<ottidmes>
clever: I dont trust third party password managers. If there is any data that even privacy skeptics (the I have nothing to hide people) agree about should remain private it is passwords, and I dont want to have to trust someone else to make sure it remains private
<ottidmes>
fiatjaf: Hmm, I now read about the option to self host bitwarden, thanks for the tip!
<arianvp>
ottidmes: it's not pretty through. You need Microsoft SQL Server
<pie_>
ottidmes, idk, to avoid temptation i guess
dbmikus_ has joined #nixos
<ottidmes>
pie_: then alias it to give you a scolding when you do?
Rusty1 has quit [Quit: Konversation terminated!]
Mr_Keyser_Soze99 has joined #nixos
<pie_>
haha
<pie_>
alternatively, to ensure that it hasnt been used :P
<pie_>
i guess mounting read only would be a better way to do that :P
<ottidmes>
pie_: environment.systemPackages = [ (hiPrio (pkgs.writeShBin "nix-env" "echo 'How dare you go back to your imperative ways! Declerative configuration FTW!'")) ];
<ottidmes>
pie_: I am actually keeping that on my system :P I only added "Use `nix-shell` instead if you want to test something.", so I am reminded that it often is just as easy to test some package with nix-shell as it is with nix-env
<arianvp>
clever: woohoo PR filed. thanks for the yarn2nix help
Mr_Keyser_Soze99 has quit [Ping timeout: 246 seconds]
jperras has joined #nixos
<pie_>
i messsed up my nix_path, how do i get the default value back?
<pie_>
well, rolling back seems to have fixed it but..
Rusty1 has joined #nixos
Rusty1 has quit [Read error: Connection reset by peer]
oldandwise has joined #nixos
<oldandwise>
i have cloned 18.09. I have an upstream=https://github.com/NixOS/nixpkgs.git, and then I have channels=https://github.com/NixOS/nixpkgs-channels.git. What is `git remote update channels` for? I am already regularly synching my fork's master,release-18.09 to upstream's master,release-18.09 accordingly. What is this channel URL for?
<pie_>
ottidmes, it didnt error but im pretty sure it doesnt do what i think it does
<ottidmes>
pie_: btw, your if should be replaced by the optional function
lassulus_ has joined #nixos
<ottidmes>
pie_: what do you think goes wrong? you mean using a config attribute you set and read in the same file? as long as you do not have recursive definitions this should work fine
<pie_>
yes thats what i mean
<pie_>
weird. why does that work?
<ottidmes>
pie_: lazy evaluation
<ottidmes>
pie_: thats probably incorrect of me, but that at least helps, in this case it is probably just the recursive attribute sets
lassulus has quit [Ping timeout: 240 seconds]
lassulus_ is now known as lassulus
ghasshee has joined #nixos
<oldandwise>
ottidmes: how is nixos-release-18.09 different from release-18.09. Both gets updated regularly
silver_ has quit [Read error: Connection reset by peer]
<oldandwise>
i mean, channel-release-18.09
<oldandwise>
ie, upstream/release-18.09 vs channels/nixos-18.09
<pie_>
ottidmes, ok i changed it to (lib.optional config.virtualisation.docker.enable "docker")
<ottidmes>
pie_: I should go to bed, because I am starting to talk nonsense, attribute sets are not recursive by default in Nix, but it does have to do with how things are evaluated in Nix
<pie_>
ottidmes, good night :p
<ottidmes>
oldandwise: that is a big difference
<ottidmes>
oldandwise: nixos is the latest version, the release is the latest succesful build of that nixos
<ottidmes>
oldandwise: so release-18.09 will always tends to lag behind nixos-18.09, because things take time to build, and that is why you want to base things on release-18.09, since then you will be able to leverage the binary cache better
<oldandwise>
ottidmes: ok. So, it's only a time difference then?!
<ottidmes>
pie_: before I go, in this case it basically comes down to config being a big value, but evaluated on demand (lazy evaluation), and since the value of enable does not depend on anything you can safely use it here
<ottidmes>
pie_: you cannot for example use config to determine what imports there should be, because to build the value of config, you need imports (so infinite recursion)
<pie_>
yeah the infinite recursion makes sense
<ottidmes>
oldandwise: they are pointing to different commits, yes, but its not "only" a time difference, its very significant. You do not generally want to build e.g. chromium on your local machine, which might happen if you dont wait on the build queue to succeed building first
kim0 has joined #nixos
endformationage has joined #nixos
justanotheruser has quit [Ping timeout: 244 seconds]
<ottidmes>
oldandwise: if you were to build everything locally, you might still even want to stick with the nixpkgs-channels, because there is no guarantee things commited to 18.09 will not break
<ottidmes>
and the releases on nixpkgs-channels check for additional things, reducing the chance things break
jasongrossman has joined #nixos
ghasshee has quit [Ping timeout: 244 seconds]
sigmundv has joined #nixos
thc202 has quit [Ping timeout: 272 seconds]
boxscapeR has joined #nixos
<boxscapeR>
if I have two sets, how can I combine them into one?
<boxscapeR>
although I suppose I could solve my problem by using map instead...
<boxscapeR>
maybe
<ottidmes>
boxscapeR: with //
<boxscapeR>
ah, thank you
<ottidmes>
{ foo = 5; } // { bar = 10; }
<boxscapeR>
huh, I looked at the operators section in the manual but somehow missed that
<boxscapeR>
ok
Supersonic has quit [Ping timeout: 252 seconds]
acarrico has joined #nixos
Supersonic has joined #nixos
<{^_^}>
[nixpkgs] @akritimore opened pull request #49083 → Darwin sandbox → https://git.io/fxMPm
<aleph->
With the nginx service, is there a way to configure /etc/nginx/sites-available/enabled?
wpcarro has joined #nixos
<wpcarro>
a few people on this channel have recommended that I use VMs for testing sys-admin and linux things. Does anyone have a good recommendation for VM software that runs on NixOS?
<{^_^}>
wpcarro: 3 days, 10 hours ago <samueldr> [about wifi on resume] with a link to said instructions, many more experienced users could probably help figure out how to port them to nixos :)
jtojnar has joined #nixos
<wpcarro>
I've used Virtual Box in the past and didn't have too many problems with that. Maybe I'll just use that again
<{^_^}>
[nixpkgs] @magnetophon opened pull request #49087 → wolf-shaper: init at 0.1.6 → https://git.io/fxMDX
<aleph->
Wonder if something is broken on the vlc package
<aleph->
Damn chromecast rendering still doesn't work
<pie_>
so i just tried to boot into single user mode so i can copy my root partition, but sulogin complains that the root user is locked
* pie_
pokes clever? ^
<pie_>
uh ok i might have run into this https://github.com/NixOS/nixpkgs/issues/4955#issuecomment-62814720 but still, how are you supposed to get into recovery mode if it needs a login? iirc you used to be able to get into some kind of recovery mode without logging in
<aleph->
elgoosy: Ever figure out the vlc issue with chromecast rendering?
astronavt_ is now known as astronavt
ryantrinkle has joined #nixos
Theuni2 has joined #nixos
dbmikus_ has quit [Ping timeout: 250 seconds]
Theuni2 has quit [Client Quit]
<{^_^}>
[nixpkgs] @Fuuzetsu pushed to master « or-tools: new package (v6.9.1) »: https://git.io/fxM7c
<pie_>
tfw you lock yourself out of the root user by acccident when setting mutable users = false, luckily i still had ssh access
<pie_>
maybe there should/could be some kind of warning for that?
<aleph->
Okay, what fecking combo of ports do I need?
<aleph->
Never had this problem on any other damn distro
hamishmack has joined #nixos
vk3wtf has joined #nixos
<pie_>
ok booting int single/emergency doesnt have / unmounted or mounted as ro
<pie_>
whats the point of even having the root in fstab then
<clever>
stage-1 is responsible for mounting the rootfs
<clever>
how can it read fstab?
<pie_>
clever, yeah ok that makes sense
<clever>
systemd gets upset if / isnt in the fstab :P
<pie_>
shouldnt it just be some dummy entry then
<clever>
line 450 of stage-1 is where it mounts all boot filesystems
<pie_>
(fml i should have just made a separate live sysstem or something)
<clever>
by reading the fsinfo file
<clever>
pie_: this is the exact kind of situation that led to me writing rescue_boot.nix above :P
<pie_>
clever, f****** hell though?
<pie_>
so how do i plug in your rescue boot
<clever>
just add ./rescue_boot.nix to your imports section, rebuild switch, and your done
<pie_>
oh cool, makes sense
<clever>
also, in the pre-systemd days, on linuxfromscratch, i have forced a machine back into single-user mode, and mounted the rootfs read-only, and imaged it
<clever>
without causing any outage of its services, lol
<pie_>
at least nixos makes things easy if you have leet friends
<clever>
yep, lol
<pie_>
jesus christ. idk why but im really worked up about this
<pie_>
---NOTE TO SELF----
<pie_>
clever, how did you do that without a service outage?
<pie_>
or do you mean not for the extent of actually rebooting
mayhewluke has quit [Ping timeout: 252 seconds]
<pie_>
not, except for
<pie_>
nixos is the best thing ever, it should excel at things like this too :P
mayhewluke has joined #nixos
<clever>
pie_: in my case, it was on a router
<clever>
so it still had to perform NAT
<clever>
but thats a purely in-kernel job
<clever>
so you just have to prevent any shutdown type scripts from trying to shutdown NAT
<pie_>
wait
<pie_>
wat xD
<pie_>
oh...you didnt actually reboot
<clever>
correct
* pie_
scratches head 0_0 xD
<clever>
i just did 99% of a shutdown, by hand
<clever>
and then remounted root read-only
<pie_>
i mean i can see how thats possible
<clever>
then i just booted every service back up when i was done
<clever>
pie_: this lists most of the things keeping files open for writing
<clever>
pie_: surprisingly few things, chrome, docker, nscd, it might still be possible
* pie_
waits for system to come back up
<pie_>
clever, what were you thinking might stop it/
<pie_>
inb4 systemd needs rw root access :P
<clever>
pie_: just systemctl stop every service that has files open write
<clever>
i dont think systemd actually has any such files
<CMCDragonkai>
I'm confused as to how to use `package_data` in setup.py when building Nix packages for Python programs
<CMCDragonkai>
I'm trying to put something in a python project rooted at `PROJECT/libexec/foobar`
<pie_>
clever, sure, i just mean why wouldnt it have worked (maybe)
<CMCDragonkai>
And then putting that has `package_data: { '': 'libexec/foobar'}`
<clever>
pie_: i was expecting systemd to resist more :P
<CMCDragonkai>
The `nix-build`, but it the resulting file is not there!
<Lears>
Why do I get infinite recursion when I try to set `services.xserver.videoDrivers = mkIf (config.networking.hostName == "ahostname") [ "nvidia" ];`? Every other setting I've defined like this has been fine.
<clever>
Lears: can you pastebin the backtrace when its ran with --show-trace ?
<pie_>
clever, hm your rescue thing mentions ssh, what does "module" take? is it just like configuration.nix ?
jackdk has quit [Ping timeout: 246 seconds]
Lisanna has joined #nixos
<pie_>
the squashfs image seems really slow to build...
<clever>
yeah, that step is always slow
<clever>
nix somehow disables the progress meter
<clever>
pie_: yeah, module is just the contents of configuration.nix
<clever>
so you can do module = { pkgs, config, ... }: { ... };
<pie_>
ok cool
<pie_>
" 31.13% of uncompressed filesystem size (1606558.86 Kbytes)" thats pretty big
hyper_ch2 has joined #nixos
<pie_>
clever, you do a lot of neat stuff
<pie_>
clever, uhh is $drive1 a grub thing? :P
<clever>
oh, and rescue_boot needs a decent chunk of your /boot
<clever>
yeah
<pie_>
because its an embpty variable for me afaict
<clever>
its a grub variable
<pie_>
complains about not being able to find /rescue-kernel
<clever>
it will be set by grub.cfg
<clever>
oh, it also depends on how your /boot is setup
<pie_>
gotta wait another 20 seconds for the tinc service to fail to start (gotta love how thats blocking boot...)
<pie_>
(might have been something i messed up)
alex``` has joined #nixos
aleph- has quit [Ping timeout: 246 seconds]
Thra11 has joined #nixos
<pie_>
clever, lmao sigh, @bootRoot@ got replaced with ($drive1)/boot
<clever>
Lears: skipping thru that, i can see the following nixos options, system.build, boot.loader.systemd-boot.enable, networking.hostName, nixpkgs.config, services.xserver.videoDrivers,
<pie_>
well the error this time was (hd0,msdos1) not found
<pie_>
which is reasonable since its an ext partition i guess?
<hyper_ch2>
so, everybody's in London now or on the way :(
<Lears>
My original goal was to move all settings other than the hostname from my machine specific configs into my base config module. So that doesn't really do it, but I guess I'll just leave the videoDriver setting in the machine's config.
<clever>
Lears: i make a reverse tree out of my configs
vk3wtf has joined #nixos
<clever>
Lears: so machine1.nix defines stuff unique to that machine, and has imports = [ ./core.nix ];
<clever>
so you start at a root specific to that machine, and it works its way up the tree to core.nix
<Arahael>
Wow, keepassx is *ancient* on nixos, how do I update it?
<Arahael>
It's on version 0.4.4
<Arahael>
Oh, wait - there's keepassxc, which is better.
<Lears>
That's how I had it before clever, but I found it fragile and harder to tweak when I wanted to change how config depended on the machine or other factors.
<pie_>
clever, ok so when the dd finishes and i want to reboot onto the new machine, i have two drives with both partitions of the same uuid, wat do?
<pie_>
if avoidable i dont want to delete the old root partition yet
<pie_>
clever, uhhh...if i did a rebuild switch while in the recovery thing did i just nuke my grub?
Itkovian has joined #nixos
<srhb>
Are the nixcon streams going to be on the nixcon youtube channel? :) soon^tm
emily has quit [Ping timeout: 252 seconds]
Ariakenom has joined #nixos
<samueldr>
srhb: I believe the word is: if everything goes fine it will :)
ThatDocsLady_ has joined #nixos
<samueldr>
(things are being set up)
<srhb>
samueldr: Right, I know the caveats when live streaming, been there done that. :D
<gchristensen>
srhb: I'm still bummed you're not here, though!
<srhb>
Me tooooo D:
<srhb>
But I wish you an awesome conference :)
<srhb>
I'll be leeching the vibe from over here :-P
<clever>
pie_: i dont think the recovery has grub configured, so it shouldnt write to any MBR's
<pie_>
clever, ok
<clever>
pie_: there should also be utils to regenerate the uuid on ext4, google around for that
<pie_>
clever, i guess that makes sense
<clever>
then you can scramble the uuid of the old disk
<pie_>
clever, i feel dumb now, dont i just change the disk boot order?
<clever>
grub and linux will search all drives, and ignore the order
<pie_>
ah well thats *also*an option but it would need an mbr on the other disk
<pie_>
so yeah changing grub is best because im not sure i know my bios password
<pie_>
hm but that assumes leaving the old grub intact. anyway based on the above i think i can figure it out, but bbl
<pie_>
thanks as usual \o/
<clever>
yep
palo_ has quit [Ping timeout: 244 seconds]
jasongrossman has joined #nixos
Tucky has joined #nixos
Theuni2 has quit [Ping timeout: 276 seconds]
lonokhov has joined #nixos
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<gchristensen>
''like this: ''${TEST} should do it''
<symphorien>
,escape''
<{^_^}>
'' two single quotes: ''' bash curly bois: ''${} newline: ''\n tab: ''\t any character x: ''\x ''
schopp0r has joined #nixos
<Enzime>
hey that's cool
<Enzime>
\${} doesn't seem to work though... :\
<Enzime>
wait no it does seem to be working
<Enzime>
:)
<Enzime>
thanks
<schopp0r>
Hi.
Lisanna has joined #nixos
<Lisanna>
kind of a strange question: is it possible to trick nix into thinking that a build is done, by e.g. sending a particular signal to the nix-daemon process running the build?
<ocharles>
Hi all. I'm absolutely baffled by this. All of a sudden, our Hydra server has gone from taking 160s to evaluate a jobset to over 1000
<Lisanna>
ocharles does the evaluation perform any import-from-derivations?
<ocharles>
I have no idea why. All jobsets are doing this, and if I find the last 160s evaluation and compare against the next evaluation for a jobset, there is literally one line of code changed
<ocharles>
yes, but it did before as well and was absolutely fine
<ocharles>
it's just all of a sudden ground to a halt
<ocharles>
I can't share hydra, but let me share some screenshots
wmertens has joined #nixos
<Lisanna>
ocharles "journalctl -fu hydra-evaluator.service" might give some clues
<Lisanna>
but, FWIW, I've had miserable luck with any kind of IFD in hydra evaluations, so much so that I've purged all of them from my evaluations.
<clever>
ocharles: if you manually run this on your release.nix, can you reproduce the slowness?
<ocharles>
you mean locally on or hydra?
<clever>
locally
<ocharles>
I'll try.
<ocharles>
Oh, I ran hydra-eval-jobs earlier, and timestamped everything
<ocharles>
let me paste that
<Lisanna>
I used to use an IFD to pin a specific version of nixpkgs, but that was giving me occasional lockups in hydra eval, so I had to switch it to having nixpkgs as a submodule in my private channel's git repo
<ocharles>
there's the diff, and a 10x evaluation time 30 minutes later
<clever>
ocharles: is the nix code public?
<ocharles>
afraid not
<ocharles>
there's quite a lot of it...
<clever>
ocharles: can you reproduce the issue with nix-instantiate?
<ocharles>
i'm running it now
<ocharles>
let me try something other than githubstatus though
<ocharles>
actually, maybe it is that one that's the problem
<ocharles>
that and shell (which also builds a load of stuff) seem to be the ones where the time is wildly higher
<ocharles>
i'm not entirely sure why but that instantiate right now is building the haskell foundation library rather than just pulling it from my hydra
<ocharles>
(that'll be IFD stuff)
<clever>
ocharles: ocharles oh, i notice shell is part of gitHubStatus
<clever>
to speeding up shell will also speed up gitHubStatus
<ocharles>
githubstatus is all jobs in that jobset
<ocharles>
i'll paste jobs.nix
<clever>
the restarting on line 50 could be part of whats making it slower
<schopp0r>
my god ... nix (and nixos) are nice, but when trying to compile it you notice what a big compostheap this software is. it is *really* badly written :(
<ocharles>
interesting
<ocharles>
thanks srhb, never thought about configuring this
<srhb>
schopp0r: I think criticism like that is best phrased in a more constructive fashion. :)
eskytthe has joined #nixos
ixxie has joined #nixos
<schopp0r>
srhb: dunno what to say constructive about it :( it is a grown piece of software that nobody can fully understand anymore. it's not like other software was different.
<srhb>
schopp0r: My main point was really to avoid calling peoples' work compost. I'm sure you can work it out. :-)
<schopp0r>
srhb: the thing is: compilation is not really documented anywhere, and asking questions about compilation errors is mostly useless (even though I tried it in this new discourse (hail capitalism!) thingy) because nobody seems to compile nix anymore or something.
orivej has joined #nixos
scribbler has joined #nixos
<srhb>
schopp0r: That post from 15 minutes ago?
<schopp0r>
srhb: but ok, constructive criticism: give an option to *not* compile the documentation (I needed to sledgehammer this out of the makefiles)
* srhb
nods
scribbler has quit [Remote host closed the connection]
<srhb>
schopp0r: Is there a reason you're not compiling nix using Nix? Your question doesn't seem to preclude that option.
<schopp0r>
srhb: I cannot use Nix without having Nix.
<ocharles>
clever: still running btw, so that doesn't seem to have been the fix
<schopp0r>
srhb: or do you mean some cross-compiling-stuff?
<ocharles>
clever: I'm just going to rebuild an actual build that has a known duration and see if that still builds in the same duration. I also saw a nix-build go from 6 minutes to 24 minutes, so I'm wondering if it's the machine itself...
<srhb>
schopp0r: I was thinking about something like that, but I'm not too sure how to proceed with that line of thought really.
<schopp0r>
srhb: I mean, I have a computer with nixos. I could probably write a shell with a nix with another store-dir. but how would I get this shell to another computer? nix-copy-closure only works with an already running nix...
<symphorien>
schopp0r: it is enough to tar the whole store
<schopp0r>
symphorien: so … create a shell with overridden configure-flags (--with-store-dir=...), compile it on nixos. then use *this* nix to compile nix in the other store-dir. and then copy this store-dir?
<ocharles>
clever: that hydra-eval-jobs call produced "Oct 25 10:52:39 restarting hydra-eval-jobs after job 'gitHubStatus' because heap size is at 7479033856 bytes"
<ocharles>
so maybe it can't take that option
<__monty__>
schopp0r: Is installing a binary an option?
<Lisanna_>
schopp0r the compilation is documented exactly in the form of its nix derivation. it tells you what all the dependencies are. it should be very straightforward to map from its nix derivation to whatever your local distro's package equivalents are. beyond that, it's just an autotools project.
<symphorien>
copy it overt
<symphorien>
*over
<symphorien>
and then you have a workning nix on the destination
<srhb>
schopp0r: From your errors, I also get the feeling you're not building a maintenance branch..
<betaboon>
schopp0r: about the documentation. how about ./configure --disable-doc-gen ?
<schopp0r>
Lisanna_: the compilation fails though the configure script runs through and all dependencies are met. Compiling the documentation fails because of xmllint-crap. Compiling libexpr fails because of bison-crap (who the hell uses bison when there is boost::spirit). And compiling the rest now fails because of what I pasted minutes ago.
<schopp0r>
srhb: I build from the tarball that can be downloaded.
<schopp0r>
betaboon: I didn't find that option in ./configure --help, must have overread it, will try, thx
patrl1 has joined #nixos
<betaboon>
schopp0r: I'm on master and ./configure --help documents that option (after running ./bootstrap.sh ofc)
<schopp0r>
betaboon: ok, maybe I should try out the git version
<betaboon>
schopp0r: what version are you trying to build anyway ?
<schopp0r>
betaboon: the tarball that can be downloaded from nixos.org. I think 2.1.3
zarel has joined #nixos
<betaboon>
schopp0r: if you dont want to clone using git, you can still download a zip-file from maintenance branches from github
<schopp0r>
betaboon: I wanted a stable version. but now I try master.
<betaboon>
schopp0r: in github just change to a mainenance branch (eg 2.1-maintenance) and the use the download-as-zip button
<schopp0r>
betaboon: cloning is not the problem. I just thought to use the tarball because normally source tarballs are from stable versions.
<schopp0r>
meh :( from the git version with --disable-doc-gen: src/libexpr/parser.y:14.1-5: invalid directive: `%code'
<ixxie>
I have been recently introduced to semantic versioning, and I had this funny thought about what it would be like for Nix to be aware of semantic versions
<sphalerite>
ixxie: shlevy talked about related stuff at the pub yesterday :p
<ixxie>
dang I wish I could be there
<ixxie>
it does go against build purity somehow
<ixxie>
but it would result in more secure builds
silver has joined #nixos
<ben>
"semantic versioning" is the scheme where everybody prefixes their version string with "0.", right? I think I've seen that in the rust community
<gchristensen>
exactly
<ben>
anyone know if there's plans to make opengl work out of the box for nixpkgs-installed apps outside of nixos?
<ixxie>
"MAJOR.MINOR.PATCH - increment the MAJOR version when you make incompatible API changes, MINOR version when you add functionality in a backwards-compatible manner, and PATCH version when you make backwards-compatible bug fixes. Additional labels for pre-release and build metadata are available as extensions to the MAJOR.MINOR.PATCH format."
<gchristensen>
ben's joke is poignant since youf you stick a 0 in the front you can do anything at any time and no semver rules apply
<ixxie>
lol
<gchristensen>
Major version zero (0.y.z) is for initial development. Anything may change at any time. The public API should not be considered stable.
<ixxie>
I see
<ixxie>
didn't catch that
<ben>
Yeah, and the rust community is super hesitant to declare anything "stable", so you get widely-used libraries in perpetual 0.x.y state
iyzsong has quit [Read error: Connection reset by peer]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<schopp0r>
hm. ok, another constructive criticism: as nix is using boost anyway, get rid of the bison-depencency and use boost::spirit instead. and try to get rid of perl.
iyzsong has joined #nixos
<ocharles>
we are trying to get rid of perl
<schopp0r>
ocharles: (y)
<ocharles>
that is an ongoing effort
<schopp0r>
very good.
<alex```>
How to change /bin/sh?
<alex```>
I want set it to dash
<alex```>
lrwxrwxrwx 75 root 25 Oct 8:09 /bin/sh -> /nix/store/v0x81k3riiq2h1wmb327n7l4v5y4acdz-bash-interactive-4.4-p23/bin/sh
strobelight has joined #nixos
<dtz>
that is not likely to work very well
<dtz>
but
<dtz>
`environment.binsh` is the setting you're looking for. But it's likely to cause lots of things to fail if not bash, which is why it's not in the manual I think.
<dtz>
I want to say that if it's statically linked a few things get grumpy, although I think they're easy to fix
<dtz>
and not only should be bash but intentionally pulls in ncurses + readline, lol
<dtz>
:(
<dtz>
so I don't know a lot about dash, maybe it'll work better than I'm saying but if not you should be warned :)
worldofpeace has joined #nixos
<elvishjerricco>
Can compton be restarted without re-logging?
rprije has quit [Remote host closed the connection]
<dtz>
think so--if it's a user service you can just do 'systemctl --user restart compton' or so
<elvishjerricco>
dtz: Thanks, works perfectly
<dtz>
\o/
rprije has joined #nixos
haitlah has quit [Remote host closed the connection]
schopp0r_ has joined #nixos
schopp0r has quit [Ping timeout: 256 seconds]
viric has quit [Remote host closed the connection]
viric has joined #nixos
viric has quit [Remote host closed the connection]
viric has joined #nixos
viric has quit [Remote host closed the connection]
viric has joined #nixos
rprije has quit [Remote host closed the connection]
rprije has joined #nixos
<dtz>
alex``: (see above for some info about your /bin/sh -> dash question)
zarel has joined #nixos
<dtz>
alex``: could be a fun project, but just ... it may be a project :).
jtojnar has joined #nixos
b1000101 has joined #nixos
viric has quit [Remote host closed the connection]
<rycee>
Anybody involved with the streaming here? I tried embedding the live stream into the #Nix Matrix room but it says "Playback on other websites has been disabled by the video owner".
<srhb>
rycee: goibhniu
<goibhniu>
I think YouTube won't allow me to set that option ... but I'll check again soon, thanks!
<rycee>
Great. Thanks. No big worry. Just figured it would be neat to embed the video in the chat room 😃
lonokhov has quit [Quit: WeeChat 2.2]
iyzsong-x has quit [Ping timeout: 252 seconds]
eskytthe has quit [Ping timeout: 252 seconds]
Mr_Keyser_Soze99 has quit [Read error: Connection reset by peer]
Mr_Keyser_Soze99 has joined #nixos
vaibhavsagar has joined #nixos
obadz has quit [Ping timeout: 252 seconds]
shlevy has joined #nixos
<shlevy>
Getting a bunch of "assertion 'buf != NULL' failed" from gdk_pixbuf and then a fatal "Could not load cached PNG image for slide 0: Unrecognized image file format" when trying to run pdfpc. Anyone seen this?
<shlevy>
Looks like some missing support in gdk pixbuf or something?
<jabranham>
How do I request that the R CRAN packages get updated?
obadz has quit [Ping timeout: 252 seconds]
patrl1 has joined #nixos
<Synthetica>
There's probably an update script somewhere
johanot has quit [Ping timeout: 240 seconds]
eskytthe has quit [Ping timeout: 272 seconds]
<Phillemann>
Simple question: I'd like to have a simple boolean to switch two hardware configurations.
<Phillemann>
Is there a way to do an "if mybool hardware.foo.bar = "x";"?
<Myrl-saki>
I wasn't able to watch the last 2 talks until now.
<Phillemann>
Ahhh wait, I probably can use: hardware.foo = if mybool then {} else {...}
<Myrl-saki>
`extend` is pretty similar to `override*`, is there anything that's unrepresentable(and also useful) if we use recursive merge instead of the current merge?
zarel has joined #nixos
vaibhavsagar has quit [Quit: Leaving]
<Phillemann>
Damn, is there "nixfmt" or something I can use as a pretty printer?
sir_guy_carleton has joined #nixos
vaibhavsagar has joined #nixos
<vaibhavsagar>
Phillemann: closest thing I'm aware of is HNix, which does do this
<haitlah>
clever: With the current user, whenever I try to nix-env -i -f default.nix I get error: opening lock file '/nix/var/nix/profiles/default.lock': Permission denied
<va1entin>
Is it possible to use nvidia drivers from unstable in 18.09? I tried using unstable in services.xserver.videoDrivers but it didn't work
endformationage has joined #nixos
<{^_^}>
[nixpkgs] @bjpbakker opened pull request #49115 → Enable GFX in xf86-video-vmware → https://git.io/fxykZ
haitlah has quit [Remote host closed the connection]
betaboon has quit [Quit: WeeChat 2.0]
<exarkun1>
`nixops deploy ...` re-sends and re-starts key services every time even when keys haven't changed. Dependent services are then also restarted as a consequence. It would be nice not to do this if keys are unchanged.
<ocharles>
dependent services are only restarted if you encode the dependency that tightly. requires & after won't cause a restart
<ocharles>
of course deploying a new key won't restart the service either, so i do agree with you
<jophish>
,locate libX1.so
<{^_^}>
Couldn't find in any packages
<jophish>
,locate libX1.so.6
<{^_^}>
Couldn't find in any packages
<jophish>
,locate libXi.so
<{^_^}>
Found in packages: xlibs.libXi
<exarkun1>
ocharles (IRC): Hmmm. I only use `requires` and `after` but I seem to be seeing restarts.
<ocharles>
oh sorry, wants & after
<exarkun1>
But, yes. I don't want no restarts at all. I want minimum required restarts. :)
sigmundv has quit [Remote host closed the connection]
<exarkun1>
I was imagining a nixops change vaguely resembling a check of the deployed keys and the local keys and skipping re-deployment of them (and service restart) if they are unchanged
<ocharles>
I think it would be better to have the service do the check
<ocharles>
basically what happens right now is a service gets started that uses inotify to wait for the file to change, but doesn't filter out no-ops
<exarkun1>
The key service? Or my service that depends on the key service?
<ocharles>
the key service
jperras has quit [Ping timeout: 246 seconds]
genesis has joined #nixos
<exarkun1>
seems plausible I guess
va1entin has quit [Quit: Leaving]
jperras has joined #nixos
Ariakenom has joined #nixos
<{^_^}>
[nixpkgs] @orivej-nixos pushed 2 commits to master: https://git.io/fxyqp
<buckley310>
Do nvidia drivers typically get backported on NixOS? specifically master branch has 410 but release-18.09 is still at 390.
rprije has quit [Read error: Connection reset by peer]
rprije has joined #nixos
johnnyfive has quit [Quit: I go now.]
billsun has joined #nixos
<hyper_ch>
I don't think there are backports on stable - only when it's a security issue
Thra11 has joined #nixos
<buckley310>
ah okay. my friend decided to try out nixos but he has a 2080ti which wont run on 390, so I cobbled together a weird overlay thing that sets up the 410 driver xD
rprije has quit [Remote host closed the connection]
rprije has joined #nixos
<Lisanna>
,locate bin cgexec
<{^_^}>
Found in packages: libcgroup
mmercier has joined #nixos
aleph- has joined #nixos
boxscapeR has joined #nixos
<boxscapeR>
is there an operator or a builtin function for the composition of functions? I can't find one
Gohla has quit [Ping timeout: 240 seconds]
<simpson>
There's probably a library function, I'm guessing in lib/trivial.nix.
<boxscapeR>
thanks, I'll take a look
<simpson>
I can't find it, and your time at this point might be better-spent open-coding it.
<boxscapeR>
you might be right
aleph- has quit [Ping timeout: 245 seconds]
Phillemann has quit [Quit: WeeChat 2.2]
GiGa has joined #nixos
<GiGa>
Hello
<GiGa>
all
<GiGa>
I've got a fresh NixOS build with an Nvidia card. The config has worked on this device before (it's my main machine). Annoyingly, when I boot it can take over 5 minutes before the graphical environment starts. SystemD's logging doesn't show any delays. Sometimes no graphical environment loads at all. Any advice please?
<GiGa>
Happy to get logs, if someone can advise where?
<{^_^}>
[nixpkgs] @Profpatsch opened pull request #49116 → Add dhallPackages and add current Prelude → https://git.io/fxyWn
truh has joined #nixos
<GiGa>
ocharles: I get a 404 for that link
<ocharles>
hm
<ocharles>
weird, it works in chrome for me
<GiGa>
ocharles: No I don't, sorry, Hexchat copied the ': on
<ocharles>
and nix-prefetch-url gave me that sha
<GiGa>
File downloads ok
<GiGa>
In a browser
<ocharles>
does that nix expression work for you?
<GiGa>
can I test it with nix-build and pointing it at a file?
<ocharles>
yep
<GiGa>
yep
<GiGa>
I get a result that's a binary file
<ocharles>
actually, here's the repro:
<ocharles>
nix-instantiate a file with that, which will output a .drv file
<ocharles>
then nix-store -r that .drv file
<ocharles>
it should be /nix/store/di2qv0w02rdwvbcq7nm966dkbh7q153a-busybox-x86_64.drv
<ocharles>
arghh, that's working too. losing my mind!
<truh>
Hallo everyone, I'm currently trying to upgrade to 18.09 (from 18.03), I replaced the nix-channel and rebuilded the system but after a reboot I'm still on 18.03. Anyone got any ideas what could be going on?
<aanderse>
truh: can you list the commands you took to upgrade?
<{^_^}>
[nixpkgs] @alexeymuranov opened pull request #49117 → [18.09] flatpak: 1.0.2 -> 1.0.4 (cherry-picked from master) → https://git.io/fxylf
<aanderse>
anyone: need help developing a nixos test. i want to create a zfs file system and then have it in the machine configuration. i'm not sure how to run code to setup a specific file system so it can be used in a test script machine configuration. any hints?
<GiGa>
ocharles: /nix/store/g9jp0nfqzkqi1h6j4fnv3z02xlwn4mq1-busybox-x86_64 is what I get to "nix-store -r"
<ocharles>
yea, thanks - it must be me doing something weird
<GiGa>
Rusty1: any ideas? I'm thinking "Unable to kill session worker process" happening a lot of the time isn't a good thing
<aanderse>
truh: and if you cat /etc/os-release it says 18.03?
<jabranham>
Can someone confirm that the SHA256SUMS file on https://www.cs.helsinki.fi/u/tmtynkky/nixos-arm/installer/ contains the correct sha256 sum for sd-image-aarch64-linux.img? It fails for me even after downloading that file multiple times.
<truh>
aanderse, yes, it says 18.03, same with nixos-version
<aanderse>
truh: that is odd.. sorry, not sure
Itkovian has joined #nixos
aleph- has joined #nixos
truh_ has joined #nixos
truh has quit [Ping timeout: 264 seconds]
<truh_>
aanderse, mh, I just rebuilded again, nixos-rebuild says that it's building 18.09 but in the bootloader there are only entries for 18.03, maybe the new system got build but updating the bootloader failed?
<aanderse>
but no output with an error eh
<truh_>
aanderse, no errors
<aanderse>
grub?
<truh_>
aanderse, well there was some error about perl complaining about my locale
<aanderse>
can you take a look at the boot entries from your boot loader
<aanderse>
under /boot or whatever
zolk3ri has quit [Quit: Lost terminal]
revtintin has quit [Quit: WeeChat 1.9.1]
truh has joined #nixos
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<aanderse>
truh_: and your system is booting generation 118?
<aanderse>
vaibhavsagar: the release managers briefly mentioned that situation today during their talk. they're working on an RFC on what to do with that, IIRC.
truh_ has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 240 seconds]
<vaibhavsagar>
oh, good to know, thanks aanderse
<vaibhavsagar>
fwiw I try to stay as far away from Stack as possible, but other people use it in a way that makes it my problem :(
drakonis1 has quit [Quit: WeeChat 2.2]
<truh>
now that's a good question aanderse, the Generation number in the boot menu is actually 69... is it possible that I configured one mechanism in the configuration.nix but my machine actually uses something else?
truh_ has joined #nixos
haitlah has joined #nixos
<haitlah>
Hey Hey again ! :D
<vaibhavsagar>
hi there haitlah, did you get to the bottom of your Docker issue?
<truh_>
is the configuration of previous generations backed-up somewhere?
<haitlah>
vaibhavsagar: Well, I completly rewrote the entire dockerfile to instal nix in a non root user and stuff
<haitlah>
I am trying to actually make postgres working with my nix-shell
drakonis_ has joined #nixos
<haitlah>
vaibhavsagar: i feel like nix inside docker is not so great :/
<mdash>
that's a very polite way to put it :)
<vaibhavsagar>
there are other ways of getting what you want :)
<haitlah>
vaibhavsagar: please tell me :)
<haitlah>
mdash: Its more like the docker part that I hate :P
<srhb>
(I think it's a relatively new addition, and definitely non-critical :-P)
<Ericson2314>
dtz: linux-headers makes this `include/asm/..install.cmd` which contains the bootstrap tools reference because I use more `SHELL=` now with `$makeFlags`
markuskowa has quit [Ping timeout: 256 seconds]
<bsima>
is 'nix run' supposed to take the place of 'nix-shell'?
<boxscapeR>
I want to symlink my configuration.nix to a file in a local git repo. Is it inadvisable to make the file writable without sudo? It would make things much easier...
<infinisil>
,locate bin/gcc
<symphorien>
bsima: no nix-shell gives you an environment suitable to build software when nix run only augments the PATH
<{^_^}>
Found in packages: gcc, gcj, gccgo, gfortran, gcc_debug, ccacheWrapper, gcc-unwrapped, distccWrapper, distccMasquerade, gcc-arm-embedded-4_7, mentorToolchains.armEabi, mentorToolchains.armLinuxGnuEabi
<{^_^}>
[nixpkgs] @primeos pushed 7 commits to release-18.03: https://git.io/fxyr3
<bsima>
symphorien: ok thanks
acarrico has quit [Ping timeout: 252 seconds]
<bsima>
boxscapeR: my configuration.nix is really minimal, most of my configs are in ~/config/whatever.nix and then in /etc/nixos/configuration.nix i just import that file
<boxscapeR>
bsima for the most part I actually do the same, except I use home-manager to handle the .config directory. I mostly just want the configuration.nix to also be symlinked for completeness sake
<bsima>
and I keep my ~/config directory under version control
<bsima>
you can put it wherever you want, i just find imports to be more elegant than symlinks
<bsima>
i use home-manager too fwiw
<boxscapeR>
ohh
<boxscapeR>
I think you meant it much more minimally than I thought at first
<boxscapeR>
that's a good idea, I'll do that
siers has joined #nixos
elgoosy has quit [Remote host closed the connection]
strobelight has quit [Ping timeout: 246 seconds]
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jperras has quit [Ping timeout: 245 seconds]
sir_guy_carleton has joined #nixos
Wharncliffe has joined #nixos
truh_ has quit [Remote host closed the connection]
<amitoj>
I'm trying to do nix-copy-closure from a hydra instance to my (macOS) machine, but running into a hash mismatch. Anyone have any pointers?
<{^_^}>
nix#2032 (by volth, 29 weeks ago, closed): nix-copy-closure: hash mismatch should not be a fatal error
<amitoj>
whoops
dbmikus_ has quit [Quit: WeeChat 2.2]
truh has quit [Read error: Connection reset by peer]
Thra11 has quit [Ping timeout: 246 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
<schmittlauch[m]>
dtz: Hey, you did the minor version bump of tor-browser-bundle-bin. Did you plan to also backport that to the current stable release? I just did that but want to get familiar with how and when people backport stuff, and if you forgot that or planned to do it later.
adamantium_ has quit [Ping timeout: 246 seconds]
Peetz0r has quit [Ping timeout: 252 seconds]
haitlah has quit [Ping timeout: 252 seconds]
<Baughn>
When there's a good reason for it.
siers has quit [Ping timeout: 268 seconds]
<Baughn>
Breaking bugs, security holes, etc. Not for new features.
<Baughn>
But "it doesn't build" should be a good enough reason.
GiGa has quit [Quit: Leaving]
<schmittlauch[m]>
"It doesn't build" is alway an issue with -bin packages where old versions are gone from the servers. But also for tor-browser-bundle, I would consider the whole package as especially security sensitive even without searching for security patches.
jperras has joined #nixos
boxscapeR has quit [Remote host closed the connection]
jackdk has joined #nixos
aanderse has quit []
BlessJah_ has joined #nixos
erasmas has quit [Quit: leaving]
jperras has quit [Ping timeout: 252 seconds]
<BlessJah_>
is it possible to replace (override?) deriviation with other deriviation? as if make pkgs.prometheus return what pkgs.prometheus_2 would return?
nckx- is now known as nckx
drakonis_ has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @das-g opened pull request #49126 → minor fixes in Nixpkgs stdenv docs → https://git.io/fxyXo
aanderse has joined #nixos
<jackdk>
sounds like you might want an overlay, but it's hard to tell without more context
<BlessJah_>
I want overlay. However what I actually need is new modules for prometheus 2.x (syntax and commandline flags did change slightly)
jperras has joined #nixos
<jackdk>
are you using nixos or nix in some other os?
__monty__ has quit [Quit: leaving]
<BlessJah_>
nixos, for now I'll simply write own service that takes new flags into account
<BlessJah_>
I'll think about updating module and creating MR tomorrow
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<bsima>
hmm, i wonder if I can do a "docker push" in just nix code? has anyone done this before?
<andi->
how would that work with network access? You can probably do that without the sandbox... not sur eif that is a good idea tho
<bsima>
i have all my deps, my base image, and my deployable images specified in nix already
b1000101 has joined #nixos
<bsima>
right now all i use docker for is "docker load" and "docker push"
carlosdagos has joined #nixos
<andi->
I don't think there are many sensible ways you could use nix to do network I/O to push the derivation.. I'd probably create a script as output that pushes the image to my registry
mbrgm has joined #nixos
b1000101 has quit [Quit: Page closed]
<bsima>
well nix has things like 'fetchgit' that does network IO
<bsima>
you're probably correct though, a small script would be simplest
<alphor>
I'm a rust newbie but I've been using nixos for a while. I want to jump to using nixos on my local machine for rust development. Is there anything I should be familiar with beyond basic cargo abstractions that nix may break?
Peetz0r has quit [Ping timeout: 264 seconds]
jperras has joined #nixos
<Travankor>
can nix work on a musl-libc host system? I tried to install vscode but nix erros out with 'error: unexpected-end-of-file'
<Travankor>
I did not get that error with a glibc host system, although I thought nix would use it's own bundled glibc