<{^_^}>
svanderburg/node2nix#8 (by gilligan, 3 years ago, open): expose function to create `node_modules` only
<pie_[bnc]>
grrr but its still the same thing with just the json files! i must be doing something wrong.
<Ovsyanka>
Hello! I just started to discover what is NixOS and installed it in the VirtualBox using demo appliance for VirtualBox. I run it and noticed that virtualbox additions version is outdated (6.0.14). I have VirtualBox 6.1.2 r135662. So, my question is - how I can update it?
<Ovsyanka>
I need it because my NixOS window isn't streaching and shared clipboard doesen't work and etc.
mbrgm has quit [Ping timeout: 246 seconds]
mbrgm_ has joined #nixos
mbrgm_ is now known as mbrgm
mrSpec has quit [Ping timeout: 252 seconds]
mrSpec has joined #nixos
mrSpec is now known as Guest13424
Soo_Slow has joined #nixos
sbetrium has quit [Quit: sbetrium]
fresheyeball has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
sbetrium has joined #nixos
<pie_[bnc]>
joepie91: which is to say lib/node_modules like in the above example doesnt exist for me..
<cole-h>
Make an overlay that appends the location of your patch to the `patches` list, maybe?
sbetrium has joined #nixos
sbetrium has quit [Client Quit]
<bhipple>
selfsymmetric-mu: I have both a 4th gen and 6th gen X1 Carbon on NixOS, but I'd imagine the 7th probably works decently well out of the box. The NixOS hardware stuff is only making minor touchups to power management and shouldn't be hard to run on the 7th gen
<buckley310>
cole-h, i tried to write an overlay but it didnt work. is there anything obviously wrong with this? http://ix.io/2bt6
orivej_ has quit [Ping timeout: 265 seconds]
sbetrium has joined #nixos
rogue_koder has joined #nixos
orivej has joined #nixos
<cole-h>
I'm still new to Nix myself, but if it were me, I would just overlay xf86inputlibinput and not xorg as well. I don't know how well that would work though
rogue_koder_ has quit [Ping timeout: 268 seconds]
sbetrium has quit [Client Quit]
<buckley310>
yeah i tried to just target that one thing, but i dont think you can overlay a submodule alone
<cole-h>
Additionally, shouldn't `pkgs` references be replaced with `super` in overlays? (`pkgs.fetchurl`)
<buckley310>
perhaps. ill try that in case that is somehow breaking it
<cole-h>
Additionally #2: `super.xorg.xf86....` should probably be `superB`, no?
<gchristensen>
https://github.com/NixOS/nixpkgs/pull/79967 the bug clever was helping me with : /dev/stderr was pointing to a socket file which wasn't available inside nixos-enter
<{^_^}>
#79967 (by grahamc, 20 minutes ago, open): nixos-enter: redirect to fd2 instead of a file named /dev/stderr
<gchristensen>
I ... don't think ... it matters .... but sure, I can do that
<cole-h>
It doesn't matter, but it looks nicer :P
<gchristensen>
sure :)
<gchristensen>
thanks, cole-h!
<cole-h>
Sorry, I nitpick everything I come across (see my comments on alacritty pulls lol) ;-)
<buckley310>
cole-h, yeah i still get the error "error: attribute 'xlibsWrapper' missing, at /nix/store/...../gstreamer/legacy/gst-plugins-base/default.nix:43:9"
<gchristensen>
to --check, you have to first build it
<gchristensen>
and so you probably built, and then --check'd it, all on the same day
<gchristensen>
and it was comparing your two builds, not your build against upstream
<bhipple>
ahhhh ok, that makes sense. Yeah --check passed, but I did it on the same day as my rebuild :)
<bhipple>
What a hassle. Maybe we should just delete the docs :D
<gchristensen>
:P
<gchristensen>
gnupg will only be slightly less usable if you delete th edocs :P
<bhipple>
Of all the packages to not take reproducibility seriously . . . haha, yeah
<pie_[bnc]>
so i still havent done it since the last two times i asked, here i am again
Maxdamantus has quit [Ping timeout: 272 seconds]
<pie_[bnc]>
how should i get nixos on hetzner cloud?
<gchristensen>
I think they have an iso?
v88m has joined #nixos
<{^_^}>
[nixpkgs] @dtzWill closed pull request #62920 → xpra: 2.5 -> 2.5.2, mark winswitch broken as discussed → https://git.io/fjgTr
<evanjs>
hrm. Does what's the best way to get a build for linuxPackages_4_1? linuxPackages_custom is complaining about pkgs/os-specific/linux/exfat/default.nix:5:1 (assertion that says EXFAT needs > 4.2)
<evanjs>
not in nixpkgs explicitly, at any rate, e.g. in rtlwifi (just 8188ee there)
<pie_[bnc]>
gchristensen well i found the iso part of the management panel, i kind of dislike how they dont show you any of this stuff unless you already paid for something
<pie_[bnc]>
evanjs btw create_ap is nice
<evanjs>
pie_[bnc]: *checks* oh look, not in nixpkgs :P but yeah also a work thing so little flexibility atm, unfortunately. though I mean if it takes less time to implement... *shrugs* I think the big problem right now is just getting the adapter to work with AP functionality, period
dansho has quit [Ping timeout: 265 seconds]
<evanjs>
we got it working on gentoo and can compile it with 4.1 sources and the drivers from TP Link but I'm trying to jam it into the build system (Jenkins on NixOS) so we don't have to manually do much crap going forward
johnw has quit [Remote host closed the connection]
jluttine_ has joined #nixos
drakonis2 has joined #nixos
mexisme has quit [Ping timeout: 265 seconds]
boxscape has quit [Remote host closed the connection]
<pie_[bnc]>
evanjs check the issue tracker i think there might have been something. or maye i created it. im not sure.
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JvWEh
<pie_[bnc]>
does getting a bigger hetzner machine save me the prive of the smaller one i already paid for?
<emily>
2 gigs of ram / 20gb disk is not going to be fun with nix
<selfsymmetric-mu>
bhipple: Thank you! Unfortunately after an hour and three browsers I can confirm that Lenovo's website is BAD. Checkout always spins out at the end.
<selfsymmetric-mu>
I am considering just buying one from Amazon at the 140USD markup. Alternatively, maybe I don't NEED a Thinkpad.
<{^_^}>
[nixpkgs] @bhipple opened pull request #79978 → treewide: change fetchCargoTarball default to opt-out → https://git.io/JvWuo
fusion809 has joined #nixos
<pikajude>
well, my hydra is refusing to run the hydra vm-tests
<pikajude>
and i'm not sure why, because it's not a required-features problem
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Ridout has quit [Quit: Lost terminal]
<bhipple>
selfsymmetric-mu: do you by any chance have a protonmail email? I spent hours in checkout on the phone and had to get help from a sales rep, only to find out they have "known" but undocumented issues handling any protonmail acct :/
<pikajude>
machine 'localhost' does not support step '/nix/store/abdiqpql2xv3d111sq40q32wm6dymmj9-vm-test-run-unnamed.drv' (system type 'x86_64-linux')
<pikajude>
yeah, not sure about that one hydra
<bhipple>
If you can use paypal as an intermediary that also can reduce issues. Kinda comical a company of Lenovo's size wouldn't have this ironed out ...
<selfsymmetric-mu>
bhipple: Incredible. I believe it. In my case I was checking out as a guest with a gmail account.
<selfsymmetric-mu>
I'm not sure if I saw a PayPal option. I did try Amazon Pay. No dice there either.
<bhipple>
If they didn't make really nice hardware I would've given up after the first hour, since it was an awful checkout experience
<selfsymmetric-mu>
It truly is embarrassing. They should use some of the money they made selling PPI and hire themselves a web developer.
NoctisLa1 has quit [Read error: Connection reset by peer]
NoctisLa1 has joined #nixos
sbetrium has joined #nixos
sbetrium has quit [Client Quit]
sbetrium has joined #nixos
sbetrium has quit [Client Quit]
instinctive has joined #nixos
nikola_i has joined #nixos
<instinctive>
hi all - noob nixos user... just installed 19.09 onto a 2nd ssd following the manual, rebooted, appears to work, except I do `$ nix-env -qa` and get nothing... I'm sure I'm missing something obvious?
butterthebuddha has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<eoli3n>
simple question of the day : what is the benefit of the fact that nix is a functional language ? why that choice ?
Guest13424 has joined #nixos
Guest13424 has quit [Changing host]
timor has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @siriobalmelli opened pull request #79996 → mscgen: change platforms to 'all' after successful compile on macOS → https://git.io/JvWV0
TwoNicToupper has quit [Remote host closed the connection]
sbetrium has joined #nixos
sbetrium has quit [Client Quit]
<Yaniel>
eoli3n: it is stateless, like the package manager aims to be (as far as possible)
<eoli3n>
ok :)
<eoli3n>
thx
m0rphism has joined #nixos
<Yaniel>
and functional code is often easier to reason about than imperative, once you know the basics
<thongpv87>
When building with nix, I got this error: a 'x86_64-linux' with features {kvm} is required to build '/nix/store/x1c50mvix7fgw12y3h5k71akjq1l4w6h-docker-layer-myapp.drv', but I am a 'x86_64-linux' with features {benchmark, big-parallel, nixos-test}
<eoli3n>
thongpv87: FYI there is an "ix" package
xkapastel has joined #nixos
<thongpv87>
I found that, if the command `runAsRoot` is commented out, then that error disapear
<thongpv87>
But my app need to find zoneinfo in `/usr/share`, so I need those command.
orivej has quit [Ping timeout: 265 seconds]
<thongpv87>
Why does it happen, and how can I solve it.
<thongpv87>
Thanks @eoli3n
Tucky has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos
<Yaniel>
looks like it needs root permissions (duh) but nix-build can only have those in a virtual machine or something?
dm9 has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ashkitten>
Rovanion: nix packages don't have service files. you can either write your own or use nixos which provides the services.slurm configuration option
nikola_i has quit [Ping timeout: 240 seconds]
thongpv87 has quit [Remote host closed the connection]
FRidh2 has quit [Quit: Konversation terminated!]
<ashkitten>
nix itself does not depend on systemd and is mostly os-agnostic so it wouldn't make sense to provide systemd service files anyways
sbetrium has joined #nixos
<Rovanion>
ashkitten: Could it not make sense for nix to provide init scripts/service files if available and compatible with the underlying OS. I mean, it does have knowledge of OS and architecture to provide OS and architecture specific binaries/build recipies.
sbetrium has quit [Client Quit]
<Rovanion>
But then I know: Nix provides no service files for no package.
cfricke has joined #nixos
<ashkitten>
Rovanion: consider that most packaged programs do not provide their own service files and distros have to write them themselves. we only support service definitions on nixos for this reason, but i'm sure you could hack together something that uses the slurm nixos module code to generate and install a systemd service file on ubuntu if you want
<ashkitten>
or you could just write your own
sbetrium has joined #nixos
<Rovanion>
I don't know Nix (the language) well enough to go the first route so I'm going to write my own.
<eoli3n>
in a custom script how to found nix-store path ?
<eoli3n>
i dynamically need absolute path of a file which isn't in PATH
sbetrium has quit [Quit: sbetrium]
o1lo01ol1o has joined #nixos
<ashkitten>
eoli3n: you should be able to get the path of a built derivation with `nix-build --no-build-output --no-out-link '<nixpkgs>' -A hello` for example
cfricke has quit [Ping timeout: 240 seconds]
<ashkitten>
do note that this will build the derivation if it's not built already
o1lo01ol1o has quit [Remote host closed the connection]
graf_blutwurst has quit [Remote host closed the connection]
graf_blutwurst has joined #nixos
lorimer has quit [Read error: Connection reset by peer]
gxt has quit [Ping timeout: 240 seconds]
mexisme has quit [Ping timeout: 246 seconds]
<{^_^}>
[nixpkgs] @basvandijk opened pull request #80006 → expect: remove autoreconfHook since that causes a wrong path to stty → https://git.io/JvWKq
<gentauro>
I need to install a version less than `0.13` of `joblib`. For that I have made the following `default.nix` in my dev folder and I run it with `nix-shell`
<Ashy>
hmm, paraview seems to be broken on 19.09 atm with this error: https://nixos.wiki/wiki/Qt#qt.qpa.plugin:_Could_not_find_the_Qt_platform_plugin_.22xcb.22_in_.22.22
<eso>
Hello there, I was asking myself: what is used under the hood to build docker images using pkgs.buildDocker ? is it docker itself (i hope not) or another tool ?
<rajivr___>
I was wondering it there is a derivation equivalent to debian "build-essentials" in NixOS? I tried `nix-shell -E '(import ./nixpkgs.nix {}).stdenv'` but that does not seem to work.
<evanjs>
e.g. assert lib.versionAtLeast kernel.version "4.2" || lib.versionOlder kernel.version "4.0";
<evanjs>
though the bigger question for this one is, how might I support both 4.1 and any other version? Like if conditional branch based on kernel version or etc
<evanjs>
is there an easier way of testing a new kernel module than referencing a local nixpkgs repo? an overlay, perhaps? Haven't dealt with kernel modules in one before
anohigisavay has quit [Quit: Konversation terminated!]
Soo_Slow has joined #nixos
justanotheruser has joined #nixos
sbetrium has joined #nixos
sbetrium has quit [Client Quit]
<j4m3s__>
question, I'm developing some C on my machine, How do I do to make gdb able to debug my program ? (it fails because it's missing debug information, but symbols are there in the glibc / ld.so). Strace shows that gdb fails to write in memory with Ptrace (Gettext) )
<j4m3s__>
Do you guys have any idea ?
<multun>
just compile with -g
<multun>
works for me
bvdw has quit [Read error: Connection reset by peer]
<Soo_Slow>
privacy question. Snap creates unique machine id (officially: for statistics). Iirc flatpak has something like that too. Does nix package manger suffer from that too?
iyzsong has quit [Read error: Connection reset by peer]
domogled has joined #nixos
<Taneb>
I don't believe so
orivej has quit [Ping timeout: 240 seconds]
__monty__ has quit [Quit: leaving]
<gchristensen>
nope
sbetrium has joined #nixos
sbetrium has quit [Client Quit]
jethroksy has joined #nixos
fendor has quit [Read error: Connection reset by peer]
domogled has quit [Quit: domogled]
<{^_^}>
[nixpkgs] @vcunat pushed 2 commits to release-20.03: https://git.io/JvWXM
<clever>
evanjs: is this a module (in imports) or a package (loaded with callPackage)
domogled has joined #nixos
<clever>
evanjs: its in imports, so it must set boot.kernelPackages, not return a set of linuxPackages
noudle has joined #nixos
<pie_[bnc]>
does stuff like forwardPorts = [ { containerPort = 8080; hostPort = 27777; protocol = "tcp"; } ]; not work with extra-container?
<evanjs>
clever: right, just noticed that, and updated accordingly
<evanjs>
clever: was more trying to get it to return a set that I can import, and then set like I would normally, e.g. boot.kernelPackages = linuxPackages_4_1;
gxt has quit [Ping timeout: 240 seconds]
<clever>
then you would want to load it with callPackage, and change line 1 to just linuxPackages, linuxPackagesFor
<clever>
rnhmjoj: something is -9'ing the entire eval
dingenskirchen has quit [Remote host closed the connection]
dingenskirchen has joined #nixos
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
veske2 has quit [Quit: This computer has gone to sleep]
<rnhmjoj>
ah, i see. so whatever happened is affecting both unstable and release-20.03
sbetrium has joined #nixos
sbetrium has quit [Client Quit]
Guest13424 is now known as mrSpec
cjpbirkbeck has joined #nixos
orivej has joined #nixos
<tilpner>
rnhmjoj: If I understand #nixos-dev backlog, it's either using custom Nix from the flakes branch, or trying to build a non-master nixpkgs
<gchristensen>
neither are related
fusion809 has quit [Remote host closed the connection]
* tilpner
doesn't understand #nixos-dev backlog
<gchristensen>
master and 20.03's evaluations take a lot of memory -- reaching 34G before being OOM'd
<gchristensen>
recenlty, several tests were added in one swoop which probably pushed it over: each nixos test vm requires a few hundred mb to evaluate
dirtwrangler has quit [Remote host closed the connection]
<gchristensen>
there is some looking at a more effective GC for Nix to see about fixing that, and another option might be removing some of those tests for the time being
<clever>
gchristensen: i was also experimenting with a method of forking, to reduce heap and speed up eval, when we where at nixcon, its currently stalled in segfaults and concurrency bugs
<evanjs>
removing those assertions in my local nixpkgs seems to work... though I'd rather not do that. Any way to skip specific packages in kernelPackages? I mean I guess I shouldn't be building them all in the first place, right?
<eoli3n>
it seems to be normal that packages doesn't work after a few...
<clever>
evanjs: something must be adding it to extraKernelModules
<eoli3n>
tried to install genymotion, needed to open issue. just tried to install steam, not working
<eoli3n>
how is it possible has packages are stateless and has every deps in it ??
dansho has joined #nixos
<eoli3n>
as i understand, where a package is "created" it should work everytime
<eoli3n>
packages could become outdated but not broken
hyper_ch2 has quit [Remote host closed the connection]
<immae>
eoli3n: Say at a given point you have package foo that depends on package bar, and it works. Then later you upgrade bar. If you don’t try to build foo after upgrading bar, then foo may be incompatible with the new version of bar and you didn’t notice that it broke, because you didn’t try build it. To ensure that foo works forever, you would need to point to a precise version of bar which cannot change
<immae>
(and recursively if bar has dependencies too)
<eoli3n>
immae: and nix doesn't do that ?
<immae>
you can do that if you feel like it, but the set of derivations that you’ll find in nixpkgs doesn’t do that no
<eoli3n>
so it "is normal" that a package break after a few ?
<immae>
I mean: you can write a set of derivation that is garanteed to work until the end of times, but that’s not what nixpkgs do
<immae>
Yes it is
<justanotheruser>
for some reason, when I install python3Packages.pykeepass by including it in environment.systemPackages in configuration.nix, then running `nixos-rebuild switch` I fail to import pykeepassx. However, if I run `nix-shell -p python3Packages.pykeepass`, I can import it fine within the shell. Any idea why this might be?
<eoli3n>
immae: that's sad
<eoli3n>
nix magic is gone in my head :)
<Taneb>
eoli3n: the specification of the package is the entire of nixpkgs
<immae>
eoli3n: no, packages evolve with time. What is nice though is that it’s very easy to pin everything to a given working state if you need
<Taneb>
If you keep that the same then nothing will change
<Taneb>
But if you ever change any of nixpkgs, things may change
<eoli3n>
the only thing i see is that, as end user, i can't trust packages.
v88m has quit [Ping timeout: 265 seconds]
<eoli3n>
i install them and often need to create an issue
<eoli3n>
an nix learning curve doesn't let me tweak it to workaround
gustavderdrache has joined #nixos
<eoli3n>
so i'm stuck with a package which doesn't work waiting for a dev to take a look at this
<immae>
eoli3n: being able to workaround (and understand what you do) will come quickly
<eoli3n>
immae: i'm reading pills since some days, and its really hard to get, even if pills are great
mac10688 has joined #nixos
<immae>
eoli3n: at some point you should start working on your own derivations, nix-pills may be great but doing things by yourself will get you where you want
<{^_^}>
[nixpkgs] @bennofs pushed commit from @andir to master « nix-prefetch-git: add gawk to PATH »: https://git.io/JvWD7
nikola_i has joined #nixos
<eoli3n>
as nix let us have multiple version of same packages, it should be default to pin versions to let packages be infinitly functionnal. Packages becomes outdated anyway so, upgrading packages is required, it seems better to upgrade the package AND its deps
<immae>
eoli3n: I disagree with you there, since it’s very easy to actually pin a package in the way it’s currently done you don’t need to enforce it for "latest" snapshot (which would become very big otherwise)
<immae>
eoli3n: if you want a package foo that you know is working in some nixpkgs commit abcd, then you can add: nixpkgs-foo-working=https://url/to/commit to the NIX_PATH, and then install (import <nixpkgs-foo-working> {}).foo in your derivation
<immae>
This way you’re garanteed to have the foo that was working at that time
<immae>
with all its dependencies down to glibc itself
<simpson>
eoli3n: It's easy to imagine a Nix-like system in which every branch modification is checked somehow to verify that all leaves are still viable. Such a system would take a *very* long time for certain changes, since some leaves have extremely dense builds.
<simpson>
Hydra's "leaky integration" (sorry for the pun!) is a standard trick: Check lots of branches at once. There's so many branches to check that we wouldn't otherwise have time to do checks.
<eoli3n>
i get what you mean immae, and yes simpsons, maybe nix packages needs a huge infrastructure to make CI on the whole thing :)
<eoli3n>
but in my exemple, i can't determine which branch to use for a certain package to work as i just discovered that latest is broken. And i don't want to pin ALL nixpkgs
<eoli3n>
s/which branch/which commit
<immae>
then, look at your package derivation in the nixpkgs repo, see when it last changed, and you can be quite assured that this commit will work
<simpson>
Ah. For that, I'd use git-bisect. nixpkgs isn't always bisectable, but often enough it is.
<eoli3n>
if package foo need commit n°1, and bar need commit n°3 to work, that's problematic
<simpson>
Or immae's git-blame technique is also good, for rarely-changed leaves. You'll want to bisect to find the branch package that changed and broke things.
<eoli3n>
i don't know git-bisect
ryanartecona has joined #nixos
NoctisLabs has joined #nixos
<immae>
simpson: well, the goal here is to get a commit that you know will work. When such derivation changed last is a good optoin to look at to assume that it will probably work
<eoli3n>
i was thinking that nix solved problem in front of other distros, but here, it seems to me a regression. A end user should be sure that installing a package works
<simpson>
No worries. Learning git in anger/frustration is a possibility, but not the only one. So, there's two answers in that case. One answer is to use *two* nixpkgs in two different builds, and then mix the resulting packages using (either) nixpkgs' standard environment-builders on the resulting paths explicitly.
<simpson>
The other is to cherry-pick one commit onto another, which is still git-in-anger, but relatively minor git.
NoctisLa1 has quit [Read error: Connection reset by peer]
<simpson>
eoli3n: FWIW, most packages *do* work, for reasonable definitions of "most", "packages", and "work"; this is what Hydra gives us.
cjpbirkbeck has quit [Quit: cjpbirkbeck]
<simpson>
Steam is a good example of a special package: It's unfree, requires the complexity of FHS, and has to change often. It's very hard to know *why* it is broken.
cjpbirkbeck has joined #nixos
<eoli3n>
ahh didn't know about hydra, nice
<eoli3n>
who pays for this ?
<eoli3n>
Delft University of Technology ?
<gchristensen>
we pay for some of it by donations, and a lot of it is contributed very generously by Packet.com
<simpson>
There's a NixOS Foundation; they manage resources donated by many different groups. Corporate support gives the capital, and volunteer support gives the labor.
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « python27Packages.av: 6.2.0 -> 7.0.1 »: https://git.io/JvW9l
<evanjs>
clever: yah not much here unless I've been missing it the whole time---on top of which... looks like the build for 4.1.8 fails with something about gcc, wondering if I should start looking at a different thing
<clever>
evanjs: what does --show-trace show?
<evanjs>
goal: get this dumb chip (r(tl?)8188eu) working with the initrd stuff)
graf_blutwurst has quit [Remote host closed the connection]
eso has quit [Quit: WeeChat 2.6]
<jared-w>
So I'm trying to figure out how I can import a nix store and then run nix build without needing channels, internet access, or anything else. No matter how large of a closure or arcane incantation I figure out to use to grab more stuff from my store, nix keeps trying to download the channels first
<tilpner>
jared-w: What's your nix-build call?
mexisme has joined #nixos
<jared-w>
tilpner: on the system with internet I run: nix build. The default.nix uses a pinned nixpkgs, and runs buildEnv with a few packages.
<tilpner>
jared-w: How do you determine the closure to transfer?
<jared-w>
To dump everything I've tried: nix-store --export $(nix-store -qR --include-outputs $(nix-store -q --deriver ./result)) > full.nar (which was a much larger closure than the first attempt)
sbetrium has quit [Quit: sbetrium]
<clever>
jared-w: --export doesnt generate a nar, it generates a stream containing many nar's
<clever>
jared-w: nix copy --to local?root=/mnt/ might be simpler to manage
justanotheruser has joined #nixos
<clever>
jared-w: you can also --option substituters 'local?root=/mnt https://cache.nixos.org' to make nix look in there, and grab what it can
NoctisLabs has joined #nixos
<jared-w>
clever: ah okay, that makes sense. I was looking for a nix v2 way to do it but couldn't find anything
<tilpner>
jared-w: How does default.nix pin nixpkgs, and how are you transferring the pinned nixpkgs to the other host?
<clever>
jared-w: and why do you think its trying to download a channel?
<jared-w>
tilpner: I'm using niv to pin nixpkgs, so it should be a sha256+fetchurl I believe?
<clever>
jared-w: i use code like this, to bundle all of the things from niv into a single derivation (called ifd-pins here), and then you can nix-build and copy-closure that
NoctisLabs has quit [Read error: Connection reset by peer]
NoctisLa1 has joined #nixos
vandenoever has joined #nixos
<jared-w>
will nix-copy-closure work if I just want to dump the result into a file that I can manually upload to the destination machine?
<clever>
jared-w: thats where i prefer `nix copy --to local?root=/dir`, then you can tar it up or rsync it over
<clever>
jared-w: although, how are you uploading it to the machine?
NoctisLa1 has quit [Read error: Connection reset by peer]
<clever>
jared-w: there can be more optimal ways
NoctisLa1 has joined #nixos
<marek>
I'm thining of adding a git revision to a specific version file, but I do not want to trigger unnecessary rebuilds, so I thought of adding it directly to the release.nix arg attrset, what is the best way to do that?
<jared-w>
I unfortunately have to upload it by way of the browser. This may or may be coincidentally related to my rants on bad file upload UI patterns in #nixos-chat
<clever>
marek: hydra passes the git revision to release.nix as an argument
NoctisLabs has joined #nixos
<clever>
jared-w: step 1, fix whatever is forcing you to jump thru such hoops :P
<jared-w>
heh, wouldn't that be nice
<marek>
clever: how could I get that value and write that out to a specific file inside my derivation (outsite hydra, let's say, for example VERSION file in the release tarball)?
<jared-w>
ifd-pins is great
cjpbirkbeck has quit [Quit: cjpbirkbeck]
<jared-w>
No way would I have figured that out in any sane amount of time lol
* jared-w
makes a mental note to just read through all of IOHK's nix files
NoctisLa1 has quit [Ping timeout: 268 seconds]
NoctisLa2 has joined #nixos
<marek>
clever: or even better, actually adding it somewhere in the releaseTools.channel
<clever>
marek: if you have an input named foo in hydra, then hydra will pass release.nix an attrset named foo, that contains the revision
<clever>
marek: this code is providing a default, for when hydra wasnt in use
<jared-w>
Hmm... Any reason a 'writeShellScript" would be causing nix copy to break due to lacking a valid signature?
NoctisLa2 has quit [Read error: Connection reset by peer]
<marek>
clever: oh I see, nice, thank you
<clever>
jared-w: the --export lacks signatures by default, and you need to either sign everything yourself, or disable signatures when importing
veske has joined #nixos
veske2 has joined #nixos
<jared-w>
nix copy --yolo it is, then
NoctisLabs has quit [Ping timeout: 246 seconds]
<clever>
marek: you can then route that string down to whatever is building the version file
NoctisLa2 has joined #nixos
<clever>
jared-w: one idea i have had in the past, to deal with air-gapped machines, is to `nix copy --to local?root=/mnt/usb` then physically move the usb stick over the air gap, and --from on the other end
<clever>
jared-w: the benefit that has over --export, is that it only copies what is missing from the usb stick, so changes are incremental
<jared-w>
clever: ah nice
<clever>
jared-w: but if your crappy UI lacks recursive upload, thats going to be even worse :P
ixxie has joined #nixos
<jared-w>
Yeah it lacks recursive upload. Isn't that nice? I miiight be able to do some sort of shennanigans with sticking the nix store inside a shared "repo" thingy but that also scares me a bit because people can pull and push at any time, including when the store is in an inconsistent state, and that sounds like a recipe for utter disaster lol
Chiliparrot has joined #nixos
<tilpner>
jared-w: Why do you want to build something on a system this restricted?
<jared-w>
tilpner: define "want"
knupfer has joined #nixos
<jared-w>
I enjoy my salary? :p I'm doing this so that other people having to actually use these machines don't have to deal with an OS that lacks every possible tool and is 10 years out of date? Various reasons I suppose lol
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « postman: 7.16.1 -> 7.18.0 »: https://git.io/JvWH6
<tilpner>
jared-w: Why do you need to instantiate locally? That sounds like you could just transfer the built tools?
<clever>
that also sounds like an improvement
<jared-w>
Assuming they're all statically linked, sure, but the real overall scheme here is to be able to give people a chroot-like environment so things like python work correctly
<clever>
jared-w: they dont have to be staticly linked, just import them the same way your importing this closure
<jared-w>
so it's really a build environment, not just a few binaries
<rnhmjoj>
nobody: it appears to be packaged in nixpkgs since april 2018. so you should just add python3Packages.jupyterlab to your configuration.
<clever>
jared-w: once it lands in /nix/store, the dynamic stuff will keep working
<marek>
clever: I'm trying to think about this the other way, what if I can do a version file somewhere in /etc that contains lib.commitIdFromGitRepo but I do not want hydra to cause rebuilds on each commit (which would change the derivation), do you think this is something sane to do?
<jared-w>
True. I might be shooting myself in the foot here a bit with thinking I need to build the profile before I actually use it
<clever>
marek: haskell solves this with file-embed, build it once, with a magic string in the binary, then have a cheap derivation that will basically sed a new version into the raw binary
<Ilya_G>
I am creating a qcow2 image using `nixos-generate` and I end up with the following issue when build process tries to reload that image: `qemu-system-x86_64: CPU model 'host' requires KVM`
<Ilya_G>
any ideas as to why this occurs?
sbetrium has joined #nixos
zupo has joined #nixos
<marek>
clever: interesting! looking thru some code, I found it might be possible to add my versionSuffinx directly into system.nixosVersion
<doublej472>
ilya_g: maybe hardware virtualization is turned off in the bios?
smatting has quit [Ping timeout: 260 seconds]
<clever>
marek: the binary patching, is basically just replacing a char *foo="biglongstring"; with "small\0\0\0\0\0" (of the same length)
<clever>
marek: as long as you ensure you only replace the right bytes, and dont change the lenght of the thing your replacing, you can patch almost any compiled binary like that
<Ilya_G>
doublej472 I am building this inside nixos docker image and I was able to do this before no problem.
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « piper: 0.3 -> 0.4 »: https://git.io/JvWHN
<clever>
ilya_g: you either need to tell qemu to not use kvm, or ensure that /dev/kvm exists within the docker container
<Ilya_G>
I remember struggling with this one and not sure what cause it to blow up on me again. :(
<gustavderdrache>
ilya_g: last time i had that issue, it was because my user wasn't part of the 'kvm' group (on a non-NixOS system) despite having the necessary kvm packages
<clever>
ilya_g: thats weird, lines 1-8 show part of qemu did work, so why is it failing on 12
justanotheruser has quit [Read error: Connection reset by peer]
<clever>
ilya_g: what does `nix show-derivation /nix/store/cz3pxdglzyfq6nf6mxd8hcpdck9x6m3v-nixos-disk-image.drv` reveal?
NoctisLa1 has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Ilya_G>
@clever I ran a re-biuld will get back to you shortly.
NoctisLabs has quit [Ping timeout: 260 seconds]
NoctisLa2 has joined #nixos
<OR[m]>
Hello! I just started to discover what is NixOS and installed it in the VirtualBox using demo appliance for VirtualBox. I run it and noticed that virtualbox additions version is outdated (6.0.14). I have VirtualBox 6.1.2 r135662. So, my question is - how I can update it?
<OR[m]>
I need it because my NixOS window isn't stretching and shared clipboard doesn't work and etc.
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<Yaniel>
fork nixpkgs on github, update the package and make a pull request
<Yaniel>
oh yeah and that in case your image is old
<clever>
OR[m]: if the additions within the latest nixpkgs are still too old, you may need to do what Yaniel said
<{^_^}>
[nixpkgs] @edolstra opened pull request #80036 → Disable most VM tests → https://git.io/JvW7U
<{^_^}>
[nixpkgs] @davidtwco opened pull request #80037 → snipe-it: init at 4.8.0 + module → https://git.io/JvW73
<gchristensen>
davidtwco: ^ cool :)
v88m has quit [Ping timeout: 272 seconds]
<OR[m]>
clever Yaniel Thank you, I'll try it. Offtop: I am new in the Matrix too and am I right that you both are from the freenode irc channel and don't see my reactions to your messages?
<Yaniel>
we don't see reactions, correct
NoctisLabs has joined #nixos
zupo has joined #nixos
apriluwu has joined #nixos
NoctisLa2 has quit [Ping timeout: 268 seconds]
psyanticy has joined #nixos
<psyanticy>
Hi,
<psyanticy>
Im trying to run a custom telegraf check using a python script
<psyanticy>
anyone have an idea what i m i m missing
<psyanticy>
i added telegraf the the nixbld group but same issue
chloekek has joined #nixos
<davidtwco>
gchristensen: thanks, laravel made it a pain but I managed.
NoctisLa1 has joined #nixos
NoctisLabs has quit [Read error: Connection reset by peer]
<gchristensen>
davidtwco: I wonder if the patches you've made could be solved more elegantly with help from snipe / her team?
<davidtwco>
gchristensen: almost certainly, there’s some stuff that was pretty laravel specific but the vast majority of the patches was modifying snipe so that I could change the upload directory.
veske2 has quit [Quit: This computer has gone to sleep]
veske has quit [Quit: This computer has gone to sleep]
<davidtwco>
I intend to file some issues upstream now that I know what needs changed.
<gchristensen>
cool
<gchristensen>
I told her about the PR already, because I think she'll like it
<davidtwco>
Oh, cool!
NoctisLa1 has quit [Read error: Connection reset by peer]
NoctisLa1 has joined #nixos
<Ilya_G>
@clever it came back with nothing sadly
<Ilya_G>
{ "/nix/store/cz3pxdglzyfq6nf6mxd8hcpdck9x6m3v-nixos-disk-image.drv": {}error: opening file '/nix/store/cz3pxdglzyfq6nf6mxd8hcpdck9x6m3v-nixos-disk-image.drv': No such file or directory}
nikola_i has quit [Ping timeout: 260 seconds]
<clever>
ilya_g: the exact path, depends on the hash of all of the inputs, so it might change when you edit the source
<clever>
ilya_g: you need to run the query in the container, after the build fails
NoctisLa1 has quit [Read error: Connection reset by peer]
ryanartecona has quit [Quit: ryanartecona]
apriluwu has joined #nixos
NoctisLabs has quit [Read error: Connection reset by peer]
apriluwu has quit [Client Quit]
NoctisLabs has joined #nixos
Seto_Kaiba has joined #nixos
claudiii has quit [Quit: Connection closed for inactivity]
Hunterkll has quit [Ping timeout: 240 seconds]
noudle has quit []
NoctisLa1 has joined #nixos
Neo-- has quit [Ping timeout: 245 seconds]
Chiliparrot has joined #nixos
NoctisLabs has quit [Ping timeout: 240 seconds]
justanotheruser has joined #nixos
ddellacosta has joined #nixos
konobi has quit [Quit: konobi]
mexisme has joined #nixos
nek0 has quit [Remote host closed the connection]
NoctisLa1 has quit [Read error: Connection reset by peer]
NoctisLa1 has joined #nixos
<jared-w>
clever: alright so turns out I was two parts dumb and on the right track
vidbina has quit [Read error: Connection reset by peer]
vidbina has joined #nixos
<jared-w>
clever: the dumb bit--I never made the connection that I could just do `nix-env -p $some_path -i /nix/store/my-kewl-env` which is dumb because I was doing it earlier with the result symlink... which is just the store path. *sigh* ok whatevs.
nikola_i has joined #nixos
Blessjah has joined #nixos
nek0 has joined #nixos
xkapastel has quit [Quit: Connection closed for inactivity]
NoctisLa1 has quit [Read error: Connection reset by peer]
NoctisLa1 has joined #nixos
mexisme has quit [Ping timeout: 240 seconds]
<jared-w>
clever: so, that works; don't need build dependencies, I can get away with the tinier store closure and simpler command, all is welmz. That being said, if I ever *do* want to rebuild the env for any reason, I'm screwed because I don't have the build dependencies. But thanks to your nifty ifd-pins, I should have all of that now and be able to dump that. `nix copy` combined with `nix from` should be all I need. As a bonus,
<jared-w>
I can treat it like "db migrations" where I have a series of tar bundles, each one being the incremental changes, so that if we ever do a thing like update a python lib, people don't need to import 3GB of store for no reason.
* jared-w
narrator voice: At least, in theory...
NoctisLa1 has quit [Read error: Connection reset by peer]
<clever>
jared-w: --export and `nix copy --to local?root=` both produce the full closure, not incremental
<clever>
jared-w: things only get incremental if you can rsync the result from local?root= and copy the differences
<jared-w>
yeah but I can do some nonsense with rsync in the middle to figure out what the minimal amount of files to copy over would be. Have rsync do a dry run and tell me what it'd copy over and then grab that and bundle that rather than everything
* jared-w
needs to faster type type
wavirc22 has quit [Ping timeout: 265 seconds]
linarcx has quit [Quit: WeeChat 2.7]
<jared-w>
either way, I don't expect to actually have to do that, but it's nice to know I can do that easier than the other way
<clever>
boogiewoogie[m]: step 3, read lines 30-77, and see that it never adds anything to systemPackages, it just configures a usermode timer, to run it at a given interval
<boogiewoogie[m]>
clever: yep, did both of them! I'm not fluent in nix though.
<boogiewoogie[m]>
clever: oh, I see. I guess there's a reason for that then :)
<clever>
boogiewoogie[m]: line 34-49 defines a psd service, that will always be running, and it runs sync on startup, and unsync on shutdown
<clever>
boogiewoogie[m]: and 51-73 defines a psd-resync, that runs at a set interval
nobody has quit [Ping timeout: 260 seconds]
<clever>
and then it "just works" without poluting your $PATH with things
o1lo01ol1o has joined #nixos
Tucky has quit [Quit: WeeChat 2.7]
wildtrees has quit [Quit: Leaving]
<Ilya_G>
clever: I was not able to do it by just bashing into the docker instance and running it
<{^_^}>
[nixpkgs] @tfc opened pull request #80048 → nixosTests.docker-tools: Port to Python → https://git.io/JvWNG
<clever>
ilya_g: you have to run it in a container that has attempted to build the disk image
<Ilya_G>
I am going to rerun the script with that command in it right after fail.
<clever>
ilya_g: manually run the same nix-build in the container, to create the file
<clever>
yeah
NoctisLabs has joined #nixos
<{^_^}>
[nixpkgs] @tfc opened pull request #80049 → nixosTests.docker-containers: Port to python → https://git.io/JvWNZ
NoctisLabs has quit [Read error: Connection reset by peer]
<clever>
ilya_g: a: the path i gave you was the output, not the .drv file, that one is much cleaner
scoates has joined #nixos
<Ilya_G>
:( sorry
<clever>
ilya_g: b: run `nix-store --query --binding buildCommand /nix/store/hash-vm-run.drv`
<clever>
ilya_g: that will extract 1 field for you
<clever>
ilya_g: but we want the output, not the buildCommand, which is in line 1 of your gist, now /nix/store/7cf2ziyfwd7r4dpxy46mf8fpcblgscgk-vm-run
<clever>
ilya_g: the gist contains a script that generates a script
<scoates>
hey; I'm having a *lot* of trouble with nix + pipenv. I've even tried wiping the /nix/store + /nix/var (this is on Mac). if I do `nix-shell -p 'python3' -p 'pipenv'` I get the system `pip`, which has a bad import. Any ideas? I've read through this and thought it was resolved, but I'm still in a mess. https://github.com/NixOS/nixpkgs/issues/71178
<{^_^}>
#71178 (by mawis, 17 weeks ago, closed): Difficulty using pip in a venv
<Ilya_G>
clever: will do. I ended up formatting it so new lines are normal and not \n and it looks like it is running quem-kvm and expects $QUEM_OPTS at the end. Do I just possible need to some how feed `--no-kvm` into those and that would solve my problem?
<clever>
ilya_g: the shell script qemu-kvm will test if /dev/kvm exists, and automatically omit -kvm if it doesnt
<clever>
ilya_g: the problem is the -cpu host, which isnt as smart, you may be able to use -cpu again to overwrite it
<clever>
ilya_g: you will also want to read the original nix expr for this
<Ilya_G>
clever: I am still running the command you asked me with the correct `hash`, right?
<clever>
$ vi pkgs/build-support/vm/default.nix
<clever>
195 qemuCommandLinux = ''
<clever>
197 -nographic -no-reboot \
<clever>
196 ${qemuBinary qemu} \
<clever>
ilya_g: i dont see a -cpu host, on nixpkgs master
<evanjs>
currently it's pulling the host system's build, but what I want is the version of the linuxPackages/kernel passed into this module's expression
vandenoever has quit [Read error: Connection reset by peer]
<Ilya_G>
clever: sorry what does "move it to Haswell" mean?
<clever>
evanjs: and this lets you add another custom out-of-tree thing to a set
<evanjs>
clever: yup I'm doing it via rtl8188eu = kernelPackages.rtl8188eu; in default where it links to my kernel override in nixos.nix
<evanjs>
got that all figured out, TP-Link driver broadcasting, now I'm trying to see if the realtek one from https://github.com/lwfinger/rtl8188eu/ works
<evanjs>
wow I misread that. thanks samueldr ... here I am trying to replace KSRC... when what you linked me sets it.... right XD
<samueldr>
:D
<samueldr>
there *may* be other patterns up there in `pkgs/os-specific/linux`, so feel free to ag/grep/ripgrep/whatever and pick your favourite
<samueldr>
that's literally the first example of KSRC I found there
<evanjs>
glad you added ripgrep there :P and ah alright
nikola_i has quit [Ping timeout: 240 seconds]
alexherbo29 has joined #nixos
<samueldr>
the few times I had to deal with out of tree modules, that's how I got things working without having to think much
alexherbo2 has quit [Ping timeout: 240 seconds]
alexherbo29 is now known as alexherbo2
travelion7 has quit [Quit: bye]
<{^_^}>
[nixpkgs] @jonringer opened pull request #80056 → R ryantm/auto update/python2.7 sphinx jinja → https://git.io/JvWA9
<{^_^}>
[nixpkgs] @jonringer pushed commit from @Ma27 to master « mono: fix build w/glibc-2.30 »: https://git.io/JvWxn
<clever>
evanjs: basically, the algo is /nix/store/${hash sha256 name}-${name}
<clever>
evanjs: so if the name changes, the hash totally changes, and it cant find the old $out, even with the same sha256
<ajs124>
Minijackson[m]: no idea. Ma27 originally removed it. Seems like something went wrong when master was merged into staging-next.
<clever>
evanjs: previously, fetchFromGitHub included the rev in the name, but that caused some issues where nix-prefetch-git doesnt use the same name, and the sha256 isnt right
<evanjs>
well that's why then :D
<clever>
evanjs: now most things in nixpkgs use name = "source"; which i feel makes things worse
<evanjs>
I named them both rtl8188eu soooooooo :P
<evanjs>
and definitely agree with that one
<clever>
evanjs: fixed-output and non-fixed use a different algo to compute $out
<evanjs>
10 consecutive files in my rebuild with source ...
Yannik_Sc has joined #nixos
sbetrium has quit [Quit: sbetrium]
<m1cr0m4n>
Hey folks. How would I override a python package in an overlay? Tried python37Packages.$pkgname = <standard overlay override> but that didn't do it
leotaku has joined #nixos
<m1cr0m4n>
the pkg is hyperkitty in this case. It doesn't even build. Judging by the fact no one has a ticket for this I'm assuming I'm the only person to use mailman with hyperkitty :P
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « python27Packages.webassets: 0.12.1 -> 2.0 »: https://git.io/JvWxl
Jackneill has quit [Remote host closed the connection]
<evanjs>
m1cr0m4n: you sure it's not in there? nix-locate brings up e.g. /nix/store/lfnhw242q0xxljrjxwwj5ic7xmc247nb-python3.7-mailman-hyperkitty-1.1.0/lib/python3.7/site-packages/mailman_hyperkitty
<{^_^}>
[nixpkgs] @jtojnar pushed to apeh-anyso « autoPatchelfHook: Support arbitrary paths in runtimeDependencies »: https://git.io/JvWx4
<evanjs>
or is that just interop or something?
<evanjs>
OH
<evanjs>
it's under pythonPackages37 lol
<evanjs>
m1cr0m4n: e.g. python37Packages.mailman-hyperkitty
<evanjs>
***python37Packages ...
<m1cr0m4n>
evanjs: it is there yeah it's just that version doesn't build
vidbina has joined #nixos
<m1cr0m4n>
so yeah that's what Im trying to do I have python37Packages.hyperkitty = an overrideAttrs but it's still using the old version. I've done overlays before in this very repo I'm working on
<evanjs>
ohh you just need a bump
<m1cr0m4n>
yeah. I will PR it in if it works
gagbo has joined #nixos
<evanjs>
i typically just do it in a local nixpkgs repo :P then e.g. `nix build -f . python37Packages.mailman-hyperkitty` from the repo
KeiraT has quit [Remote host closed the connection]
KeiraT has joined #nixos
<{^_^}>
[nixpkgs] @jonringer merged pull request #80044 → neovim-remote: add now needed setuptools as input → https://git.io/JvWbn
<{^_^}>
[nixpkgs] @jonringer pushed commit from @doronbehar to master « neovim-remote: add now needed setuptools as input »: https://git.io/JvWxu
apriluwu has quit [Quit: WeeChat 2.6]
<m1cr0m4n>
eeeeeh ok I guess I'll keep a local nixpkgs ;P
<immae>
m1cr0m4n: doing overlays on pythonPackages is a little trickier than on base but it’s still doable, if you don’t want to keep a local nixpkgs (which has other drawbacks)
<claudiii>
when trying to package an electron application should I try to build id with npm or just unpack the .deb? wat's the "best practice" here, I see both used?
<Ilya_G>
clever: I have it along side with another overlay I have made earlier. Am I even on the right track?
<immae>
m1cr0m4n: I’m doing it like that: https://git.immae.eu/?p=perso/Immae/Config/Nix.git;a=tree;f=overlays/python-packages;h=13769f5518f30a1ef136d730b863401332f1afe2;hb=HEAD
Yannik_Sc has quit [Quit: Yannik_Sc]
<immae>
which is... kind of similar to the comment you’re posting actually :p
linarcx has quit [Ping timeout: 240 seconds]
Yannik_Sc has joined #nixos
o1lo01ol1o has joined #nixos
<clever>
ilya_g: 11 with import ../../../nixos/lib/qemu-flags.nix { inherit pkgs; };
<clever>
ilya_g: crap, pkgs/build-support/vm/default.nix directly imports that file, you have no way to overlay it
<Ilya_G>
clever: is there a way to use the older version in this flow? It has kvm64 in it
<clever>
ilya_g: use an older nixpkgs rev, or use a fork of nixpkgs with it fixed
philr has joined #nixos
<{^_^}>
[nixpkgs] @arianvp closed pull request #68953 → nixos/boot: Make sure interfaces are renamed in stage-1 → https://git.io/JeOzK
<Ilya_G>
if I have `/nix/store/58l27cl4pks1v25c8q7yfpfq0hnzc1v6-nixpkgs-20.03pre194957.bef773ed53f/nixpkgs/nixos/lib/qemu-flags.nix` and `/nix/store/gv4m5kg1d7z5k1f5s3a3cr8158ng2kky-nixpkgs-20.03pre212770.cc1ae9f21b9/nixpkgs/nixos/lib/qemu-flags.nix`
<Ilya_G>
is there a way to downgrade just that bit?
bb010g has quit [Quit: authenticating]
konobi has joined #nixos
peelz has joined #nixos
<m1cr0m4n>
immae: wow that's a LOT of overlays
<peelz>
Does anybody know what's going on here? My xdg mime mappings are generated through home-manager: https://i.imgur.com/jnsr7HR.png
nisstyre has joined #nixos
<immae>
m1cr0m4n: yes, I remmeber that I spent quite some time to figure out how to do it correctly
<m1cr0m4n>
immae: I haven't looked through too many but why have you needed so many?
<immae>
Honestly I don’t remember, and I may have done too much
<m1cr0m4n>
haha, ok :)
<immae>
What I know is that it works though, since I use it every day
<m1cr0m4n>
that's what counts!
<Ilya_G>
clever: is it possible to downgrade just that package? ^^^
<immae>
yes :p
<clever>
ilya_g: import a second copy of nixpkgs, and get runInLinuxVM from that
<{^_^}>
[nixpkgs] @tfc opened pull request #80063 → nixosTests.solr: Port to python → https://git.io/JvWpJ
Yannik_Sc has quit [Remote host closed the connection]
jgeerds has joined #nixos
<Ilya_G>
clever: I am sorry I don't know what that means in this context. :(
<clever>
ilya_g: figure out where your code is calling runInLinuxVM, within nixpkgs, and override that
Yannik_Sc has joined #nixos
<pbb>
"cannot coerce null to a string, at /var/src/sources/nixpkgs/pkgs/stdenv/generic/make-derivation.nix:191:19"
<pbb>
Any ideas? I can not see a link between the system parameter and this line.
<pbb>
I need to build a NixOS system for an i586 cpu, so I set the system parameter to "i586-linux". i686 works, but with i586 I get:
psyanticy has quit [Quit: Connection closed for inactivity]
<clever>
pbb: i686-linux is the codename for all 32bit x86 cpu's
<pbb>
This does not boot on that device: "Unable to boot - please use a kernel appropriate for your CPU."
<pbb>
"This kernel requires an i686 CPU, but only detected an i586 CPU."
<pbb>
So I need to use i686-linux as system, but set an i586 kernel?
<pbb>
Won't there be incompatible instructions in there?
<pbb>
(in the userspace parts too?)
<clever>
pbb: youll need a custom linux .config file, or at least a few custom config entries, to force it down to 586
<pbb>
Okay, thanks. I'll try
Yannik_Sc has quit [Remote host closed the connection]
<Aleksejs>
Hello! I'm trying to make a script that will use shebang, nix-shell and nodejs. I've already did some scripts like this but with nix-shell+runhaskell. But with node I'm getting an error
<{^_^}>
[nixpkgs] @tfc opened pull request #80065 → nixosTests.nsd: Port to python → https://git.io/JvWpi
<immae>
(ugly, I admit :p )
<Aleksejs>
nice, thank you :) at least it works
<immae>
you’re welcome :)
TroelsAgergaardJ has joined #nixos
<NemesisD>
hi all, i noticed when i boot into a nix shell and run `set -o` it shows that `vi` is on and `emacs` is off. what determines that? i'd like to default to emacs
<NemesisD>
i've even tried `set -o emacs` to the shellHook but to no avail
apriluwu has quit [Quit: WeeChat 2.6]
apriluwu has joined #nixos
<multun>
NemesisD: I'm pretty sure emacs is the default
<multun>
NemesisD: no idea why it's changed in your system though
WhittlesJr has quit [Ping timeout: 260 seconds]
<m1cr0m4n>
immae: I nearly have this override working but I can't resolve self.fetchPypi from within the override :P
<NemesisD>
multun: the plot thickens. outside of nix shell if i run bash it is in vi mode for some reason. i wonder if it inherits? because on another host where this isn't an issue, bash defaults to emacs and so does nix-shell
<NemesisD>
i'm not sure where vi is being set though. i've checked ~/.bashrc, ~/.profile, /etc/profile, /etc/profile.d/*
<immae>
scoates: the fact that it’s not able to fetch pipenv from the nixos cache suggests that either the package is broken (might happen if you are on unstable) or that you use a modified version (overlay / private code) that changes it
Naptra has quit [Remote host closed the connection]
<{^_^}>
[nix] @edolstra pushed to precise-gc « Add some GC diagnostics functions »: https://git.io/JvWhW
<m1cr0m4n>
oh my god.... It's already 1.3.2 on master LOL. I haven't updated my channel!!!!
<scoates>
the actual failure (`from pip._internal.cli.main import main', "ModuleNotFoundError: No module named 'pip._internal.cli.main'"`) is the same as https://github.com/NixOS/nixpkgs/issues/71178
<{^_^}>
#71178 (by mawis, 17 weeks ago, closed): Difficulty using pip in a venv
<scoates>
I thought it was working after that, but now I'm a mess again. Doesn't seem to be for everyone. /-:
ryanartecona has quit [Quit: ryanartecona]
<immae>
scoates: did you try the workaround in the first comment?
<scoates>
I've been through all of that. I can't even get a working pipenv to run that right now.
<scoates>
(the actual nix-shell fails to provision pipenv)
<scoates>
perhaps I should blow everything away and just switch to stable… I wonder…
<immae>
In my case NIX_PATH=nixpkgs=https://nixos.org/channels/nixos-19.03/nixexprs.tar.xz works (in the sense: I entered a nix shell correctly, I didn’t run anything from there), if you want to try it (it’s a bit old)
casaca has joined #nixos
<scoates>
(trying that)
kvda has joined #nixos
<scoates>
that gets me a shell but then pipenv itself fails because of the same problem in that issue… the workaround might work, but I don't have the luxury of reworking everything I have that uses pipenv. )-:
<NemesisD>
found the issue. had an old .inputrc laying around
<scoates>
what's really a bummer is that this seems to be isolated to me somehow, and I have no idea where to find out how to take the next step.
<scoates>
thanks for trying.
<immae>
scoates: just in case it might help, I had something like that: https://paste.ee/p/k7sLQ that permitted me to run pip inside a nix-shell
<immae>
(I don’t know if pip and pipenv are similar or not, so please check first)
* scoates
nods
<scoates>
thanks… the issue I linked very specifically says to not install pip from nix if you want to use it from pipenv (though pipenv does install pip)… it's all a mess )-:
<immae>
Ha ha... good luck then, maybe someone will know better than me here
<scoates>
I think I'm going to go make dinner. And flip the breaker. And wirecut the coax.
* scoates
waves (-:
* immae
waves
magnetophon has quit [Ping timeout: 240 seconds]
morr has quit [Quit: WeeChat 2.6]
morr has joined #nixos
morr has quit [Client Quit]
gustavderdrache has quit [Quit: Leaving.]
morr has joined #nixos
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #80066 → nixos/mate: use upstream session → https://git.io/JvWhM
lsix has quit [Quit: WeeChat 2.6]
<{^_^}>
[nixpkgs] @tfc opened pull request #80067 → nixosTests.plotinus: Port to Python → https://git.io/JvWhS
vidbina has quit [Ping timeout: 272 seconds]
endformationage has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
<musteresel>
Hi ... I have a program which crashes due to a NULL pointer; to debug this I need debug symbols for glibc. How can I get glibc with debug symbols? I'm working in a nix-shell with a nix file with the usual stdenv.mkDerivation
<enteee>
qyliss: yep, could connect the dots on my own ;) thank for you help
justanotheruser has quit [Ping timeout: 246 seconds]
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #80067 → nixosTests.plotinus: Port to Python → https://git.io/JvWhS
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JvWje
<noonien>
i'm trying to do some cross-compilation, i'm calling a package with `crossPkgs.callPackage`, and i need python inside the package for the build process
<noonien>
however, it seems to want to recompile python
fusion809 has joined #nixos
<noonien>
i'm adding python as a dependency as such: `depsBuildBuild = with pkgs; [ python3 ];`