<tilpner>
You have three good options: 1. upgrade system to 17.09 (not a bad in general, but more work), 2. pick new mitmproxy from a newer channel without upgrading base system, 3. override old mitmproxy to use newer python
<tilpner>
*not a bad idea
Arcaelyx has joined #nixos
<Alling>
Tried 3 for several hours without luck. Don't understand what 2 means in practice. I guess 1 is the way to go. Does that mean downloading NixOS and reinstalling?
<Alling>
I mean, it seems to be pyflakes (and possibly other dependencies) that fails.
asuryawanshi has quit [(Ping timeout: 248 seconds)]
<tilpner>
You would do 1. by removing the nixos-17.03 channel, and adding nixos-17.09 under the same name, then rebuilding
<tilpner>
But you can freely mix channels. If you don't wish to upgrade, you could add a new channel and use one for the base system, and just pick a few applications from the newer channel
<samueldr>
tilpner: couldn't you directly --add the channel, using the same name (no need to remove iirc)
<tilpner>
samueldr - Maybe, I'm not very familiar with nix-channel
<Shados>
samueldr: you can, yes
<Alling>
Tbh I don't know what a channel is and I think I may just as well use only 17.09. There should be no obvious reason to "keep" 17.03, right?
<tilpner>
Only if you have to do an important thing with your machine tomorrow
<tilpner>
*+to do
<Alling>
You mean it will take a couple of hours or it might fail badly?
<tilpner>
(But you could always rollback, so it's not terrible)
<samueldr>
when failing, it probably won't be catastrophic
<Alling>
No critical services running so I can basically experiment wildly.
<samueldr>
long, it depends on the internet speed and the number of packages
<Alling>
But I would absolutely appreciate some further guidance.
<tilpner>
It should be smooth, and not take long
<Shados>
Alling: It'll download a bunch of stuff, and there's implausibility little chance of anything going really wrong. Real worst case is that you have to reboot back into your previous configuration and set your channel back to 17.03.
<Alling>
OK, so how do I do this?
<samueldr>
the worst upgrade I had was on a container not having system.stateVersion
<samueldr>
keeping an older stateVersion is generally not an issue, but if you're absolutely sure there's nothing that can be lost, it wouldn't be an issue
* tilpner
doesn't use channels, should probably not try to help
<Shados>
Nah, it's already only a handful of keypresses to begin with
<tilpner>
Alling - You're not supposed to touch stateVersion, in general
<Alling>
OK, I left it at "17.03" and ran this: sudo nix-channel --add https://nixos.org/channels/nixos-17.09 nixos && sudo nix-channel --update && sudo nixos-rebuild switch
<Shados>
Is there a summary of things that change between state versions?
<Alling>
I'm perfectly aware I'm such a noob at this. :D
<tilpner>
Alling - I always reboot after that, just in case
<samueldr>
not sure if it's all the things affected, but release notes generally have stateVersion summaries
<Alling>
tilpner: I wanna do that now that it's finished, but I don't trust my motherboard to boot, and I only have ssh access right now ...
<Shados>
Alling: It shouldn't be a problem to not reboot, you'll probably just be behind on kernel version more or less.
<Shados>
samueldr: Looks like the "Backward Incompatibilities" section has a bit explicitly about stateVersion changes, so that's pretty much what I wanted, thanks
<samueldr>
I don't think that's an exhaustive list
<Alling>
You gotta be kidding me ... Suddenly it just works. And I've been sitting all day trying all kinds of weird Nix tricks. You guys are absolutely fantastic! Shados tilpner samueldr
<tilpner>
Alling - FWIW, something like mitmproxy.override { python3 = python36; } should have worked too
<tilpner>
(But upgrading is good in general)
<Alling>
tilpner: Tried stuff like that; could not get it to build for various reasons (e.g. pyflakes failing).
hellrazor has joined #nixos
<tilpner>
I don't see a mention of pyflakes in .nix, and I don't really know what it is. I guess it's not important anymore
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] samueldr opened pull request #31810: documentation: python3Packages is python36Packages (release-17.09...patch-2) https://git.io/vFHl1
NixOS_GitHub has left #nixos []
hellrazo1 has quit [(Ping timeout: 240 seconds)]
<Alling>
What usually happens for me is that some dependency fails to build, so I write a Nix expression for that dependency ... and then one of its dependencies fails, so I have to write an expression for THAT ... ^.^'
<Alling>
And by "write" I mean "copy and modify".
<samueldr>
anyone has experience with building a react app using nix, using *something* to track dependencies through nix?
<samueldr>
like, clone project, nix-build, get react app out
<samueldr>
Alling: I may seem knowledgeable, but don't ask me to write a derivation for anything more complicated than the usual configure make make install software
<samueldr>
I'm mostly still a user, and used to the user-experience
ryantm has joined #nixos
<Alling>
samueldr: I've never even built anything in React. Elm and Haskell though. :p
mekeor has joined #nixos
fragamus has joined #nixos
<Alling>
Bonus context: This is the script I was trying to get running:
<NixOS_GitHub>
[nixpkgs] eqyiel opened pull request #31812: nixos/matrix-synapse: make home directory configurable (master...matrix-synapse-configurable-home) https://git.io/vFHRu
NixOS_GitHub has left #nixos []
pie_ has quit [(Remote host closed the connection)]
pie_ has joined #nixos
noobineer has quit [(Read error: Connection reset by peer)]
slack1256 has quit [(Remote host closed the connection)]
pie__ has joined #nixos
noobineer has joined #nixos
pie_ has quit [(Read error: Connection reset by peer)]
rogue_koder has quit [(Remote host closed the connection)]
ertes has quit [(Ping timeout: 248 seconds)]
asuryawanshi has quit [(Ping timeout: 260 seconds)]
<hyper_ch>
I'm still debugging my zfs issue with tcaputi... he told me to gdp the zfs.ko module... but the problem is that it will look for files on /tmp - I think those files were created during compiling.... how to run gdb when those files have vanished or how can I make them not vanish after compiling? - https://github.com/zfsonlinux/zfs/issues/6821#issuecomment-345466742
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 2 new commits to master: https://git.io/vFHgs
<NixOS_GitHub>
nixpkgs/master f8d232e Samuel Dionne-Riel: documentation: python3Packages is python36Packages...
<NixOS_GitHub>
nixpkgs/master c1ac125 Vladimír Čunát: documentation: python 2.6 is no more
<hyper_ch>
I wanted to ask you something but I forgot :)
<hyper_ch>
anyway, do you happen to be familiar with nixos container and bridged networking setup?
Itkovian has quit [(Remote host closed the connection)]
Itkovian has joined #nixos
<vcunat>
no, not at all
<hyper_ch>
ok :) well, I have bridge network running with some kvms.... works good but then when I tried to also use nixos container, the container couldn't get internet access and everything I tried didn't help :)
<vcunat>
I'll want to look in a way of getting close to running "nixos containers" on Turris Omnia
<hyper_ch>
still not having tried your new suggestions on the TO....
<hyper_ch>
there this one thing I still need to get to run on my now nixos office server
<vcunat>
well as long as your current setup works the way you want it, I understand there's little motivation to change
<hyper_ch>
doesn't TO already have LXC or something containers?
<hyper_ch>
the problem is, no updates with my current setup
<vcunat>
yes, it has LXC support
<vcunat>
and people do use it with Debian, Arch, etc. official containers
<vcunat>
(official images)
<hyper_ch>
also another issue is with not using dns-masq --> how to setup email or tell where the mail server is for a given domain
<vcunat>
oh, like /etc/hosts but for MX record?
Itkovian has quit [(Client Quit)]
<hyper_ch>
yeah, in dnsmasq.conf I can set mx-host=jus-law.ch,mail.jus-law.ch,10
<vcunat>
there's no such direct option ATM
<hyper_ch>
well, at the office, the mail server is hosted on the lan
<hyper_ch>
(it's a kvm)
<hyper_ch>
so, mail programs need to know where to look for it so the mx entry in dnsmasq is the way to get this done
<hyper_ch>
so, let's crash zfs again ;)
<vcunat>
well if it's a mail server for a public mail domain, it already has a name and I'd expect to just redirect the name (pointed to by MX record in public DNS) to the LAN address
<hyper_ch>
that doesn't really work for some reason....
<hyper_ch>
not when sending from the lan
<hyper_ch>
but it's a good question why.... hmmm....
<hyper_ch>
it's been a long time since I tried last
<hyper_ch>
it seems vbox builds again :)
erictapen has quit [(Ping timeout: 248 seconds)]
knedlsepp_ has joined #nixos
filterfish has quit [(Ping timeout: 248 seconds)]
deepfire has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vFHam
<NixOS_GitHub>
nixpkgs/master a1573bb Orivej Desh: python.pkgs.pybfd: move to python-modules
NixOS_GitHub has left #nixos []
JosW has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to staging: https://git.io/vFHaZ
<NixOS_GitHub>
nixpkgs/staging b089c21 Orivej Desh: Merge branch 'master' into staging...
<vcunat>
That stuff isn't too difficult, really, if looking at policy code around and knowing some general stuff about DNS.
<vcunat>
but the APIs could still see some improvements
<vcunat>
(and the docs)
<hyper_ch>
quantum physics isn't too difficult either for physicicsts....
ertes has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej opened pull request #31818: python.pkgs.pybfd: support split libbfd and libopcodes (staging...pybfd) https://git.io/vFHVf
NixOS_GitHub has left #nixos []
asuryawanshi has joined #nixos
goibhniu has joined #nixos
nil has joined #nixos
asuryawanshi has quit [(Ping timeout: 240 seconds)]
hotfuzz_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vFHV8
<ixxie>
can someone with python packaging experience walk me through my first package? At this point I am trying to figure out if there is some helper function to find the pypi sha256 hash
euniarte has joined #nixos
hiroshi- has joined #nixos
hiroshi has quit [(Ping timeout: 248 seconds)]
hiroshi- is now known as hiroshi
hotfuzz has joined #nixos
hotfuzz_ has quit [(Ping timeout: 255 seconds)]
* knedlsepp_
ixxiee: While this is probably not the best way to do it, I usually just write a wrong hash, try to build the thing and copy the sha256 from the error message that says: „output path ´…´ has r:sha256 hash ´...´ when ´...´ was expected
<vcunat>
that's probably the most commonly used way :-)
<vcunat>
but in most cases you can also nix-prefetch-url -A pkg.attribute.path.src
ansero1234 has quit [(Quit: Leaving)]
<vcunat>
(That has a small advantage that you only download it once, not twice.)
<sphalerite>
knedlsepp_ ixxie vcunat: nix 1.12 also has nix-build --hash, which builds any fixed-output derivation and saves it into the store regardless of whether it matches, and outputs the hash
<vcunat>
yes, I know that
<vcunat>
it should work in all the cases, but most people don't have 1.12 yet
* knedlsepp_
I watched the video a couple of days ago and was quite looking forward to it. On my last try to install nixUnstable however it quite messed up my nix install. So I guess I should try again.
<vcunat>
(quite hindering my workflow, so I'm using both versions at once)
deepfire has joined #nixos
* knedlsepp_
How is it broken? And how do you use both versions at once?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] rittelle opened pull request #31819: pythonPackages.pyqt5: Add optional support for qtconnectivity (master...pyqt-qtconnectivity) https://git.io/vFHoi
<NixOS_GitHub>
[nixpkgs] pierrechevalier83 opened pull request #31820: gdm: start gdm on tty7 rather than default tty1 (master...master) https://git.io/vFHKd
NixOS_GitHub has left #nixos []
adamt__ has joined #nixos
pierre__ has joined #nixos
pierre__ is now known as pierrechevalier8
<ixxie>
srhb: I based the backage on jupyter_core
adamt_ has quit [(Ping timeout: 248 seconds)]
<viric>
lall
<hyper_ch>
ixxie: got it to run now?
<hyper_ch>
hi viric
orivej has quit [(Ping timeout: 255 seconds)]
<ixxie>
hyper_ch: not yet; got a weird build error atm, its complaining about oauth2 version but as far as I can tell it shouldnt be an issue
<NixOS_GitHub>
[nixpkgs] primeos pushed 1 new commit to master: https://git.io/vFHPe
<NixOS_GitHub>
nixpkgs/master 5147d8b Michael Weiss: android-studio-preview: 3.1.0.2 -> 3.1.0.3
NixOS_GitHub has left #nixos []
Wizek_ has quit [(Ping timeout: 250 seconds)]
zzamboni has joined #nixos
tgunb has joined #nixos
Thra11 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to release-17.09: https://git.io/vFHPa
<NixOS_GitHub>
nixpkgs/release-17.09 ebb7e37 Peter Hoeg: kodi: 17.4 -> 17.6 and move to cmake...
NixOS_GitHub has left #nixos []
Wizek_ has joined #nixos
roberth has joined #nixos
Neo-- has joined #nixos
JosW has quit [(Quit: Konversation terminated!)]
glion has joined #nixos
hydraz has quit [(Remote host closed the connection)]
hydraz has joined #nixos
hydraz has quit [(Changing host)]
hydraz has joined #nixos
eqyiel[m] has left #nixos ["User left"]
<NickHu>
Is anyone else having libGL errors on NixOS with steam recently?
Neo-- has quit [(Ping timeout: 248 seconds)]
jensens has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vFHXL
<NixOS_GitHub>
nixpkgs/master 1b594f0 Peter Hoeg: mime-types: init at 9
NixOS_GitHub has left #nixos []
<adisbladis>
NickHu: Did you enable 32 bit dri
<jluttine>
syncthing service allows one to set up only one syncthing service. i have multiple users on my computer, and they all want to use syncthing..
<jluttine>
this applies to surprisingly many services in nixos: one can enable only one instance. in some cases that makes sense, but i think that in many cases there's no reason to limit the instances to one, because a user may want to set up many instances of the service
pierrechevalier8 has quit [(Remote host closed the connection)]
zzamboni has quit [(Quit: Leaving.)]
<danbst[m]>
jluttine: have you tried declarative containers for service isolation?
asuryawanshi has joined #nixos
pie__ has joined #nixos
asuryawanshi has quit [(Ping timeout: 248 seconds)]
Supersonic112 has quit [(Ping timeout: 255 seconds)]
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
zzamboni has joined #nixos
Phillemann has joined #nixos
<Phillemann>
I'm trying to use the nix Docker image as suggested on docker hub. However, it fails in hubstaff/default.nix because of an unexpected $undefined.
<joepie91>
jluttine: there's an open issue about this somewhere on the issue tracker I believe
<toppler>
Is there a good way to combine sets like { x.y = { z = 1; }; } ++ { x.y = { w = 2; };} to get { x.y = { z = 1; w = 2 }?
hellrazor has quit [(Ping timeout: 240 seconds)]
hellrazor has joined #nixos
rihards has quit [(Quit: rihards)]
Isorkin has joined #nixos
knedlsepp_ has joined #nixos
<clever>
toppler: lib.recursiveUpdate
<toppler>
Sweet, cheers.
hamishmack has quit [(Quit: hamishmack)]
phreedom has quit [(Read error: Connection reset by peer)]
jensens has joined #nixos
jtojnar has quit [(Quit: jtojnar)]
adamt__ has joined #nixos
jtojnar has joined #nixos
jtojnar has quit [(Read error: Connection reset by peer)]
jtojnar has joined #nixos
orivej has joined #nixos
erictapen has quit [(Ping timeout: 248 seconds)]
myrl has joined #nixos
<myrl>
Stupid question.
<myrl>
Is 4GB enough for Nix on Debian?
Itkovian has joined #nixos
<sphalerite>
myrl: nix itseslf, certainly.
<sphalerite>
Depends on whaty ou're installing with it though :)
<joepie91>
myrl: RAM or disk space?
<myrl>
...
vcunat has joined #nixos
<myrl>
sphalerite: Damn, you're right.
<myrl>
sphalerite: Just remembered the ghc on ARm process..
<myrl>
On ARM, (and other systems with no bincache), ghc has to cascade through a lot of versions, right?
<myrl>
joepie91: 4 GB disk space, 1 GB RAM, if that's relevant.
<myrl>
I guess ghc is a no go? :C
<joepie91>
myrl: right. I don't know specifically about GHC, but keep in mind that for a lot of packages, Nix will be pulling in all the deps recursively, even if an equivalent dep already exists on your system that was installed through apt... so for sufficiently complex applications, I would not be surprised if you hit the 4GB disk limit
<joepie91>
(assuming you're using Nix on it for the first time)
<joepie91>
for simpler applications you're probably fine
<myrl>
joepie91: I went with a clean install of debian, just for the sake of installing Nix.
<myrl>
Do you think it's safe to use nixos-in-place on a chroot?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 2 new commits to master: https://git.io/vFHyl
<NixOS_GitHub>
nixpkgs/master 1ec13ff Tuomas Tynkkynen: pythonPackages.pyamf: init at 0.8.0
<NixOS_GitHub>
nixpkgs/master 7d32169 Tuomas Tynkkynen: yle-dl: 2.20 -> 2.27
NixOS_GitHub has left #nixos []
knedlsepp_ has quit [(Quit: knedlsepp_)]
<myrl>
I should test that out on a VM.
<joepie91>
frankly I have no idea, but I expect that it will run into trouble for the same reason NixOS runs into trouble on OpenVZ / LXC: it wants to control the kernel, and it can't
<Dezgeg>
well, it mostly works on lxc
<sphalerite>
myrl: I essentially pivoted my Chromebook from a debian chroot with nix to NixOS
<sphalerite>
that worked pretty nicely
<joepie91>
Dezgeg: on a non-NixOS host?
<gchristensen>
would anyone mind reading the TOS for Github's Gists to see if there are rules / limits on automated useE?
<Dezgeg>
yes
<myrl>
sphalerite: NixOS chroot?
<sphalerite>
myrl: no, actual NixOS :)
<myrl>
sphalerite: Aww. :C I want to use Android as a "hypervisor," so it's probably a no-go.
<joepie91>
Dezgeg: have any reading materials on how to make this work for me? :)
<joepie91>
gchristensen: one sec
* joepie91
ToS scanner mode engage
<Dezgeg>
the thing that doesn't work at all is that the build sandbox can't be enabled, and nixos-rebuild switch throws some errors that don't seem to cause problems
<gchristensen>
thank you :D
<sphalerite>
myrl: well I started with a sort-of-nixos chroot, never tried running systemd in there though
<joepie91>
gchristensen: "While using GitHub, you agree that you will not under any circumstances: [...] attempt to disrupt or tamper with GitHub's servers in ways that could harm our Website or Service, to place undue burden on GitHub's servers through automated means [...]"
<joepie91>
gchristensen: scraping not allowed unless within a specific category of usecases, but otherwise doesn't address automation in general
<myrl>
sphalerite: Ah. I wonder if those suckless advocates were able to make progress on the alternate init system.
<joepie91>
"If we determine your bandwidth usage to be significantly excessive in relation to other GitHub customers, we reserve the right to suspend your account or throttle your file hosting until you can reduce your bandwidth consumption."
<gchristensen>
ok I'll try it until they ask me to stop (cc domenkozar)
<joepie91>
yadda yadda copyright yadda yadda privacy yadda
<clever>
builder for ‘/nix/store/wla89i87v87pbwq5kyn5g2qvmr40a4dx-fail-4.9.0.0.drv’ failed with exit code 1
<clever>
lol
<myrl>
sphalerite: root + linux deploy(chroot) + XSDL(x server). I'm not sure how xsdl works with proot(chroot emulation by syscall tracing, I think).
<sphalerite>
clever: myrl : I think there are sort of stages of caring. I have a friend who deletes files obsessively (often to the point of removing data they actually end up needing again later). I think the next step is notifications, then browser tabs
orivej has quit [(Ping timeout: 248 seconds)]
<chris_>
Hello! I'm trying to install a haskell package to be available in ghci - could anypoint point me to what I should be looking for? Would the library be installed with cabal maybe?
<sphalerite>
myrl: hmm ok. I'd quite like to take out the android part for mine but not sure how feasible that is
rihards has joined #nixos
<clever>
sphalerite: ive found old harddrive backups, in old harddrive backups
<sphalerite>
myrl: and the screen on the phone I'm planning to use is very broken — touch input works but the output is almost completely gone
<clever>
sphalerite: and i have at least a dozen harddrive images from old installs
<sphalerite>
clever: :D
<myrl>
sphalerite: touch input fully works?
<clever>
sphalerite: and yes, i have rescued 4 year old projects from those backups
<sphalerite>
myrl: yep, just can't really see what I'm touching xP
<myrl>
sphalerite: Did you have USB debugging before it died? Or, well, the easier option: Can anyone repair it?
<myrl>
sphalerite: Either way, broken screen means you're probably not taking Android out. :P
<clever>
sphalerite: some phones have hdmi output, even if you think its just USB
<sphalerite>
myrl: yeah I did, but then I broke the boot xD as for repairing, not really keen on that option. I originally broke it by dropping it in the airport, which broke the touch input (but the display was still working fine). So I got it repaired, and the repair guy did a really shoddy job of it that apparently put a lot of mechanical strain on the screen, and it cracked spontaneously in my pocket
<clever>
sphalerite: they got tricky, and made the "micro usb" socket deeper, so special connectors can reach the hdmi pins
<hodapp>
hmmm. git bisect has told me that c3255fe8ec326d2c8fe9462d49ed83aa64d3e68f broke pkgs.python35Packages.tensorflowWithCuda
<sphalerite>
so I'm not too keen on another repair, plus they're not cheap and the main reason I want to use this phone in particular is that I already own it => no cost
<sphalerite>
what I'm hoping to do is build my own OS image that I can fiddle with arbitrarily. I can install it through the USB debugging in the recoviery
<myrl>
clever: I don't even have OTG on my phone... feelsbadman. :(
<toppler>
myrl: I'm running nix on a VM and I'm currently using 7.2G of disk space. I have 1GB of RAM, and definitely need to go into swap for some installations.
<toppler>
(it's also headless)
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<myrl>
toppler: What do you have installed?
<myrl>
toppler: And are you using a bincache?
<sphalerite>
clever: apparently it has slimport, not MHL
jtojnar has joined #nixos
<toppler>
myrl: Mostly just a headless emacs, GHC, various haskell stuff and a few simple servers like email.
<toppler>
I'm not using a bincache.
<myrl>
toppler: How did you bootstrap GHC?
knedlsepp_ has joined #nixos
<toppler>
myrl: I didn't need to.
<myrl>
toppler: What arch?
<myrl>
toppler: unstable or release?
<toppler>
Not sure. QEMU Virtual CPU? And unstable.
<toppler>
Oh, i686
<myrl>
toppler: Tanks.
<myrl>
toppler: How long did it take for all those to build?
<toppler>
myrl: The one that needs the most time after an update is my main emacs shell, which can take twenty minutes or so, I guess.
<toppler>
I'd like to be able to run Nix on a raspberry pi, but had some pain last time. And besides, does GHCi support ARM yet?
__monty__ has joined #nixos
<toppler>
Probably not that long, actually. I haven't had any complaints about build times yet, but I'm doing this over SSH and don't need it to be super fast.
<hodapp>
toppler: I believe GHCi has supported ARM for at least a year or two now
<toppler>
hodapp: Ah sweet. I should have guessed I was out-of-date on that one.
<__monty__>
I created an environment for haskell development but I get errors when loading it and I have no idea what they mean: http://ix.io/Cab
<myrl>
toppler: How about ghc?
<toppler>
myrl: Again, it's nothing too bad. We're talking tens of minutes at most. The compiler needs more than 1GB for my purposes, but it wasn't horrible going to swap.
<toppler>
If you only have 4GB of hard disk, that might be a problem.
<myrl>
toppler: Oh wait. Single threaded?
<toppler>
It's one core.
<myrl>
Yeah, that makes sense. Thanks.
<srhb>
Which package does nixos-install come from?
<joepie91>
it's really irritating how github doesn't send notifications for comments to gists...
<chris_>
Is there a way to see modified / created files by nix-env -i ?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] grahamc reopened pull request #30777: DO NOT MERGE: testing to see if we got assimilated (master...do-not-merge) https://git.io/vdjox
<aanderse>
i have a laptop running nixos+kde and when i put the laptop into suspend mode the fan never turns off and the keyboard backlight remains on. if i press any keys the laptop does not wake up, even if i press the power button the laptop does not wake up. i have to hard reboot the laptop. after a hard reboot i check journalctl from the prior boot and the last thing the log says is something about the system about to suspend. gone through a bunch of
<aanderse>
stuff online but so far no solutions....
<aanderse>
anyone have any ideas at all?
ma27 has quit [(Ping timeout: 240 seconds)]
<clever>
aanderse: my laptop sometimes locks up as its going into suspend, the difference is if the power light goes out fully
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vFHHt
<NixOS_GitHub>
nixpkgs/master 7b86918 Jon Banafato: corebird: 1.7.1 -> 1.7.2
<aanderse>
the laptop is pretty new and also runs windows 10. when i first got the laptop the suspend/resume did work on nixos... i'm not sure when it stopped working because i hadn't tried using it in a while. i don't have the full nixos history on my boot screen as i cleaned out old versions (oops)
<aanderse>
i'm not even sure how to troubleshoot this
asuryawanshi has quit [(Ping timeout: 240 seconds)]
<aanderse>
i've tried isolating different components which could cause the issue
<clever>
suspend is fairly tricky to debug, because you cant easily get logs out of it
<aanderse>
like kde powerdevil... i completely disabled powermanagement but the issue remained
<aanderse>
yeah
<aanderse>
the logs haven't given me much
<aanderse>
the laptop has 16gb of ram so i never bothered to make a swap so can't test hibernate to see if it works
<aanderse>
i'm thinking to load a live cd/usb and create a swap so i can test hibernate
<clever>
my laptop has 32gig of ram, and i tried to make 40gig of swap for hibernate
<myrl>
Apparently, I'm not using systemd under chroot, but `service` still works. This is wow.
toppler has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] johbo opened pull request #31830: minikube: Add vmnet as buildInput on darwin (master...fix-minikube-darwin) https://git.io/vFHQ1
NixOS_GitHub has left #nixos []
<katona>
Why doesn't nixos-rebuild downloads cached binary, even though I have the exact same revision number?
<katona>
Download*
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] johbo closed pull request #31830: minikube: Add vmnet as buildInput on darwin (master...fix-minikube-darwin) https://git.io/vFHQ1
<vcunat>
hyper_ch: a caching layer for a block device
<vcunat>
e.g. you have a large rotating drive and a small fast SSD
<vcunat>
combined into a single block device
<hyper_ch>
vcunat: nothing to worry about for me then :)
<vcunat>
frequently accessed blocks get cached onto the SSD
<vcunat>
It's not used in default installs
<clever>
vcunat: so its like the L2 of zfs?
<hyper_ch>
clever: what's l2 of zfs? :)
<vcunat>
I'm afraid I'm not familiar with how ZFS works
<clever>
zfs has an L2 ARC
<clever>
where you can use an SSD as a read-cache for your magnetic drives
<hyper_ch>
anyway, that user reported that 4.14's bcache ate his data :)
<vcunat>
bcache does write caching as well (optionally)
<joepie91>
hmm, is there some sort of cache issue? I've seen a few people mention now that they're not getting prebuild binaries
<clever>
zfs has a log device for speeding up writes
ma27 has quit [(Ping timeout: 255 seconds)]
<hyper_ch>
how does the log help speeing up writes? I mean on spinning disk you can output it into one stream and alter "assign" it correctly... but on ssds?
<clever>
hyper_ch: with the zfs log device, it will write all data to the log, in a circular buffer
<clever>
hyper_ch: and once it has been flushed to the log device, the FS considers it saved and the syscall can return
<clever>
so it can skip waiting for the magnetic drives to seek
<hyper_ch>
ah log device
<clever>
under normal operation, the log device is 100% write, and it will never read, and the only seek operation is to return to the start
<clever>
and log is only read at recovery, after an improper shutdown
<clever>
it will keep a 2nd copy of the pending writes in ram, and send them to the main storage at a later time
jtojnar has quit [(Remote host closed the connection)]
jtojnar has joined #nixos
iyzsong has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
<sphalerite>
myrl: any commit hash in the nixpkgs repo, yes
<sphalerite>
myrl: also for symbolic refs like branch names and tags
<myrl>
Cool. Thanks.
<euniarte>
Is there a way to make nix create symlinks in /usr/bin for programs other than env?
<sphalerite>
myrl: any URL where you can get a tarball really. Github specifically has those convenient /archive/foo.tar.gz
<euniarte>
(on nixos)
<myrl>
sphalerite: Can I use nixpkgs-channels instead?
<myrl>
They have the same directory structure, right?
<sphalerite>
myrl: yeah, although that only has a subset of the nixpkgs commits and is longer to type :p
<myrl>
sphalerite: Thanks. :P I needed the fork of nixpkgs-channels, since it has ghc binaries for arm.
<sphalerite>
myrl: oh, it won't use any extra binary caches though
<myrl>
sphalerite: Yeah. I only need the binaries, since there are no binaries for ARM and I can't even get ghc to work.
<sphalerite>
No difference in behaviour
<myrl>
(OTOH, I wonder if you can bootstrap the very first ghc from hugs then cascade from that.)
<myrl>
That should be fun.
<sphalerite>
Oooh only just saw it was dhess/nixpkgs-channels, not nixos/nixpkgs-channels
<myrl>
I'm trying to think of a way to make this permanent.
<myrl>
Set NIX_PATH to the archive?
<clever>
if you point NIX_PATH directly to an archive, then <nixpkgs> will check the archive for a nixpkgs directory (it wont be there)
<clever>
and then it just fails
<clever>
so you need to do NIX_PATH=nixpkgs=archiveurl
peacememories has joined #nixos
<clever>
which maps the nixpkgs path to the root of that archive
<myrl>
clever: Thanks so much.
jensens has quit [(Ping timeout: 240 seconds)]
<trikl[m]>
Every time I try to build something it fails with `error: getting status of '/nix/store/ijrxfh0bq3ssldkrcgkm8asqffkwqnpn-python3-3.5.3/bin/pip': No such file or directory`
<clever>
trikl[m]: are you using nixUnstable?
<myrl>
trikl[m]: Did you have to force fsck earlier?
<trikl[m]>
myrl: I did not
<trikl[m]>
clever: I did install nixUnstable, but I'm using nix-env
<clever>
trikl[m]: behind the scenes, nix-channel uses nix-env to install the new channel into your channel profile
<hodapp>
"collision between `/nix/store/3ivy4wirbf9xpc93z448gvdv5kxkn106-python3.5-html5lib-0.9999999/lib/python3.5/site-packages/html5lib/__init__.py' and `/nix/store/xm03cfb3w3wpd5h77047x7s6ilnxlxak-python3.5-html5lib-0.999999999/lib/python3.5/site-packages/html5lib/__init__.py'"
erictapen has joined #nixos
* hodapp
stabs Python and html5lib in the face
<trikl[m]>
Is that version python3.5-html5lib-0.999999999 some kind of joke?
<myrl>
Is it normal that my ARM device is doing "download-from-binary-cache.pl"?
<hodapp>
I don't know, I just know I've had to deal with this bullshit several times
<sphalerite>
myrl: yeah I think it's querying it regardless
<hodapp>
it appears that one html5lib version is a transitive dependency of one thing I'm using, and something else pulls a conflicting version - and pip cannot handle this
<myrl>
sphalerite: So... disable bincache.
w-spc-gir has joined #nixos
jacob has joined #nixos
<sphalerite>
Since you won't find any binaries though you can disable it
<sphalerite>
Yeah
<clever>
myrl: some font packages are specially written so they can go into the binary cache
<myrl>
Sigh. Stupid question. How again?
<hodapp>
if I could find a way to make Jupyter run a kernel that is isolated in its own environment (nix-shell or something) it probably wouldn't be a problem
<myrl>
I haven't used just nix at all.
<clever>
myrl: change the value of binary-caches = in /etc/nix/nix.conf
<sphalerite>
If it doest exist, create it
<myrl>
clever: Thanks, there's no equivalent to rebuild?
<myrl>
I mean
<sphalerite>
Nope
<myrl>
Is setting it enough to change the behavior?
<clever>
myrl: you can override any nix.conf flag with --option
<clever>
nix-build --option binary-caches ""
chris_ has joined #nixos
dywedir has quit [(Remote host closed the connection)]
<hyper_ch>
it's the problem I have with kvm/virt-manager... it also uses virtio and works fine from the ISO... but when I reboot I get the same... so I changed from virtio to sata
<Dezgeg>
doesn't nixos-generate-config recognize virtio? (or the vmware one for that matter)?
<hyper_ch>
I guess not
astronav_ has quit [(Remote host closed the connection)]
aminechi1haoui has joined #nixos
<myrl>
Err... what does ARM use to build gcc?
<Dezgeg>
nothing special on ARM
<sphalerite>
It's a bootstrap process like any other
<Isorkin>
thanks, sata - worked
<Dezgeg>
that's probably slower than vmw_pvscsi
myrl has quit [(Quit: WeeChat 1.9.1)]
asuryawanshi has joined #nixos
Thra11 has quit [(Ping timeout: 250 seconds)]
<clever>
Dezgeg: that reminds me, nixos-generate-config can find luks devices and register them in the config properly
<clever>
Dezgeg: but it cant find the luks under an lvm
kaba has joined #nixos
<clever>
because the luks isnt in the output of "mount"
myrl has joined #nixos
<sphalerite>
lsblk gets the hierarchy though iirc doesn't it?
hellrazor has quit [(Ping timeout: 248 seconds)]
<Dezgeg>
yeah probably it could be made smarter
aminechi1haoui has quit [(Client Quit)]
<kaba>
Hi, I've asked it already, sorry if it looks as if I'm spamming the chat, but I would really like to solve this problem.. I'm using this command: nixos-rebuild -I "nixpkgs=https://github.com/jtojnar/nixpkgs/archive/830dc3d34b4583d7e6c1a49681583744f6354ede.tar.gz" switch to change to gnome 3.26
<sphalerite>
Yeah but does it even make sense to talk about purity in this context?
<clever>
its a variable that changes depending on state, that could be improved
nordkap has joined #nixos
<sphalerite>
But it's all state
<clever>
oh, another more crippling bug, the zram stuff
<clever>
zram adds some special block devices as swap, to compress the data and store it back in ram
<kaba>
Maybe it uses a different version of nix? But I've also tried commenting out almost every line in configuration.nix building based on the wanted rev (it builds without compilation), and starting the build of the gnome shell after changing into that environment, so it can't possibly be a separate nix version problem. Somehow it forgets about the cache
<clever>
nixos-generate-config adds those devices to hardware-configuration.nix, without the compression config, and then adds with zram, and tries to enable the same devices twice
astronavt has joined #nixos
asuryawanshi has joined #nixos
adamt__ has joined #nixos
kaba has quit [(Quit: Page closed)]
astronavt has quit [(Remote host closed the connection)]
asuryawanshi has quit [(Ping timeout: 264 seconds)]
asuryawanshi has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
asuryawanshi has quit [(Ping timeout: 240 seconds)]
nordkap_ has joined #nixos
Itkovian has joined #nixos
nordkap has quit [(Read error: Connection reset by peer)]
asuryawanshi has joined #nixos
mrkgnao has quit [(Ping timeout: 248 seconds)]
mrkgnao has joined #nixos
noobineer has joined #nixos
noobineer has quit [(Max SendQ exceeded)]
nordkap_ has quit [(Quit: Leaving)]
<Acou_Bass>
query - does anyone use the VPN service 'privateinternetaccess' with nixos? ive only ever set it up using their own scripts, so not 100% sure if itll just work with network manager
<drakonis>
PIA?
<Acou_Bass>
yar
jb55 has quit [(Ping timeout: 240 seconds)]
<Acou_Bass>
nevermind, it seems like doing it purely via networkmanager is simple enough :P
asuryawanshi has quit [(Ping timeout: 260 seconds)]
<katona>
Acou_Bass: but you will likely have to enable autoloadconntrackhelpers
<Acou_Bass>
hmm is there any reason that needs to be enabled? i dont think ive ever enabled that on my fireawll on another distro
<Acou_Bass>
i will bear it in mind though, thanks :D
asuryawanshi has quit [(Ping timeout: 248 seconds)]
astronavt has joined #nixos
asuryawanshi has joined #nixos
<obadz>
gleber_: you just did it ;-)
Thra11 has quit [(Ping timeout: 255 seconds)]
asuryawanshi has quit [(Remote host closed the connection)]
asuryawanshi has joined #nixos
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
jb55 has quit [(Ping timeout: 248 seconds)]
erictapen has joined #nixos
pie_ has joined #nixos
astronavt has quit [(Remote host closed the connection)]
Thra11 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 4 new commits to master: https://git.io/vFHx2
<NixOS_GitHub>
nixpkgs/master 58218d4 Franz Pletz: libsndfile: add patches for multiple CVEs...
<NixOS_GitHub>
nixpkgs/master 439fc35 Franz Pletz: owncloud705: remove vulnerable version...
<NixOS_GitHub>
nixpkgs/master 1d66d64 Franz Pletz: zabbix20: 2.0.11 -> 2.0.21 for multiple CVEs...
NixOS_GitHub has left #nixos []
<gleber_>
obadz: thanks! Should this test be added to `release.nix` (which I did not know existed :) ) ?
orivej has joined #nixos
erictapen has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 4 new commits to release-17.09: https://git.io/vFHxi
<NixOS_GitHub>
nixpkgs/release-17.09 c4ff51e Franz Pletz: libsndfile: add patches for multiple CVEs...
<NixOS_GitHub>
nixpkgs/release-17.09 e728baf Franz Pletz: owncloud705: remove vulnerable version...
<NixOS_GitHub>
nixpkgs/release-17.09 26118d2 Franz Pletz: zabbix20: 2.0.11 -> 2.0.21 for multiple CVEs...
NixOS_GitHub has left #nixos []
<obadz>
gleber_: only if it should prevent a nixos commit from being released to a channel…
<fpletz>
obadz: no, only if it's a constituent of the tested job
<fpletz>
please add it to release.nix to have it built on hydra
<fpletz>
the tested job is defined in release-combined.nix
<gleber_>
obadz: if it starts to fail, it means that mutableUsers mechanism is somehow broken. Currently it tests only a part of it. I want to extend it to test that an assertion preventing #4990 from happening to users holds
<obadz>
fpletz is correct of course :)
<fpletz>
gleber_: sounds useful to have channels releases fail if that test fails :)
<obadz>
I'm getting rusty on nixos stuff :-/
<fpletz>
:)
<gleber_>
fpletz: should I add it both to release.nix and add a mention of it to release-combined.nix?
<gleber_>
or just release.nix?
<fpletz>
gleber_: both please :)
<gleber_>
fpletz: ack
myrl has quit [(Ping timeout: 248 seconds)]
alhariel has joined #nixos
<gleber_>
fpletz: how do I test that the test is correctly included?
<Isorkin>
how to fix boot with hardened mode? error - could not insert 'zfs': Unknown symbol in module, or unknown parameter (see dmesg)
asuryawanshi has quit [(Remote host closed the connection)]
<fpletz>
gleber_: for release.nix, you can just do: nix-build nixos/release.nix -A tests.yourtestname
chrishill has joined #nixos
<fpletz>
gleber_: doing the same with the tested job from release-combined.nix is not that easy, hydra will parse the constituents option iirc
<toppler>
Has anyone reached out to youtubers and asked if they'd be willing to have their content on IPFS?
<Isorkin>
added - boot.kernelModules = [ "spl" "zfs" ]; - not work
<toppler>
Whoops, wrong channel
<fpletz>
Isorkin: unfortunately zfs is not compatible with some of the hardening features iirc
asuryawanshi has joined #nixos
astronavt has joined #nixos
<fpletz>
Mic92: do you know that for sure? :)
92AACG0BR has joined #nixos
<92AACG0BR>
[nixpkgs] orivej pushed 2 new commits to staging: https://git.io/vFHpv
<92AACG0BR>
nixpkgs/staging 130bf93 Orivej Desh: python.pkgs.pybfd: support split libbfd and libopcodes...
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
Oida has quit [(Ping timeout: 240 seconds)]
Isorkin has quit [(Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)]
oida has joined #nixos
acarrico has quit [(Ping timeout: 248 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ttuegel opened pull request #31835: Iosevka: build from source, custom font sets (master...iosevka/source) https://git.io/vFHpP
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vFHp1
<NixOS_GitHub>
nixpkgs/master 4e57256 Franz Pletz: openjpeg: 2.1.2 -> 2.3.0 for multiple CVEs...
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to release-17.09: https://git.io/vFHpy
<NixOS_GitHub>
nixpkgs/release-17.09 4370e93 Franz Pletz: openjpeg: 2.1.2 -> 2.3.0 for multiple CVEs...
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] bjornfor pushed 3 new commits to master: https://git.io/vFHpS
<NixOS_GitHub>
nixpkgs/master 37875c4 Pierre Chevalier: gdm: start gdm on tty7 rather than default tty1...
<NixOS_GitHub>
nixpkgs/master aed114e Pierre Chevalier: Revert "gdm: don't start getty@tty1 if enabled"...
<NixOS_GitHub>
nixpkgs/master c46d4da Pierre Chevalier: gdm: remove old workaround for conflicts with getty...
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] bjornfor closed pull request #31820: gdm: start gdm on tty7 rather than default tty1 (master...master) https://git.io/vFHKd
NixOS_GitHub has left #nixos []
nallar has joined #nixos
Ross has quit [(Ping timeout: 240 seconds)]
nallar is now known as Ross
<catern>
what's the best language to write simple scripts around Nix? is there a Nix API for Python?
<catern>
I guess the best language is Perl isn't it :)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to staging: https://git.io/vFHpb
<NixOS_GitHub>
nixpkgs/staging fa824e6 Franz Pletz: Merge branch 'master' into staging
NixOS_GitHub has left #nixos []
bgamari has quit [(Ping timeout: 240 seconds)]
knedlsepp_ has joined #nixos
<obadz>
catern: I don't think there's a best languange. There are perl, python and shell scripts out there and probably a bunch of other langs too. What are you trying to do?
<catern>
no, I say Perl because there is a Perl API for Nix
<catern>
and not one for any other language
<toppler>
Isn't much of Nix written in perl?
<catern>
some of it, yes
<clever>
toppler: the perl has been purged
ixxie has joined #nixos
<toppler>
clever: Apologies for not just googling my answers, but what's replaced it?
<toppler>
Or are you self-hosting :)
<clever>
toppler: c++
tmaekawa has joined #nixos
<clever>
toppler: within nix itself
jb55 has joined #nixos
<clever>
nixpkgs still uses a perl and bash everywhere
<toppler>
Were there major performance issues to justify that move?
<fpletz>
gleber_: in #31805 you also need to add the test to release.nix
<toppler>
On funding, I'm contributing a regular payment to the foundation. Is that not going to development stuff like this?
<fpletz>
gleber_: in this case, the tested job also makes sense - thanks for all those basic tests :)
<gleber_>
fpletz: I am enjoying working on Nix a lot :) I've subscribed to the code triaging things and am trying to improve things here and there as I learn
<gleber_>
fpletz: I will add it too
tmaekawa has quit [(Client Quit)]
bgamari has joined #nixos
zzamboni has quit [(Quit: Leaving.)]
andrewrk has joined #nixos
<andrewrk>
I get this error when using nox with the newest unstable: https://clbin.com/VzhuB
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vFHhR
<NixOS_GitHub>
nixpkgs/master 68d05c0 Franz Pletz: owncloud705: fix evaluation
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to release-17.09: https://git.io/vFHh0
<NixOS_GitHub>
nixpkgs/release-17.09 3c51620 Franz Pletz: owncloud705: fix evaluation...
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to staging: https://git.io/vFHhz
<NixOS_GitHub>
nixpkgs/staging 7faaade Franz Pletz: Merge branch 'master' into staging
NixOS_GitHub has left #nixos []
<catern>
obadz: in particular I want to extract information from store derivations
<catern>
in particular, the dependencies of those store derivations (as store derivations)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz closed pull request #31569: nixos/tests: add a test for config.users.mutableUsers. (master...add-mutable-users-test) https://git.io/vFrYh
<disasm>
I set networking.defaultMailserver in configuration.nix. I verified which sendmail points to ssmtp. But if I run `echo test|mail -s "foo" disasm@gmail.com it says mail: cannot open mailer: no such file or directory mail: cannot send message no such file or directory
<catern>
oh, perfect
<catern>
nixUnstable has a show-derivation command which turns a derivation into JSON!
<disasm>
errr I mean I verified the sendmail wrapper is in the path when I said I verified it points to ssmtp
<ixxie>
hyper_ch: not sure if that is relevant; I am packaging a python package
Arcaelyx has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<imagio>
I'm having trouble trying to load module-echo-cancel in pulseaudio. I have in my configuration.nix for pulseaudio extraconfig = '' load-module module-echo-cancel'' but when I try to start pulseaudio I get "Assertion 'PA_SINK_IS_LINKED(s->state)' failed at pulsecore/sink.c:1707, function compute_reference_ratios(). Aborting."
<imagio>
Any tips about modules in pulseaudio or the echo-cancel module in particular?
<ixxie>
is there a python-version-agnostic way to specify a python package? Can I just write pythonPackages.redis
<andrewrk>
ok it wasn't EFI boot space that was the problem. I'm still getting c5badb123a when I boot, but b8f7027360 after I run rebuild-switij
jb55 has quit [(Ping timeout: 240 seconds)]
<clever>
andrewrk: is /boot mounted when you do a switch?
<andrewrk>
no
<andrewrk>
does that mean something is wrong with my /etc/nixos/configuration.nix?
<clever>
andrewrk: yeah
<clever>
ixxie: i had to do this: configFile = pkgs.runCommand "config.pa" {} "cat ${./default.pa} ${./extra.pa} > $out";
<clever>
ixxie: and append the default pulse config with my custom one
<clever>
imagio: oops, that was for you
<ixxie>
clever: oh that makes more sense xD
<ekleog>
hmm, how do you make a service wait for complete network initialization? I've got static network configuration on all interfaces, yet when depending on either the .device's or network-setup.service, opensmtpd refuses to start :/
<clever>
ekleog: there is also a network.target and network-online.target
<ekleog>
hmm, according to my current opensmtpd.service, network.target is already in the dependance list... will try adding network-online, thanks!
noobineer has quit [(Read error: Connection reset by peer)]
vcunat has quit [(Ping timeout: 246 seconds)]
acarrico has joined #nixos
Wizek__ has quit [(Ping timeout: 240 seconds)]
noobineer has joined #nixos
noobineer has quit [(Max SendQ exceeded)]
<ixxie>
Its so weird, I can specify some python package dependencies without any pythonPackages prefix, but others don't work
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz closed pull request #29269: Test opening encrypted filesystem with keyfile (master...test_crypt_with_keyfile) https://git.io/v5MR8
NixOS_GitHub has left #nixos []
<imagio>
clever: the content of /etc/pulse/default.pa is as expected. If I try to manually load the module with "pactl load-module module-echo-cancel" pulseaudio crashes with the error I listed above
tv1 is now known as tv
tv has quit [(Changing host)]
tv has joined #nixos
<ekleog>
clever: looks like this one worked! thanks :D
<ixxie>
maybe its because redis for example is not only a pythonPackage but has other variants
seanparsons has quit [(Ping timeout: 250 seconds)]
euniarte has quit [(Ping timeout: 258 seconds)]
euniarte has joined #nixos
<Mic92>
fpletz: I am not so familiar with hardening patches. I am general suspicious with changes not accepted by upstream
goibhniu1 has joined #nixos
goibhniu1 has quit [(Remote host closed the connection)]
goibhniu1 has joined #nixos
<catern>
Is there a way to instruct Nix to realize a derivation *only* with substitutes? i.e. don't try to build at all? and if there are no substitutes, then just fail?
goibhniu has quit [(Ping timeout: 240 seconds)]
seanparsons has joined #nixos
<catern>
Or to otherwise implement logic like that? I just want to check if substitutes are available for a path (and actually realizing the path in the process is fine)
tgunb has joined #nixos
<Ralith>
I recall there being such an option
<catern>
I don't see it
<catern>
There's --fallback which does the opposite
<symphorien>
-j0 maybe
<catern>
(er, well, does an opposite in a sense - it enables falling back to local builds. but falling back to local build is the default for any path that doesn't have a substitute. i want to disable that fallback)
<catern>
symphorien: hah, that would be amusing if it works :)
tgunb has quit [(Remote host closed the connection)]
<catern>
(an option to *disable* substitutes would also be useful, though not for me)
<catern>
(unrelatedly, I think I have figured out what mechanism would be really useful for Nix: "untrusted build offload servers", where you can offload builds of fixed-output derivations. and any user can add to this list of servers for their builds!)
jb55 has joined #nixos
drakonis has quit [(Quit: Leaving)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vFQf2
<NixOS_GitHub>
nixpkgs/master c32cf00 Jon Banafato: corebird: 1.7.2 -> 1.7.3...
<symphorien>
catern: I just checked, -j0 works, unless you use distributed builds
<catern>
symphorien: oh neat, thank you!
yegortimoshenko has quit [(Quit: Page closed)]
<doublehp>
I was able to build firefox 57 2 months ago, now, I got "FATAL: kernel too old"; how can I Revert to an old package that will build like 2 months ago ?
<symphorien>
you can use nix-env to install packages as on nixos-17.09
<symphorien>
and they shouldn't use glibc2.26
MinceR has joined #nixos
<symphorien>
so you won't have the FATAL kernel too old pb
<ixxie>
what do I make of it when a build failes because several 'requirement[s are] already satisfied'?
liketechnik has joined #nixos
asuryawanshi has joined #nixos
<doublehp>
... I think it worked; it's now grabbing glibc-2.25
<liketechnik>
Hello, I have a question concerning package overrides in /etc/nixos/configuration.nix. I want to use a different version of firefox, but I'm always getting errors when running nix-rebuild switch
katyucha has joined #nixos
hariel has joined #nixos
timbod7 has joined #nixos
jensens has joined #nixos
<ixxie>
liketechnik: how are you overriding it?
alhariel has quit [(Ping timeout: 268 seconds)]
Itkovian has joined #nixos
<liketechnik>
I took the way described in the tutorial: ixpkgs.config.packageOverrides = pkgs: { firefox = pkgs.firefox.override { version = "58.0b4"; src = fetchurl { url = "mirror://mozilla/firefox/releases/58.0b4/source/firefox-58.0b4.source.tar.xz"; sha512 = "ecf4e22781849194df9bf60d66ce57aba7ec8d39ed4a50b2f51d1dbc33242a420adcf0a6301975667deef28af1b9ae5585af7f208814bde613e29dc9ca5c88dd"; }; }; };
<liketechnik>
Ooops, sorry my bad, i overlooked the first overrideAttrs part
<liketechnik>
Now I get the following error: error: attribute ‘fetchurl’ missing, at /nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs/lib/modules.nix:163:28
<clever>
liketechnik: you need to use pkgs.fetchurl
<tilpner>
Is builtins.fetchurl worse?
<clever>
tilpner: in the stable nix, it doesnt support a sha256
<liketechnik>
this works now, thank you guys
<tilpner>
Oh, pkgs.fetchurl does? (I'm using nixUnstable)
<liketechnik>
i'm using the latest stable version, it seems to work (although it didn't install anything new)
<liketechnik>
and firefox --version still reports version 56
myrl has joined #nixos
<clever>
tilpner: pkgs.fetchurl also supports mirrors and a few other options
<tilpner>
Do you have firefox in your systemPackages, and did a nixos-rebuild switch?
<NixOS_GitHub>
[nixpkgs] bluescreen303 opened pull request #31839: firewall: fix rpfilter blocking dhcp offers when no ip was bound yet (master...fix-rpfilter) https://git.io/vFQk9
NixOS_GitHub has left #nixos []
jensens has joined #nixos
justanotheruser has quit [(Quit: WeeChat 1.7.1)]
<ekleog>
hmm, any way to mark a filesystem as optional? currently nixos tries to mount a fs it doesn't manage to mount and resets in « press ^D to continue » every few minutes :/
erictapen has quit [(Ping timeout: 248 seconds)]
zzamboni has joined #nixos
justanotheruser has joined #nixos
justanotheruser has quit [(Client Quit)]
justanotheruser has joined #nixos
<gchristensen>
but nofail in the mount options ekleog
jmc_fr has quit [(Remote host closed the connection)]
erictapen has joined #nixos
<joepie91>
gchristensen: do you happen to have any particular ideas on how to approach partial rebuilds, conceptually? I'm thinking of a scenario where eg. a build for a 1GB game is stuck on a really slow download, but everything else can be installed fast, and the user does not want to wait for the 1GB game to finish downloading since it's not a dependency of anything anyway
<joepie91>
gchristensen: (assuming a declarative system config, no imperative nix-env)
<gchristensen>
no but someone here is doing a phd on it
<gchristensen>
or a masters? I forget
<joepie91>
promising answer :P
<joepie91>
who's the someone?
hiratara has quit [(Ping timeout: 258 seconds)]
<gchristensen>
I forget their nick :/
hiratara has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<clever>
joepie91: one option, is to build the engine in a single derivation, that has no dependency on the data, then have a second derivation, that depends on a 1gig fetchurl, and throws a bash script around the engine, saying where to find the data
katyucha has quit [(Ping timeout: 240 seconds)]
<clever>
joepie91: so you can then use normal -j to build the engine while downloading
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] bjornfor pushed 1 new commit to master: https://git.io/vFQIr
<NixOS_GitHub>
nixpkgs/master bf338ab Bjørn Forsman: nixos: run parted with --script option (2nd)...
NixOS_GitHub has left #nixos []
Itkovian has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] bjornfor pushed 1 new commit to master: https://git.io/vFQIM
<NixOS_GitHub>
nixpkgs/master aedba5c Bjørn Forsman: nixos: run parted with --script option (3rd)...
NixOS_GitHub has left #nixos []
<joepie91>
clever: right, but I'm thinking more of a nixos-rebuild; say that I want to rebuild because of a security patch, but a new version of the 1GB game also happens to have been released; now the security patch is 'blocked' by the 1GB game since the environment won't be switched until *all* build jobs complete
<joepie91>
clever: even though the security patch doesn't require the new 1GB game (and it can reliably be determined that it doesn't)
<obadz>
joepie91: i usually just comment out the 'game' out of the packages
<obadz>
can probably do without it for a few minutes/hours..
<ekleog>
gchristensen: ok thanks! :D
<joepie91>
right, I'm looking for a more generalizable approach that doesn't require the user to comment it out and remove it from their env :)
<ekleog>
now, trying figure out why since the latest upgrade LVM VGs seem to no longer be auto-enabled on boot
<joepie91>
I guess my real goal is "allow for a temporarily outdated expression if it can be determined that it would be safe to do so"
<obadz>
joepie91: create an env with everything, and an env with the game, and then update them in that order?
<joepie91>
do you have a concrete example of that?
<joepie91>
code-wises
asuryawanshi has joined #nixos
<joepie91>
wise *
katyucha has joined #nixos
acarrico has joined #nixos
<catern>
aaaaaaaaaaargh
<clever>
obadz: also, you can grab the storepath of the game, and run "nix-env -i /nix/store/foo" on that
<obadz>
joepie91: mainEnv = self.buildEnv { name = "main-env"; paths = with pkgs; [ all your packages … ]; ignoreCollisions = true; }
* knedlsepp_
liketechnik: I think you’ll need to override: firefox-unwrapped = super.firefox-unwrapped.overrideAttrs
<catern>
why is the build output in nix 1.12 horribly wrapped and annoying?!?
<catern>
is this because I'm using nix 1.12 with a 1.11 daemon?
<obadz>
joepie91: inject that into pkgs via overrides…
<clever>
joepie91: obadz: the exact (old) version of the game you already have built, is now in the users profile, and you wont loose it after taking it out of systemPackages
myrl has joined #nixos
<obadz>
joepie91: then nix-env -iA mainEnv
<obadz>
joepie91: then nix-env -iA yourGame
<joepie91>
obadz: I'm looking for something that integrates with nixos-rebuild, though, ie. based on a configuration.nix
<obadz>
it's based on configuration.nix but doesn't get upgraded with nixos-rebuild
<ekleog>
oh, that was an issue with thin-pools not being enabled in the kernel. wonder how that worked before
<clever>
joepie91: the issue there, is that nix makes the updates atomic, you either get no changes, or you get everything changing at once
<obadz>
(you probably shouldn't put all your stuff in systemPackages anyway, envs are nicer for that)
asuryawanshi has quit [(Ping timeout: 240 seconds)]
<obadz>
clever: yeah but with nix-env you can use the old game anyway
<clever>
yeah
<obadz>
I put the bare minimum in systemPackages and only update that when I reboot
<obadz>
everything else in an env (aka a profile)
<obadz>
after all if you have /nix/var/nix/profiles/system, you can also have /nix/var/nix/profiles/whatever-else
<ekleog>
there is one issue with envs, that is if I ever upgrade then it's only when trying to access the environment again that I'll notice it needs re-downloads -- which may be an issue if not having internet access
<ekleog>
maybe it'd need something like environments referenced from configuration.nix, for which rebuild jobs are executed after the end of the activation scripts
<ekleog>
(the alternative solution being having a different channel between the envs and the configuration.nix, but that doesn't really solve the issue unless using a different channel per env, and then upgrades become a pain)
Thra11 has joined #nixos
<obadz>
joepie91: essentially just move the game to a profile that gets updated atomically, but separately with everything else
<obadz>
and change your update routine to nixos-rebuild switch; nix-env -p /…/whatever-env -u …
<catern>
aieeee
<obadz>
instead of just nixos-rebuild switch
<catern>
symphorien: you lie! -j0 doesn't work to do what I wanted :(
<obadz>
I don't know a good way to make nixos-rebuild switch not atomic nor do I think that's really desirable…
<joepie91>
obadz: it'll have to do :P would still be interested in a way to work this into nixos-rebuild though
<catern>
does anyone know how to get behavior of nix-env --prebuilt-only, with nix-store --realise?
<symphorien>
I may have misunderstood, sorry
<catern>
it's okay, a good attempt :)
<ekleog>
joepie91: maybe adding a systemd.services.upgradeEnv = {enable = true; script="nix-env -p /…/whatever-env -u …"; unitConfig.Type = "oneshot";}; could do it? (not tested)
Thra11_ has joined #nixos
Thra11 has quit [(Ping timeout: 260 seconds)]
<joepie91>
ekleog: is upgradeEnv a magical service name or is it just something you came up with as essentially an alias?
simukis has quit [(Quit: simukis)]
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
mekeor has joined #nixos
<Thra11_>
If I have a small fast disk (e.g. SSD or eMMC), and lots of slightly slower disk space, is there a good way to configure NixOS so that the current generation is all on the fast disk for speed, but I can take advantage of the larger slower disks to keep several older generations around?
<clever>
Thra11_: nix requires that the entire store exist on the same filesystem
<clever>
Thra11_: but some filesystems like ZFS can use an SSD as a read-cache for another larger disk
<ekleog>
joepie91: just something I came up with
<clever>
the only issue with the read-cache part of ZFS, is that the cache doesnt persist across reboots
<catern>
I really find it hard to believe that there's no way to check to see whether a derivation is substitutable
p4cman has quit [(Quit: Connection closed for inactivity)]
myrl has quit [(Ping timeout: 240 seconds)]
<joepie91>
right, okay :)
<catern>
but that seems to be the case
<ekleog>
it's basically defining an upgradeEnv.service file in systemd (btw, I should have called it like upgrade-env or similar, camelCase is not really encouraged by systemd)
<joepie91>
infinisil: did you ever solve your magnet-links-in-firefox issue?
agjacome has quit [(Quit: leaving)]
globin has joined #nixos
astronavt has quit [(Remote host closed the connection)]
astronavt has joined #nixos
astronavt has quit [(Remote host closed the connection)]
<catern>
argh!
<Thra11_>
clever: Ok, thanks. Maybe I just need to remember to collect garbage more often.
<catern>
is there a Nix store client available in anything other than C++ or Perl?
katyucha has quit [(Ping timeout: 268 seconds)]
Thra11_ has quit [(Quit: IRC for Sailfish 0.9)]
kkini has joined #nixos
<catern>
aaaaaaaaaargh
<catern>
okay no I guess I have to write a perl script
remexre has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
sigmundv has quit [(Ping timeout: 248 seconds)]
<catern>
how does one write a script using the Perl Nix library?
mekeor has quit [(Remote host closed the connection)]
<catern>
does anyone here know Perl and can guide me in this?
ssmike has quit [(Read error: Connection reset by peer)]
kkini has quit [(Ping timeout: 248 seconds)]
<catern>
ha ha oops nevermind looks like even the Perl Nix library doesn't offer an interface to query substitutability, blargh
sigmundv has joined #nixos
benzrf has left #nixos ["WeeChat 1.9.1"]
<catern>
okay whatever I don't need this, I'll just use brute force
remexre has joined #nixos
myrl has joined #nixos
ssmike has joined #nixos
<catern>
wow
<catern>
wooow
<catern>
can I even check whether a store path is valid?!?
<aminechikhaoui>
catern: in a binary cache ?
<catern>
aminechikhaoui: sure?
<aminechikhaoui>
you can use nix path-info in general
<aminechikhaoui>
add --store if you want to specify a different store like a binary cache
roberth has quit [(Ping timeout: 240 seconds)]
<catern>
path-info appears to output nothing when passed a store derivation
asuryawanshi has joined #nixos
ma27 has quit [(Ping timeout: 246 seconds)]
<catern>
in fact it exits with 0 even if it's passed a completely made up store derivation that doesn't exist
<catern>
i guess as long as it ends in .drv, it doesn't care??? ¯\_(ツ)_/¯
ssmike has quit [(Ping timeout: 250 seconds)]
<aminechikhaoui>
huh, indeed :/
<catern>
i supposed that doesn't actually hinder what I am doing, but, kinda weird
<catern>
but yes thanks nix path-info works fine
asuryawanshi has quit [(Ping timeout: 268 seconds)]
<myrl>
Yep
<myrl>
nix-prefetch-url odesn't seem to work but curl does.