<clever>
Dezgeg: heh, can you pastebin it to remind me?, or give a date
<Dezgeg>
I see: 2017-07-31 02:41:51 clever gchristensen: found an oddity in grubs support of xfs
<clever>
hmmm, thats barely a month ago
<clever>
and due to timezones, that was on the 30th here
<clever>
2017-07-30 20:42:14< clever> gchristensen: when /boot is on the same fs as /, grub has to navigate the horrors of /nix/store to find kernels, and spits out many warnings about invalid inodes
<clever>
i believe i was passing on information i heard from another user
<clever>
xfs and zfs have the same issue in grub
freeman42y has joined #nixos
<Dezgeg>
ah ok. I just debugged it and apparently they forgot to handle the case where directories are sparse (due to deleting files from the middle) so the bad inode warnings are harmless I think
gnuhurd has quit [(Remote host closed the connection)]
goibhniu has quit [(Ping timeout: 240 seconds)]
gnuhurd has joined #nixos
LnL has joined #nixos
<clever>
Dezgeg: ahh
freeman42x has quit [(Ping timeout: 255 seconds)]
<clever>
Dezgeg: oh, and i think my netbook has an xfs / without /boot
<clever>
i had used a nixos-in-place variant, and couldnt change the FS to what i prefer
ee has joined #nixos
<clever>
so those inode warnings probably where on my end
<gchristensen>
FRidh: but not like debian where they do it on purpose
<FRidh>
you mean on stable?
<gchristensen>
yeah
<FRidh>
ah, ok
<oscarvarto>
gchristensen, I'm trying Nixos, because it's supposed to be safer when you make changes. I've used Arch Linux in the past, and sometimes I've screwed my OS...
<gchristensen>
oscarvarto: you should find nixos to be quite safe
<dylanjust[m]>
If nix-env only searches nix-pkgs, what's the point of the -f flag?
<clever>
dylanjust[m]: that gives it a different set of nixpkgs to search within
<gchristensen>
well you can nix-env -i <path>
<dylanjust[m]>
So, on the target machine, I need to know the path of what was installed? I have no means to search for a package in my local nix store?
<clever>
it has no way of knowing which one you just copied over and want to use
<dylanjust[m]>
But it should know I have a package called "dynamodb-local" installed.
phinxy has left #nixos ["Leaving"]
<gchristensen>
there is a difference of "installed" and "in the store"
<gchristensen>
it is in the store but it is not "installed" in to your user environment
<dylanjust[m]>
Is there a command for querying the store?
eacameron has joined #nixos
joehh has joined #nixos
<gchristensen>
I think we have a case of the X-Y problem
<dylanjust[m]>
My intuition was that nix-env -qa would search my installed store and channels.
<dylanjust[m]>
And the error doesn't really point out my mistake.
<clever>
it will only search what is available in the nixpkgs on your channels
<dylanjust[m]>
Also, I got the same error when trying to make a nix channel out of this repo.
<dylanjust[m]>
Is there a nice set of instructions somewhere on how to create a channel?
<gchristensen>
I think we have a case of the X-Y problem
<dylanjust[m]>
What's the X-Y problem?
<dylanjust[m]>
Maybe I should dial back to my goals.
<dylanjust[m]>
I wish to build something from source on a CI server in my office. I wish to then run this software on a machine in AWS. The machine in AWS will not have access to my source repositories. I wish to point it to a channel, or ship binaries and source to it - however I need to do it.
eacameron has quit [(Ping timeout: 240 seconds)]
Nobabs27 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] eqyiel opened pull request #28428: awsebcli: init at 3.10.5 (master...awsebcli) https://git.io/v5faF
NixOS_GitHub has left #nixos []
<Nobabs27>
how do I set the resolution? like, I change it with xrandr but that does not save it.
<clever>
Nobabs27: are you using a desktop manager?
<Nobabs27>
clever: lightdm
<gchristensen>
dylanjust[m]: the X-Y problem is basically you're trying to have us solve you problem without you telling us what your problem is, so you ask questions around your problem but don't give us the opportunity to actually address the core of it
<clever>
Nobabs27: the one that comes up after you login
<gchristensen>
dylanjust[m]: you could do a two-step deploy, nix-copy-closure and then create a runner script which points to the path
<gchristensen>
the runner script is in a "known" location that points to the unknown location in the store
<clever>
Nobabs27: lightdm is the display manager, there is also a desktop manager that runs after you login
<Nobabs27>
clever: as in, i3?
<clever>
gchristensen: id say thats going to fail when garbage collection runs and eats things
<clever>
Nobabs27: yeah
<Nobabs27>
mk
<gchristensen>
dylanjust[m]: if this approach works for you, we can address clever's valid point
<clever>
Nobabs27: a quick google says that i3 doesnt support setting the resolution, so you just need to arrange for the xrandr to run every time you login
proteusguy has joined #nixos
<dylanjust[m]>
gchristensen: That might work, thanks. Would you recommend creating a channel for the machine to pull from?
<Nobabs27>
clever: lightdm doesn't start in the right resolution either - there is supposed to be a way to make it run a start scrip
<dylanjust[m]>
It'll be for building images for microservices. One issue is that we'll be within docker containers on AWS instances, so ssh'ing into the container may be tricky. Solvable, but in case we don't like that, what's the easiest way to make a channel?
rpifan has joined #nixos
<gchristensen>
dylanjust[m]: what if you used nix to make a docker image?
<dylanjust[m]>
I don't think any garbage collection issues will be a problem for us, since we're doing immutable deployments.
<gchristensen>
niiice I used to be big on immutable deployments. gave some talks about it.
<clever>
Nobabs27: so you need to figure out if the config belongs under [LightDM] or [Seat:*], then either use extraSeatDefaults or patch the module
<gchristensen>
I stopped after learning Nix
justanotheruser has quit [(Ping timeout: 240 seconds)]
freeman42y has quit [(Read error: Connection reset by peer)]
freeman42y has joined #nixos
freeman42y has quit [(Read error: Connection reset by peer)]
freeman42x has joined #nixos
eacameron has quit [(Remote host closed the connection)]
abcrawf has joined #nixos
<gchristensen>
dylanjust[m]: I strongly encourage immutable deployments fwiw, huge kudos: it shows great rigor and attention to detail. I'd bet your network is better than most.
<dylanjust[m]>
gchristensen: With my replication case, I should point out that it works fine if we don't have the pinned nixpkgs. nix-env -qaP -f default.nix dynamodb-local finds the package we're trying to install. Any ideas why that would work, but it wouldn't with the pinned nixpkgs?
<benkolera>
The docs make those two out to be equivalent (if we haven't missed a dep). We ought to fix those docs if they are not. :)
<clever>
while evaluating the file ‘/tmp/dyno/nixpin.nix’:
<clever>
cannot import ‘/nix/store/v2w72ydhc2sygmaqnzhfw7b04z3i54mv-nixpkgs-channels-25f4906da6387e132823417bc54ea86040fb9bd5-src’, since path ‘/nix/store/pkb1w9v39kz26fm109si5d6hxb699q32-nixpkgs-channels-25f4906da6387e132823417bc54ea86040fb9bd5-src.drv’ is not valid, at /tmp/dyno/nixpin.nix:5:3
<clever>
gchristensen: i ran into a similiar weird bug a week ago, where it claimed things where not valid for no reason at all
<gchristensen>
yeah, what does that mean?
hanslo-W has quit [(Remote host closed the connection)]
<gchristensen>
dylanjust[m]: I'm not sure why you can't SSH to the target ,when nix-copy-closure is already using SSH
proteusguy has joined #nixos
<dylanjust[m]>
I can ssh into the target machine now, for my experiments. I was just saying that this may not be the case when we do final deployments.
<gchristensen>
how do you do immutable deployments?
<gchristensen>
do you boot a machine and deploy to it? or make disk images?
<dylanjust[m]>
At the moment we use packer to spin up an instance, use ansible to tell it to make a single docker image on this host, then we cut an AMI of the whole thing.
<gchristensen>
nice, I've done the same
<benkolera>
Ideally, we probably want a channel in s3 for each of our services so building that image is a simple channel add and an install.
<clever>
dylanjust[m]: nixos already has tools to generate a AMI with nixos pre-installed, including any apps you want
<gchristensen>
5min unplanned down-time over 3 years ... works great :D
<dylanjust[m]>
The idea was to be able to define the project's runtime state with a Dockerfile.
<benkolera>
(If that's a sensible thing to want, that is)
thc202 has quit [(Ping timeout: 248 seconds)]
<gchristensen>
benkolera, dylanjust[m]: do you two work together?
<dylanjust[m]>
Yeah - nix is a new thing for us. We're trying to figure out if it's suitable for us. Obviously it has a lot of capabilities that we can take advantage of. We just need to get around our current stumbling blocks.
<dylanjust[m]>
Yeah
eacameron has joined #nixos
<gchristensen>
may I PM?
<dylanjust[m]>
sure
eacameron has quit [(Ping timeout: 240 seconds)]
LnL has quit [(Quit: exit 1)]
<gchristensen>
publishing a channel could be done, but gets a bit annoying. I'd be inclined to use an approach like nix-copy-closure + a wrapper with the path
<gchristensen>
especially for immutable deploys, not sure what the channel _really_ gets you
LnL has joined #nixos
<Nobabs27>
clever: apparently settings display-setup-script to anything just reks lightdm
<dylanjust[m]>
The main thing I was looking for with the channel, was to be able to pull rather than push.
<clever>
Nobabs27: check the journal for display-manager for more hints
b has quit [(Quit: Lost terminal)]
stepcut has joined #nixos
<clever>
dylanjust[m]: you can pull the same way without a channel, if you have a binary cache setup
<clever>
dylanjust[m]: just nix-store -r /nix/store/foo and it will download foo from a cache
<gchristensen>
dylanjust[m]: the other thing is you can export an archive of a closure. put that up on a webserver, and pull that
<dylanjust[m]>
An archive of the closure + a binary cache = a channel, right?
<clever>
the binary cache is just a collection of smaller archives (not full closures), for every single package
gnuhurd has quit [(Ping timeout: 255 seconds)]
<dylanjust[m]>
Do I need the nix expression I used to build the package in order to access a binary cache?
<gchristensen>
the archive contains, all dependencies, a binary cache has many small archives, one per dependency
<clever>
dylanjust[m]: yes
neeasade has joined #nixos
<dylanjust[m]>
The archive of a closure - that's with nix-store .... > closure
<clever>
dylanjust[m]: or an absolute path into the store, that the nix expressions computed
<dylanjust[m]>
I read about that.
<dylanjust[m]>
It made a 2GB file for me.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/v5fwb
<NixOS_GitHub>
nixpkgs/master 96457d2 Frederik Rietdijk: aspellWithDicts: fixes #28318
NixOS_GitHub has left #nixos []
<gchristensen>
whether its one big file or many smaller files, it'll be ~2g
<gchristensen>
though you may accidentally be including stuff you don't need, like references to the compiler
<dylanjust[m]>
If I keep pushing to a binary cache, it'll be incremental, right?
<gchristensen>
true :)
<clever>
i believe nix-push is incremental
<gchristensen>
so yes, that could work
<dylanjust[m]>
Is there native s3 support on nix-push? Is that a planned feature? Or would I export then "aws s3 sync" or something?
<gchristensen>
a channel is a binary cache + a set of expressions. you can use a binary cache without the expressions with `nix-store -r` if you know the path you're fetching. you'd s3 sync
johnw has quit [(Ping timeout: 255 seconds)]
sellout- has joined #nixos
<dylanjust[m]>
Using a binary cache without the expressions sounds ideal for our immediate purposes, actually.
<gchristensen>
cool, so if you can feed in to the ansible config the path in /nix/store, you could use that
<gchristensen>
or write the path "next to" the binary cache ... like:
<dylanjust[m]>
I think I've found my next experiment.
<dylanjust[m]>
Thanks heaps!!!
<gchristensen>
(to state the obvious, you'd want to fetch that fooapp-latest once and cache it locally so you don't get a mixed view of what "latest" means during a build :) )
<gchristensen>
cheers! I hope you find Nix workable, let us know if you need more help
<Nobabs27>
clever: no useful info, I guess ill just set it via the i3 config
johnw has joined #nixos
<Nobabs27>
unless someone has the magic fix xd
Supersonic112 has quit [(Disconnected by services)]
pxc has joined #nixos
<Nobabs27>
um now I have a new issue: lightdm won't start
<clever>
Nobabs27: does the journal give any errors?
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
<Nobabs27>
clever: an authentication error and a G_IS_OBJECT fail
<clever>
odd
Wizek_ has quit [(Ping timeout: 240 seconds)]
pxc has quit [(Ping timeout: 240 seconds)]
mrkgnao has joined #nixos
mrkgnao has quit [(Client Quit)]
mbrgm has quit [(Ping timeout: 248 seconds)]
mbrgm has joined #nixos
eacameron has quit [(Remote host closed the connection)]
eacameron has joined #nixos
<Nobabs27>
do I have to have lightdm installed for it work being enabled?
<clever>
nixos installs it automatically when enabled
eacameron has quit [(Ping timeout: 240 seconds)]
MarcelineVQ has joined #nixos
<Nobabs27>
thought so - then why is lightdm non-goating rn!?
mrkgnao has joined #nixos
kitten[m] has joined #nixos
mrkgnao has quit [(Client Quit)]
mrkgnao has joined #nixos
brynedwardz has joined #nixos
brynedwards has quit [(Disconnected by services)]
brynedwardz is now known as brynedwards
sellout- has quit [(Quit: Leaving.)]
<Nobabs27>
omg
<Nobabs27>
I t hink I might know what the issue was
<clever>
?
<Nobabs27>
i had the other thing that didnt work still in the config -,-
<clever>
lol
oscarvarto has quit [(Quit: Leaving)]
ison111 has quit [(Quit: WeeChat 1.7.1)]
ona has quit [(Read error: Connection reset by peer)]
ona has joined #nixos
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
proteusguy has quit [(Ping timeout: 246 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/v5fKC
<NixOS_GitHub>
nixpkgs/master 6bbc3a0 Frederik Rietdijk: Merge commit '3b29468313bc8604fe8f85c8d9316fd276d3985c' into HEAD
NixOS_GitHub has left #nixos []
proteusguy has joined #nixos
Nobabs25 has joined #nixos
justanotheruser has quit [(Quit: WeeChat 1.7.1)]
Nobabs27 has quit [(Ping timeout: 240 seconds)]
justanotheruser has joined #nixos
mrkgnao has quit [(Quit: WeeChat 1.7.1)]
<joehh>
How do I put a file into /usr/lib/systemd/system-sleep/ (or nixos equivalent)?
<joehh>
wanting to run a script before sleep/hibernate
<clever>
joehh: its at this path on my system, so you would need to apply a packageOverride to systemd, and then rebuild everything that depends on systemd
<schoppenhauer>
what's a good filesystem for the nix store, when having a rotational disk? I have ext4 and it is often very slow. seems to me like something that is more optimized for many small files could be better?
<joehh>
clever: not ideal, but workable
<joehh>
could I not just create my own package that puts a file there?
<joehh>
as long as the paths don't clash?
<clever>
joehh: nix doesnt allow that
<clever>
joehh: every package gets its own directory, so the files can never clash, and it will never read something from another package
<dylanjust[m]>
shoppenhauer: zfs? It has some nice in-memory caches.
<edef>
joehh: I'm guessing it reads that from etc too?
<edef>
joehh: if so, you can just use environment.etc."systemd/system-sleep/foo".source
<schoppenhauer>
dylanjust[m]: i had bad experience with zfs on linux
<edef>
joehh: you can also just define a service that is wantedBy = [ "sleep.target" ]
<schoppenhauer>
what about btrfs?
* edef
screams
<joehh>
edef: based on man systemd-sleep, it just says the /usr/lib path
<edef>
Note that scripts or binaries dropped in /usr/lib/systemd/system-sleep/ are intended for local use only and should be considered hacks. If applications want to react to system suspend/hibernation
<edef>
and resume, they should rather use the Inhibitor interface[1].
nslqqq has quit [(Quit: Leaving.)]
<joehh>
interesting - mine just says "there are much nicer interfaces available"
<joehh>
I'm trying to disable the lid switch from waking my laptop
<edef>
oh hm
<joehh>
maybe a service is the best way to do it
<edef>
I wonder if there's an ACPI setting for that or something
<joehh>
there is
<edef>
but?
<joehh>
I can run: grep LID.*enabled /proc/acpi/wakeup > /dev/null && echo LID > /proc/acpi/wakeup
<joehh>
to prevent it, but (haven't yet tested) I believe that it gets reset after each resume/boot
<edef>
mhm
<joehh>
so thought best way was to disable either after each resume or before each suspend/hibernate
<clever>
joehh: that sounds like it only gets cleared on reboot
<edef>
so like, systemd-suspend.service is just a unit that runs those things
<edef>
you can make a unit of your own that runs like that, orrr maybe it does look in etc
<joehh>
so then - what is the best way to run that command on boot in a nixos/systemd world?
<Enzime>
what's the best way to add some text to a pam.d/ file
<edef>
also, protip: grep -q
<edef>
instead of redirecting to /dev/null
<joehh>
ahh nice :)
<edef>
you could stuff that thing in the activation script tbh
<joehh>
activation script?
<Enzime>
is there any easy way to str replace on Nix values
<Enzime>
is I feel like it should be usable for allt he different types of auth
<Enzime>
and if not, how do I make it for each specific type of auth
<edef>
mmh
<edef>
the PAM stuff is worryingly inflexible tbqh
<Enzime>
probably because it tries to support so much, but in a hardcoded manner :\
mounty has joined #nixos
mounty has quit [(Ping timeout: 240 seconds)]
ona has quit [(Ping timeout: 240 seconds)]
canndrew has quit [(Quit: WeeChat 1.0.1)]
joehh has quit [(Ping timeout: 248 seconds)]
joehh has joined #nixos
Wizek has quit [(Quit: Connection closed for inactivity)]
reinzelmann has joined #nixos
rpifan has quit [(Remote host closed the connection)]
stepcut has quit [(Remote host closed the connection)]
ShalokShalom has joined #nixos
Nobabs25 has quit [(Quit: Leaving)]
rpifan has joined #nixos
reinzelmann has quit [(Quit: Leaving)]
eacameron has quit [(Remote host closed the connection)]
FRidh has quit [(Ping timeout: 240 seconds)]
MercurialAlchemi has joined #nixos
FRidh has joined #nixos
FRidh has quit [(Read error: Connection reset by peer)]
pxc has joined #nixos
vandenoever has quit [(Ping timeout: 248 seconds)]
<dylanjust[m]>
Is there a simple way to access a binary cache stored on a private s3 bucket? s3 support in nix-env, for instance?
pxc has quit [(Ping timeout: 248 seconds)]
leat has quit [(Ping timeout: 248 seconds)]
hamishmack has quit [(Quit: hamishmack)]
freusque has joined #nixos
justan0theruser has joined #nixos
justan0theruser has quit [(Client Quit)]
justan0theruser has joined #nixos
oida_ has joined #nixos
justanotheruser has quit [(Ping timeout: 240 seconds)]
oida has quit [(Ping timeout: 255 seconds)]
FRidh has joined #nixos
<mog>
does anyone use nixos as a router? i am having trouble dnsmasq is handing out a lease but my device can only talk to the ap not any other address
vandenoever has joined #nixos
proteusguy has quit [(Remote host closed the connection)]
FRidh has quit [(Ping timeout: 255 seconds)]
reinzelmann has joined #nixos
ebzzry_ has quit [(Ping timeout: 255 seconds)]
ebzzry_ has joined #nixos
DrWaste has joined #nixos
<dylanjust[m]>
is dnsmasq handing out a default gateway?
ShalokShalom has quit [(Remote host closed the connection)]
Ivanych has joined #nixos
DrWaste has joined #nixos
freusque has joined #nixos
FRidh has joined #nixos
slyfox has joined #nixos
FRidh has quit [(Ping timeout: 248 seconds)]
ShalokShalom has joined #nixos
phreedom has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dfordivam opened pull request #28430: openfst: init at 1.6.3 (master...master) https://git.io/v5fSv
NixOS_GitHub has left #nixos []
<freeman42x>
Is there any icu-dev or libicu-dev or something on NixOS? Since just installing icu is only giving lib folder without any include folder. http://i.imgur.com/z7ypaL8.png
<freeman42x>
I need icu include ince I need to run stack install like this: `stack install --nix --extra-lib-dirs=/nix/store/k81nxq0j14wdl6yp1imdwyy3fi54k1qz-icu4c-58.2/lib --extra-include-dirs=/nix/store/k81nxq0j14wdl6yp1imdwyy3fi54k1qz-icu4c-58.2/include`
vandenoever has quit [(Ping timeout: 255 seconds)]
hiratara has joined #nixos
FRidh has joined #nixos
leat has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
<viric>
niksnut: do you know the status of nixpgks on cygwin64? Is there any maintainer?
proteusguy has joined #nixos
hamishmack has joined #nixos
nwuensche has joined #nixos
proteusguy has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
<nwuensche>
Hello! So I want to configure my printer with cups. I added the .ppd file manualy, but now it says: /nix/store/h...kmv-cups-progs/lib/cups/filter/brlpdwrapperMFC5440CN not availabe: no such file. I have this file, but I (of course) can't cp it into this folder. So how can I add it? I think services.printing.drivers could help me here, but I don't know how to add a path to the file in the config, as there are
<nwuensche>
just packages allowed
<nwuensche>
nwuensche: Or can I add a absolute path to services.printing.drivers?
eacameron has quit [(Ping timeout: 240 seconds)]
<mog>
thanks Guest7639 i was able to get it working
<viric>
niksnut: ok. :) Last time I checked, master didn't even evaluate stdenv on cygwin64
<viric>
niksnut: (I fixed that and I used that a bit). We use nixpkgs on cygwin64, but we are happy users of a revision one year old :)
<niksnut>
viric: ah cool
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vbgl pushed 1 new commit to master: https://git.io/v5fHp
<NixOS_GitHub>
nixpkgs/master 2a62e93 Vincent Laporte: ocamlPackages.qtest: 2.2 -> 2.6...
NixOS_GitHub has left #nixos []
<__monty__>
I'm looking into setting up nix to do all builds remotely (on another machine on the lan). If I understand correctly using the default build-remote build hook would do this. I'm actually not sure if it would still use the local machine for parts of the build or not. There's still an issue though, it would upload the dependencies from the localhost. I'd like to avoid this since these would be on the
<__monty__>
remote anyway, am I mistaken?
<LnL>
it only uploads what's missing
<__monty__>
Ah, ok. That makes sense.
<__monty__>
What'd happen if you provide more than max_jobs? Would the overflow be built locally?
<__monty__>
Also, a missing package would be missing on both machines in this case. How would it be handled? The local machine builds/substitutes and then sends it to the remote?
<LnL>
indeed that would make every build wait for a slot on the remote machiens
joehh has quit [(Ping timeout: 240 seconds)]
hiberno has joined #nixos
goibhniu has joined #nixos
proteusguy has joined #nixos
<__monty__>
So that's a yes on the local project as well? Sounds like a really easy way to set up a local build server. Thanks, definitely going to try setting that up today.
thc202 has joined #nixos
_ts_ has joined #nixos
_ts_ is now known as ts468
bkchr has joined #nixos
takle has joined #nixos
jellowj_ has joined #nixos
gillmanash has joined #nixos
jellowj_ has quit [(Ping timeout: 255 seconds)]
ertes has quit [(Ping timeout: 240 seconds)]
gillmanash has quit [(Client Quit)]
reinhardt has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
reinzelmann has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 2 new commits to master: https://git.io/v5fFF
<NixOS_GitHub>
nix/master 2c75945 Matthew Bauer: Remove nix-mode.el from Nix....
<makefu>
funny, where is nixos.wiki coming from? too bad it has not been announced yet
* goibhniu
also didn't know there was another wiki
<makefu>
me neither
nwuensche has quit [(Remote host closed the connection)]
nwuensche has joined #nixos
roundhouse has joined #nixos
jellowj_ has quit [(Ping timeout: 248 seconds)]
xd1le has joined #nixos
<makefu>
i try to get in contact with user fadenb to join efforts
<nwuensche>
Hi! How can I set chmod +x for a file inside of a nix package I created?
<makefu>
nwuensche: override the postInstall phase of the upstream package which generates the file
betaboon has joined #nixos
betaboon has quit [(Changing host)]
betaboon has joined #nixos
Lisanna has quit [(Ping timeout: 248 seconds)]
<nwuensche>
makefu: Thank you! Why does nix don't allow me to do this in the installPhase?
Arcaelyx has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<makefu>
ah you can of course also do this in the installPhase if you write the derivation yourself
<nwuensche>
Still, I have the problem that my cups filter doesn't have the right permissions. How can I give it all permissions? chmod 0777 filter/brlpdwrapperMFC5440CN?
<makefu>
you can chmod $out/filter/brlpdwrapperMFC5440CN after you've copied it over
<makefu>
it is called postInstall, not postInstallPhase :)
<makefu>
but better put the chmod into the installPhase
<nwuensche>
And in my configuration.nix: printing.drivers = [ ((pkgs.callPackage /home/nwuensche/printer2.nix) { })];
<nwuensche>
makefu: Dammit! :(
<nwuensche>
makefu: Dammit! :)
jellowj_ has joined #nixos
<nwuensche>
makefu: I will try it out. Thanks for finding the typo.
<makefu>
or use 'install -m755 $srcPPD <dst>'
<nwuensche>
Now it says while rebuild: chmod: changing permissions of '/.../lib/cups/filter/brlpdwrapperMFC5440CN': Operation not permitted
nextloop has joined #nixos
eacameron has joined #nixos
<makefu>
nwuensche: the filter is a symlink, so you will need to chmod the origin
<makefu>
ohh i see
<makefu>
instead of 'ln -s $srcLPD' use 'install -m 755'
eacameron has quit [(Ping timeout: 240 seconds)]
<nwuensche>
makefu: Now it's -r-xr-xr-x. Should be right, isn't it?
<makefu>
yep
<makefu>
brb
magnetophon has joined #nixos
<nwuensche>
makefu: Thanks, but now the next problem occurs: Job stopped due to filter errors; please consult the error_log file for details. Where is this error_log?
jellowj_ has quit [(Ping timeout: 240 seconds)]
jellowj_ has joined #nixos
<nextloop>
hello! is there any existing configuration for a full firewall/router available, so i can get startet easily? i'll need dhcp openvpn also.
helpusobi has joined #nixos
seppellll has quit [(Remote host closed the connection)]
seppellll has joined #nixos
mpcsh has quit [(Quit: THE NUMERICONS! THEY'RE ATTACKING!)]
mpcsh has joined #nixos
Neo-- has joined #nixos
pxc has joined #nixos
pxc has quit [(Ping timeout: 255 seconds)]
erictapen has joined #nixos
nico202 has joined #nixos
ShalokShalom has quit [(Ping timeout: 255 seconds)]
eacameron has joined #nixos
<nwuensche>
makefu: And again, I get almost the same errors: ixos cupsd[6201]: execv failed: Exec format error
<nwuensche>
nixos cupsd[6201]: PID 6396 (.../lib/cups/filter/brlpdwrapperMFC5440CN) stopped with status 108 (Exec format error)
eacameron has quit [(Ping timeout: 240 seconds)]
jellowj_ has quit [(Quit: WeeChat 1.7.1)]
jellowj has joined #nixos
Capprentice has joined #nixos
bennofs has joined #nixos
<krav_>
what's the correct way of getting usr/share of a derivation? better than globbing /nix/store/*package*/
<clever>
krav_: which derivation?
<clever>
nextloop: i have nixos on my router, let me grab an old copy of its config
<nextloop>
i am getting a bit tired with pfSense...
<clever>
nextloop: the dns and dhcp can possibly be configured for HA, but the router part cant really do that
jtojnar has quit [(Ping timeout: 248 seconds)]
<clever>
any form of NAT results in state held in ram, and loosing that state breaks every single connection
<nextloop>
yeah, so you have to sync state ofc.
<clever>
and even if you didnt have NAT, most devices only accept 1 gateway, so it has no way to fallback
joehh has joined #nixos
<nextloop>
also true, that's why there is CARP
<clever>
ah, that would work, but i havent seen anything about support for it in nixos
<clever>
but the conntrack module doesnt allow inserting state, so no way to sync them
<clever>
it would only work if your ISP gives you several publicly routable IP's and you just dont NAT
ebzzry_ has joined #nixos
<nextloop>
i don't do nat, at least not really. just some port forwarding...
<nwuensche>
clever: Do you know what's going wrong? I try to install a cups driver for my printer, but right now it just says: nixos cupsd[6201]: execv failed: Exec format error nixos cupsd[6201]: PID 6396 (.../lib/cups/filter/brlpdwrapperMFC5440CN) stopped with status 108 (Exec format error)
<nwuensche>
Where can I find the error log from cups after I set the LogLevel to Debug?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 4 new commits to master: https://git.io/v5JJk
<NixOS_GitHub>
nix/master 9400cb3 Eelco Dolstra: Disallow accidental copy construction
<NixOS_GitHub>
nix/master 4c6a265 Eelco Dolstra: Remove debug line
<NixOS_GitHub>
nix/master 4af2611 Eelco Dolstra: Allow builders to create activities...
NixOS_GitHub has left #nixos []
<clever>
nwuensche: try "ls -l /proc/<pid>/fd/" with the pid of cups
<clever>
nwuensche: that will show every file it currently has open
<clever>
nextloop: what IP's are inside the lan?
<nextloop>
it's withing the 10.0.0.0/8 block
<nextloop>
my provider does the NAT
<clever>
ah, thats a bit weird
jtojnar has joined #nixos
<clever>
nextloop: at that point, do you even need a router?, just link the modem directly to the systems with a swich?
<clever>
nextloop: or maybe 1 (or more) systems configured in bridge mode, with SPAN enabled, so there is no gateway configured
<clever>
the dhcp/dns doesnt have to be on the same machine, and can just be a normal system on the LAN
<nextloop>
i need a firewall
<krav_>
clever: any? the latest, i just want to access the doc folder
<clever>
nextloop: something configured in bridge mode could do the same job and firewall, though the example i linked may not work as well there
<nextloop>
i can't really trust WAN (10.0.0.0/8 is still quite big), i need to block incoming v6...
<clever>
krav_: which derivation?
<clever>
krav_: and from what context? a shell? another nix expression?
<krav_>
from the shell, just now i wanted to get the example config for polybar, which is in there, but not linked anywhere afaik
<clever>
nextloop: yeah, my ISP's router is pretty fishy as well, in more hidden ways
<clever>
krav_: one way, "nix-build '<nixpkgs>' -A polybar -o polybar"
<clever>
krav_: that would create a symlink called polybar in the current dir, pointing to the polybar path
<nextloop>
clever: and the uplink gateway would be in another ip block (192.168) than my private network
<nextloop>
most OSes don't like that
<clever>
nextloop: my ISP has 2 uplinks over the same cable using 802.1q, one uplink gives a single public ip
<clever>
nextloop: the other uplink gives an ip in the 10.0.0.0/8 range, and has multicast service
<krav_>
clever: cool, i didn't know that!
<clever>
nextloop: but, there is also a 3rd uplink, that is BRIDGED into the private lan, effectively bypassing the firewall
<nextloop>
clever: i also have a public v4 addressblock. it's just another route...
<nextloop>
it's a campus network. best ISP i could imagine ;-)
<clever>
ah
<clever>
nextloop: how major is the HA setup?, my power goes out more often then the nixos router crashing
<nextloop>
i just thought nixos would be a great thing for a professional firewall/router distribution, appearently the technology isn't quite there yet...
<nextloop>
i use pfsense, but having everything in one file, and beeing able to manage it using git is a nice idea...
sauyon has quit [(Ping timeout: 248 seconds)]
<nextloop>
so far no HA setup, in my configuration HA is not trivial.... i need real IPs in another block as the CARP virtual IPs and so on...
<nextloop>
it's really great for doing router updates
<clever>
i generally just dont upgrade it until it fails hard, lol
<nextloop>
;-)
<clever>
no extra downtime then
<nwuensche>
clever: I could only find cupsd. There it just gives /dev/null, some sockets and pipes and anon_inode:[eventpoll]. What does this says?
<clever>
nwuensche: ls -lh /proc/*/fd/* | grep 211478
Wizek_ has joined #nixos
<clever>
nwuensche: this may find the other end of the socket
Wizek has joined #nixos
<clever>
nwuensche: netstat and lsof may also work better
Neo-- has quit [(Ping timeout: 255 seconds)]
<nwuensche>
clever: I just get nosuch file or dir errors. How can I use netstat/lsof?
<clever>
lsof -p <pid>
<clever>
netstat -anp | grep <pid>
simukis_ has joined #nixos
<srhb>
nwuensche: Are you still debugging the "exec format error" thing?
<srhb>
Just out of curiosity.
<nwuensche>
srhb: Yup
<srhb>
nwuensche: Are you aware that those are 32 bit executables, not 64 bit ones?
<srhb>
Not sure if that's the issue at hand, but it seems surprising.
<nwuensche>
clever: lsof gives me a lot of files, but nothing like '*.log'. netstat gives me more sockets, all of which don't show anything when I ls them
<nwuensche>
srhb: No, I didn't noticed that. Which executables do you mean? And can I fix this?
<srhb>
nwuensche: I actually meant everything in $out/usr/bin and $out/usr/lib
<srhb>
I'm sure it can be fixed, not sure how yet though. :-)
erictapen has quit [(Ping timeout: 240 seconds)]
<srhb>
For instance, I think you need a 32 bit glibc
<bennofs>
nwuensche: do you know the what file it tries to exec? can you try running it manually perhaps?
<nwuensche>
srhb: However, it still would be so nice to find the log from cups for that...
<srhb>
nwuensche: Yeah, absolutely, haven't looked at that. :)
<srhb>
nwuensche: Just wanted to give a heads-up -- that's probably at least ONE issue.
<bennofs>
nwuensche: last time i dealt with cups, i believe i found the logs in journalctl
ShalokShalom has joined #nixos
Ivanych has joined #nixos
erictapen has joined #nixos
<nwuensche>
bennofs: I can just find: Job stopped due to filter errors; please consult the error_log file for details. in journalctl
kiloreux has joined #nixos
jellowj has quit [(Ping timeout: 248 seconds)]
iyzsong has joined #nixos
jellowj has joined #nixos
<nwuensche>
So noone has a clue what I could do now?
<nwuensche>
Is it possible to create a VM with Ubuntu and run the printer from there? Or do I need to set up the printer in NixOS too to make this work?
<srhb>
You could pass through USB or whatever to a VM, but I don't think that's super trivial either, though I haven't tried recently.
<clever>
nwuensche: hmmm, i think you need the drivers working even if your sharing via another cups?
<clever>
nwuensche: the last printer i used has google cloud print sharing built right into it, i linked it to my google account
<clever>
nwuensche: now i can print directly from chrome, even if the printer has left the house
<srhb>
This model is rather old.
hotfuzz_ has quit [(Ping timeout: 240 seconds)]
goibhniu has quit [(Ping timeout: 248 seconds)]
<nwuensche>
clever: Jup, it's really old. However, it's still does what it should, except this driver...
jellowj has quit [(Ping timeout: 240 seconds)]
<srhb>
It would be really nice if we could find that error_log. Did you try just searching for that file in your file system?
<srhb>
If we could see what it was trying to execute, that would make things somewhat easier.
<srhb>
If it's just brprintconfij2 then I'm fairly sure that it's just the 32 bit thing.
<srhb>
You could simply try fixing that up.
<srhb>
Something like patchelf --set-interpreter ${i686_glibc}/lib/ld-linux.so.2 $out/usr/bin/brprintconfij2
<nwuensche>
So I did sudo find / -name "*error_log*", and it found nothing!
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gebner pushed 2 new commits to master: https://git.io/v5JZJ
<NixOS_GitHub>
nixpkgs/master c19d06b Andrew Cann: microcode-intel: 2017-05-11 -> 2017-07-07
<NixOS_GitHub>
nixpkgs/master 0c36f85 Gabriel Ebner: Merge pull request #28420 from canndrew/intel-microcode...
NixOS_GitHub has left #nixos []
<bennofs>
nwuensche: try cupsctl LogLevel=debug2
ntinos has joined #nixos
<nwuensche>
bennofs: I executed this in the cmd and tried to print a testpage again. Where is the extra Log? In Journalctl?
eacameron has joined #nixos
Tucky has joined #nixos
<ts468>
rycee: would you be interested in joining #nixos-dev as I'm just asking about NixUP there? As you're responsible for Home Manager, it might be good to have you in the discussion as well!
FrozenCow has joined #nixos
<disasm>
why would this command recompile systemd? Is that because there's a commit on master that hasn't completed building in hydra? nix-shell -I nixpkgs=/home/sam/nixpkgs -p atom
eacameron has quit [(Ping timeout: 248 seconds)]
<ben>
thats generally the reason for rebuilds when youre living on master
<srhb>
nwuensche: Yeah I would expect journalctl -u cups
<mbrock>
disasm: you might like to keep another checkout of nixpkgs synced with the nixpkgs-unstable branch, which usually points to a somewhat recent commit that has been almost completely built on Hydra (I'm not sure exactly the conditions)
<nwuensche>
srhb: This shows me nothing. I used journalctl --since=today | grep -i cups , which shows a bit more, but there is still no difference.
aloiscochard has joined #nixos
<srhb>
nwuensche: Meh. :/
<mbrock>
disasm: although I dunno why atom would depend on any systemd components? maybe something about udev or something
<disasm>
probly the gvfs dependency
goibhniu has joined #nixos
<mbrock>
if you really want the atom version in master, but not other stuff from master, you can do stuff like cherrypicking the atom update commit onto nixpkgs-unstable, etc etc
takle has quit [(Ping timeout: 255 seconds)]
simukis_ has quit [(Remote host closed the connection)]
<disasm>
mbrock: yeah, I just stashed my change and switched to channels/nixos-unstable and did a pop, thanks!
<disasm>
I didn't think of testing it from unstable channel branch :)
leat has joined #nixos
arianvp2 has joined #nixos
Ivanych has quit [(Ping timeout: 240 seconds)]
<nwuensche>
So are there still some ideas left? Or do we surrender?
mguex has quit [(Ping timeout: 255 seconds)]
<bennofs>
nwuensche: hmm, have you tried restarting the cups service after cupsctl LogLevel=debug ?
eacameron has joined #nixos
<bennofs>
sorry, i don't have my NixOS machine right now so i cannot check what i used when I debuged a print driver
mguex has joined #nixos
arianvp2 has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] disassembler opened pull request #28438: fixes #28436 - atom git integration (master...master) https://git.io/v5JcD
NixOS_GitHub has left #nixos []
<astsmtl>
nwuensche, which driver you configured for your printer? I'm pretty sure hplip is broken for some time.
arianvp2 has joined #nixos
eacameron has quit [(Ping timeout: 255 seconds)]
mguex has quit [(Client Quit)]
<srhb>
astsmtl: They are trying to package a new (oldish) Brother driver.
<nwuensche>
astsmtl: I try to make my own out of the brother .deb package
<NixOS_GitHub>
[nixpkgs] domenkozar pushed 1 new commit to master: https://git.io/v5JCC
<NixOS_GitHub>
nixpkgs/master 37c3407 Domen Kožar: servant-auth: tests pass for a while now
NixOS_GitHub has left #nixos []
eacameron has quit [(Ping timeout: 248 seconds)]
jellowj has quit [(Ping timeout: 248 seconds)]
arianvp2 has joined #nixos
takle has joined #nixos
<gchristensen>
goodwill: ping
joehh has joined #nixos
ixxie has joined #nixos
Rajigo has joined #nixos
arianvp2 has quit [(Client Quit)]
<Rajigo>
Hi, I am using Nix (alone, without NixOS for the time being). I set up my .config/nixpkgs/config.nix as follows: https://pastebin.com/GDqamqaD to override the default gcc definition. The syntax is what I have coherently found in both the manuals and on blog posts. Yet I get the following error: https://pastebin.com/ccX0RJcX
jophish has quit [(Ping timeout: 240 seconds)]
<Rajigo>
(function called with unexpected argument)
<srhb>
Rajigo: Otherwise you probably wanty something like wrapCC (gcc7.cc.override { langFortran = true };)
<nwuensche>
That's what I think could be a problem too. There are paths to /usr/bin, which of course don't work. However, I don't really know how to replace them inside the script without knowing the hash of the cups package
<srhb>
nwuensche: You can use any tool you're comfortable with, for instance substituteInPlace or sed or something like that
eacameron has quit [(Ping timeout: 248 seconds)]
<Rajigo>
srhb: well, yes I need it though I know there is a package for that... My purpose is really to learn how everything works (and changing attributes is really something I need). I am going to try with wrapCC
<srhb>
nwuensche: The cups package is simply ${cups}
<srhb>
nwuensche: That kind of substitution, Nix will handle for you
<srhb>
nwuensche: You can refer to any package that you are passing in as an argument and ${pkgname} will point to its store path
<nwuensche>
srhb: Yes, but this is a cups-filter, not a .nix file
<srhb>
Rajigo: The important thing to note is really the gcc7.cc.override part, but you'll need the wrapper regardless, I think.
<srhb>
nwuensche: In the nix derivation where you install the cups filter, you modify that file.
<srhb>
nwuensche: So for instance if you wanted to replace /usr/bin/cups with the path to /nix/store/.../bin/cups you would do something like substituteInPlace --replace /usr/bin/cups "${cups}/bin/cups"
bfrog has quit [(Quit: WeeChat 1.9)]
<srhb>
nwuensche: Er, and point it at the file in $out where you want to make the replacement
<nwuensche>
srhb: Oh, ok, cool :)
<srhb>
nwuensche: So, substituteInPlace $out/file --replace "before" "after"
<srhb>
nwuensche: And then use the nix substitution mechanic to get at the nix store locations. I hope that was clearer. :-)
<Rajigo>
srhb: neat! Thanks a lot. In that case, is wrapCC applied to my override or is it pattern matching? (Sorry, I am still learning)
<srhb>
Rajigo: wrapCC is just a function that is applied to (gcc7.cc [...])
<srhb>
Rajigo: You can read about it in build-support/cc-wrapper/default.nix
<Rajigo>
srhb: okay thanks! :)
<Rajigo>
(that's precisely what I am reading now)
<srhb>
:)
<srhb>
Rajigo: You might also be interested in reading the definition of eg gfortran7 in top-level/all-packages.nix
FrozenCow has quit [(Remote host closed the connection)]
<m0rphism>
Hi, when I start chromium from the terminal no window appears until I hit ctrl-c. I've checked out the chromium wrapper script in which a comment says "libredirect causes chromium to deadlock on startup". Is there any way to prevent this?
sellout- has quit [(Read error: Connection reset by peer)]
hke has joined #nixos
<copumpkin>
oh, I guess we're gettng 503s from the avatar site
<disasm>
just curious if anyone else is having issues with github right now? Just hangs on connecting 90% of the time. Everywhere else I've tried is fine.
sellout- has joined #nixos
<m0rphism>
disasm: yep, same here
<disasm>
ok, going to stop troubleshooting my local network then :)
<nwuensche>
srhb: I have a new problem. It now says: chmod: changing permissions of '/nix/store/lyhx71g1cwi1vvj5bmm25fi2wby1dj1q-brlpdwrapperMFC5440CN': Operation not permitted, although I don't have chmod in my package. Just install, but this was no problem before:wq
<disasm>
nwuensche: the nix store is immutable, a package shouldn't try to change the mode of a file after it's already copied to the store.
reinhardt_ has joined #nixos
<srhb>
nwuensche: Can you regist your derivation?
<disasm>
if gist works :)
<srhb>
nwuensche: It sounds like you may be trying to use chmod on a symlink to another derivation in the store.
joachifm has quit [(Quit: Lost terminal)]
<srhb>
Er, another path.
joachifm has joined #nixos
leat has quit [(Ping timeout: 248 seconds)]
<srhb>
I'm starting to use derivation as a magical word that means "anything which I'm not quite sure what is formally named" -- sorry :-P
<nwuensche>
srhb: Sorry, I tried to substitude a fetchedURL directly. That seemed to be the problem. Solved it now by installing first and substituding second
<srhb>
nwuensche: OK, good that you figured it out :)
Infinisil has quit [(Ping timeout: 255 seconds)]
reinhardt has quit [(Ping timeout: 240 seconds)]
<nwuensche>
srhb: Is it normal that a script, which is human readable, is not human readable after I copied it into a package a used it there?
Shados has quit [(Quit: No Ping reply in 180 seconds.)]
<nwuensche>
srhb: Or a better question, what is the correct way to fetch something on my HDD? I used fetchurl with a local url
<clever>
nwuensche: src = ./foo.txt;
<xd1le>
Hi all, I've run into a bit of a situation. Basically, running `pkg-config --cflags libevdev` in buildPhase works, it points to /nix/store/...-libevdev... . But running it in preBuild makes it point to /var/empty instead (which is empty). So it this expected (and if so how do I remedy this), or am I possibly doing something wrong?
Shados has joined #nixos
Infinisil has joined #nixos
<xd1le>
*is this expected
<nwuensche>
clever: Where is foo then? In the same dir as the *.nix file?
Shados is now known as Guest2883
<clever>
nwuensche: yeah
<xd1le>
clever: does that work for directories?
<clever>
xd1le: yes
<xd1le>
clever: thanks
<clever>
i use src = ./.; a lot
<clever>
that just pulls in the entire current directory
<xd1le>
clever: also my question above, I believe you helped me with libevdev before!
<xd1le>
cause iirc you're the one who suggested to use pkg-config :)
<clever>
xd1le: cant think of anything obvious as to why pkg-config wouldnt find it when ran earlier like that
<xd1le>
clever: libevdev headers couldn't be found because libevdev puts it's headers in a non-standard place, if you remember
<xd1le>
clever: or are you saying you don't know why it wouldn't work in preBuild ?
<clever>
xd1le: the 2nd
<xd1le>
clever: cool, thanks :/
<xd1le>
clever: man I've been looking for src = ./.; for so long. Can't believe it was so simple... :|
<xd1le>
just been using fetchgit with a local path
nixo202 has joined #nixos
nico202 has quit [(Read error: Connection reset by peer)]
MercurialAlchemi has quit [(Ping timeout: 255 seconds)]
<disasm>
https://github.com/NixOS/nixpkgs/issues/28387 -- would the correct way to handle that until the next release of polybar be to just apply the patch from the commit that fixes the issue?
<nwuensche>
clever: Atleast the file is now readable, however, I still get the same error_log message in journalctl...
Neo-- has quit [(Ping timeout: 240 seconds)]
<xd1le>
disasm: just override polybar's src attribute to one that points to a more recent git rev
vandenoever has quit [(Ping timeout: 240 seconds)]
<xd1le>
clever: it seems like nix-build reuses the previous build stuff.. is there a way to nuke everything, and make nix-build build from a fresh src directory?
takle has quit [(Ping timeout: 240 seconds)]
<xd1le>
actually, that's a general question, so it didn't need to be specifically aimed at you, sorry
gnuhurd has quit [(Ping timeout: 240 seconds)]
<disasm>
xd1le: wouldn't that change the stability of the package though? If you use a newer package that isn't a release, you'd need to change the version to unstable.
<srhb>
nwuensche: Do you think you can figure out how cups is calling the script and try to do the same thing manually?
<xd1le>
disasm: you're correct, I just meant overriding it in your personal config.
<nwuensche>
srhb: I don't really think so. I don't know what cups does internally
<disasm>
xd1le: yeah, I'm not using the package, just going through issue queue trying to knock out some easy fixes.
Ivanych has joined #nixos
<nwuensche>
srhb: When I tcsh brlpdwrapperMFC5440CN, I get: "PPD: Undefined variable."
<xd1le>
disasm: yeah that's what I just figured, can't really think of anything elegant except for maybe asking upstream for a release.. but that's probably not ideal either. :/
reinhardt_ has quit [(Quit: Leaving)]
<xd1le>
as in, not ideal from a packagers perspective
<xd1le>
(i.e. what if upstream does not want to release, which is 100% fair for them)
jellowj has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
<disasm>
I'm going to apply the patch and push a PR and see if anyone complains. Patch removes one #include line :)
<xd1le>
disasm: yeah, in thise case I think this can be a fair exception :)
<xd1le>
Anyway, so the way I'm figuring out that nix-build is using stuff from previous builds, is just by running cat on a src file in preBuild. It's /different/ from the actual source code version! How can I force nix-build to start building from fresh source code??
FRidh has quit [(Quit: Konversation terminated!)]
<disasm>
nwuensche: did you try systemctl status -l -n 50 cups.service?
<xd1le>
(the src attribute is fetchurl'ing an online tarball, not some local directory or anything)
<disasm>
xd1le: change the hash to something else, that'll cause it to fail, download, hash will mismatch, and give you the correct new hash to use.
<xd1le>
disasm: wait but I already have the correct hash?
<disasm>
xd1le: you said the file changed, upstream, right?
Ivanych has quit [(Ping timeout: 248 seconds)]
<disasm>
xd1le: oh, you're using a local src
<gchristensen>
copumpkin: where are you staying for nixcon?
felipedvorak has quit [(Ping timeout: 248 seconds)]
<nwuensche>
disasm: I found " HTTP_STATE_WAITING Closing for error 32 (Broken pipe)" But what does this tell me?
<xd1le>
disasm: no, it's the same tarball, I haven't changed the src attribute. But nix-build is not using the same source code as the same tarball, probably as a side effect of previous failed builds
thblt has quit [(Remote host closed the connection)]
<xd1le>
disasm: so I'm not using a local src
<xd1le>
*as the tarball
<gchristensen>
Profpatsch: are you near where nixcon will be?
<mrkgnao>
how do I get vim to find zsh? a plugin I'm using appears to check /bin/zsh.
<mrkgnao>
can I patch the paths somehow by using "nix-shell -f SOME_MAGIC_OPTS --run nvim" instead?
<nwuensche>
How can I get the path of the cups package in another package? ${pkgs.cups} didn't work
arximboldi has joined #nixos
<copumpkin>
gchristensen: not sure, haven't bought anything yet
<gchristensen>
cool
<gchristensen>
BOS to MUC is like $550 on Lufthansa right now
<goibhniu>
Are you both looking for a place to stay? You're welcome to stay in my place if you like.
takle has joined #nixos
<gchristensen>
nice! I'll be bringing my wife along with me, so I reckon we'll find a hotel but that is really generous. thank you :)
* goibhniu
has a guestroom and a sofa bed, plenty of room
<goibhniu>
your wife would be welcome too, of course ... but if you'd prefer a hotel, go for it
<disasm>
nwuensche: pkgs.cups should be correct, what's not working?
eacameron has joined #nixos
<nwuensche>
disasm: It just ignores this when I want to substitudeInPlace and just writes the stuff after this variable into the file
<mog>
whats the right way to have a service enabled but not started by default
<gchristensen>
don't have anything want it
<mog>
like i want to have the normal httpd but i want to manually start it rather than have it happen at boot
<nwuensche>
disasm: Sorry, I said it wrong. It displays: error: undefined variable ‘pkgs’ when I rebuild
<Rajigo>
What is the way to change the default GCC version used by nix in all packages? I have dumbly tried `stdenv = pkgs.overrideCC pkgs.stdenv gcc;` but it obviously create an infinite recursion (as gcc is defined in terms of stdenv) and nix fails.
bkchr has joined #nixos
<srhb>
nwuensche: I don't remember your gist exactly, but didn't you bring in cups as an argument? In which case you probably want ${cups} not ${pkgs.cups}
felipedvorak has joined #nixos
eacameron has quit [(Ping timeout: 276 seconds)]
<nwuensche>
srhb: Is there a difference between $cups and ${cups}?
pxc has quit [(Ping timeout: 240 seconds)]
<srhb>
nwuensche: Assuming you're modifying installPhase or something, $cups is an environment variable and ${cups} is a nix substitution.
<xd1le>
disasm: I think I've actually just had a misunderstanding actually, apologies. I think nix has actually just already patched the file before preBuild. And I've verified this by cat'ing the file in prePatch. Been a long day, my bad.
Sonarpulse has joined #nixos
<srhb>
Rajigo: Shouldn't that work as long as you refer to one of the explicitly versioned gcc attributes which aren't defined in terms of stdenv? I may be mistaken here, but you can try.
FRidh has joined #nixos
<Rajigo>
srhb: oh oh oh, I see...
<srhb>
Rajigo: No promises!
joehh has quit [(Ping timeout: 240 seconds)]
bennofs has quit [(Ping timeout: 240 seconds)]
nevermin_ has joined #nixos
<Rajigo>
srhb: nope, not working just like that (I suppose because the gcc package cannot not depend on stdenv)... However it might just work by using the let ... in syntax.
arximboldi has quit [(Quit: arximboldi)]
<nwuensche>
srhb: I see that I don't want to go into the cups-2.1.4 package, but inside the cups-progs package. How do I get this one?
<srhb>
nwuensche: I don't think we have such a package. Which file are you looking for?
<nwuensche>
srhb: I have a folder "/nix/store/...-cups-progs/ ^^
<clever>
/home/clever/apps/nixpkgs/nixos/modules/services/printing/cupsd.nix: name = "cups-progs";
<srhb>
nwuensche: Oh.
<clever>
thats a buildEnv
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
<clever>
and confusingly, there are 2 with the same name
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 3 new commits to master: https://git.io/v5JrZ
<NixOS_GitHub>
nixpkgs/master dd809b9 Eric Sagnes: pytorch: 0.1.12 -> 0.2.0
<NixOS_GitHub>
nixpkgs/master 14ce421 Eric Sagnes: torchvision: init at 0.1.9
<NixOS_GitHub>
nixpkgs/master b7b815b Frederik Rietdijk: Merge pull request #28386 from ericsagnes/pkg-update/pytorch...
NixOS_GitHub has left #nixos []
<nwuensche>
clever: But that's where I store the filter + .ppd files.
DrWaste has quit [(Quit: Run away!)]
<nwuensche>
clever: And I'd like to reference the ppd File
<srhb>
nwuensche: The ppd file really ought to be just a simple derivation itself.
<nwuensche>
clever: Have I done something wrong and this should be part of cups?
<srhb>
nwuensche: In fact, didn't you already have srcPPD be that file?
<clever>
brb
<srhb>
nwuensche: Again, it's hard to follow along unless you're updating some gist of all your files.
<srhb>
nwuensche: But if you did just bring it in like that, then its path in the store is ${srcPPD}
pxc has joined #nixos
bkchr has quit [(Ping timeout: 276 seconds)]
seppellll has quit [(Quit: Leaving)]
Infinisil has quit [(Quit: I hate studying)]
<nwuensche>
srhb: You're right, sorry for that.
pie_ has quit [(Ping timeout: 248 seconds)]
<srhb>
nwuensche: No reason to apologize! :-)
ebzzry_ has quit [(Ping timeout: 248 seconds)]
magnetophon has quit [(Ping timeout: 276 seconds)]
leat has joined #nixos
takle has quit [(Remote host closed the connection)]
<gchristensen>
given a lambda is there a way to see where it came from or what arguments it expects?
<gchristensen>
my kingdom for a docfn which tied a `.doc` to the lambda or something
<clever>
gchristensen: part 1: nix-repl> builtins.unsafeGetAttrPos "fetchgit" pkgs
<gchristensen>
nix-repl> builtins.unsafeGetAttrPos x
<clever>
nix-repl> builtins.functionArgs ({ a, b, c ? "default" }: "foo")
<clever>
{ a = false; b = false; c = true; }
<gchristensen>
doesn't work where the import returns the fn
<clever>
it needs a string and an attrset
<gchristensen>
yeah I don't have an attrset
<clever>
that makes things harder
<gchristensen>
I have a lambda unexpectedly returning a lambda and I want to know where that one is coming from
<clever>
to start with, what does builtins.functionArgs say about it?
<clever>
and if you just throw random values at it, it will probably barf with a location
<nwuensche>
srhb: Now I get: PID 8867 (/../cups/filter/brlpdwrapperMFC5440CN) stopped with status 1
<gchristensen>
{ } :)
<nwuensche>
The permissions of the file are -r-xr-xr-x
<gchristensen>
good call clever, at this point I've solved my problem and am exploring this like a usability issue
<srhb>
nwuensche: Sounds like progress, it's running but exiting with a failure code.
<clever>
nwuensche: what if you manualy run the filter in a shell?
eacameron has joined #nixos
vandenoever has quit [(Ping timeout: 240 seconds)]
<nwuensche>
clever: It just runs endless without any output
<srhb>
nwuensche: cat something to it
<nwuensche>
srhb: What do you mean by that?
<srhb>
nwuensche: I mean it's probably waiting for standard input
<srhb>
nwuensche: You want to echo "foo" | thefilter
<srhb>
nwuensche: Or something like that, i don't know exactly what it expects on stdin :)
<nwuensche>
srhb: So I echo "asdas" | tcsh /nix/store/y50i6a733vc4q96irjzhx7idp2vcsaav-mfcj47dd0dw-cupswrapper-3.0.0-1/lib/cups/filter/brlpdwrapperMFC5440CN, and now it says: /usr/local/Brother//lpd/rastertobrij2: No such file or directory. Progress! Thanks. Why does this work when I pipe something
<srhb>
nwuensche: Try running cat without any arguments on the command line
<srhb>
nwuensche: Compare to running cat somefile
<srhb>
nwuensche: The first one waits indefinitely for stdin to produce something (and get closed)
<nwuensche>
srhb: Ok, I see. I just thought it "works"^^
bennofs has joined #nixos
<srhb>
Actually the second example is better explained with echo "foo" | cat
<clever>
srhb: up next, libredirect on /usr/local/Brother//lpd/rastertobrij2
jellowj has joined #nixos
<srhb>
clever: What's that?
<srhb>
I mean, I can almost guess from the name...
<nwuensche>
srhb: One error gone, one new error :) : error while loading shared libraries: libbrcompij2.so.1: cannot open shared object file: No such file or directory
<clever>
srhb: an LD_PRELOAD library that can redirect most filesystem access
<srhb>
clever: Hah, nifty.
<clever>
srhb: the teamspeak3 client has a good example
<srhb>
nwuensche: Either what clever suggested (probably) or, perhaps patching RPATH on the caller with patchelf. You may have some examples handy.
<disasm>
nwuensche: nwuensche that's going to be a patchelf to the location of that library
<disasm>
what srhb said :)
<clever>
"/usr/local/Brother//lpd/rastertobrij2" sounds more like an executable then a library
<clever>
checking what syscall it used would confirm that, run brlpdwrapperMFC5440CN under strace
<srhb>
I think rastertobri2 is attempting to find that library
<srhb>
libbrcompij2.so.1: cannot open shared object
<srhb>
But yeah, strace will make things clearer.
<nwuensche>
srhb: So I get some low-level code from that. Now what?
<clever>
do any lines mention rastertobrij2 ?
<nwuensche>
rt_sigsuspend([INT], 8/nix/store/656q3wybdk2ainlmizy8d3v0k03a3q44-mfcj47dd0dw-cupswrapper-3.0.0-1/usr/local/Brother//lpd/rastertobrij2: error while loading shared libraries: libbrcompi j2.so.1: cannot open shared object file: No such file or directory
<clever>
yeah, its an executable, and the path looks to already be corrected
erasmas has joined #nixos
pie_ has joined #nixos
<clever>
"strace -f -s 3000" will give longer strings in this log
<disasm>
clever: doesn't that executable need rpath patched to find that so file?
<clever>
then you can see exactly what arguments rastertobrij2 is ran with, and can manually run it
nevermin_ is now known as nevermind
roundhouse has quit [(Ping timeout: 255 seconds)]
<clever>
disasm: rastertobrij2 needs its own rpath and interperter, so it can run stand-alone
<disasm>
clever: yeah, that's the executable I was talking about
<nwuensche>
So what exaclty should I do?
<clever>
nwuensche: make sure rastertobrij2 is also patchelf'd fully
<disasm>
nwuensche: where is libbrcompij2.so.1? Is that in the same package?
Arcaelyx has joined #nixos
<nwuensche>
disasm: Yep, it is.
<disasm>
if so, I'd say add $out/path/to/libbrcompij2.so.1 to the rpath for $out/usr/local/Brother/lpd/rastertobrij2
iyzsong has quit [(Ping timeout: 240 seconds)]
<disasm>
nwuensche: you can also do readelf -d /nix/store/656q3wybdk2ainlmizy8d3v0k03a3q44-mfcj47dd0dw-cupswrapper-3.0.0-1/usr/local/Brother//lpd/rastertobrij2 to get a full list of what so files it's looking for dynamically.
<nwuensche>
disasm: Could you please explain exaclty what I should do? I'm sorry, but I have no clue what I should do now. Should I patchelf? And if so, how does this command work? The two important files are $out/usr/lib/libbrcompij2.so.1.0.2 and $out//usr/local/Brother/lpd/rastertobrij2 I guess
<disasm>
nwuensche: are you already running patchelf on that file or not? could you gist the nix file building this?
<disasm>
if that's the only so file it needs to run (unlikely) command would be patchelf --set-rpath ${out}/path/to/libbrcompij2.so.1 usr/local/Brother/lpd/rastertobrij2
<disasm>
err $out/usr since your copying it before running patchelf
<disasm>
typically I'd patch everything in the tmp dir, then after it's all good, copy it to $out
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 1 new commit to master: https://git.io/v5J1d
<NixOS_GitHub>
nixpkgs/master 056a389 John Ericson: ios cross cc-wrapper: Fail assertion to avoid breaking tarball job...
NixOS_GitHub has left #nixos []
eacameron has quit [(Remote host closed the connection)]
civodul has quit [(Quit: ERC (IRC client for Emacs 25.2.1))]
<nwuensche>
disasm: I did this. However, the problem is that the file's called libbrcompij2.so.1.0.2, but the error is: /usr/local/Brother//lpd/rastertobrij2: error while loading shared libraries: libbrcompij2.so.1: cannot open shared object file: No such file or directory. Should I just rename the file?
<disasm>
nwuensche: now you need that REDIRECT clever mentioned above
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 2 new commits to staging: https://git.io/v5JDU
<NixOS_GitHub>
nixpkgs/staging ad8d564 Bojan Nikolic: set +u workaround for empty arrays...
<NixOS_GitHub>
nixpkgs/staging b6014cc John Ericson: Merge pull request #28415 from bnikolic/fix/ld-wrapper-set-u...
<nwuensche>
disasm: Sorry, but I have no clue what's going on there. Do you mean I need a line like: patchelf --set-rpath $(pwd):${stdenv.cc.libc}/lib:${stdenv.cc.cc.lib}/lib forticlientsslvpn/64bit/helper/setup
<nwuensche>
Just for the llibbrcompij2.so.1
<disasm>
/usr/local/Brother//lpd/rastertobrij2 is the wrong path
digitus has joined #nixos
<nwuensche>
disasm: FIxed it to $out/usr....
pie_ has quit [(Read error: Connection reset by peer)]
pie_ has joined #nixos
<disasm>
so whatever is executing that so whatever is executing that needs wrapProgram lines at the end of the gist there, with NIX_REDIRECTS=/usr/local/Brother//lpd/rastertobrij2=$out/usr/local/Brother//lpd/rastertobrij2
<disasm>
nwuensche: are brlpdwrapperMFC5440CN and MFC.ppd redistributable? Is there a place I can download them to test this?
<woffs>
A newbie question: how to make terminfo/termcap entries available? Actually missing rxvt-unicode. Just take TERM=rxvt as workaround?
Ivanych has quit [(Ping timeout: 255 seconds)]
Havvy has quit [(Read error: Connection reset by peer)]
takle has joined #nixos
<nwuensche>
disasm: So how's it going?
Havvy has joined #nixos
tmaekawa has quit [(Quit: tmaekawa)]
<srhb>
woffs: I'm not sure I follow. What are you missing and how did you conclude this?
<tilpner>
woffs - st uses installPhase = '' TERMINFO=$out/share/terminfo make install PREFIX=$out ''; (nix edit is sweet)
<woffs>
srhb, I'm just logging in via ssh into a rather bare nixos and my TERM=rxvt-unicode and w3m complains about that: "Can't find termcap entry rxvt-unicode"
obadz has quit [(Ping timeout: 240 seconds)]
nevermind has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<srhb>
woffs: Did you install rxvt_unicode?
erictapen has quit [(Ping timeout: 255 seconds)]
<woffs>
srhb, no :-)
<srhb>
woffs: That should help, then. :)
<woffs>
ok, works now. I hesitated to install X components, but this is not too fat.
nico202 has joined #nixos
<srhb>
woffs: You can also just copy the terminfo definitions to the remote home dir, but I think that feels less nice personally.
<srhb>
A better solution might be to separate out the terminfo from the urxvt-unicode derivation and make it entirely independent.
<srhb>
Depends how much you care :-)
eacameron has joined #nixos
obadz has joined #nixos
SOO7 has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
<woffs>
not too much atm. ;) thank you!
eacameron has joined #nixos
sauyon has joined #nixos
<disasm>
nwuensche: ok, built
eacameron has quit [(Ping timeout: 240 seconds)]
<nwuensche>
disasm: Cool! So would have you done?
propumpkin has joined #nixos
magnetophon has quit [(Ping timeout: 240 seconds)]
mudri has joined #nixos
contrapumpkin has quit [(Ping timeout: 248 seconds)]
eacameron has joined #nixos
<disasm>
nwuensche: one sec...
<rycee>
ts468: Ok, thanks for the heads up. I've joined now. Sorry for the delay :-)
<nwuensche>
disasm: No problem
eacameron has quit [(Ping timeout: 240 seconds)]
nevermind has joined #nixos
Ivanych has joined #nixos
<Ralith>
woffs, srhb: isn't there a rxvt_unicode.terminfo derivation?
eacameron has joined #nixos
<srhb>
Ralith: There might be!
<srhb>
Ralith: Yes indeed!
eacameron has quit [(Ping timeout: 248 seconds)]
<srhb>
woffs: ^ using that instead will allow nix to get rid of the unused dependencies :)
<woffs>
Ralith, srhb, how to find this!
<srhb>
woffs: rxvt_unicode.terminfo
<woffs>
in unstable?
<srhb>
woffs: Yup.
<woffs>
and not in 17.03?
<srhb>
woffs: I can check.
<srhb>
woffs: It's there in 17.03 as well.
sellout- has quit [(Quit: Leaving.)]
<nwuensche>
disasm: Everything ok?
<woffs>
srhb, note that I'm new to nixos :) "nix-env -qaP" does not show anything containing "terminfo", and looking into nixpkgs.git didn't bring me much further
<Ralith>
there's a bunch of crap that nix-env can't find for various reasons
<Ralith>
(via search)
<Ralith>
(it can still install it fine)
<mbrgm>
fpletz: let's continue here then... ;)
<srhb>
woffs: It's a multi output derivation, I don't think any of the alternatives are ever shown, actually.
<srhb>
If they're not named explicitly in all-packages.nix
<mog>
how can i explicitly tell a systemd service to not start on boot besides disabling it totally. i tried setting systemd.services.unifi.wantedBy = [""]; but that didnt seem to do anything
<srhb>
mog: Would the systemd.services.name.enable attribute work?
<mog>
i think i tried that as well but let me try that too
erictapen has joined #nixos
<aminechikhaoui>
mog: I don't think it should start by defualt after boot unless you specify "multi-user.target" in wantedBy or something
<mog>
oh i did that srhb the problem with that is you cant start it if that is set
<mog>
Failed to start unifi.service: Unit unifi.service is masked.
<mog>
is what happened when i tried to manual start it
<fpletz>
mbrgm: I have found the issue: we're using networking.defaultMailServer which uses ssmtp… that tool doesn't need suid \o/
<mog>
AmineChikhaoui i have tried setting the wantedby to blank and too bs value it still starts with multi-user.target
<woffs>
thanks srhb. is there a nix-env option to list the possible outputs?
<srhb>
woffs: Not as far as I know.
<fpletz>
mbrgm: I think we should just remove NoNewPrivileges for now
<aminechikhaoui>
mog: can you probably paste the service config ?
<disasm>
nwuensche: that nix file you gisted me builds for you?
<nwuensche>
Jup
<nwuensche>
But I have the printer2.nix package in the same dir
<disasm>
nwuensche: ah, I think there's an extra \ at the end of a line...
<disasm>
I'm getting substitute(): ERROR: Invalid command line argument: sed
ryantrinkle has quit [(Ping timeout: 240 seconds)]
<disasm>
yeah, line 27 has a '\' trailing at the end
<srhb>
nwuensche: For the record, if you make a gist with all the necessary files, you can update it and we can just clone all the necessities and test and reiterate much faster :)
<srhb>
nwuensche: I think a lot of time is just spend endlessly trying to stay up to date with what you've changed here :)
Thra11 has joined #nixos
<srhb>
spent*
Ivanych has quit [(Ping timeout: 240 seconds)]
rauno has joined #nixos
<sauyon>
erm
tokudan has joined #nixos
<nwuensche>
srhb: I didn't know that. I thought a gist is final. Thanks!
<sauyon>
system rebuilds are now failing because my /boot is full
<rauno>
hey, if i have changed package config in nixpkgs, how was the rebuild/reinstall made ?
<srhb>
nwuensche: Cool! I only learned so recently.
<sauyon>
not sure what I'm supposed to do about that
<fpletz>
mbrgm: I added the hardening and NoNewPrivileges but obviously didn't test it with a proper sendmail, I'll just remove it
<srhb>
sauyon: sudo nix-collect-garbage -d, then reboot, then do it again, iirc.
<sauyon>
oh I have to reboot
<fpletz>
mbrgm: maybe we can do something with CapabilityBoundingSet
<aminechikhaoui>
mog: just checking did you make sure to have lib.mkOverride 0 for example ?
<srhb>
rauno: I don't understand the question, can you elaborate?
phinxy has joined #nixos
nico202 has quit [(Ping timeout: 240 seconds)]
<rauno>
Like if i have one library isntalled and I now modified it's default.nix under nixpkgs, which command it was to rebuild it / reinstall ?
<aminechikhaoui>
mog: also not sure if partOf is also triggering the service
<srhb>
rauno: You mean nixos-rebuild but pointing to your modified nixpkgs repository?
* aminechikhaoui
always confused with systemd options
<rauno>
so it picks automatically up the change ?
<sauyon>
that didn't work D:
<srhb>
rauno: Depends how you're using nixpkgs. if you're using channels, it won't find it automatically or anything like that.
<rauno>
I have local repo of it
jellowj has joined #nixos
<rauno>
and defined in nix path
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/v5Jby
<NixOS_GitHub>
nixpkgs/master 775a794 Franz Pletz: kanboard: 1.0.44 -> 1.0.46...
vandenoever has quit [(Ping timeout: 276 seconds)]
<srhb>
sauyon: I can never remember this either. Logically there might be a rebuild in there somewhere in order to regenerate /boot without the now-unnecessary generations
phinxy has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to release-17.03: https://git.io/v5Jbj
<sauyon>
maybe it's just because I'm keeping too many generations around
takle has quit [(Remote host closed the connection)]
<srhb>
rauno: If you have that, and it's installed in your system profile, then nixos-rebuild switch will automatically pick up any changes.
<rauno>
ok, seems like it worked yes, just running nixos-rebuild did it :D
<rauno>
thx bro
takle has joined #nixos
<srhb>
*cringe* you're welcome.
<fpletz>
mbrgm: thanks for nagging me about that :) now we have to do gitlab… we have other issues with gitlab on master, though, have you seen https://github.com/NixOS/nixpkgs/pull/27159
<rauno>
just wondered if it would do it automatically, if its already installed.
<srhb>
As soon as you ask for the rebuild, yes :)
eacameron has joined #nixos
<nwuensche>
disasm: Still problems with the config?
<nwuensche>
disasm: I created a PR. Their are some things missing in the repo
<disasm>
should the interpreter prefix really be ${stdenv.glibc.out} or should it be ${stdenv.cc}?
<srhb>
disasm: I think it should just be ${stdenv.glibc}/lib...
<srhb>
At least if it's built with something like nix-build -E 'with (import <nixpkgs> {}); callPackage_i686 ./printer2.nix {}'
<disasm>
srhb: can you rebase off of my changes? merge conflicts
<srhb>
Uh, if I've pushed anything that was certainly unintentional.
<disasm>
ooh, callPackage_i686 that might be why things are breaking...
<disasm>
meh, nwuensche can you rebase your PR?
<disasm>
actually nm, I'll pull down and merge manually, it's a hiary one.
Arcaelyx_ has joined #nixos
Arcaelyx has quit [(Ping timeout: 240 seconds)]
<nwuensche>
disasm: Sorry, I haven't seen the merge conflict. I didn't want to do even more trouble...
<nwuensche>
disasm: My build command in configuration.nix is: printing.drivers = [ (( pkgs.callPackage_i686 /home/nwuensche/printer2.nix) { })];
Arcaelyx_ has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
Ross has quit [(Ping timeout: 260 seconds)]
Ross has joined #nixos
<disasm>
haha, my mouse cursor just froze, glad I use i3 and can still interact with my session...
<__monty__>
I'm on arch linux and there's an aur package for nix but there seem to be some issues with it. If I follow these instructions will I be able to keep nix updated using nix itself?
<ambro718>
Anyone familiar with Qt5? I have a problem getting qtvirtualkeyboard to work. I defined a package (much like many other qt5 packages that were already defined) and put into systemPackages. But there is no indication of anything working. SDDM still complains in the logs: sddm-greeter[921]: file:///run/current-system/sw/share/sddm/themes/breeze/components/VirtualKeyboard.qml:20:1: module "QtQuick.VirtualKeyboard"
freeman42x has joined #nixos
Arcaelyx has joined #nixos
<ambro718>
the interesting part is that I see /run/current-system/sw/lib/qt-5.9/qml/QtQuick/VirtualKeyboard/qmldir which does in fact define "module QtQuick.VirtualKeyboard".
<ambro718>
oh, only the -bin output got installed and not the main output which actually has the library, how to fix this?
<disasm>
great job srhb :)
<ambro718>
I mean apparently there is no reference to the main output since it didn't get copied by nixops
kiloreux has quit [(Ping timeout: 248 seconds)]
<ambro718>
never mind I miss a moveToOutput thing
jellowj has quit [(Ping timeout: 240 seconds)]
pie_ has quit [(Ping timeout: 248 seconds)]
<ambro718>
fixed it but still get the same error. This is my nix expression for qtvirtualkeyboard.nix: http://pasteall.org/521164
<nwuensche>
srhb: Yes, I think so. But how can I log it? Now we have the old problem again.
<srhb>
nwuensche: Unfortunately I have no idea, and no printer to start testing cups with.
<nwuensche>
srhb: disasm: THank you so much for the progress we made until now! It's amazing how supportive you and the community are
ixxie has quit [(Quit: Lost terminal)]
<srhb>
nwuensche: Glad to help, and I do hope you make progress and can get it packed up properly. :)
<disasm>
nwuensche: check the cups log via systemctl or journalctl
<nwuensche>
disasm: journalctl --since=today | grep -i cups | less gives me not much things worth it mentioning besides this line
<nwuensche>
disasm: Maybe: CreateDevice failed: org.freedesktop.DBus.Error.ServiceUnknown:The name org.freedesktop.ColorManager was not provided by any .service files?
MP2E has joined #nixos
<bennofs>
nwuensche: oh, i just found something more: cupsctl --debug-logging
<bennofs>
nwuensche: you can also run cupsctl without args to view the current configuration
eacameron has joined #nixos
nixos-users-wiki has quit [(Remote host closed the connection)]
takle has joined #nixos
<disasm>
nwuensche: do journalctl -f -u cupsd.service and replicate
reinzelmann has quit [(Quit: Leaving)]
<nwuensche>
disasm: I tried journalctl -f -u cupsd.service and made a test print in chromium, but nothing happend in the log
<nwuensche>
I'm back in 20 Minutes
<__monty__>
srhb: Seems like it doesn't exist yet. How do I find out where I should create one?
eacameron has quit [(Remote host closed the connection)]
<srhb>
__monty__: You can strace nix-env and grep for nix.conf
<srhb>
__monty__: strace nix-env --version |& grep et
<srhb>
Er
<srhb>
grep nix.conf
<disasm>
nwuensche: I'd start with probly something like `echo test|lpr -P <printer name>`
<__monty__>
srhb: Thanks, seems like /etc/nix is the right place.
<srhb>
__monty__: Great :)
<__monty__>
srhb: Should root own /etc/nix or no?
takle has quit [(Remote host closed the connection)]
<srhb>
__monty__: I don't think it matters on a single user system.
<srhb>
__monty__: But tbf I have very little experience with Nix outside of NixOS.
goibhniu has joined #nixos
eacameron has joined #nixos
nevermind has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<nwuensche>
disasm: Again, it says: Job stopped due to filter errors; please consult the error_log file for details.
Thra11 has quit [(Quit: IRC for Sailfish 0.9)]
* shapr
cleans his filters
<nwuensche>
disasm: This was journalctl -f -u cups.service.
<nwuensche>
disasm: cupsd.service gives me nothing at all
<srhb>
nwuensche: Just curious, does using sudo on that command help any?
LnL has quit [(Quit: exit 1)]
nevermind has joined #nixos
<nwuensche>
srhb: Unfortunately not
<nwuensche>
srhb: I also found: PID 6343 (/nix/store/1cfzc5fjpn2jcbd2zm6d00wb10m46s2y-cups-progs/lib/cups/filter/brlpdwrapperMFC5440CN) stopped with status 1.
<srhb>
That's at least more helpful.
<srhb>
But not by much ._.
<nwuensche>
Oh, and: /tmp/br_cupsfilter_debug_log: Permission denied. This is the log file specified in the wrapper!
<nwuensche>
How can I give it access to write?
<srhb>
nwuensche: Delete it and let it create it on its own
<srhb>
It should have access to /tmp
<srhb>
(I think)
LnL has joined #nixos
posco has quit [(Ping timeout: 240 seconds)]
<nwuensche>
I just chmod 0777'd the file, and I've got something! I create a gist.
eacameron has quit [(Remote host closed the connection)]
<srhb>
Looks like brlpdwrapperMFC5440CN hasn't been patched appropriately in order to find brprintconfij2
<srhb>
Check if the script refers to it directly
<nwuensche>
Could it be that I have to set add the path to BRPRINTCONFIG in brlpdwrapperMFC5440CN?
<srhb>
Use the usual substitutions to find it
<disasm>
nwuensche: same thing to brprintconfij2
<srhb>
Or a wrapper, sure.
<srhb>
Looks like you can just replace brprintconfij2 with its full path.
<tokudan>
firefox on master is on 55.0.2, but release-17.03 is still on 54, are there any plans to update firefox in 17.03?
eacameron has joined #nixos
<gchristensen>
17.03 should be updated
<srhb>
17.03 has 54.0.1, yeah
LnL has quit [(Quit: exit 1)]
<nwuensche>
disasm: srhb: We have a little bit of progress! The printer already shows that it's receiving data now, but it still doesn't print. But progress nevertheless
<srhb>
Holy carp. :-)
<tokudan>
i don't see any open pull requests for that, so what would be the best plan of action for that? I have no idea how packages move from master to release-17.03
<nwuensche>
I updated the last gist (I learn!) with the new output of /tmp/br_log
eacameron has quit [(Ping timeout: 248 seconds)]
kiloreux has quit [(Ping timeout: 255 seconds)]
<nwuensche>
Dammit! I read it now, I think I made a mistake there
eacameron has joined #nixos
LnL has joined #nixos
<srhb>
tokudan: I think they are cherry picked.
<gchristensen>
tokudan: find the patches in master and `git cherry-pick -x commit commit commit
<gchristensen>
then push to a branch PR to release-17.03
<tokudan>
oh, ok... I'll see if i can create a pull request then. can't build it on my laptop though, as i'm running on batteries
<srhb>
nwuensche: I think it just needs patching of the interpreter
<srhb>
nwuensche: Try to run the file manually.
<nwuensche>
srhb: echo "test" | tcsh lib/cups/filter/brlpdwrapperMFC5440CN gives me some weird unicode stuff, but no error messages
<srhb>
nwuensche: I mean, try to run brprintconfij2 manually :)
eacameron has quit [(Remote host closed the connection)]
<srhb>
nwuensche: You should get a very special error message that you should learn to recognize when you're making nix derivations of binaries from elsewhere
<nwuensche>
srhb: Is it ok when I find all of this very magical at the moment? ^^
nevermind has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<gchristensen>
nwuensche: yes :)
<srhb>
Absolutely. This is the sort of thing I've just memorized.
<gchristensen>
!m srhb
<[0__0]>
You're doing good work, srhb!
<srhb>
When I see this: /nix/store/q99m20nz28z92512vlhq12y85844lskp-mfcj47dd0dw-cupswrapper-3.0.0-1/usr/bin//nix/store/q99m20nz28z92512vlhq12y85844lskp-mfcj47dd0dw-cupswrapper-3.0.0-1/usr/bin/brprintconfij2 does not exist
<srhb>
I think "uh oh, wrong interpreter"
<gchristensen>
srhb: incredible incredible work
Neo-- has quit [(Ping timeout: 240 seconds)]
<srhb>
Thanks :-)
jbgi has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 opened pull request #28448: cc-wrapper: Newstyle setup vars (master...cc-wrapper-setup-var) https://git.io/v5U3C
NixOS_GitHub has left #nixos []
eacameron has joined #nixos
<nwuensche>
srhb: However, I updated the gist. It still gives me almost the same log
<srhb>
nwuensche: Please update the github repository as well so I can see what you have.
<nwuensche>
srhb: I don't really understand why there are two packages in this one link
<nwuensche>
In the WARNING in the log. It's two times cupswrapper-3.0.0.1
<srhb>
nwuensche: Can you run that path it warns about manually now?
nevermind has joined #nixos
<srhb>
/nix/store/yqfb19idiib7ssp4p43xgzzhnwckk1c3-mfcj47dd0dw-cupswrapper-3.0.0-1/usr/bin//nix/store/yqfb19idiib7ssp4p43xgzzhnwckk1c3-mfcj47dd0dw-cupswrapper-3.0.0-1/usr/bin/brprintconfij2 <-- that one
<nwuensche>
srhb: /nix/store/yqfb19idiib7ssp4p43xgzzhnwckk1c3-mfcj47dd0dw-cupswrapper-3.0.0-1/usr/bin/brprintconfij2 gives me a nice output, the long one doesn't work
<srhb>
Oh yeah, I see it now
<srhb>
Derp
<nwuensche>
srhb: But I can't see the error in the config
eacameron has joined #nixos
<hodapp>
gaaack. I finally made this unholy bastardization produce Python bindings of existing C++ code which are capable of interfacing to OpenCV's Python bindings
<srhb>
nwuensche: Because there's two substitutions going on
freusque has quit [(Quit: WeeChat 1.7.1)]
<hodapp>
I feel like this should go online but it's so disgustingly ad-hoc that it should probably just go on my blog, not into Nixpkgs
<srhb>
nwuensche: One substitutes /usr to $out/usr
<hodapp>
OpenCV's bindings generators weren't so much "batteries not included" as "electronics not included"
<pxc>
so the Nix installer is failing for me on CentOS 6 and 7 (both in VMs). It hangs on 'unpacking channels', and additionally the channel appears not to fully be downloaded before the 'unpacking channels' message is printed. Is this a known issue?
<srhb>
nwuensche: Yeah it is. check line 441 of the resulting wrapper
<srhb>
nwuensche: That would wrap the wrapper, I think that would go poorly
<srhb>
(Or maybe just overwrite one of them? Not sure)
<nwuensche>
srhb: I just did it like you said :)
<srhb>
Goodie.
<nwuensche>
Now it rebuilds...
JosW has quit [(Quit: Konversation terminated!)]
johnw has quit [(Ping timeout: 240 seconds)]
eacameron has quit [(Ping timeout: 240 seconds)]
<nwuensche>
srhb: My Printer got the command to clean itself! That's usually a good sign. Now wait until 5 minutes and look if it prints something after the cleaning:D
<srhb>
:o
<nwuensche>
srhb: It prints! IT PRINTS!!!
<srhb>
Magic!
<pxc>
k, so it looks like the real issue I'm having is not with the Nix installer but nix-channel
<pxc>
can anyone think of why nix-channel might just hang when it's supposed to be unpacking the tarball?
<srhb>
pxc: Maybe strace can help?
Profpatsch has joined #nixos
johnw has joined #nixos
<nwuensche>
srhb: Thank you so much! I really appreciate the help and time you spend for me! I won't do much in the next days, but then I will clean the package up and push it into nix-pkgs. This community is amazing!
<srhb>
nwuensche: Thank you, I was just about to ask you to do that once you've had some time to work with it. :D
<srhb>
nwuensche: And I'm glad you got it to work, good job sticking with it! :)
stepcut has joined #nixos
<nwuensche>
srhb: But bad news: Next monday I have to set up another Brother Printer... But I hope that the things I learned today will help me there too.
<m0rphism>
I'm having the problem that running chromium deadlocks before even showing a window. According to the chromium wrapper script "libredirect causes chromium to deadlock on startup". Is chromium currently broken or is this due to some configuration option?
<srhb>
nwuensche: I'm sure it will! :)
magnetophon has joined #nixos
nwuensche has quit [(Quit: Leaving)]
<pxc>
srhb: I figured out what specific process is hanging. It's the one that uses nix-env to unpack the channel tarball. I don't really know what to make of the strace output but the last line is clearly where it hang
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to master: https://git.io/v5UWE
<NixOS_GitHub>
nixpkgs/master 163e648 Roman Volosatovs: mopidy-iris: 3.1.3 -> 3.2.0
<NixOS_GitHub>
nixpkgs/master ea06446 Joachim F: Merge pull request #28414 from rvolosatovs/update/mopidy-iris...
NixOS_GitHub has left #nixos []
alx741_ has joined #nixos
<pxc>
strace output for the command called by nix-channel which hangs nix-channel --update: https://pastebin.com/990CZawL
tokudan has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] WilliButz opened pull request #28451: Version bumps: tcpdump, valgrind, qjson, rmlint, rawdog (master...version-bumps) https://git.io/v5UW7
NixOS_GitHub has left #nixos []
alx741 has quit [(Ping timeout: 248 seconds)]
erictapen has joined #nixos
<srhb>
pxc: I'm not sure what to make of that either. Hm..
<pxc>
srhb: the open file handles are pipes. I'm not sure which commands it's waiting on
magnetophon has quit [(Ping timeout: 248 seconds)]
lambdamu has quit [(Remote host closed the connection)]
<srhb>
I wonder which circumstances does that, since I've seen it work earlier today.
<gchristensen>
srhb: incredible dedication, thank you for getting their printer working
<srhb>
gchristensen: Team effort! :-)
<gchristensen>
who else? I missed it
<srhb>
At least disasm, clever and bennofs. It was a multi-day thing.
<gchristensen>
incredible
<srhb>
I'm impressed they stuck so long, good sign!
<gchristensen>
thank you disasm, cleve, bennofs for helping too
michaelpj has joined #nixos
<tokudan>
gchristensen, firefox 55 needs a new libpng, should I try to cherry pick the libpng from master to release-17.03 or just setup an extra derivation used only by firefox for 17.03?
<gchristensen>
tokudan: what is the version diff?
<tokudan>
1.6.28 to 1.6.29 (minimum) or current master is on 1.6.31
<gchristensen>
either one should be fine
<tokudan>
updating libpng would be a big rebuild?
takle has joined #nixos
<gchristensen>
it would but se la nixos
hamishmack has quit [(Quit: hamishmack)]
yegods__ has quit [(Ping timeout: 246 seconds)]
eacameron has joined #nixos
yegods has joined #nixos
michaelpj__ has joined #nixos
yegods_ has joined #nixos
hiratara has quit [(Ping timeout: 255 seconds)]
michaelpj has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
yegods_ has quit [(Client Quit)]
<michaelpj__>
adding groups to `users.users.ME.extraGroups` doesn't seem to take effect, even after a logout and login
<michaelpj__>
however, if I do `su ME` then I *do* see the right groups
<michaelpj__>
anyone know anything about this?
<tokudan>
ok... so I'll add that to the firefox PR
<srhb>
michaelpj_: You need to log out harder!
nevermind has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<michaelpj__>
srhb: I'm not even sure if you're joking ;)
phinxy has quit [(Quit: Leaving)]
eacameron has quit [(Ping timeout: 240 seconds)]
<srhb>
michaelpj__: I'm not, that's the only thing I can think of and I think it's probable. :-P Ensure you've nothing left using pgrep -u ME as root or something like that
<srhb>
michaelpj__: (tmux perhaps?)
<michaelpj__>
I mean, I've also rebooted the machine since then, so I don't think it *can* be that
<srhb>
oh, what.
<srhb>
That's weird.
<michaelpj__>
guess I can try again
<michaelpj__>
yeah
hiratara has joined #nixos
<srhb>
michaelpj__: The group in question exists?
<tilpner>
michaelpj__ - Try adding some groups to users.extraGroups
<michaelpj__>
`groups ME` also returns the right result, but `groups` is missing it
yegods has quit [(Ping timeout: 255 seconds)]
<michaelpj__>
all the entries in `/etc/groups` are fine
michaelpj has joined #nixos
michaelpj__ has quit [(Remote host closed the connection)]
michaelpj has quit [(Read error: Connection reset by peer)]
<seanparsons>
Is there a "Not being a dick" way of getting a PR (https://github.com/NixOS/nixpkgs/pull/28380) shoved down the pipeline? I suspect I'm the only person who actually cares about this package. :)
michaelpj has joined #nixos
<michaelpj>
nope, reboot and it's still wrong
<michaelpj>
wtf
<srhb>
michaelpj: What does id -a give you?
<michaelpj>
doesn't have the groups
gnuhurd has left #nixos ["Killed buffer"]
<michaelpj>
id -a ME does, though
<srhb>
Just to be clear, whoami?
<michaelpj>
michael
<srhb>
(ME?)
<michaelpj>
s/ME/michael/ in above messages
<srhb>
I mean
<michaelpj>
yes
<srhb>
Thanks :D
<srhb>
That was about to get very confusing.
<the-kenny>
seanparsons: my only point would be that the name `releasePath` isn't that clear. Maybe you can add a comment (either in all-packages.nix or in the expression) to document how to get the "full" version from your copy of the game?
<michaelpj>
logging in at a tty gives me the right groups
<the-kenny>
seanparsons: other than that: Just ask, no need to feel bad for that.
<srhb>
michaelpj: How do you normally log in?
<michaelpj>
lightdm
<seanparsons>
the-kenny: Yeah, that's cool. Having been on the other side of the fence I don't want to be a burden. :)
<srhb>
michaelpj if you newgroup (oneofthosegroups) does it show up?
<the-kenny>
seanparsons: usually asking here at some point in the day or dropping a mail in the mailing list helps :) I'll merge your PR as soon as you think it's ready, wether you want to follow my suggestion or not :)
<srhb>
michaelpj: How does lightdm start your user session?
<michaelpj>
um
<srhb>
michaelpj: Custom .xsession?
<michaelpj>
no, whatever the default is
<srhb>
michaelpj: Which display environment do you use?
<michaelpj>
xfce
<michaelpj>
`desktopManager.xfce.enable = true; displayManager.lightdm.enable = true;` is basically it
<srhb>
So, services.xserver.desktopManager.xfce.enable ?
<srhb>
Yeah...
<srhb>
Oh, I wonder, can this be some sort of saved session thing
<srhb>
Does xfce do that?
<michaelpj>
hm, there is a session thing
<michaelpj>
let me have a poke at that
takle has quit [(Remote host closed the connection)]
<srhb>
I don't think it's really plausible tbh.
<seanparsons>
the-kenny: How does that look now?
michaelpj has quit [(Remote host closed the connection)]
<the-kenny>
seanparsons: awesome! Merging it
<seanparsons>
the-kenny: Amazing, does that also include sneaking it into 17.03? :)
<ilpianista>
should I prefer displayManager.sessionCommands over ~/.xprofile?
michaelpj has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] the-kenny pushed 1 new commit to release-17.03: https://git.io/v5U0B
<NixOS_GitHub>
nixpkgs/release-17.03 c4abd6f Sean Parsons: renoise: 3.0.1 -> 3.1.0 (#28380)...
NixOS_GitHub has left #nixos []
* seanparsons
internet high fives the-kenny
<michaelpj>
the plot thickens: 1) reboot, groups are wrong, 2) logout/login, groups are wrong 3) nixos-rebuild switch, groups are wrong 5) logout/login, groups are *right*
<srhb>
michaelpj: o_o
<michaelpj>
I think this must be an xfce thing, because I had also noticed it was losing my settings on reboot, but they also reappear in that final session where the groups are right
eacameron has joined #nixos
<srhb>
That's... So weird.
<__monty__>
When querying multiple packages if any of them doesn't match nix-env returns an error. Is there a way around this, i.e. I'd like it to keep querying for the other selectors?
<the-kenny>
seanparsons: you're welcome :) If you have anything else in the next days, just ping me or ask in the channel here
<srhb>
__monty__: No, I think you have to loop over them then.
<hodapp>
anyone happen to know why both torchPackages.torch and torch exist?
<srhb>
__monty__: Why do you need this though? Perhaps there's a better way to achieve what you're attempting.
<__monty__>
I have a list of packages and I want to see which ones are available and at what versions.
<srhb>
__monty__: Yeah, okay, loop over them.
<__monty__>
Nothing complicated. TBH nix-env's behaviour for this is unexpected though. Why does a single error abort the entire query?
betaboon has quit [(Quit: This computer has gone to sleep)]
<srhb>
__monty__: I find it surprising too. I never use names though, too slow. Simpler to just go by attribute names or check all-packages.nix
<srhb>
__monty__: You may also want to give nox a shot
aanderse-laptop has joined #nixos
<srhb>
It might know how to do this kind of thing
<__monty__>
How would one go by attribute names?
<srhb>
__monty__: In this case it doesn't help specifically, but you could just nix-env-qaP | grep -e foo -e bar -e steam
michaelpj has quit [(Remote host closed the connection)]
<srhb>
Oh, nox is broken on unstable right now. meh.
<neonfuz1>
is there anything like nix for android? lol
<neonfuz1>
by that I mean, not just running nix on android, which is probably possible
<neonfuz1>
I mean like, apps and stuff
<sphalerite>
neonfuz1: there is in my dreams :p
<neonfuz1>
lol right
ee has left #nixos []
<sphalerite>
I want to one day look to see if I can implement something like that because it would be super awesome
<neonfuz1>
a nix android rom could be really damn cool
<sphalerite>
But to my knowledge no such thing exists
<neonfuz1>
one thing is, for adding packages, would you have a gui on the phone for that?
<sphalerite>
Probably not, for myself
<neonfuz1>
yea
<sphalerite>
I'd want something nixops-like so I can just push config onto it from my laptop
<neonfuz1>
is there a way to automate installing many apps from the play store?
<neonfuz1>
besides the little thing that restores your phone from google
<neonfuz1>
IE, give it a list, and install them
<sphalerite>
I doubt it
<neonfuz1>
afaik to install apk files on android you usually have to click confirm, and wait for each one to install, but I mean, if you had root you could probably bypass that somehow
<neonfuz1>
so you could design a rom around it, or maybe just a tool you can install rather than afull rom
<sphalerite>
I don't have any google stuff on my phone and just get my apps through fdroid so that's not a use case I'd cover
<neonfuz1>
oh yeah
<neonfuz1>
nah I embrace the botnet
Sonarpulse has quit [(Ping timeout: 240 seconds)]
<sphalerite>
Anyway I should sleep. Gnight
<neonfuz1>
uh oh
<hodapp>
well, that's a strange arrangement in tensorflow. it has argument 'cudnn', but if you actually use the package called 'cudnn', it is unusable because it's the wrong version, whereas the actual expressions that use tensorflow override it to cudnn51_cudatoolkit80
<neonfuz1>
I got "Failed to check file system type of "/boot": No such file or directory"
<neonfuz1>
then I looked under /mnt/ and there's no boot partition in there
mkoenig has quit [(Ping timeout: 240 seconds)]
<neonfuz1>
boot directory*
<neonfuz1>
am I supposed to manually make this directory, or set something somewhere to make it be made?\
jbo has quit [(Quit: Lost terminal)]
rpifan has joined #nixos
<srhb>
neonfuz1: Yes, you're supposed to partition it, make the directory and mount it there
<srhb>
neonfuz1: You should also regenerate the hardware-config after doing so, or it will be missing from your fstab
<neonfuz1>
wat you guys require boot be a fat partition?
<srhb>
neonfuz1: It's in the manual :)
<neonfuz1>
that's dumb, grub2 can boot from ext4
<srhb>
neonfuz1: Technically no, but since the EFI partition must be, it's easier to do it like that.
<srhb>
neonfuz1: Yes, but then you need a separate EFI partition.
<srhb>
neonfuz1: Besides, what's the problem with using FAT?
<neonfuz1>
I have an efi partition, are the kernels stored in /boot on nixos?
<srhb>
Yes.
<neonfuz1>
why would I want to keep my kernels in my efi partition?
<srhb>
Well, the bzImage
<hodapp>
I made a habit of having /boot as FAT for a long while because of hypothetical cases where I might have to recover my Linux system from just DOS, or some garbage like that.
<srhb>
neonfuz1: Why not?
<neonfuz1>
hodapp: lmao
<neonfuz1>
idk
<srhb>
neonfuz1: idk either.
<neonfuz1>
I mean my last os I just didn't do it this way
<srhb>
neonfuz1: That sounds plausible. :-P
<hodapp>
you folks may not remember the days of syslinux and loadlin
<srhb>
It's just simple like this.
<srhb>
Why complicate matters.
<neonfuz1>
ugh, I'm gonna have to mount my efi partition and make sure there isn't any shit in there
<neonfuz1>
idk what's in there, I think just efi stuff
<hodapp>
does anyone remember the UMSDOS filesystem from the days of dual-booting DOS & Linux?
<neonfuz1>
I used to have my efi partition in /boot/efi/
<neonfuz1>
hodapp: >tfw never used dos
<neonfuz1>
at least until the last few years
<hodapp>
because DOS was so stupid and dull, you could basically use it like a bootloader for Linux if you wanted
<neonfuz1>
okay well here's my real reason as to why not have boot on the efi partition
<srhb>
hodapp: I must have missed that :D
<neonfuz1>
there's already a bunch of efi shit on there, why would I want to risk file conflicts?
<srhb>
neonfuz1: Luckily that problem doesn't apply to NixOS.
<neonfuz1>
I already have "boot" "bootsect.bak" "efi" "mach_kernel" on there
<srhb>
Since it's all managed for you
<neonfuz1>
okay so, if I have a file conflict, it will tell me?
erictapen has joined #nixos
<srhb>
You shouldn't reuse it for anything, I'd say. I expect it'll get nuked from orbit.
Arcaelyx has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<hodapp>
srhb: that's what loadlin let you do. syslinux would also let you load stuff (at boot time) from a FAT partition.
<hodapp>
srhb: with UMSDOS you could let a Linux filesystem reside on FAT without destroying everything
<hodapp>
and coexist with DOS/Windows
<srhb>
hodapp: That... Vaguely rings a bell.
<srhb>
From Mandrake days.
* srhb
shudders
<neonfuz1>
so it's normal for other people to just have their efi partition mounted on /boot?
<neonfuz1>
that sounds insane to me
<srhb>
neonfuz1: It's normal in NixOS
<srhb>
neonfuz1: Probably nowhere else.
<neonfuz1>
efi and boot are 2 different concepts
<srhb>
I don't know why you think it's insane.
<neonfuz1>
because the efi is firmware
<srhb>
neonfuz1: I... Are you arguing philosophy? :-P
<srhb>
No it's not.
<neonfuz1>
and /boot is for the bootloader
<srhb>
You're misunderstanding what the efi partition is for.
<neonfuz1>
what's stored in /boot in nixos?
<srhb>
It's for efi boot loaders. The _firmware_ loads them. Just as it (or BIOS) loads the bootloader from MBR
<srhb>
Or similar concepts.
<neonfuz1>
so the efi bootloaders are in there, of course
<neonfuz1>
but the kernels are too?
<srhb>
neonfuz1: Kernel, bootloader and bootloader configs.
<srhb>
neonfuz1: The bootloader loads the kernel, yes.
<hodapp>
To me, this is no different than making a partition for /boot which is bootable and contains a bootloader, while the filesystem on the partition contains both files the bootloader may need and what it's booting (kernels, initrd and the derivatives)
<neonfuz1>
so are the bootloader configs just placed into /boot, or are they in a subdirectory
<srhb>
neonfuz1: A subdirectory.
<neonfuz1>
which one?
__monty__ has quit [(Quit: leaving)]
<srhb>
neonfuz1: For systemd-boot, /boot/loader/entries
<srhb>
neonfuz1: I don't know if it's the same for grub, haven't used it in a long while.
<hodapp>
originally, this was because bootloaders at that stage of the boot process couldn't always access certain parts of a disk or certain filesystems, but it makes sense regardless
<neonfuz1>
alright, so long as it stores it in a subdirectory that's chosen well, and not just right on /boot, that seems reasonable
<neonfuz1>
I just know a lot of distros assume they can do shit to /boot
<neonfuz1>
like a lot of distros just drop the grub config at /boot/grub
<srhb>
neonfuz1: You'll be happy to hear that the kernel resides in /boot/EFI then :-P
<neonfuz1>
and if you did that with multiple OSes they would screw eachother up
<srhb>
Well, /boot/EI/nixos
<srhb>
EFI*
<neonfuz1>
yeah that's fine
<hodapp>
every distro I've ever used has been rather courteous about not clobbering /boot without making it very clear that it's going to
<hodapp>
it was mostly Windows that would clobber things at random
<neonfuz1>
ehh
<neonfuz1>
yeah windows for sure
<neonfuz1>
but other distros? idk
<neonfuz1>
I mean, I never shared /boot though
<srhb>
hodapp: Do you know for a fact that NixOS doesn't?
<srhb>
I would expect it to.
<neonfuz1>
well, if something else was in /boot/efi/nixos maybe
<hodapp>
srhb: I thought in NixOS you explicitly issued the command to set things up on the partition
aanderse-laptop has quit [(Ping timeout: 248 seconds)]
<srhb>
Sure, creating the partition and mounting it initially.
<srhb>
Everything else is handled by rebuild.
<srhb>
So I'm thinking if something is in EFI or loader, it'll get nuked.
<srhb>
I may be wrong.
<srhb>
Never tried.
<neonfuz1>
srhb: also, I'm going down the manual now, and I see nothing about making a boot partition
<neonfuz1>
maybe it is further in, but under "Chapter 2. Installing NixOS" it just has an example of making a root partition
<srhb>
Perhaps UEFI installs should be considered first-class-citizens these days.
<neonfuz1>
yeah maybe
<neonfuz1>
I guess the thing I don't really think about is, /boot is treated differently in efi installs really
<neonfuz1>
like, /boot is your efi partition if using efi, otherwise it's your bootloader config
mudri` has joined #nixos
stepcut has quit [(Remote host closed the connection)]
<srhb>
neonfuz1: It's more like: In order to treat /boot (the directory) the same for both non-efi and efiboots, it must be vfat on efiboots.
<srhb>
In non-efi you just happen to not need any other partition at all.
<neonfuz1>
true, though historically a lot of people used vfat or ext2 partitions for boot
* srhb
nods
rpifan has left #nixos ["Leaving"]
mudri has quit [(Ping timeout: 255 seconds)]
ntinos has quit [(Remote host closed the connection)]
ntinos has joined #nixos
<neonfuz1>
brb, hopefully booting into my new install
takle has joined #nixos
<srhb>
neonfuz1: Got networking and everything set up? ;-)
gm152 has joined #nixos
<neonfuz1>
No, though it worked in the live environment
<srhb>
Should be OK then.
<neonfuz1>
I forgot, I have irc on my phone lol
chris|_ is now known as chris|
kiloreux has quit [(Ping timeout: 255 seconds)]
<neonfuz1>
I helped a friend try out nixos, he did it on his macbook and had no trackpad drivers
<srhb>
That sounds scary.
stepcut has joined #nixos
<neonfuz1>
I'm getting "a start job is running until udev wait for complete..."
<neonfuz1>
Is that normal?
<srhb>
Upon reboot?
<pxc>
I wouldn't advise people to try out GNU+Linux on Apple hardware, tbf
<neonfuz1>
I missed the full message, but it waited about 2 minutes
<pxc>
they just make no efforts at compatibility whatsoever. You're better of with some whitebox manufacturer you've never heard of
michaelpj has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
<hodapp>
I ran Linux for a bit on my Macbook and enough stuff only-mostly-worked that using it was a bit of a drag
michaelp- has joined #nixos
<pxc>
hodapp: I worry that in many respects, the Apple desktop/laptop ecosystem is (not necessarily intentionally) structured so as to give a very negative impression of free software.
<michaelpj>
srhb: in case you were wondering: I somehow lost the entry that mounted my /boot partition, which resulted in me having a /boot folder full of new boot entries that were doing nothing, and a real /boot partition that was pointing at old configs. So I was always booting into an old config, which got fixed when I switched. Frankly, I'm astonished it wasn't more broken
<pxc>
Part of it is bad hardware compatibility, since Apple use custom/niche hardware here and there and generally have no concern for drivers on any OS but their own. But also free software is harder to get on Mac, and there are fewer high quality open-source applications. It seems like the expectation is on macOS that if you want a reliable, full-featured desktop app, you should be buying it as a commodity (which
<pxc>
means proprietary software). :-\
<neonfuz1>
Yeah
<neonfuz1>
Okay I'm getting ~2 minute hangs on boot and shutdown
<neonfuz1>
The last thing it said was "Started save hardware clock"
<pxc>
srhb: I am the worst. The issues I ran into with our installer were due to me assuming it was finally hung when it was really just waiting for disk because VirtualBox has really bad I/O performance unless you enable passing through the host i/o cache
<neonfuz1>
Still waiting
<pxc>
I was killing it partway through, which caused mysterious behavior
<neonfuz1>
Oh god that's bad
<neonfuz1>
Any ideas what's up with this?
<neonfuz1>
My computer booted in like 10 seconds with my last distro,
<pxc>
neonfuz1: I've not run into that one. On my multimedia PC, NixOS sometimes takes a long time to boot because a couple of my hard drives take a long time to mount. But on my laptop, for example, NixOS boots in like 7 seconds.
<pxc>
neonfuz1: the behavior you've observed is not normal or a typical boot time for NixOS
<neonfuz1>
This is a laptop, with one ssd
<neonfuz1>
And that's it
<neonfuz1>
Yeah I know lol
erasmas has quit [(Quit: leaving)]
tokudan has quit [(Quit: Leaving)]
<neonfuz1>
I was just messing with you guys
<neonfuz1>
Now it says "reached target shutdown"
<neonfuz1>
And its done nothing since
<pxc>
I've never seen that one except on a system where some ACPI-related hardware was damaged
<neonfuz1>
Oh damn the next boot was fast
michaelp| has joined #nixos
michaelpj has left #nixos ["Leaving"]
<clever>
pxc: my machine has a mix of ssd, nvme, and spinning rust, but the bulk of / is on ssd's, [root@amd-nixos:~]# systemd-analyze
<freeman42x>
gchristensen, I think that might be the main one on Gitter, but it is kinda deserted
joehh has quit [(Ping timeout: 255 seconds)]
<gchristensen>
freeman42x: over 1.5yrs, this is the first time I'm hearing of anyone being in there
<freeman42x>
gchristensen, Gitter only recently caught my attention because the vast majority of #scala moved there from IRC. I am assuming that for some reason they must have thought it is better than IRC. That... or maybe some god told them to move there.