the package-name ?
<goibhniu>
I'm having trouble with dolphin lately, (on unstable) ... No files are listed and I see io-slave errors at the top e.g. "Unable to create io-slave. klauncher said: Error loading '/nix/store/770bfm8526nnvkvrwq2gmxpjih4mh8rs-kio-5.33.0/lib/qt5/plugins/kf5/kio/file.so'." ... any idea what that might be?
* goibhniu
suspects it's a store path in some plasma config/kde file, but hasn't tried using a fresh user account
<Baughn>
I'd give it a shot, but the Nix code scares me. A Closure compiler-like approach might be better...
<Baughn>
Someone will have to do it eventually, I don't think I want it to be me.
<Baughn>
..
<Baughn>
joachifm: Ah.
<Baughn>
^^;
<Baughn>
*SO* starred.
indi_ has quit [()]
<joachifm>
Baughn: as I recall, the dynamic attrs of nix makes it a little tricky to type; there's a discussion of the difficulties somewhere but I can't recall where offhand
<NixOS_GitHub>
[nixpkgs] georgewhewell opened pull request #25740: dosfstools: fix build on darwin (master...dosfstools/fix-darwin) https://git.io/v997h
NixOS_GitHub has left #nixos []
vitalii has quit [(Remote host closed the connection)]
goober2 has joined #nixos
ertes-w has quit [(Ping timeout: 240 seconds)]
goibhniu has quit [(Ping timeout: 260 seconds)]
digitus has joined #nixos
snikkers has joined #nixos
mudri has joined #nixos
bkchr has quit [(Quit: Konversation terminated!)]
ZoomZoomZoom has quit [(Read error: Connection reset by peer)]
aneeshusa has joined #nixos
CcxWrk has quit [(Ping timeout: 240 seconds)]
johnsonav has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ehmry opened pull request #25741: nixos/mlmmj: fix newline problems with multiple lists (master...mlmmj) https://git.io/v99N1
NixOS_GitHub has left #nixos []
endformationage has joined #nixos
bhipple has joined #nixos
bennofs has joined #nixos
johann__ has quit [(Quit: Leaving.)]
<msd>
Mic92: fixed
<msd>
Mic92: pushed
ambro718 has quit [(Remote host closed the connection)]
pratch has quit [(Remote host closed the connection)]
snikkers has quit [(Quit: Lost terminal)]
pratch has joined #nixos
ryantrinkle has joined #nixos
pratch has quit [(Max SendQ exceeded)]
msd has quit [()]
pie_ has joined #nixos
<pie_>
man plasma is buggz
<pie_>
buggy
<ryantrinkle>
is it intentional behavior that giving fetchurl sha256=null results in it working without apparently checking any hash?
pratch has joined #nixos
<pie_>
on an unrelated note, anyone been able to completely do their personal system configuration via nix?
<pie_>
so like firefox etc whatever are all more or less reproducible?
<ryantrinkle>
pie_: yep, i've been using nixos as my main OS for about 4 years
<simpson>
pie_: I have several machines which are specified completely via Nix expressions, but none of them are my personal laptop.
<ryantrinkle>
pie_: i just add a buildEnv-based "package" to ~/.nixpkgs/config.nix
<simpson>
pie_: OTOH this is because my personal laptop has many active git repos and it's useful for development to be able to add ad-hoc stuff to my rollout.
<ryantrinkle>
and then whenever i change it, i just run: nix-env -iA nixpkgs.myEnv --keep-going --show-trace
<ryantrinkle>
and the nixpkgs i use is a git checkout
ryanartecona has joined #nixos
derjohn_mobi has joined #nixos
<unlmtd>
if nix can link stuff in /etc then surely it can link things in ~/
<pie_>
hm
<ryantrinkle>
this seems really impure, is it supposed to work?
<pie_>
also has anyone thought to expose firefox's about:config in nix?
<dash>
pie_: scary
<pie_>
dash, why
<makefu>
i like the idea :)
<dash>
pie_: web browsers have lots of mutable state, you want to manage bookmarks, history etc that way too? :-)
Filystyn has joined #nixos
Filystyn has quit [(Changing host)]
Filystyn has joined #nixos
<simpson>
pie_: Sounds tricky.
<pie_>
oh the scary was at the reproducible personal mahcine stuff
<pie_>
simpson, i wouldnt know why, so just asking
<pie_>
dash, well i guess for practical purposes you would integrate with some other backup tool
Kingsquee has joined #nixos
primeos has joined #nixos
zeus_ has joined #nixos
betaboon has joined #nixos
betaboon has quit [(Changing host)]
betaboon has joined #nixos
<makefu>
pie_: it was the same for my window manager. normally it would use a mutable file in ~ , i've overwritten the package to overwrite the "default" file coming with the package.results in a consistent window manager config
civodul has quit [(Quit: ERC (IRC client for Emacs 25.2.1))]
zeus_ has quit [(Ping timeout: 272 seconds)]
justelex_ has joined #nixos
SuprDewd has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<gchristensen>
trying to force mutable state things in to immutable paradigms can be just hell, I tried to do this super hard for a couple years
<pie_>
well, stuff like addons and configuration should be ok though no?
<ryantrinkle>
gchristensen: hey, if you have a sec, would you mind giving me a quick sanity-check on this thing?
<ryantrinkle>
i feel like i must be missing something
<marc__>
i am marcnnn on github
<ryantrinkle>
it looks like fetchurl can easily be invoked impurely: cat $(nix-build --no-out-link -E '(import <nixpkgs> {}).fetchurl { url = "http://www.time.gov/actualtime.cgi"; sha256 = null; }')
<gchristensen>
ryantrinkle: sure, taking a look
<gchristensen>
ryantrinkle: so builtins.fetchtarball can be impure too
ertes has joined #nixos
<gchristensen>
I wasn't aware you could pass null to fetchurl though
<ryantrinkle>
yeah, that seems like it shouldn't be allowed
spinus has joined #nixos
<gchristensen>
fetchurl, fetchtarball, or both?
<pie_>
requiring a hash sounds overly restrictive
<ryantrinkle>
i haven't tried fetchtarball
<ryantrinkle>
just fetchurl
<pie_>
i mean what do i know
<ryantrinkle>
pie_: not at all; without the hash, there's no reason to think the derivation is pure
<gchristensen>
pie_: its a fairly fundamental assumption in nix that all inputs are known at evaluation time ...
<pie_>
i guessss
marsel has quit [(Ping timeout: 240 seconds)]
<spinus>
it's not that hard to write own fetcher as well if you need specific feature
<pie_>
impureFetchURL
<pie_>
* impurefetchurl
<pie_>
:P
<pie_>
power to the people
<pie_>
kek
<pie_>
(i just made that up)
<gchristensen>
ryantrinkle: very interesting ......
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 1 new commit to master: https://git.io/v9HIB
<NixOS_GitHub>
nixpkgs/master c9f8fb4 Thomas Tuegel: fontconfig-penultimate: 0.3.3 -> 0.3.4
NixOS_GitHub has left #nixos []
<simpson>
pie_: If you don't want Nix, don't use Nix.
<pie_>
who wouldnt want nix
<ryantrinkle>
pie_: haha, well the power *i* want is to be able to rely on a given hash of nixpkgs, and never have anything surprise me when i try to rebuild :P
<simpson>
Well, if you want impure builds, you probably don't want Nix.
<ryantrinkle>
and i've gotta say, it's been phenomenal how well that's worked
justelex_ has quit [(Quit: Konversation terminated!)]
<pie_>
impure language extensions for the heretics \o/
acarrico has joined #nixos
justelex_ has joined #nixos
<gchristensen>
not sure if trolling ...
<pie_>
only a little
<ryantrinkle>
the only time in the last 4 years that i've had an issue with that was when the hash format for fetchGit changed
<pie_>
<3 nix but i still havent gotten around to using it
<pie_>
err, learning it is what i meant
<ryantrinkle>
pie_: it's actually very easy to do impure stuff with nix: just make the output of your derivation be a bash script, and then run it
<ryantrinkle>
you can even invoke nix some more from there
<pie_>
been using it for a few months on a "lets ask how to do this specific thing" basis
dannyg has quit [(Quit: dannyg)]
<gchristensen>
ryantrinkle: this seems like a bad bug.
<pie_>
anyway, what you guys said makes sense
<ryantrinkle>
gchristensen: that's what i thought, but i was hesitant to think i would be the first to run into it
<pie_>
gchristensen, how does this even work if its not intentional though
<ryantrinkle>
one of the guys here at obsidian just mentioned it to me
<gchristensen>
bugs?
<pie_>
yeah the bug
<gchristensen>
ryantrinkle: what are you running?
acertain has joined #nixos
<ryantrinkle>
» nix-env --version
<ryantrinkle>
nix-env (Nix) 1.11.8
justelex_ has quit [(Ping timeout: 240 seconds)]
<gchristensen>
ryantrinkle: running Nix 1.11.4 on OSX, the daemon, and build-use-sandbox = false, I can reproduce. on 1.11.8 on NixOS and the sandbox, I can't reproduce
<ryantrinkle>
gchristensen: i don't think i'm using the sandbox
<ryantrinkle>
should i generally enable that?
<gchristensen>
IMHO it should always be enabled by default
<ryantrinkle>
ok, i'll switch to that
<gchristensen>
I think separately it is probably a bug that you can specify a hash algo and a null hash
<gchristensen>
bennofs: I recommend building in retries :)
<ryantrinkle>
gchristensen: what's your github handle?
<gchristensen>
grahamc
Itkovian_ has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
justelex_ has quit [(Client Quit)]
hellrazo1 has quit [(Quit: WeeChat 1.7.1)]
<bennofs>
gchristensen: i have retries. but I'd have to wait a huge amount of time before retrying, because the error only goes away after a variable time (~2 or 3 to 5 minutes perhaps?)
<bennofs>
just retrying again in 1 second will cause the same error
justelex_ has joined #nixos
<gchristensen>
bennofs: yuck ...
<bennofs>
I could probably change the accept-encoding header to "force" a refresh but meh
<bennofs>
it's probably because cloudfront will cache the 500 response from s3
<bennofs>
so you have to wait till the page gets evicted from cloudfront cache or you'll just get another cached 500?
<gchristensen>
IRC 500's are uncacheable
<ryantrinkle>
gchristensen: how big is the perf impact from sandboxing?
<gchristensen>
ryantrinkle: IMO? not enough to worry about, it used to be bad
justelex_ has quit [(Remote host closed the connection)]
acertain has quit [(Ping timeout: 255 seconds)]
<gchristensen>
ryantrinkle: how do you feel about the perf. info from domen's last comment?
<ryantrinkle>
gchristensen: well, i think it may discourage people from splitting up builds more finely
<gchristensen>
oh?
<ryantrinkle>
in haskell, e.g., it's kind of difficult to work on projects with multiple cabal packages with nix, because each package has to rebuild entirely
<ryantrinkle>
so one might want to, instead, have nix build each source *file
<ryantrinkle>
*
<ryantrinkle>
and then make the link step be a separate nix derivation
<gchristensen>
omg
<ryantrinkle>
the more constant overhead there is per derivation, the less things like that make sense
svij2 has joined #nixos
<pie_>
ugh i want nix on openbsd so bad
<gchristensen>
what if we could convert a Makefile to a nix based on the ordering rules (lol)
<ryantrinkle>
haha yep, that'd be awesome
<gchristensen>
I mean it'd be something :P
<ryantrinkle>
tbh, nix is basically a system-level function memoizer
<ryantrinkle>
obviously there'll always be a level at which it's too granular
<gchristensen>
my argument though is people choose nix not because it is extremely fast, but because it is correct
<ryantrinkle>
yes, that's absolutely right
<pie_>
(fast cant hurt though :P)
<ryantrinkle>
obsidian uses nix for everything
<ryantrinkle>
but our dev cycles would get faster with these kinds of improvements
<gchristensen>
fast can't hurt, sure, but turning on this feature which makes nix _more_ correct by default seems important for nix's core raison d'etre
<ryantrinkle>
agreed
<ryantrinkle>
i just turned it on :)
<gchristensen>
despite the modest decrease in speed
<ryantrinkle>
but i definitely do think the concern is valid
<gchristensen>
ryantrinkle: I didn't know you were part of the cardassian union
<ryantrinkle>
we're a dev shop that does mostly haskell and nix
<pie_>
i just realized i missed a ton of scroll so no idea what i commented on
<gchristensen>
nice
justelex has quit [(Ping timeout: 240 seconds)]
<ryantrinkle>
8 people, currently
<gchristensen>
"Founded by two Harvard Law School graduates" where are you?
<ryantrinkle>
yup, that's me
<ryantrinkle>
and my business partner
<ryantrinkle>
we were roommates at HLS
justelex has joined #nixos
<gchristensen>
are you in Cambridge?
<ryantrinkle>
nah, we're in NYC now
justelex has quit [(Read error: Connection reset by peer)]
<gchristensen>
gotcha, cool -- I visit NYC 1-2wks a quarter, we should grab drinks next time I'm in town
ravloony has quit [(Ping timeout: 240 seconds)]
<ryantrinkle>
gchristensen: that sounds great :)
<gchristensen>
(paging ikwildrpepper who will be there soon)
acertain has joined #nixos
Tucky has quit [(Remote host closed the connection)]
<gchristensen>
ryantrinkle: I think it'd be neat to have a tool to regenerate nix configuration based on location, to implement timezone switching / redshift updating / "location aware settings" sort of thing.
<ryantrinkle>
gchristensen: yeah, that'd definitely be cool :)
<ryantrinkle>
if only nix supported functional reactive programming... :P
<gchristensen>
hehe
<gchristensen>
my "vision" is it'd basically be a cron and on config change, prompt to rebuild
<ryantrinkle>
time.timeZone :: Dynamic OlsonId
<gchristensen>
:o
<ryantrinkle>
in the frontend apps we build, we do exactly that
<ryantrinkle>
we've got a big ReaderT that passes timezone (and other config) everywhere
<ryantrinkle>
and it's all dynamic
<ryantrinkle>
so if the user changes their timezone setting, all the time displays in the app immediately update
<gchristensen>
is this React?
<ryantrinkle>
this is Reflex, the haskell-based FRP engine we built
pdobrogost has quit [(Quit: Connection closed for inactivity)]
pie_ has quit [(Changing host)]
pie_ has joined #nixos
bennofs has joined #nixos
<betaboon>
gchristensen you were right, I added a duplicate package. just fixed that on the branch
<sphalerite>
ryantrinkle: just some random unsolicited website opinion: I was very confused, couldn't find where to navigate the site. The stark contrast between the navigation bar and the top made me think the navigation bar was part of browser chrome, not the actual site
MichaelRaskin has joined #nixos
mizu_no_oto has joined #nixos
<pie_>
gchristensen, someone told me that according to that it really does look like the external drive got mounted over /
<pie_>
wtf
<pie_>
so does that mean my open file descriptors are currently interacting with which drive?
mudri has joined #nixos
<Isorkin>
How can i access the /var/www folder in a container from the internet?
<pie_>
gchristensen, so my currently open file descriptors are writing to which drive?
<pie_>
????
<pie_>
this is so bad
<gchristensen>
that is weird, man
<sphalerite>
Isorkin: by running an HTTP server in it and exposing the port
<gchristensen>
open ones are probably writing to the old drive
<sphalerite>
that sounds bad.
riclima has joined #nixos
<Isorkin>
sphalerite: need file access, ftp/sftp
<pie_>
gchristensen, i mean one can do this just the same manually
<pie_>
but this happening by default seems baaaaad
<sphalerite>
Isorkin: then run an SSH server in it and expose that.
acertain has quit [(Ping timeout: 272 seconds)]
<gchristensen>
cstrahan: ping?
<Isorkin>
sphalerite: port forwarding?
<pie_>
gchristensen, got time and energy to try reproducing?
<pie_>
gchristensen, im guessing its basically just obeying the mountpoint
<pie_>
ok
<Isorkin>
sphalerite: thanks
pie_ has quit [(Quit: Leaving)]
<gchristensen>
anyone know how github handles colorizing?
<gchristensen>
oh, d'oh, figured it out :)
eacameron has quit [(Ping timeout: 240 seconds)]
takle has quit [(Remote host closed the connection)]
matthewbauer has joined #nixos
acertain has joined #nixos
justelex_ has quit [(Excess Flood)]
justelex_ has joined #nixos
<ryantrinkle>
sphalerite: you're absolutely right, and we're already working on a new site
<ryantrinkle>
it just keeps getting preempted by more client work :P
<betaboon>
gchristensen, argh... just done with the pr-branch, trying to build from master, now pyopenssl tests fail and I cant finish my system build
riclima has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<sphalerite>
ryantrinkle: a whole new site? I thought the current one is very pretty, just the navigation is confusing
<betaboon>
any advice ?
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
Filystyn has quit [(Quit: Konversation terminated!)]
Filystyn has joined #nixos
Filystyn has quit [(Changing host)]
Filystyn has joined #nixos
<nh2>
what's the best way to package something that want's to copy things to `/var` during `make install`?
sigmundv__ has quit [(Ping timeout: 260 seconds)]
<gchristensen>
nh2: patch it so it doesn't
<betaboon>
nh2, patch the makefile during patchPhase so that it is copying ti $out/var ?
<gchristensen>
or change its configuration so it doesn't
<nh2>
gchristensen: hmm, it is kind-of nontrivial, as patching glusterfs can increase risk of data loss quite a bit. I'm wondering if I should make it a systemd `preStart` script to copy from the store location to `/var`
<gchristensen>
you could but if the `make install` step is trying to copy to /var, you need to make that not happen
<bennofs>
gchristensen: https://gist.github.com/bennofs/7cbd52a35ee61609bbc053b0c49d108c it's happening again. I think it really is cloudfront caching 500 responses :/ (according to the docs, they passthrough the Accept-Encoding header, but they only check if gzip is listed or not and then set the header accordingly)
<nh2>
gchristensen: it doesn't seem to try it currently, as it builds all fine. I'm not actually sure where it currently ends up
<gchristensen>
what is it putting in /var?
<ryantrinkle>
sphalerite: thanks :) i do like parts of the current one, and we'll probably reuse some of it, but it's a bit small, i think
<ryantrinkle>
we need to have a bit more content
<ryantrinkle>
we made this site right around when we founded the company (beginning of 2015) and we didn't have as much to show
<gchristensen>
seems like the sort of thing you wouldn't want in /var
<nh2>
gchristensen: I imagine that they imagine that the user may want to modify these hook scripts at run-time
<clever>
id make it an option within nixos, that sets up env vars when the systemd unit gets ran
<clever>
if you have a systemd unit already
<clever>
then patch it to obey the env var
<nh2>
(I personally agree that `make install` should never put things into /var, instead they should put it in `/usr/lib` or what, and change their logic so that it checks if there's a user-script in /var, and if not, fall back to the default script in `/usr/lib`
<gchristensen>
in nixos, I think users would want to modify the hooks at nixos-rebuild time
pie_ has joined #nixos
<clever>
and if the nixos module just ran buildEnv over a list of hooks, and then stuck the resulting path into the systemd unit, it would be able to do that
<nh2>
hmm OK so the current `make install` puts it into `/nix/store/4qk6iy95x6jqga0ixck46v3f8r07rf1r-glusterfs-3.10.1/var/lib/glusterd/hooks`, but at run time it seems to check only `/var/lib/glusterd/hooks`
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/v9HwE
<NixOS_GitHub>
nixpkgs/master 093130f Matthew Daiter: openmvs: init at current master (#25672)...
NixOS_GitHub has left #nixos []
Itkovian has joined #nixos
justelex_ has quit [(Ping timeout: 240 seconds)]
AndreasO has joined #nixos
leat has quit [(Ping timeout: 272 seconds)]
<AndreasO>
Hello, I'm trying out the nixos OVA in virtualbox. The config file just have an import and for some reason the enviroment.systemPackages line wont work.
<Filystyn>
x
<clever>
AndreasO: what error is it giving?
<Filystyn>
./join ##linux
<Filystyn>
damn sry for that ;-)
takle_ has joined #nixos
marsel has joined #nixos
pie__ has joined #nixos
<pie__>
oh my god. new drive reduced boot time to like 10 seconds
<gchristensen>
what was it before?
moet has joined #nixos
pie_ has quit [(Ping timeout: 240 seconds)]
<pie__>
idk, probably like 30 to a minute
<AndreasO>
error: The option `enviroment' defined in `/etc/nixos/configuration.nix' does not exist. (use ‘--show-trace’ to show detailed location information) building Nix...
<clever>
AndreasO: there is a n missing, environment.systemPackages
<pie__>
gchristensen, it dashes through the linux boot diagnostics now
hellrazor has joined #nixos
<AndreasO>
clever: Oh, does it really spell like that... ops.
<AndreasO>
clever: by the way, when a collision is detected i asume it get resolved?
<clever>
AndreasO: systemPackages is set to ignore collisions, it will print a warning at build time, and i think it will use whichever package provided a given file first
<AndreasO>
clever: ignoring a collision? does it mean there is problem. I'm totally new to nix, would be nice to know how things are handled.
<pie__>
huh, gksudo isnt accepting my password
<clever>
AndreasO: if you try to install 2 different things that provide the same binary, nixos will build, and print a warning as it does so
<pie__>
it works but gksudo doesn accept it
<clever>
AndreasO: and only one of those programs will be usable
<gchristensen>
what type of collision are you thinking about, AndreasO?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/v9HPy
<clever>
i usualy just nix-build -E 'with import <nixpkgs>'; callPackage ./default.nix {}'
<clever>
but i supposed you could just stick that into a bash function
zeus_ has quit [(Ping timeout: 245 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
tokudan has quit [(Quit: Leaving)]
romildo has quit [(Quit: Leaving)]
Itkovian has joined #nixos
moet has joined #nixos
acarrico has quit [(Ping timeout: 260 seconds)]
<cstrahan>
gchristensen: what's up
pie_ has joined #nixos
marsel has quit [(Ping timeout: 240 seconds)]
bhipple has quit [(Ping timeout: 258 seconds)]
<pie_>
clever, updating fixed xauthority :
<pie_>
:)
<clever>
pie_: nice
bhipple has joined #nixos
preyalone has joined #nixos
<sphalerite>
steveeJ: did you manage to work around those issues with the cross compiler?
bhipple has quit [(Ping timeout: 240 seconds)]
<gchristensen>
cstrahan: if I update https://github.com/wmertens/sublime-nix, does the linguist repo need to change (I saw you sent a PR to it) to update how github handles nix?
dridus has quit [(Quit: znc? wassat?)]
<cstrahan>
I think so. It's been a while though. Updates would be appreciated -- the syntax highlighting on GH is a little finicky, as I'm sure you've seen :)
<gchristensen>
yeah
ryanartecona has quit [(Quit: ryanartecona)]
<gchristensen>
it always incorrectly highlights some of my code wrong, that I'm trying to fix:
riclima has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
bhipple has quit [(Ping timeout: 260 seconds)]
bennofs1 has quit [(Quit: WeeChat 1.7)]
betaboon has joined #nixos
betaboon has quit [(Changing host)]
betaboon has joined #nixos
johnsonav has joined #nixos
svij3 has joined #nixos
acertain has joined #nixos
svij2 has quit [(Ping timeout: 240 seconds)]
matthewbauer has quit [(Ping timeout: 255 seconds)]
matthewbauer has joined #nixos
<gchristensen>
is adelbertc here?
<nh2>
dammit, I won't get around patching glusterfs source apparently. It has hardcoded to look for xfsprogs binaries in /usr/sbin and /sbin
<gchristensen>
:)
<gchristensen>
this should be fine, I don't think you should be too worried about "increas[ing] risk of data loss quite a bit."
<gchristensen>
you're not making any weird patches, should be fine
<nh2>
gchristensen: in C any character can be a death sentence
<LnL>
lol
<gchristensen>
nh2: have you considered taking up farming in rural Bangladesh?
hiratara has quit [(Ping timeout: 272 seconds)]
moet has joined #nixos
dridus has quit [(Quit: znc? wassat?)]
hiratara has joined #nixos
<iMatejC>
I do not quite get it, but rotfl
<gchristensen>
farming in very wealthy nations usually involves lots of computers and automation, and AFAIK Bangladesh has very little computers involved in most farming operations
ilyaigpetrov has quit [(Quit: Connection closed for inactivity)]
<Guest59096>
Hello, just starting out with nixops. I am trying to deploy a basic example via ssh with deployment.targetEnv = "none"; However it tries to ssh via the root user. How can I specify a different user?
justbeingglad has joined #nixos
MoreTea has joined #nixos
<nh2>
Guest59096: I don't know this for sure but I'd guess it's required to be root to use nixops?
<nh2>
as it's supposed to switch system-wide configuration
justbeingglad has left #nixos []
<Guest59096>
really? I always disable root ssh
<clever>
i disable password, but allow keypair auth
<iMatejC>
+1
<nh2>
Guest59096: there's an sshd option to disable password logins for root
<LnL>
^ that's the default on nixos
<Guest59096>
yes I have that: services.openssh.permitRootLogin = "no";
<gchristensen>
Guest59096: I use = "with-password"
<Guest59096>
So how do I add a keypair for root? I can ssh via another user
<iMatejC>
anyway, crap, nh2 is right, but there is another issue, not only that glusterfs looks xfsprogs binaries inside /usr/sbin, but quick search for sbin keyword inside glusterfs repository... I found 52 files and 30 of that are in C and shell, making a patch and later maintaining it would be a nightmare, better solution would be to replace each occurence with substituteInPlace bash function and with a bit of
<iMatejC>
scripting to iterate through files, done that already a few times myself
civodul has quit [(Quit: ERC (IRC client for Emacs 25.2.1))]
<adelbertc>
gchristensen: oo alright i will look into qaP
<adelbertc>
and thanks for the reference to Nix Pills - i did not link it at the time because i hadn't read it (still haven't .. should get around to doing that)
<adelbertc>
but thanks!
<adelbertc>
glad you liked it :-)
<gchristensen>
adelbertc: I didn't read it for over a year
<adelbertc>
it was one of those opportunities where i figured i should take notes as i was learning Nix and decided to turn it into a blog post
<gchristensen>
finally go around to reading it to pass on to my coworkers to read, and was glad I read them
<adelbertc>
will be sure to read it :)
zraexy has joined #nixos
<Guest59096>
Alright nixos-install gives me "mount point /mnt doesn't exist". I will switch to a supported provider for the first time, it will be easier. Thanks!
<LnL>
you have to mount your root disk there before generating the config
<Guest59096>
I did that the first time, howver the config was already there this time. So I edited it $EDITOR /etc/nixos/configuration.nix and ran "nixos-install"
<Guest59096>
Anyway I will follow a guide first before trying to deviate :)
<gchristensen>
Guest59096: hmm you should be editing /mnt/etc/nixos/configuration.nix
agjacome has quit [(Remote host closed the connection)]
<Guest59096>
Oops - * goes to restore the snapshot
<LnL>
you where editing the configuration of the iso
<manveru>
guess i can just call that directly anyway
digitus has quit [(Quit: digitus)]
<ekleog>
it's about that ;) <matthewbauer> sounds like a bad idea, but it is probably possible
<ekleog>
woops
<ekleog>
<manveru> is there any way to mount a loop device in a derivation?
<copumpkin>
Oh, nope, not without VMs
<ekleog>
or using cptofs to completely avoid the going through a VM, no? :)
mojjo has quit [(Ping timeout: 272 seconds)]
<manveru>
maybe chroot?
<copumpkin>
Yeah! I was so happy when I found that
<copumpkin>
No chroot needs root
<ekleog>
learned about it while reading this commit :D
<copumpkin>
Namespaces help but you can't mount filesystems in them
<manveru>
ok
<copumpkin>
There's also more granular ways to use LKL but cptofs was super easy
<ekleog>
copumpkin: maybe with userns + mountns that would be possible? but would be crazily complex to setup I guess
<copumpkin>
Nope you can still only do simple mounts
georges-duperon has joined #nixos
<copumpkin>
It won't let you mount ext4 for example without some high capabilities
<copumpkin>
Bind mounts, overlayfs, etc. are fine
<copumpkin>
But for images they're not sufficient
<ekleog>
oh ok, didn't know
<ekleog>
thanks :)
<copumpkin>
Np!
<copumpkin>
On mobile so gonna tune out now but feel free to ping
<ekleog>
'night! (or whatever time where you are :))
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
mojjo has joined #nixos
dustmote has joined #nixos
dustmote has quit [(Client Quit)]
jsgrant_om has quit [(Ping timeout: 272 seconds)]
moet has joined #nixos
svij3 has quit [(Read error: Connection reset by peer)]
svij3 has joined #nixos
erasmas has quit [(Quit: leaving)]
markus1209 has joined #nixos
<manveru>
making the boot loader is harder...
markus1219 has joined #nixos
<cstrahan>
ekleog, copumpkin: I suppose a new option could be added to the derivation primitive, where you could specify how you'd like to mount something *prior* to the builder running; assuming the nix-build process is running as root, it could conceivably do the mounting a new mount ns prior to invoking the builder as specified in the derivation. granted, you'd then
<cstrahan>
have a dependency on whatever the host could mount.
markus1189 has quit [(Ping timeout: 240 seconds)]
markus1199 has quit [(Ping timeout: 255 seconds)]
jsgrant_om has joined #nixos
<copumpkin>
Yup although LKL makes it simpler and doesn't need as much support in Nix :)
acertain has joined #nixos
zraexy has quit [(Ping timeout: 255 seconds)]
<manveru>
i guess for now i can use qemu as bootloader...