AllanEspinosa has quit [(Ping timeout: 260 seconds)]
rcschm has quit [(Remote host closed the connection)]
<NixOS_GitHub>
[nixpkgs] mucaho opened pull request #24865: Conan (master...conan) https://git.io/vSyK6
Guest60 has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
steshaw has joined #nixos
romildo has joined #nixos
darlan has joined #nixos
alunduil has joined #nixos
<ramsdenj>
Anyone farmilliar with autofs?
<alunduil>
ramsdenj, it's been a while but I might muddle through some things. What's going on?
<alunduil>
Does anyone know of a prefix version or a way to convert to a prefix version the // operator for use with foldl'?
<ramsdenj>
I have enabled it with 'services.autofs.enable ' and I'm trying to edit the configuration with 'services.autofs.autoMaster' as mentioned in the examples but it doesn't seem to do anything. I'm trying to attach to an NFS share with a domain name that resolves when I ping it. I've done the same on arch with no problem.
<NixOS_GitHub>
[nixpkgs] copumpkin pushed 1 new commit to master: https://git.io/vSy6x
<NixOS_GitHub>
nixpkgs/master 1fd7bf9 Dan Peebles: nano: remove unnecessary Darwin-specific behavior...
<ramsdenj>
alunduil: There doesn't seem to be any configuration options so I'm not sure how I'm supposed to configure it.
<alunduil>
ramsdenj, is there a reason you're not using the systemd automount options?
snikkers has quit [(Ping timeout: 240 seconds)]
<ramsdenj>
alunduil: systemd automount units still need to be defined. With autofs as long as the server I'm trying to visit is resolvable I can access every share on that server without explicitly defining it. I have 25+ shares and it's irritating having to explicitly add every share as an automount unit.
johnw_ has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<alunduil>
ramsdenj, yep. Just checking that didn't fit your case.
<ramsdenj>
I will do that if I need to though.
johnw_ has joined #nixos
johnw_ has quit [(Client Quit)]
romildo has quit [(Quit: Leaving)]
<alunduil>
ramsdenj, can you link your current autofs configuration?
<alunduil>
ramsdenj, did you enable debug and did you get anything in the logs that would indicate an issue?
<alunduil>
ramsdenj, you also mentioned that you had this working on arch. Does this configuration match that configuration?
wak-work[m] has joined #nixos
JayVii has quit [(Ping timeout: 252 seconds)]
<ramsdenj>
alunduil: alunduil, no, i'll give that a shot. I wasn't sure if I had properly used the autoMaster setting. But yes this does match my Arch config.
schoppenhauer has quit [(Quit: Adé)]
schoppenhauer has joined #nixos
<alunduil>
ramsdenj, does autofs start with that configuration under NixOS?
thc202 has quit [(Ping timeout: 268 seconds)]
<ramsdenj>
alunduil, Yes.
<ramsdenj>
alunduil, Does it look like I am correctly using the autoMaster option though? From what you can tell?
<alunduil>
ramsdenj, from what I can tell it should be mounting your configuration under /auto/net. Is that what you're expecting?
rcschm has joined #nixos
<ramsdenj>
alunduil, Just /net is what I expected.
eqyiel has joined #nixos
eqyiel has quit [(Changing host)]
eqyiel has joined #nixos
<alunduil>
ramsdenj, ah, well then I think you're using a let expression where you don't intend to.
mekeor has quit [(Ping timeout: 255 seconds)]
<alunduil>
ramsdenj, I'm not sure that provided example is the clearest.
johnsonav has quit [(Ping timeout: 255 seconds)]
<alunduil>
ramsdenj, what if you try: autoMaster = "/net -hosts --timeout=60"; ?
<ramsdenj>
alunduil, oh. Yea, I had trouble interpreting it. This is my first time using a let expression so I was a little confused.
<alunduil>
ramsdenj, yep. Here they're using it to create a submap for autofs to associate with a containing directory.
<alunduil>
ramsdenj, looks like the file: stuff in autofs allows for loading another file to configure a whole directory. Does that sound right?
zraexy has quit [(Quit: Leaving.)]
erictapen has quit [(Ping timeout: 260 seconds)]
<ramsdenj>
alunduil, yea. What exactly does '/auto file:${mapConf}' do?
eacameron has quit [(Remote host closed the connection)]
<alunduil>
ramsdenj, I couldn't find anything on the file: operator. When you start that configuration what does mount report?
eacameron has joined #nixos
<ramsdenj>
alunduil, I'm not on Nix atm, but i'll take a look.
<ramsdenj>
alunduil, but you think if I were to do: 'autoMaster = "/net -hosts --timeout=60";' it would write into the proper file?
<alunduil>
ramsdenj, I suspect so but like I said, it's been a while since I've used autofs. Just remember that autoMaster is the literal string that is the contents of that file.
eacameron has quit [(Ping timeout: 240 seconds)]
aneeshusa has joined #nixos
<ramsdenj>
alunduil, ok, thanks!
<alunduil>
ramsdenj, no problem. Hope you keep making progress and enjoying NixOS.
johnsonav has joined #nixos
<ramsdenj>
alunduil, Will do! NixOS is awesome! So far my mind is sufficiently blown.
eacameron has joined #nixos
<alunduil>
ramsdenj, nice. Come back if you need some more help.
ramsdenj has quit [(Quit: Leaving)]
AllanEspinosa has joined #nixos
zraexy has joined #nixos
carlosda1 has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
Guest60 has joined #nixos
aneeshusa has quit [(Quit: WeeChat 1.7)]
alunduil has quit [(Quit: leaving)]
carlosda1 has quit [(Ping timeout: 260 seconds)]
eacameron has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
rui1 has joined #nixos
eacamero_ has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
zeus_ has quit [(Read error: Connection reset by peer)]
zeus_ has joined #nixos
<rui1>
How does owncloud/nextcloud work?
<rui1>
I never tried them, how do you set the server up in nixos?
<rui1>
There's no service and installing it doesn't give a bin :P
eacamero_ has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
takle has quit [(Remote host closed the connection)]
johnsonav has quit [(Ping timeout: 240 seconds)]
eacameron has quit [(Ping timeout: 255 seconds)]
davidak has quit [(Quit: Leaving.)]
eacameron has joined #nixos
sigmundv_ has quit [(Ping timeout: 245 seconds)]
<clever-also-afk>
rui1: the owncloud derivation contains files like index.php, a web-server would need to be configured to run on that derivation
<rui1>
clever-also-afk: Ah, great
takle has joined #nixos
eacameron has quit [(Ping timeout: 268 seconds)]
Arguggi has quit [(Ping timeout: 268 seconds)]
tilpner has quit [(Ping timeout: 252 seconds)]
oleks has quit [(Ping timeout: 260 seconds)]
eacameron has joined #nixos
kriztw has quit [(Ping timeout: 252 seconds)]
<rui1>
How do you import a file as a string?
takle has quit [(Read error: Connection reset by peer)]
<rui1>
config = "expects a string"; #but I want a file whose contents are the string. Any better way than wrapping the whole file contents in '' ''?
<clever>
error: Package ‘uboot-Bananapi_defconfig-2017.03’ in ‘/nix/store/mwwncy49y4b691wp527a5qh3lj9hv0cw-nixos-17.09pre105118.65be0b3a4b/nixos/pkgs/misc/uboot/default.nix:53’ is not supported on ‘x86_64-linux’, refusing to evaluate.
<clever>
well, that solves that problem, lol
darlan has quit [(Ping timeout: 268 seconds)]
eacamero_ has quit [(Ping timeout: 240 seconds)]
carlosda1 has joined #nixos
derjohn_mob has joined #nixos
takle has joined #nixos
<copumpkin>
clever: pity it doesn't run aarch64 :)
<copumpkin>
Dezgeg: can the ARM boxes build regular ARM?
edvorg has quit [(Remote host closed the connection)]
<clever>
copumpkin: the biggest benefit it has, is 5 ethernet ports and a semi-managed switch IC
mkoenig has quit [(Remote host closed the connection)]
<clever>
the A20 processor can change the switch IC config over i2c, so its basicaly a managed switch with the mangement software missing
carlosda1 has quit [(Ping timeout: 252 seconds)]
<copumpkin>
that's nice
<clever>
but that part is also slightly misconfigured by defaylt
<clever>
the WAN and LAN ports are bridged on bootup
<clever>
and you need to actively isolate them with a config command
<clever>
so you essentialy have zero firewall while booting
<clever>
copumpkin: the u-boot derivation should contain the result of concating the SPL and u-boot
marusich has quit [(Quit: Leaving)]
takle has joined #nixos
eacameron has joined #nixos
johnw_ has joined #nixos
<copumpkin>
clever: looks like the sort of thing I've tried to cut out of my life :D
eacamero_ has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
eacameron has quit [(Ping timeout: 240 seconds)]
spinus has quit [(Ping timeout: 255 seconds)]
eacamero_ has quit [(Ping timeout: 240 seconds)]
<clever>
copumpkin: i think the issue is that it doesnt force a cross-compile by default, and it only supports being built on a native arm box
<clever>
or with crossSystem fully configured
<copumpkin>
ah :/
marusich has joined #nixos
marusich has quit [(Client Quit)]
nh2 has quit [(Quit: Leaving.)]
rcschm has quit [(Remote host closed the connection)]
eacameron has joined #nixos
takle has joined #nixos
sellout- has joined #nixos
zeus_ has joined #nixos
rcschm has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
MP2E has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
<NixOS_GitHub>
[nixpkgs] jraygauthier opened pull request #24866: digikam: add video support through libqtav (master...jrg/digikam_add_video_support_through_libqtav) https://git.io/vSySM
eacamero_ has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
indi_ has joined #nixos
rcschm has quit [(Ping timeout: 268 seconds)]
carlosda1 has joined #nixos
mguentner2 has quit [(Quit: WeeChat 1.7)]
eacamero_ has quit [(Ping timeout: 245 seconds)]
proteusguy has quit [(Remote host closed the connection)]
<clever>
copumpkin: used nix-instantiate --argstr system armv7l-linux to get an arm .drv, then nix-copy-closure to get it into an arm, and nix-store -r to begin building it
hexagoxel has quit [(Ping timeout: 260 seconds)]
eacameron has joined #nixos
carlosda1 has quit [(Ping timeout: 260 seconds)]
mguentner has joined #nixos
eacameron has quit [(Ping timeout: 268 seconds)]
ebzzry has joined #nixos
hexagoxel has joined #nixos
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
systemfault has joined #nixos
systemfault has quit [(Max SendQ exceeded)]
systemfault has joined #nixos
byteflame has joined #nixos
eacameron has joined #nixos
byteflame has quit [(Ping timeout: 252 seconds)]
rcschm has joined #nixos
marusich has joined #nixos
schoppenhauer has quit [(Ping timeout: 240 seconds)]
eacameron has quit [(Ping timeout: 252 seconds)]
schoppenhauer has joined #nixos
rcschm has quit [(Ping timeout: 240 seconds)]
marusich has quit [(Client Quit)]
ryanartecona has joined #nixos
nslqqq has quit [(Ping timeout: 240 seconds)]
nslqqq has joined #nixos
eacameron has joined #nixos
takle has joined #nixos
eacameron has quit [(Ping timeout: 268 seconds)]
eacameron has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
marusich has joined #nixos
takle has joined #nixos
MP2E has quit [(Quit: reboot)]
takle has quit [(Ping timeout: 260 seconds)]
MP2E has joined #nixos
batdog has quit [(Remote host closed the connection)]
reinzelmann has joined #nixos
mguentner2 has joined #nixos
takle_ has joined #nixos
mguentner has quit [(Ping timeout: 240 seconds)]
takle_ has quit [(Ping timeout: 268 seconds)]
reinzelmann has quit [(Ping timeout: 260 seconds)]
AllanEspinosa has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
reinzelmann has joined #nixos
mguentner3 has joined #nixos
Sergio965 has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
takle has quit [(Ping timeout: 240 seconds)]
mguentner2 has quit [(Ping timeout: 268 seconds)]
mizu_no_oto has joined #nixos
Guest60 has joined #nixos
batdog has joined #nixos
Guest60 has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
mog has quit [(Ping timeout: 260 seconds)]
rui1 has quit [(Quit: WeeChat 1.5)]
mog has joined #nixos
fnljk has quit [(Disconnected by services)]
fnljk_ has joined #nixos
systemfault has quit [(Quit: Bye!)]
galen has joined #nixos
systemfault has joined #nixos
systemfault has quit [(Max SendQ exceeded)]
systemfault has joined #nixos
calvertvl has quit [(Quit: leaving)]
justan0theruser has joined #nixos
<NixOS_GitHub>
[nixpkgs] canndrew opened pull request #24867: hivemind: init at git-2017-04-13 (master...hivemind) https://git.io/vSydS
fnljk_ has quit [(Ping timeout: 240 seconds)]
justanotheruser has quit [(Ping timeout: 260 seconds)]
proteusguy has joined #nixos
Itkovian has joined #nixos
carlosda1 has joined #nixos
noam_ has quit [(Read error: Connection reset by peer)]
noam_ has joined #nixos
indi_ has quit [(Remote host closed the connection)]
carlosda1 has quit [(Ping timeout: 260 seconds)]
indi_ has joined #nixos
rcschm has joined #nixos
rcschm has quit [(Remote host closed the connection)]
rcschm has joined #nixos
takle has joined #nixos
marusich has quit [(Ping timeout: 260 seconds)]
systemfault has quit [(Quit: Bye!)]
sivteck has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
ryanartecona has quit [(Quit: ryanartecona)]
takle has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
byteflame has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
<timclassic>
I'm sad; it looks like SmartOS support is utterly broken right now.
byteflame has quit [(Ping timeout: 252 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<timclassic>
Is there a guide to how Nix bootstraps itself from nothing using existing compilers?
zeus_ has quit [(Remote host closed the connection)]
zeus_ has joined #nixos
zeus_ has quit [(Ping timeout: 240 seconds)]
katyucha has joined #nixos
marsel has joined #nixos
batdog has quit [(Remote host closed the connection)]
<chakerbenhamed>
Hey I want to package a web application (https://github.com/angular/angular-seed to be exact) with nix. When I use node2nix it transform to node module.
<chakerbenhamed>
Is that the right approach to follow?
<chakerbenhamed>
steveeJ: "To add a package from NPM to nixpkgs"
<chakerbenhamed>
I don't want to add a package from NPM
goibhniu has joined #nixos
<chakerbenhamed>
I want to create a package for my webapp
<steveeJ>
chakerbenhamed: I see, your webapp will be packaged with npm at all?
<chakerbenhamed>
No, I want to use NPM to get the dependencies but not for packaging
carlosda1 has joined #nixos
takle has quit [(Remote host closed the connection)]
<steveeJ>
chakerbenhamed: unfortunately I have no experience here, but I'm almost certain you'll need to use node2nix for the dependencies, and then for your package I have no idea :-) maybe you can look at the generated code for other node packages?
derjohn_mob has joined #nixos
cfricke has joined #nixos
<NixOS_GitHub>
[nixpkgs] elitak opened pull request #24871: Various small crossDrv fixes (master...cross-staging) https://git.io/vSSIH
carlosda1 has quit [(Ping timeout: 240 seconds)]
dlight has quit [(Remote host closed the connection)]
<chakerbenhamed>
Thanks steveeJ
takle has joined #nixos
snikkers has joined #nixos
<chakerbenhamed>
To clarify my problem. I am developing a web application. Frontend with angular. And a RESTful backend with pyramid. I want to package the whole thing with nix. For the backend I don't have a problem I can build it and I can enter a nix-shell session with the needed dependencies for dev. I want to be able to do the same thing for the frontend (build it, nix-shell session). To simplify things let suppose I am
<copumpkin>
domenkozar: thanks for requesting the niksnut review :) niksnut, if you have questions about why the hell I'm doing all this, I'm happy to go into more detail
edvorg has quit [(Remote host closed the connection)]
<NixOS_GitHub>
[nixpkgs] maggesi pushed 1 new commit to release-17.03: https://git.io/vSSET
<NixOS_GitHub>
nixpkgs/release-17.03 ff97d8b Marco Maggesi: BLCR 0.8.5b4 *do* work with Linux 3.10.
eacameron has joined #nixos
<NixOS_GitHub>
[nixpkgs] maggesi pushed 1 new commit to master: https://git.io/vSSEG
<NixOS_GitHub>
nixpkgs/master 1dee38f Marco Maggesi: BLCR 0.8.5b4 *do* work with Linux 3.10.
<domenkozar>
udp.c in the Linux kernel before 4.5 allows remote attackers to execute arbitrary code via UDP traffic that triggers an unsafe second checksum calculation during execution of a recv system call with the MSG_PEEK flag.
<MoreTea>
Has anyone here experience with generating javascript with idris? My main issue right now is that I don't fully grasp the package manager (and I want to use the "Javascript" library instead of coding everything myself using the FFI
<domenkozar>
MoreTea: soon I'll be able to review your nixops refactoring
<domenkozar>
MoreTea: is it working atm?
<gchristensen>
last I tried there was a module import error
eacameron has joined #nixos
<gchristensen>
domenkozar: I bet you know how to solve it :)
<domenkozar>
the sqlite thing?
<MoreTea>
"it works on my machine"
<MoreTea>
hehe, no, I'll spend some time on checking that I can deploy to vbox and gce.
<domenkozar>
MoreTea: I think you forgot to commit state.py
<domenkozar>
well, state folder
<domenkozar>
or __init__.py for it
<MoreTea>
I could use a break from reading about dependant types :)
<MoreTea>
will do!
<domenkozar>
hmm it's there
<copumpkin>
never break from reading about dependent types
<domenkozar>
huh
<domenkozar>
gchristensen: ok I'll give it a try :)
<domenkozar>
MoreTea: not sure what's wrong yet, so do wait
<MoreTea>
copumpkin, I know! :3 I'm reading about the implementation of Idris, it's core semantics etc for my literature study.
Itkovian has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
<copumpkin>
:)
<MoreTea>
@domenkozar, BTW I would love to work a bit on automated infrastructure for testing nixops. The main issue there would be to generate temporary credentials and a way to clean up all resources created with those credentials.
<MoreTea>
I'm afraid that this would be quite hard to integrate in Hydra.
romildo has left #nixos ["Leaving"]
<domenkozar>
MoreTea: yeah
<domenkozar>
MoreTea: don't functional tests already do that? for a few backends
<MoreTea>
Yes they do, but they are run on your local dev machine, right? Not via a CI system.
jacob_ has joined #nixos
<domenkozar>
MoreTea: I don't see how hydra could do the networking part, so far it's pure
<MoreTea>
A project that is able to automatically test nixops deployments on some cloud in a CI pipeline could be quite valuable. Besides testing nixops itself it could be used to do end-to-end tests for your own infrastructure.
nh2 has joined #nixos
<domenkozar>
MoreTea: we could use travis-ci for that with some encrypted credentials
<domenkozar>
but not sure who'd pay for that
<domenkozar>
and how safe it is :D
<domenkozar>
(in terms of allocating resources based on bugs)
<domenkozar>
it's hard to get right, what if someone creates a PR
<copumpkin>
niksnut: can we just talk on here a bit?
<domenkozar>
that creates a lot of ec2 big instances
<domenkozar>
etc
<domenkozar>
so you can really only test code once it's merged
<gchristensen>
we could (lol) use (lol) aszlig's trick of defeating md5 via a collision
<MoreTea>
domenkozar, a combination of your github bot + hydra + some service that is triggered by a webhook / polls => post result as github comment
byteflame has quit [(Ping timeout: 260 seconds)]
<copumpkin>
I think we should upgrade aszlig_'s script to sha1 now
<gchristensen>
haha for securitay
dannyg has joined #nixos
<domenkozar>
UDP remote execution, sha1 collisions in pure language
<domenkozar>
MoreTea: I think good instructions how to run the tests and how to get all credentials would work
<domenkozar>
MoreTea: so nixbot doesn't work that well yet
<domenkozar>
so there's plenty on that road to fix
jgertm has joined #nixos
<domenkozar>
MoreTea: I'm never against automation, but here I feel it's a looong road
<copumpkin>
niksnut: so all my (and I'd argue more and more folks' as time progresses) Linux usage is virtualized, and nested virtualization is either unreliable or impossibly slow, so my general goal is to reduce the amount of unnecessary VM usage across nixpkgs. If I try to run any of our qemu stuff inside my NixOS VM on my Mac, it just hangs the VM
<21WAAAYZY>
[nixpkgs] Mic92 pushed 3 new commits to master: https://git.io/vSSz2
<21WAAAYZY>
nixpkgs/master d606717 Toni Schmidbauer: davmail: 4.7.2 -> 4.8.0
<copumpkin>
niksnut: I also do a ton of pure EC2 work, and would argue that building AMIs can actually be a pleasant and super efficient affair if you run it purely on EC2, assuming you get rid of the VM-centric image build machinery
<copumpkin>
(By dd'ing to an EBS volume, snapshotting, and registering the image)
<copumpkin>
I do it all the time and it takes a minute or two
<copumpkin>
anyway, I get that using runInLinuxVM is convenient, but if it excludes entire classes of users like myself, can you at least acknowledge that the convenience comes at a cost?
<copumpkin>
it's just frustrating to have you keep telling me it's fine the way it is
<copumpkin>
I've put a ton of work into this refactor after your feedback to my last PR that I shouldn't duplicate the logic
<niksnut>
admittedly, I never build images on EC2
<copumpkin>
to the point of renting hardware for the sole purpose of refactoring this
<copumpkin>
niksnut: I've been using (code derived from) this PR in a CI environment for months now and it's a pleasure
<copumpkin>
they literally take <1m to run on EC2 now
<copumpkin>
and my local VM is happy too
<copumpkin>
it's also actually faster on physical hardware
<copumpkin>
but the diff is obviously much smaller
<domenkozar>
I hope this doesn't make you unhappy copumpkin, but would it work as a compromise to wait for Nix 1.12 and thus get rid of fakeroot?
<copumpkin>
niksnut: another thing I've noticed through hundreds of Nix VM runs, occasionally I get the dreaded "could not inform kernel of partition changes" message that just breaks the build
<copumpkin>
niksnut: which my solution wouldn't suffer from
<niksnut>
I guess the fakechroot is still necessary to run grub etc.
<copumpkin>
niksnut: yeah, although I might be able to kill that too (if I can get the rest of this stuff in)
<copumpkin>
domenkozar: I was just going to do what shlevy suggested and use nixUnstable temporarily
<copumpkin>
there doesn't seem to be much downside to building an image with one and using another inside the image, as long as the schemas are compatible
<domenkozar>
I highly appreciate this work copumpkin, we are using EC2 (with Hydra) to build AMIs as well
DutchWolfie has quit [(Quit: Konversation terminated!)]
<copumpkin>
thanks :)
<copumpkin>
anyway, it's not just about CI environments. I do a lot of iterative development on my tailored EC2 AMIs
DutchWolfie has joined #nixos
<copumpkin>
and when my development cycle is 40 minutes, that makes me sad
<copumpkin>
especially when it really isn't that hard to do it without VMs
<domenkozar>
same here, we need to launch 100 nodes, soon more and copying all stuff to them is slower than 40min cycle
<domenkozar>
copumpkin: what you could do meanwhile is to have hetzner (or other bare metal) to build AMIs
<domenkozar>
unless you have NDA that disallows that or smth
<copumpkin>
niksnut: TBC, I also think many of our VM tests (the service ones, not the bootup/installer ones) should/could use libvirt and in general I'd like to reduce our dependence on qemu all over the nixos repo
<domenkozar>
just trying to find intermediate solutions
<copumpkin>
domenkozar: yeah, I do, we have a special agreement with AWS
<copumpkin>
I can only use AWS for work
<domenkozar>
copumpkin: in terms of discount or?
<gchristensen>
not to mention having to also have a special hetzner box is :thumbsdown:
<copumpkin>
nah, data protection
<domenkozar>
gchristensen: surely, but its a thumbdown faster than 40min
<domenkozar>
anyway data protection is a good reason :D
<copumpkin>
it's mostly about getting high-ups in the company comfortable with where their data is going
<copumpkin>
I'm sure we could do it with other companies, but it's work that's hard to argue for "because the image building script uses a VM"
<gchristensen>
40min build time vs. adding a new vendor and automation and tooling and runbooks and monitoring and and and
<niksnut>
copumpkin: I don't want to have to maintain multiple testing infrastructures (qemu, libvirt, ...)
<copumpkin>
niksnut: libvirt can call qemu :)
<copumpkin>
niksnut: anyway, sorry for the distraction about tests, we can talk about that separately
<copumpkin>
right now I really just care about images
<copumpkin>
I don't know what to do, and I really think the new nixos-install.sh code is quite pretty, so if you have specific things beyond fakechroot (Which I'll happily kill if I can) I should improve about it
<copumpkin>
I'd love to hear them
<copumpkin>
or another way to factor out the common functionality?
<niksnut>
well I don't want to veto anything
<copumpkin>
:) but you know your mere opinions are super powerful
<copumpkin>
so we obviously consider them with great weight
<copumpkin>
I don't want you to hate our installer script if this goes in
<copumpkin>
so I'll happily do anything I can to make that not happen
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vSSgi
<copumpkin>
niksnut: anyway, the short of it is that I have a goal, to get a good, fast, and conceptually simple image building process that works efficiently on all platforms :) this nixos-install.sh refactor is a yak to shave on the path to that, and I'm not committed to the approach, but I am committed to the goal. So if there's a better way to get images to build pleasantly on EC2 (or at all on my local VM; not kidding when I say it just hangs),
<copumpkin>
I'm all ears. I considered adding more conditional behavior to nixos-install.sh rather than splitting into nixos-prepare-root, but the number of things that suddenly now need to be conditional started getting very ugly, so I moved to the current refactoring, which ended up feeling fairly elegant. I'm happy to try something else if it ends up being cleaner
dannyg has quit [(Quit: dannyg)]
cfricke has quit [(Quit: WeeChat 1.7)]
<copumpkin>
(the other non-EC2 angle is that I want to seamlessly build nix images on Darwin with a remote builder in a background VM, and the current design makes that impossible)
cpennington has joined #nixos
DutchWolfie has quit [(Quit: Konversation terminated!)]
<Mic92>
other parts like building docker containers would also profit from that.
<chakerbenhamed>
I am developing a web application. Frontend with angular. And a RESTful backend with pyramid. I want to package the whole thing with nix. For the backend I don't have a problem I can build it and I can enter a nix-shell session with the needed dependencies for dev. I want to be able to do the same thing for the frontend (build it, nix-shell session). To simplify things let suppose I am using the angular-seed
<chakerbenhamed>
project.
<copumpkin>
yup, that's why I was talking about nix images and not nixos images :)
<copumpkin>
although dockerTools won't specifically benefit from this particular PR, it'll benefit from the broader effort
markus1199 has quit [(Ping timeout: 240 seconds)]
indi_ has joined #nixos
<Mic92>
For dockerTools I have an even more efficent idea of an rewrite. I did something similar for rkt in the past.
<copumpkin>
I'd definitely support that
cpennington has quit [(Max SendQ exceeded)]
takle has quit [(Ping timeout: 240 seconds)]
<copumpkin>
I'd like to unify all these "image-building" processes
<copumpkin>
we have a ton of duplicated code all over the repo
cpennington has joined #nixos
<niksnut>
we have a unified builder, namely make-disk-image.nix
<copumpkin>
yeah I know
takle has joined #nixos
<copumpkin>
it's just not unified enough right now to handle the common concerns between disparate targets
<copumpkin>
and furthermore many of the builders don't use it :)
<niksnut>
but make-disk-image.nix is just intended to a front for nixos-install
<Mic92>
chakerbenhamed: what is specific in your frontend part?
<copumpkin>
yup
<socksy>
lol i just logged in to ask about dockerTools
<niksnut>
anyway, like I said, I don't want to veto anything if people thing non-VM builds are essential
<niksnut>
I just think the VM approach is much cleaner
<niksnut>
since you don't end up with a completely different code path for building images
<copumpkin>
niksnut: I got quite a few thumbs up on the original PR with the duplicated nixos-install logic :) that's all I know, really, besides obadz- and domenkozar saying they need this sort of thing
<socksy>
are there any examples of dockerTools that I can learn from? Can you get services running in them, or is it just sugar for dockerfiles?
<copumpkin>
niksnut: yeah, if I could unify them in some other way I'd love to. The current nixos-install refactor is just attempting to share as much code as possible between the two
<copumpkin>
niksnut: if it would help visualize the differences, I can also update the image building PR to use this change
<copumpkin>
(before we move forward with the nixos-install one)
indi_ has quit [(Ping timeout: 260 seconds)]
<Mic92>
socksy: there are examples in the repo next to dockerTools, if I remember correctly
nico202 has joined #nixos
markus1199 has joined #nixos
<copumpkin>
I guess it's hard to make PR dependencies
* copumpkin
longs for gerrit sometimes
<socksy>
ah i found it
mrottenkolber has joined #nixos
<mrottenkolber>
Hi
<mrottenkolber>
so how do I find out which version of nixpkgs nix-env etc query?
<socksy>
so no services?
<socksy>
just running shell images
<mrottenkolber>
and how can I get a nix-shell based on one of my services?
<Mic92>
socksy: yes. systemd inside docker is tricky
<socksy>
ah ok
<gchristensen>
I don't think systemd works in docker
<Mic92>
gchristensen: you can make it work, but it requires some effort
<copumpkin>
this is why we need a service abstraction layer :)
<Mic92>
lxc works better in my experience
<Mic92>
(with systemd)
griff_ has joined #nixos
<socksy>
unfortunately my company standardized on docker with alpine
<chakerbenhamed>
Mic92: Nothing it's a simple angular-application
<Mic92>
chakerbenhamed: so probably just need npm, no?
<socksy>
is there any real advantage of using dockerTools over just plain docker + docker-compose?
<Mic92>
socksy: you have more reproducible images and build dependencies only needs to build once.
<copumpkin>
Dockerfile is a turd
ryantrinkle has quit [(Ping timeout: 268 seconds)]
<copumpkin>
:)
<Mic92>
But at the moment I find dockerTools still too slow due the use of VMs
<copumpkin>
niksnut: ^^ :)
<hyper_ch>
LnL: still here? no luck wtih virtmanager
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 3 new commits to master: https://git.io/vSSrt
<NixOS_GitHub>
nixpkgs/master 5db1fe6 Benjamin Staffin: plasma-nm: patch in the correct path to openvpn executable...
<NixOS_GitHub>
nixpkgs/master 7c75940 Benjamin Staffin: plasma-nm: enable parallel building...
<NixOS_GitHub>
nixpkgs/master c9ba391 Thomas Tuegel: Merge pull request #24816 from benley/plasma-nm...
<chakerbenhamed>
Mic92: I am used node2nix and bower2nix
<chakerbenhamed>
but when I run `nix-shell -A shell` in the frontend folder it doesn't get the needed deps for the application (bower and node_modules)
<copumpkin>
niksnut: anyway, how's this for a way forward: 1) I try nixUnstable to get rid of fakechroot 2) I post to the mailing list asking for feedback and get more visibility on the PR, and then we see based on that if my needs are just niche (in which case I guess I'll maintain something separately) or a pain point for lots of people
<Mic92>
chakerbenhamed: you can use the `-o` parameter of nix-build to create specific symlink like `node_modules` pointing to the nix store.
flohsor has joined #nixos
<chakerbenhamed>
but in that case I will be using nix-build and not nix-shell?
<niksnut>
copumpkin: sure, sounds good
<copumpkin>
\o/
davidak has joined #nixos
jensens has joined #nixos
dridus has joined #nixos
<domenkozar>
copumpkin: can you share a script how you upload ami from a disk image?
<MoreTea>
Hm. does someone have an example of how idrisPackages is supposed to work? I tried to get the `idrisScript` to work (FFI to javscript), but it fails to compile the Idris prelude
jgertm has quit [(Ping timeout: 268 seconds)]
<copumpkin>
domenkozar: I don't have it on this computer, but I currently use two processes. One of them is pretty similar to the one niksnut uses to generate official images (upload to S3, ImportSnapshot, RegisterImage; I think Eelco uses ImportVolume -> CreateSnapshot -> RegisterImage, not sure why). The other builds the image on an EC2 instance with a small (1-3GB) EBS volume attached at e.g., /dev/xvdb. I then CreateSnapshot on that EBS volume (I
<copumpkin>
don't need to detach or anything) and RegisterImage on the result
<copumpkin>
I can throw together a more specific sketch a bit later if that isn't clear enough
jensens has quit [(Ping timeout: 240 seconds)]
<domenkozar>
copumpkin: oh I understand how it works, just woneder it you can paste source for ImportSnapshot
<domenkozar>
and the whole flow of image -> ami using that shortcut
<copumpkin>
oh, I don't have a paste right now but I can make a PR to simplify the current image script to avoid ImportVolume -> CreateSnapshot, since I think that's equivalent to just ImportSnapshot
<domenkozar>
copumpkin: I was thinking I'd do that to help you
<domenkozar>
if you can share your script that you're using at work
ryantrinkle has joined #nixos
<domenkozar>
copumpkin: but either way :)
<copumpkin>
oh okay, I'll see if I can get that out but we tend to be kind of difficult about releasing code :P thanks for the offer though
<domenkozar>
copumpkin: ah no worries then
<copumpkin>
it's kind of integrated with some other stuff that I definitely can't put out right now
<copumpkin>
(although it probably shouldn't be :))
<domenkozar>
copumpkin: I'll figure it out
<domenkozar>
was looking to save half an hour if you can do it in 1min :)
<copumpkin>
sorry!
<domenkozar>
np
<copumpkin>
unfortunately not so easy :( I don't IRC from there
noah-nix has joined #nixos
calvertvl has joined #nixos
<noah-nix>
Hi, does anyone have a suggestion to how to resolve the issue I have with my NixOS installation?
griff_ has quit [(Ping timeout: 260 seconds)]
<goibhniu>
hi noah-nix, did you try out UEFI already?
<copumpkin>
niksnut: am I understanding correctly that I should run nix-store with NIX_REMOTE=local?store=/path/to/other/store and then it'll automatically take care of the sqlite db for that other store?
ixxie has joined #nixos
<noah-nix>
goibhniu: no, I want to use BIOS since my other computers have BIOS as well, so UEFI is unfortunately not an option here.
<goibhniu>
what model is it noah-nix?
<copumpkin>
`nix copy` also looks promising, hmm
<noah-nix>
goibhniu: of laptop?
<goibhniu>
noah-nix: yep
<noah-nix>
goibhniu: HP EliteBook 2560p
jgertm has joined #nixos
<noah-nix>
A nice computer I bought for a nice price.
<noah-nix>
Refurbished
<NixOS_GitHub>
[nixpkgs] copumpkin closed pull request #24830: treewide: use boolToString function (master...refactor/boolToString) https://git.io/vSXom
Fare has joined #nixos
eacameron has joined #nixos
<LnL>
hyper_ch: what happens when you add the connection?
<goibhniu>
noah-nix: so, you can install NixOS, but not boot into it ... is that the situation?
<hyper_ch>
LnL: found it out... hat do add connection and select user qemu
<noah-nix>
goibhniu: yes
<goibhniu>
and what happens when you boot, to you get to the grub boot menu?
<goibhniu>
*do you*
<noah-nix>
goibhniu: I do not
reinzelmann has quit [(Quit: Leaving)]
<goibhniu>
noah-nix: just a blinking cursor?
<noah-nix>
goibhniu: nope, I see the laptop's "no OS installed" screen.
<goibhniu>
can you post your configuration.nix and hardware.nix to a pastebin?
<noah-nix>
goibhniu: haven't you checked it before?
<goibhniu>
thanks ... yeah, it all looks good to me, sorry :/
<goibhniu>
I presume the install finishes without any errors etc
<noah-nix>
goibhniu: yeah
<noah-nix>
Might have to switch distro then I assume..
<goibhniu>
have you installed another distro on this previously?
byteflame has quit [(Ping timeout: 255 seconds)]
<noah-nix>
goibhniu: I sure have
byteflame has joined #nixos
<NixOS_GitHub>
[nix] edolstra closed pull request #1325: Move note about float support out of the wrong release notes (master...float-rl-1.12) https://git.io/vSyEv
<goibhniu>
sorry noah-nix, I can't think of anything else
<noah-nix>
goibhniu: alright, thanks anyways!
<copumpkin>
niksnut: so it seems like local can take several parameters, "store" (which I can't divine the meaning of; is it where to think the store is vs. where to find it?), "real", which seems to be where to actually store/find the files, "root" (different from store?), "state" which is where the DB lives, and "log", which is where to store logs
ebzzry has quit [(Ping timeout: 255 seconds)]
<copumpkin>
oh, I see
<copumpkin>
so root lets me set a different root and affects all the other ones
<noah-nix>
is anyone in this IRC a NixOS dev/admin that I can ping?
<copumpkin>
noah-nix: depends on what you need, but gchristensen, ikwildrpepper, and niksnut are all operators in this channel (and a few others)
<goibhniu>
noah-nix: hrm ... you reformatted the disk too, right? ... could you post the output of `fdisk -l`?
<copumpkin>
noah-nix: but it looks like you don't need IRC help, so I wouldn't ask them
noah-nix_ has joined #nixos
<copumpkin>
noah-nix: you might try the mailing list if all else fails. It tends to be slower but less ephemeral
erasmas has joined #nixos
matumental has joined #nixos
<noah-nix_>
Sorry, left automatically because library PC time limit.
<noah-nix_>
copumpkin: don't need IRC help? What do you mean?
<noah-nix_>
Oh that is correct
<goibhniu>
noah-nix_ In case you missed it, `fdisk -l` would also be interesting
<copumpkin>
noah-nix_: the people I listed have powers in this channel, which means they can remove disruptive people and the like. I wasn't sure what you were asking
<copumpkin>
noah-nix_: but in general we don't really keep a group of super-nix-users separate from the general public who can answer the harder questions :)
<noah-nix_>
goibhniu: oh, sure, I will paste that! I am currently at a local library so I will paste it when I get home
<noah-nix_>
copumpkin: okay, that is actually pretty good
<copumpkin>
noah-nix_: as I said before, we're a community of volunteers and people who have time, willingness, and are paying attention will help
noah-nix has quit [(Ping timeout: 260 seconds)]
<copumpkin>
if you aren't getting help, one or more of those factors is hindering it :)
<noah-nix_>
or me not providing enough information :-)
<dash>
noah-nix_: the only other thing that comes to mind is checking what device your machine is trying to boot from, many systems nowadays will provide a device-selection menu at boot time if you hit the right key
<noah-nix_>
dash: have done so, it boots to the hard drive
<noah-nix_>
but there is no OS
<noah-nix_>
it says
<dash>
noah-nix_: ok. how's the drive partitioned? anything other than nixos on there?
<noah-nix_>
dash: nope
<noah-nix_>
dash: /dev/sda1 is NixOS
<noah-nix_>
Now this IRC chat suddenly stopped for me..
<noah-nix_>
Oops
matumental has quit [(Ping timeout: 260 seconds)]
<noah-nix_>
It thoght /dev/sda was a command
<noah-nix_>
dash: /dev/sda1 is NixOS
<noah-nix_>
dash: /dev/sda2 is swap
<copumpkin>
niksnut: wow, the NIX_REMOTE=local?root= is magic!!
<dash>
noah-nix_: ok. the next step would be to boot from the installer and see what's in the boot sector on that drive, but i don't remember how at the moment
<noah-nix_>
dash: I'm thinking about maybe trying a GRUB Boot medium?
<noah-nix_>
You know, this GRUB Rescue Disk Medium or what it is called
<im_ferris>
Hi, could someone give me a quick bit of help on how to setup a clang environment to link again linux headers?
spinus has joined #nixos
drasich has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vSSDN
<NixOS_GitHub>
nixpkgs/master 1f32d4b Tuomas Tynkkynen: make-bootstrap-tools.nix: Fix bzip2...
<noah-nix_>
would anyone here do me a favor, please, please, and send an email to the mailing list for me? I'd appreciate that! I already have a paste with information about my issue.
<copumpkin>
basically ask for nix-store -q --references of the .drv file, which gives you the immediate build-time dependencies as .drv files
<copumpkin>
then nix-store -r those to get the non-drvs of those
<mojjo>
hi! I have the impression that, there's something broken with my busybox package. Many functions from that multi binary don't work anymore or output errors. here's a few: `man` does not show any manpage, `top` tells me about 'stackoverflow'...
BramD_ has quit [()]
<gchristensen>
hmm also not quite what I want, but that gets me very close
<gchristensen>
though I may dig through some thesis papers to find that sort of diagram
indi_ has joined #nixos
BramD_ has joined #nixos
<mojjo>
anyone ever had problems with busybox?
mekeor has quit [(Remote host closed the connection)]
cpennington has quit [(Remote host closed the connection)]
alx741 has joined #nixos
alx741 has quit [(Client Quit)]
alx741 has joined #nixos
<copumpkin>
niksnut: does `nix copy` copy the entire closure or just the specific path requested? does it do work avoidance if the destination path is already there and valid in the destination DB?
mojjo has quit [(Remote host closed the connection)]
<dtzWill>
copumpkin: I'm not niksnut :P but looking at 'nix copy --help' there's an "-r" option to specify copying the entire closure or just the given path (although tbh I didn't realize copying a path without its deps was a useful/possible thing regardless), and pretty sure it only copies missing paths. Source: experience. trying it again now with "nix copy --to file://$HOME/cache path".
<copumpkin>
yeah I see that now
<copumpkin>
fooling around with it :)
seanz has joined #nixos
<dtzWill>
and to follow-up, looks like it gets grumpy if you try to copy not the entire closure and the deps are missing in dest store
<copumpkin>
thanks!
ashkitten has quit [(Quit: No Ping reply in 180 seconds.)]
<copumpkin>
error: cannot add path ‘/nix/store/4d6f8hg5gv20nsbq7b52qzn6bcs4fvlh-coreutils-8.26’ because it lacks a valid signature
ashkitten has joined #nixos
<davenull>
Can anyone help me decipher why my ssh keys aren't working? I have an entry for it in the user declaration in configuration.nix. and it generates the file in /etc/ssh/authorizedKeys.d/${user}, but when I try to ssh to that user account, it offers the key, as expected, but the sshd doesn't let me us it to log in. No error messages, it just doesn't log in with the key.
<copumpkin>
davenull: you've checked the sshd log?
<dtzWill>
add '&secret=path-to-signing-key' to sign it? but not sure why signing is needed, is trusted-users a thing for local stores?
<dtzWill>
TIL /etc/ssh/authorizedKeys.d was a thing xD
<copumpkin>
dtzWill: seems unlikely that I'd need to sign a local store :)
<davenull>
dtzWill, It's nifty, if I can get it working!! I am about 4 steps away from having an auto-provisioning nodejs cluster, this seems to be the roadblock of the day.
<davenull>
copumpkin, Where does nixos keep it's sshd logs at?
<copumpkin>
journalctl -u sshd if I were to guess
<copumpkin>
or just check the general journal
derjohn_mob has quit [(Ping timeout: 260 seconds)]
<dtzWill>
I just use 'users.users.$user.openssh.authorizedKeys.keys' or keyFiles
<copumpkin>
yup I use that all the time and it works fine for me
<copumpkin>
I seem to remember it being sensitive to whitespace at some point
<copumpkin>
not sure if you have any junk in yours
<davenull>
I'm declaring it with openssh.authorizedKeys.keys in the user declaration block.
indi_ has quit [(Ping timeout: 255 seconds)]
<im_ferris>
How do I import the `kernel` variable in a default.nix file?
gridaphobe_ has joined #nixos
stig_ has joined #nixos
seanparsons_ has joined #nixos
mpickering_ has joined #nixos
joepie91___ has joined #nixos
joepie91___ has quit [(Changing host)]
joepie91___ has joined #nixos
joepie91 has quit [(Disconnected by services)]
joepie91___ is now known as joepie91
tg` has joined #nixos
<davenull>
Hmmm.. the logs make no mention of key auth, not even attempt or failure.
smola__ has joined #nixos
teozkr_ has joined #nixos
<copumpkin>
it might be going somewhere else in the journal, not specfiically in sshd log
<ixxie>
hmmm - nix-prefetch-git can be used to produce JSONs - I suppose one can then feed these into fetchgit?
<copumpkin>
I think it would normally go into /var/log/secure.log on a non-journald machine
phorse` has quit [(Remote host closed the connection)]
<davenull>
Only thing I have left to really figure out is if it is possible to automate the node2nix installation of pm2 at install time. Currently, I bootstrap the system with stubs for the assiciated.nix files, then on first boot I use node2nix to generate my needed nix files, and then rebuild the system to apply the changes.
marsel has quit [(Ping timeout: 252 seconds)]
civodul has quit [(Remote host closed the connection)]
hyphon81 has quit [(Quit: No Ping reply in 180 seconds.)]
hyphon81 has joined #nixos
phorse has joined #nixos
georges-duperon has quit [(Quit: Leaving)]
snikkers has quit [(Ping timeout: 252 seconds)]
<davenull>
Can I still use ssh-copy-id to install a key on a running instance?
magnetophon has quit [(Ping timeout: 240 seconds)]
<davenull>
And can i use more than one key per user when defining in configuration.nix?
ixxie has quit [(Quit: Lost terminal)]
jmiven has quit [(Quit: WeeChat 1.7)]
<ronny>
hmm
jmiven has joined #nixos
<ronny>
the current nixos desktop experience is that on kde basic stuff like archive programs and browser integation is broken, and gnome is without maintainer
jmiven has quit [(Client Quit)]
jmiven has joined #nixos
<gchristensen>
ronny: interested community members are highly encouraged to step up :)
<dash>
too many of us don't use either, i'm afraid :)
<gchristensen>
dash: I wonder if that is survivorship bias :)
JagaJaga has joined #nixos
ixxie has joined #nixos
ixxie has quit [(Client Quit)]
<davenull>
What's a GUI? /s
indi_ has joined #nixos
zeus_ has joined #nixos
<joko_>
Hmm, is it possible to use init=/bin/sh or sth like that in NixOS? I see that there is another init parameter already set
<symphorien>
How can I build a library with a gcc which ships libstdc++.a (static) ?
<ronny>
gchristensen: i do manage some small bits in personal time, but i wont be able to provide sustained major contributions
<ronny>
its more likely i#ll have to stop using nix since its interfering with work that its gnome/kde stories are so bad atm
<simpson>
ronny: If only DEs were not bad software; then they would not require sustained major contributions over a period of time just to keep them building.
<avn>
ronny: basic WMs like fvwm/awesome/qtile/... just works ;) I believe most nix maintainers use them
<ronny>
simpson: well, software is bad
ixxie has joined #nixos
<gchristensen>
ronny: sounds like the problem is survivorship bias then
<simpson>
ronny: Sure. I merely mean that most packages, once put into Nix, tend to stay working for long periods of time. The massive heroic efforts required by DE maintainers would suggest to me that DE packages are atypical.
<ronny>
gchristensen: waht is "survivorship bias"
<gchristensen>
ronny: a difficult problem, no doubt, especially when we weren't pushing kde5 for such a long time, and then dropped kde4
<gchristensen>
ronny: "none of the maintainers contribute to kde/gnome" "kde and gome are bad" "I need kde or gnome, so I can't use nix, so I can't make it better"
<ronny>
gchristensen: well, i can only work on nix in my free time, and i have a kid on the way ^^
<dash>
ronny: the problem is that DEs have even more software than the alternatives :-[
<gchristensen>
ronny: no I know, it isn't your fault, just a problem we have
<simpson>
gchristensen: TBF I've formed my opinion watching not just NixOS, but also Arch, Gentoo, Debian, Ubuntu, Fedora...
<gchristensen>
simpson: no doubt
<ronny>
dash: i needs a full blown syste, and i cant really affrod to build one fro mscracch thats just differing a little bit from everyone else
<dash>
ronny: OK
magnetophon has joined #nixos
<ronny>
in particular im r eally in need of a "standardized full system" - reliable repeatable infrastructure thats simply not different from user to user ^^
<dash>
ronny: you can get that with nixos, just maybe not with nixos and everything kde you want :)
<ixxie>
hmm nixos-rebuild yields some weird systemd error ("Failed to start Setup Virtual Console") - full error log here https://pastebin.com/yp4dQYnx - anybody seen something like this?
<dash>
i doubt you can get it with kde anywhere else :-/
tg has joined #nixos
<dash>
i tried kde on nixos for a while, having not used it since 2.x days. I enjoyed turning its knobs and clicking its buttons
<ronny>
dash: before i just used gnome, that would work even on nix and had all the things i needed work
<dash>
But I didn't need anything it did, so I stopped using it
<ronny>
i switched to kde mostly because gnome broke on nix
<dash>
(... thus I am poorly suited to contribute to making it better)
bjarki is now known as SuprDewd
mrottenkolber has left #nixos []
sellout- has joined #nixos
__Sander__ has quit [(Quit: Konversation terminated!)]
ryantrinkle has quit [(Ping timeout: 258 seconds)]
cpennington has joined #nixos
MichaelRaskin has joined #nixos
betaboon has joined #nixos
bret has joined #nixos
<clever>
cpennington: yep, the uboot does appear to build, it has the right header at the start
<clever>
copumpkin: oops, wrong c name
<copumpkin>
hi :)
<copumpkin>
clever: I just sent an email to the mailing list you might be interested in!
<clever>
copumpkin: so i now have a u-boot that should be compatible witht he banana pi
<clever>
loadng up the mailing list
<clever>
what was the subject?
<copumpkin>
"Please provide input on my attempts to rework how NixOS builds images"
<copumpkin>
:)
<clever>
ah
<copumpkin>
obadz-, Mic92: you're probably interested too
nico202_ has joined #nixos
<clever>
copumpkin: dmj and i have also been looking into stdenv.override { glibc = pkgs.musl; }
snikkers has joined #nixos
<copumpkin>
niksnut: I fired off the email. Please chime in if I've misrepresented your concerns
<MichaelRaskin>
copumpkin: I cannot guarantee that this is positive support and not negative support because I have wrong enough reasons, but reusing two thirds of NixOS installer for something I can comfortably use would be nice.
<copumpkin>
MichaelRaskin: clever also has notOS that could probably benefit from it. I hadn't thought of that. There are probably tweaks we'd need to make to the script to work more generally, but I like the idea
<MichaelRaskin>
5B
<MichaelRaskin>
Oops
<MichaelRaskin>
Fun thing: I now have _two_ branches of system that can both generate a shared Bootloader config
<Fare>
MichaelRaskin, I want to use F#, which is in Xamarin Studio 6, but not in the monodevelop 5.9 provided by nixpkgs @7701cbca from unstable
rcschm has quit [()]
<MichaelRaskin>
Is Xamarin Studio opensource?
<MichaelRaskin>
I have negative amount of clue
<MichaelRaskin>
Would monodevelop 6.1 help
<MichaelRaskin>
Ah, Studio is MonoDevelop + mobile stuff, so should help
reinzelmann has joined #nixos
<MichaelRaskin>
Fare: are you comfortable just updating NixPkgs packages already? ping me when there is something to review (right now I am not doing much PR merging, but reviewing a few PRs is not hard, it's the firehose flow that gets people burnt out)
<MichaelRaskin>
What the hell is wrong with MonoDevelop
<MichaelRaskin>
«freshest version» on the homepage, freshest tarball and freshest GitHub release are all different versions
ambro718 has joined #nixos
byteflame has joined #nixos
nschoe has quit [(Quit: Program. Terminated.)]
jgertm has quit [(Ping timeout: 260 seconds)]
ragge has joined #nixos
bitonic has joined #nixos
contrapumpkin has joined #nixos
jensens has joined #nixos
Filystyn has quit [(Remote host closed the connection)]
digitus has joined #nixos
ryanartecona has joined #nixos
betaboon has quit [(Quit: This computer has gone to sleep)]
<Fare>
xamarin studio looks like it's free as in beer, not as in speech
<Fare>
MichaelRaskin, I'm too busy on other projects right now :-(
<Fare>
I suppose monodevelop 6.1 would help
JagaJaga has quit [(Ping timeout: 255 seconds)]
<Fare>
MichaelRaskin, they obviously don't use NixOS for release...
* Fare
is trying to integrate feedback from his talk into the phd thesis. Then to select a Lisp implementation to use as a guinea pig. I wanted Chez, but considering how underdocumented it is, looking at Gambit instead.
Guest37382 has quit [(Ping timeout: 245 seconds)]
<MichaelRaskin>
6.1 lacks a tarball, ang GitHub tags suggest 7.0
<MichaelRaskin>
Fare: do you need a large list or a minimalistic Lisp for your experiment, though?
ryantrinkle has quit [(Ping timeout: 245 seconds)]
<MichaelRaskin>
a large lisp
seanparsons has quit [()]
seanparsons has joined #nixos
freusque has quit [(Quit: WeeChat 1.7)]
<MichaelRaskin>
I would consider something lua-style-minimalistic, maybe even literally urn
HappyEnte has joined #nixos
<Fare>
MichaelRaskin, ideally, a minimalistic core.
HappyEnte has quit [(Client Quit)]
carter has quit [()]
HappyEnte has joined #nixos
<Fare>
but if possible with enough extensions / libraries to have non-trivial programs to migrate.
<MichaelRaskin>
Shen? Urn? PicoLisp?
<MichaelRaskin>
You mean preexisting programs?
carter has joined #nixos
<NixOS_GitHub>
[nixpkgs] LnL7 closed pull request #24797: pass: fix clipboard functionality for darwin (master...pass/coreutils-sb-patch) https://git.io/vS6g2
<MichaelRaskin>
PicoLisp can FFI
<MichaelRaskin>
Well, if you want a minimalistic core, there will be some FFI reliance, and I am not sure how your consistent theory interacts with this sad real world
ryantrinkle has joined #nixos
itorres1 has joined #nixos
itorres has quit [(Ping timeout: 260 seconds)]
HappyEnte has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
HappyEnte has joined #nixos
flohsor has quit [(Ping timeout: 240 seconds)]
<Fare>
MichaelRaskin, if possible existing. And yes, I'm not sure how exactly to handle FFI.
<Fare>
I mean, I can see many ways of handling it, none completely work-free.
flohsor has joined #nixos
<MichaelRaskin>
I will just stand here with a stopwatch and wait until you consider forking Nix
pbogdan has joined #nixos
<gchristensen>
3.. 2..
zeus_ has quit [(Read error: Connection reset by peer)]
<MichaelRaskin>
gchristensen: that's a countdown, it runs in the wrong direction and creates undesired pressure, and, more importantly, undesired illusion of predictable schedule
zeus_ has joined #nixos
<gchristensen>
that is quite the analysis of a throw-away joke comment :P
hamishmack has quit [(Quit: hamishmack)]
<MichaelRaskin>
Should I start over-analyzing your choice of definite article «the» in «the analysis» instead of «an analysis»?
mekeor has joined #nixos
<gchristensen>
if you choose
jensens has quit [(Ping timeout: 258 seconds)]
teknico has left #nixos []
<mekeor>
do i have to set ‘networking.wireless.enable = false’ if i set ‘networking.networkmanager.enable = true’ ?
simendsjo has joined #nixos
<mekeor>
i guess, not
byteflame has quit [(Ping timeout: 260 seconds)]
Filystyn has joined #nixos
<mekeor>
my problem is: when i do "nmcli device wifi connect FOO password BAR", i get an error saying: "connection activation failed: (7) secrets were required, but not provided."
<ToxicFrog>
mekeor: from the man page, nmcli wifi connect only supports open, WEP, and WPA-PSK (and thus by implication not WPA2-PSK or WPA{1,2}-EAP).
<ToxicFrog>
You may need to define the network first and then nmcli up
jensens has joined #nixos
ertesx has joined #nixos
ixxie has quit [(Quit: Lost terminal)]
aminechikhaoui has quit [(Ping timeout: 268 seconds)]
vaibhavsagar has quit [(Ping timeout: 240 seconds)]
hamishmack has joined #nixos
ertes has quit [(Ping timeout: 260 seconds)]
ertesx is now known as ertes
davidak has quit [(Quit: Leaving.)]
flohsor has quit [(Ping timeout: 258 seconds)]
ryantrinkle has quit [(Ping timeout: 258 seconds)]
aminechikhaoui has joined #nixos
sigmundv_ has joined #nixos
<copumpkin>
Fare: Nix does have an FFI :)
<copumpkin>
or were you not talking about that?
<copumpkin>
also, hi Fare :)
nico202_ has quit [(Ping timeout: 268 seconds)]
aminechikhaoui has quit [(Ping timeout: 252 seconds)]
byteflame has joined #nixos
indi_ has joined #nixos
ryantrinkle has joined #nixos
cpennington has quit [(Ping timeout: 240 seconds)]
MarcelineVQ has quit [(Read error: Connection reset by peer)]
<copumpkin>
can I ask nix to look for nix.conf somewhere other than the default location?
<LnL>
yes
pbogdan has quit [(Remote host closed the connection)]
indi_ has quit [(Ping timeout: 240 seconds)]
<LnL>
NIX_CONF_DIR
<copumpkin>
\o/ thanks
pbogdan has joined #nixos
<copumpkin>
niksnut: not that I really need this at all, but is there a fundamental reason that nix-cache-info includes the StoreDir? Doesn't the standard hashing format include that already?
ryanartecona has quit [(Quit: ryanartecona)]
takle has quit [(Remote host closed the connection)]
<NixOS_GitHub>
[nixpkgs] rycee pushed 1 new commit to master: https://git.io/vS92n
<NixOS_GitHub>
nixpkgs/master 034ea00 Robert Helgesson: pythonPackages.PyICU: 1.9.5 -> 1.9.6
Itkovian has joined #nixos
cpennington has joined #nixos
<pbogdan>
is there a way to make nix-env always use my fork of nixpkgs without having to specify it on command line all the time?
dannyg has quit [(Quit: dannyg)]
davidak has joined #nixos
<LnL>
you can set NIX_PATH, but for nix-env you have to add a symlink to ~/.nix-defexpr/nixpkgs
m3tti has joined #nixos
jensens has quit [(Ping timeout: 255 seconds)]
<nixy>
Is it possible to use a git repo of nixpkgs like a channel, is this something that has been proposed before?
<copumpkin>
you can point your channel directly at a github tarball URL, for example
<LnL>
oh really?
<copumpkin>
yup
<copumpkin>
or point it anywhere there's a suitable tarball of nix expressions, basically
<nixy>
But that wouldn't be able to update would it
<copumpkin>
the only other piece of metadata we really use in a channel is the location of binarycaches
<copumpkin>
but in this case the default one works fine
<copumpkin>
if you're setting up a custom one, it gets a bit trickier, but then you'll probably have a place to put a nixexprs.tar.bz2
<Fare>
copumpkin, hi!
<Fare>
MichaelRaskin, Nix is not meant for long-running computations, so there's not much to migrate at runtime.
<pbogdan>
LnL: do you mean I need to create ~/.nix-defexpr/nixpkgs symlink pointing to my fork? something like ~/.nix-defexpr/nixpkgs -> /home/pbogdan/nixpkgs/pkgs?
<LnL>
yep, if you only set NIX_PATH you have to use nix-env -f '<nixpkgs>'
<NixOS_GitHub>
[nix] edolstra pushed 10 new commits to master: https://git.io/vS9rl
<NixOS_GitHub>
nix/master e7cb284 Dan Peebles: Explicitly model all settings and fail on unrecognized ones...
<gchristensen>
the fix was from several years ago IIRC
<ekleog>
copumpkin: I'd guess it does, indeed :)
<bitonic>
clever: thanks
<gchristensen>
ekleog: have you found versions of our kernel vulnerable to it?
AllanEspinosa has joined #nixos
<ekleog>
gchristensen: haven't tested yet, I thought I'd have time to test before anyone noticed my message, looks like I was wrong
<copumpkin>
is our version of 3.10 vulnerable? it has no recent updates
<gchristensen>
all those commits are from 2015 and 2016
<copumpkin>
ah
ryantrinkle has quit [(Ping timeout: 240 seconds)]
ericsagnes has joined #nixos
eacameron has quit [(Remote host closed the connection)]
<mekeor>
how can i install firefox-esr?
<ekleog>
hmm, what are the actually maintained kernel packages, just so I can check them against http://www.securityfocus.com/bid/97397 ? (I'd guess as it has been embargoed 4 months it's not dating back to 2015-2016)
<bitonic>
Let's say I have a timer to run some service every 5 minutes, what happens if the service to be run takes more than 5 minutes some times?
<bitonic>
I guess this is more of a systemd question than a nixops question
<niksnut>
copumpkin: nix-cache-info includes the StoreDir so you can detect whether a binary cache is usable
<ekleog>
like, there are 685 packages matching "linux" and 203 "kernel" on https://nixos.org/nixos/packages.html, so I guess they're all derivatives of a small set of sources?
<copumpkin>
niksnut: it seems like in principle there's nothing stopping you from putting multiple store paths into a single cache though, right? there's no danger of overlap. It's just an optimization to avoid hitting things unnecessarily?
<copumpkin>
anyway, this line of questioning is probably a waste of my time and yours, since I don't use multiple store paths :) I just noticed it in the nix-cache-info file and was curious
<Fare>
MichaelRaskin, I see that some of your Lisp support made it to nixpkgs unstable. Can you build executables with dynamically-linked libraries yet?
<niksnut>
copumpkin: yes, it's just an optimization to bail out early
<Fare>
for the demo at ELS, I had to drop to an Ubuntu chroot, meh.
pbogdan has quit [(Remote host closed the connection)]
<bitonic>
OK, systemd.timer(5) answers it: `Note that in case the unit to activate is already active at the time the timer elapses it is not restarted, but simply left running.`
Itkovian has joined #nixos
nico202_ has quit [(Ping timeout: 240 seconds)]
reinzelmann has quit [(Quit: Leaving)]
<bitonic>
A more nixy question: let's say I'm specifying a record (in this case a nixops deployment), and I want to automatically generate a series of services -- that is, I want so set many `systemd.services.X = Y`, generating the values by mapping over some list. What is the best way to do that?
<bitonic>
As far as I can see there is no "deep merge" operator for records, e.g. if I generate a bunch of `{systemd.services.x_n = y_n}` in their own records and then merge it using \\ it'll replace the existing top-level keeys
Itkovian_ has quit [(Ping timeout: 260 seconds)]
Itkovian has quit [(Read error: Connection reset by peer)]
ryantrinkle has joined #nixos
Itkovian has joined #nixos
feepo has quit [()]
seanz has quit [(Quit: Leaving.)]
feepo has joined #nixos
ryantrinkle has quit [(Ping timeout: 258 seconds)]
MarcelineVQ has joined #nixos
<copumpkin>
lib/attrsets.nix probably has something good
<MichaelRaskin>
Fare: currently I want that SlimerJS + Parenscript stuff, so I haven't look at executables too closely
Avogadro has quit [(Ping timeout: 260 seconds)]
<MichaelRaskin>
I want reasonably-fast loading of ASDF systems, and monolithic-compile-bundle-op is not that bad
<copumpkin>
bitonic: at the worst, you can use listToAttrs
<MichaelRaskin>
Well, FFI in Nix is something to use only when everything is bad
<pie__>
aaaand cmake complains
<pie__>
CMake Error: CMake was unable to find a build program corresponding to "Unix Makefiles". CMAKE_MAKE_PROGRAM is not set. You probably need to select a different build tool.
jensens has quit [(Ping timeout: 258 seconds)]
ryantrinkle has joined #nixos
<MichaelRaskin>
copumpkin: does it force the derivation using exec to be fixed-output?
<copumpkin>
no, I don't think there are any restrictions on it, but it's disabled by default (and enabled only if you pass in a flag)
<copumpkin>
importNative has been there for a while now
<copumpkin>
exec is just a natural extension of that
<MichaelRaskin>
OK, enabled only with a flag is my first option — only when everything is bad
<copumpkin>
yup
<copumpkin>
although something we might want to do instead of aszlig_'s md5 collision :)
<copumpkin>
although I do think there's a more principled way to do that, too
m3tti has quit [(Remote host closed the connection)]
nico202_ has quit [(Ping timeout: 240 seconds)]
<MichaelRaskin>
Like, passing through impure environment variables like we already do?
<MichaelRaskin>
Ah, that's even more annoying…
nico202 has joined #nixos
byteflame has quit [(Ping timeout: 258 seconds)]
jensens has joined #nixos
simendsjo has quit [(Remote host closed the connection)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
jensens has quit [(Remote host closed the connection)]
<gchristensen>
anyone familiar with the ruby infra? I'd like to specify gem dependencies, and it seems they have to be present in ./pkgs/development/ruby-modules/gem-config/default.nix ??
jensens has joined #nixos
<gchristensen>
note to self: read the first 10 lines of the file to solve your own problem.
<MichaelRaskin>
copumpkin: I am not sure that stuff with Chromium should be just a Nix expression…
im_ferris has quit [(Quit: Leaving)]
<gchristensen>
I agree
<copumpkin>
MichaelRaskin: there are nice things we can do inside nix that get more painful outside it :) but yeah possibly not. At least not until my principled design happens :D
<pie__>
ok successfully build bonzo, now idk how to fonts
<copumpkin>
Mic92: you running the extension?
<MichaelRaskin>
compumpkin: well, are abstract unix domain socket properly isolated in Nix?
<mekeor>
how to install firefox-ESR? how to install emacs with lucid toolkit? could you maybe just drop the term used for these things so that i can search for it?
<mekeor>
(i thought maybe it's --set-flag but realized that's something different)
<Mic92>
copumpkin: what are your referring too?
<copumpkin>
Mic92: the cloud -> my butt browser extension :)
<Mic92>
copumpkin: oh, yes I think so
<copumpkin>
:)
<Mic92>
But I am so conditioned now, that my brain performs the same substition in reality.
<Mic92>
*substitution
aminechikhaoui has joined #nixos
takle has quit [(Remote host closed the connection)]
<gchristensen>
I've never read all of the nix pills, these are great
<copumpkin>
yup
<copumpkin>
we need to lure Lethalman back
<gchristensen>
gosh
<gchristensen>
and we need to mirror this content (with his permission)
takle has joined #nixos
sellout- has quit [(Quit: Leaving.)]
bennofs has joined #nixos
<dash>
yeah that series was significant in getting me to understand how all the nix pieces fit together
darlan has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to release-17.03: https://git.io/vS9b3
<NixOS_GitHub>
nixpkgs/release-17.03 2ed5b0e Joachim Fasting: grsecurity: 4.9.21-201704091948 -> 4.9.22-201704120836...
<NixOS_GitHub>
nixpkgs/release-17.03 fb39b01 Benjamin Smith: opera: 43.0.2442.991 -> 44.0.2510.857...
<Heffalump>
I'm trying to use nixpkgs on WSL (Windows subsystem for Linux) and I've run into a 32-bit binary (busybox) while trying to build something else. 32-bit binaries aren't supported on WSL. Is it normal to have those in the nixpkgs stack?
indi_ has joined #nixos
ambro718 has quit [(Ping timeout: 260 seconds)]
<Mic92>
Heffalump: not by default, but reachable via pkgs_i686
andymandias_ has quit [(Ping timeout: 240 seconds)]
<Mic92>
because very few packages do still not run on 64-bit
* pie__
cant find the bonzo config option for vsync
Wizek__ has quit [(Read error: Connection reset by peer)]
Wizek_ has quit [(Ping timeout: 255 seconds)]
dtzWill has joined #nixos
drasich has joined #nixos
yrashk has joined #nixos
<yrashk>
has anybody recently installed nixos on zfs? with or without /boot on zfs (that's not important to me) I tried the instruction on wiki but fail at different steps
<Mic92>
yrashk: me
Avogadro has joined #nixos
<gchristensen>
yrashk: me
aminechikhaoui has quit [(Ping timeout: 240 seconds)]
<Mic92>
would be great, if $somebody add this to the manual
<ToxicFrog>
Mic92: if I don't /home doesn't get mounted, which is A Problem
<ToxicFrog>
Actually, this reminds me, I need to switch /data on my laptop over to mountpoint=legacy, because without it it doesn't get mounted ever -- I'm not sure if `zfs mount -a` never runs or if it runs before the pools are imported.
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
<mojjo>
hi.. I'm getting an idea of how nix package definitions deal with build dependencies. But how does it work for runtime dependencies? Say a simple package just consist of a bash-script, which itself invokes binaries from other packages. Could anyone give me a hint how this would be designed?
simukis_ has quit [(Ping timeout: 240 seconds)]
<dash>
mojjo: the easy answer is "specify them at build time" :)
marsel has joined #nixos
<Mic92>
ToxicFrog: zfs do not mount on non-empty directories
<dash>
mojjo: but tell me more what you have in mind
<ToxicFrog>
Mic92: in the case of the laptop, /data is empty and manually running `zfs mount -a` mounts it.
<copumpkin>
Profpatsch: can't look right now, sorry
<ToxicFrog>
Mic92: in the case of the server, /home is empty at system boot, but Nix brings up services that create directories in /home before running zfs mount -a, at which point it is obviously too late
<gchristensen>
ToxicFrog: can you bang on a server for me and see if it seems good?
<ToxicFrog>
I did some messing around with unit dependencies to try to fix this and didn't really get anywhere
<ToxicFrog>
gchristensen: what kind of server?
<gchristensen>
x86-64, nixos, 24 cores and 25g ram, 2.8TB zfs
<gchristensen>
and bang on it for ~30 minutes or so
<ToxicFrog>
Ok, what do you mean by "bang on it" here
<gchristensen>
make sure it follows your expectations of what a nixos server is
<mojjo>
dash: well, I just have the feeling that I'm missing one part. I'm actually trying to do a quite simple thing: I need to write a bash script that uses netcat at one point...
<gchristensen>
ToxicFrog: I just want to make sure there aren't any glaring bugs, and copumpkin and domenkozar don't want to play
<NixOS_GitHub>
[nixpkgs] rycee pushed 1 new commit to master: https://git.io/vSHT1
<NixOS_GitHub>
nixpkgs/master de0f090 Robert Helgesson: josm: 11639 -> 11826
<copumpkin>
hey, I already played!
<copumpkin>
wait, I'm even already using your stuff!
<copumpkin>
:)
<copumpkin>
your stuff is what enabled me to finish that PR :P
<gchristensen>
:P
<mojjo>
dash: If the bash script is my 'source', and the builder would more or less just copy that to $out/bin ans set some rights it would be done. But using `netcat` in the script I guess would make it unpure..
<Profpatsch>
copumpkin: np
zraexy has quit [(Remote host closed the connection)]
<gchristensen>
mojjo: you can use netcat, but unless you know what your sha is at the end, you don't have internet access during the build
zraexy has joined #nixos
hiratara has quit [(Ping timeout: 255 seconds)]
<mojjo>
gchristensen: yeah, but netcat should be used at runtime, not at build time...
<gchristensen>
you could makeWrapper to add it to the run-time PATH
seagreen has quit [(Ping timeout: 245 seconds)]
<mojjo>
and I'm not really building something in the sense of compiling, there will finally be just a script in the $out dir: bin/something-that-uses-naetcat.sh
<gchristensen>
in your source script you can use @netcat@/bin/netcat and in your build phase do substituteAllInPlace ./your/script.sh, and do `netcat = pkgs.netcat`
<LnL>
in that case you might be able to rewrite it to an absolute path
<mojjo>
gchristensen: aha, nice, I'll check it out
<copumpkin>
shlevy: do you have a good grasp of `nix copy`? I'm trying to use it to copy from host to guest
<sophiag>
hi. i have to write some c for the first time in ages and need the header msg/msg.h that doesn't seem to exist on my system according to both gcc and find (i haven't tried locate since i get "permission denied" when trying to call updatedb from root, but that's a separate issue). does anyone know where the equivalent of msg.h is located on NixOS?
eacameron has quit [(Remote host closed the connection)]
markus1199 has quit [(Ping timeout: 260 seconds)]
<copumpkin>
yup, was already doing something like that
markus1189 has quit [(Ping timeout: 260 seconds)]
<sophiag>
alternatively, perhaps i could answer this myself if anyone can help me figure out why i'm getting permission denied calling updatedb from root?
Avogadro has quit [(Ping timeout: 252 seconds)]
<sophiag>
it seems to be a read only file system issue so perhaps i should poke around with fsck?
<copumpkin>
cstrahan: did you mean to comment on the other PR?
<copumpkin>
cstrahan: either way, I appreciate the comment :)
<clever>
sophiag: you need to start the updatedb.service in systemd, not run updatedb directly
<cstrahan>
I'm also composing a comment there (been chasing the bunny hole)
<copumpkin>
clever: do you know of an easy way to toposort the closure files?
zeus_ has quit [(Remote host closed the connection)]
<copumpkin>
cstrahan: hah okay thanks :)
stepcut has quit [(Remote host closed the connection)]
zeus_ has joined #nixos
<clever>
sophiag: systemctl start update-locatedb
<sophiag>
clever: thanks
byteflame has joined #nixos
endformationage has quit [(Quit: WeeChat 1.7)]
<cstrahan>
copumpkin: will we be able to omit the fakechroot?
<sophiag>
wait...that doesn't return anything. should i assume it's been updated?
Jackneill has quit [(Remote host closed the connection)]
<cstrahan>
copumpkin: I really appreciate the effort. I hadn't commented earlier because I didn't have the resources to give a thorough review, and I figured it would be unanimously be accepted down the road. Didn't realize there was any push back.
<sophiag>
clever: that gives me a whole load of "permission denied"s :(
<copumpkin>
cstrahan: we're working it out :)
<copumpkin>
I think it'll be good
zeus_ has quit [(Ping timeout: 258 seconds)]
<clever>
sophiag: that would be updatedb scanning everything in the system
bennofs has quit [(Ping timeout: 258 seconds)]
Guest60 has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
<cstrahan>
copumpkin: Just out of curiosity -- what's the status of boot loader installation? Is that still an open problem, or do you have already devised a workable solution?
<cstrahan>
I recall there was some question of whether GRUB could install to a regular file, I think.
<sophiag>
clever: yeah i understand. it's just they all came back that way and now it's lagging...
<copumpkin>
cstrahan: yeah, it can't as far as I can tell, but I don't think it's far off and for now I can just put that in a VM
stepcut has joined #nixos
<clever>
sophiag: it may still be updating, check top
<cstrahan>
copumpkin: Cool, that's what I suspected.
<copumpkin>
cstrahan: I get 95% of the speed gain from putting the main file manipulation stuff out of the VM
stepcut has quit [(Remote host closed the connection)]
<copumpkin>
but I do think we can get rid of the VM entirely with a couple of tweaks
byteflame has quit [(Ping timeout: 260 seconds)]
nico202 has quit [(Ping timeout: 255 seconds)]
<sophiag>
journalctl doesn't seem to be doing much
<cstrahan>
copumpkin: Right -- figured that would account for most of the total running time. Though getting rid of the VM completely would be super sweet.
<copumpkin>
yup
<copumpkin>
since it still doesn't run inside VMWare :)
<copumpkin>
but we're close
dridus has quit [(Quit: znc? wassat?)]
joneshf-laptop has quit [(Ping timeout: 240 seconds)]
<sophiag>
clever: so i terminated journalctl since it wasn't showing any activity
<cstrahan>
copumpkin: When you're at a point that you could use a review, ping me here or on GH and I'll take a look.
<copumpkin>
I think the code is almost ready to go! shlevy gave me a ton of feedback and I'm just finishing up the conversion to nixUnstable
<copumpkin>
but thanks, I'll ping you when it's actually "done"
Itkovian has joined #nixos
matthewbauer has joined #nixos
matthewbauer has quit [(Ping timeout: 240 seconds)]
Guest60 has joined #nixos
rotaerk has joined #nixos
Supersonic112 has quit [(Quit: Bye)]
thc202 has quit [(Ping timeout: 260 seconds)]
<sophiag>
clever: actually it seems like it *has* updated the dbs since locate now returns nothing rather than that error
<sophiag>
which i suppose brings me back to the actual question at hand...does anyone know where the nix equivalent of <msg/msg.h> is located?
endformationage has quit [(Quit: WeeChat 1.7)]
<simpson>
sophiag: I'm running $(nix-shell -p stdenv) to try to find one.
endformationage has joined #nixos
<simpson>
sophiag: Curses, my current Nix installation, which is single-user on Ubuntu, appears broken. Not sure what happened.