<oldandwise>
When doing `nix-shell -p gcc boost poco`, how do we enumerate or know what available stuff? Such as boost? So we get the correct spelling ...
<superbaloo>
not too sure I understand how they're supposed to be used :/
<superbaloo>
meh, nevermind
<freeman42[NixOS]>
which package do I need to install that provides: "error while loading shared libraries: libopenal.so" ? I tried openal but it did not fix the issue
orivej has quit [Ping timeout: 246 seconds]
philr has quit [Ping timeout: 244 seconds]
<freeman42[NixOS]>
so I am sure I need libopenal which is an open library for sound, I just can not find it in nixpkgs
<vika_nezrimaya>
can one build Gradle-based Android apps with Nix somehow?
<vika_nezrimaya>
Ant-based apps seem to have all died out
lambda-11235 has joined #nixos
<freeman42[NixOS]>
ok, it seems like I will be forced to install apt-get and install libopenal with it
<freeman42[NixOS]>
ivan, that is not a solution though - the solution requires knowing which package is providing the libopenal.so
abathur has joined #nixos
<ivan>
freeman42[NixOS]: what makes you think it's not openal?
<ivan>
I see /nix/store/xhk7wzdgfg92raddphb6vmhy22z8m054-openal-soft-1.19.1/lib/libopenal.so
<ivan>
if you were nix-env -iA'ing it or adding it to systemPackages, that is not gonna work
<freeman42[NixOS]>
ivan, "if you were adding it to systemPackages, that is not gonna work" why not? I added stuff before to configuration.nix to provide system level libraries
evanjs| has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
fusion809 has joined #nixos
npmccallum has quit [Quit: npmccallum]
evanjs| has joined #nixos
<ivan>
freeman42[NixOS]: nixos doesn't set LD_LIBRARY_PATH to include libraries in your systemPackages
<{^_^}>
[nixpkgs] @samueldr merged pull request #67895 → sd-image: don't use installer.cloneConfig option that is not imported → https://git.io/fjp8S
<freeman42[NixOS]>
ivan, ah, I am sorry, that is currently too advanced for me. I will have to settle with installing a different image viewer
<ivan>
freeman42[NixOS]: try feh or geeqie
<ivan>
geeqie handles color profiles properly
<tdeo>
is there a type that somehow can merge a single definition like "string" with a list of [ "string1" "string2" ] to become [ "string" "string1" "string2" ] (in any order)?
endformationage has quit [Quit: WeeChat 2.5]
<ivan>
a type?
<angerman>
zimbatm: what's the going-to-be 19.09 branch? If I'd be going to try cross compilation against it, which branch should I use?
<{^_^}>
[nixpkgs] @disassembler pushed commit from @dylex to master « fetchurlBoot: remove cycles introduced by openssl_1_1 -> coreutils (#67784) »: https://git.io/fjpun
mexisme has joined #nixos
freeman42[NixOS] has quit [Ping timeout: 245 seconds]
evanjs| has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
evanjs| has joined #nixos
reallymemorable has joined #nixos
vika_nezrimaya has quit [Ping timeout: 246 seconds]
jkachmar has joined #nixos
<superbaloo>
https://gist.github.com/baloo/96e0726c7920a74f2600e91f1b2d529b I'm trying to cross-compile a nixos for arm64 from a debian x86_64. it feels like nixpkgs/development/libraries/fontconfig/make-fonts-conf.nix is missing a dependency on fontconfig itself.
<superbaloo>
I'm not too sure how to fix that
<superbaloo>
any advice is welcome
MinceR_ has joined #nixos
dansho has joined #nixos
MinceR has quit [Ping timeout: 245 seconds]
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
<li_matrix>
superbaloo: using nixpkgs master? cross compiling stuff is all really bleeding edge
<li_matrix>
im looking at this on my todo so Ill be playing with it in the coming days on master
tjg1 has quit [Quit: if i commit suicide, at least a real nigga killed me]
<superbaloo>
master + a patch that is unrelated to it
<superbaloo>
i guess my question is more about the syntax
<superbaloo>
I dont quite understand the inherit here
<ivan>
superbaloo: try something like export PATH=$PATH:${fontconfig}/bin in the script in make-fonts-conf.nix?
dansho has quit [Ping timeout: 258 seconds]
<superbaloo>
does that means fontconfig = callPackage ../development/libraries/fontconfig { }; will be applied before?
evanjs| has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
<superbaloo>
ivan: looks like it fixes the issue indeed
<ivan>
cool
<superbaloo>
thanks!
<superbaloo>
I'll try to submit a PR for that
evanjs| has joined #nixos
<li_matrix>
superbaloo: inherit there is used to select what to return
dansho has joined #nixos
<li_matrix>
no nevermind
<li_matrix>
its to make the argument optional
<li_matrix>
fontconfig anyway
Rusty1 has joined #nixos
sb0 has joined #nixos
boxscape has quit [Ping timeout: 245 seconds]
ned64 has joined #nixos
drakonis1 has quit [Quit: WeeChat 2.4]
<ned64>
Hey, just wondering how nix manages to increment the generation name (e.g. if latest generation is 'default-59', how does it atomically create 'default-60') without the risk of a race from another process? Links to source code welcome :D
<li_matrix>
ned64: id say thats why theres a daemon
mrCyborg has joined #nixos
<{^_^}>
[nixpkgs] @worldofpeace pushed to master « light-locker: wrapGAppsHook to nativeBuildInputs »: https://git.io/fjpuA
dansho has quit [Ping timeout: 244 seconds]
<ned64>
li_matrix: I'm using single-user mode, with no daemon, so does that mean it's not safe or is there some sort of lockfile involved?
reallymemorable has quit [Quit: reallymemorable]
Okinan has quit [Quit: Konversation terminated!]
evanjs| has quit [Ping timeout: 244 seconds]
LnL has joined #nixos
LnL is now known as Guest95592
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #67915 → Add onboard to Pantheon → https://git.io/fjpux
buckley310 has quit [Quit: leaving]
felixfoertsch23 has quit [Ping timeout: 252 seconds]
polman has quit [Read error: Connection reset by peer]
polman has joined #nixos
thc202 has joined #nixos
zupo has joined #nixos
<{^_^}>
[nixpkgs] @flokli opened pull request #67926 → biber: fix currently broken tests → https://git.io/fjp21
v0|d has quit [Ping timeout: 258 seconds]
polman has quit [Read error: Connection reset by peer]
polman has joined #nixos
boxdot has joined #nixos
rauno has joined #nixos
DariusTh` has quit [Ping timeout: 245 seconds]
Makaveli7 has joined #nixos
drewr has quit [Ping timeout: 264 seconds]
dasj19 has quit [Quit: dasj19]
Makaveli7 has quit [Quit: WeeChat 2.5]
Makaveli7 has joined #nixos
_zincy has quit [Read error: Connection reset by peer]
tom__ has joined #nixos
johanot has joined #nixos
<simpson>
It appears that I can't set `serviceConfig.Type = "exec";` in a systemd service unit. Why not?
civodul has joined #nixos
<simpson>
Specifically, I get a trace message saying that the field `Type` cannot have value `exec`.
<boxdot>
Hi — I am learning nix language right now. I wonder how to reference a rec inside of its definition, i.e I would like to do something like that: rec { x = ..; y = foo this_rec; }.
shabius has quit [Remote host closed the connection]
shabius has joined #nixos
<Taneb>
boxdot: there's no "magic" way of doing that, but you can do "let x = { a = 1; b = x.a + 2;}; in x"
<Taneb>
i.e. you have to give it a name yourslef
<Taneb>
> let x = { a = 1; b = x.a + 2;}; in x
<{^_^}>
{ a = 1; b = <CODE>; }
<Taneb>
> let x = { a = 1; b = x.a + 2;}; in x.b
<{^_^}>
3
<Taneb>
Hope this helps!
vika_nezrimaya has quit [Ping timeout: 252 seconds]
knupfer has quit [Ping timeout: 264 seconds]
drewr has joined #nixos
<etu>
Also this may be revelant for boxdot
<etu>
> rec { a = "a"; b = a; }
<{^_^}>
{ a = "a"; b = "a"; }
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
lord| has quit [Read error: Connection reset by peer]
lord| has joined #nixos
polman has quit [Read error: Connection reset by peer]
DariusTh` has joined #nixos
noudle has joined #nixos
polman has joined #nixos
<boxdot>
ok, thanks.
<boxdot>
still slightly confused by let in and rec
<kraem>
I'm getting this error (https://bpaste.net/show/Xff7) when running `nixos-rebuild`, i'm on unstable. I'm unsure on how to start debugging this. Any pointers? :)
polman has quit [Read error: Connection reset by peer]
DariusTh` has quit [Ping timeout: 252 seconds]
mexisme has joined #nixos
DariusTh` has joined #nixos
chloekek has joined #nixos
polman has joined #nixos
philr has joined #nixos
<maralorn>
kraem: In the nixpkgs repo in file nixos/pkgs/development/libraries/vulkan-loader/default.nix at line 8 is apparently an "assert" statement. You can have a look there, what it asserts.
freeman42[NixOS] has joined #nixos
<simpson>
Aha, "exec" is relatively new and the system I'm working with is freshly installed from ISO.
Giorgio has joined #nixos
asymptotically has joined #nixos
Giorgio has quit [Client Quit]
mjisme has quit [Ping timeout: 246 seconds]
joshuagl has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<kraem>
maralorn: it's asserting my vulkan-headers.version
<teto>
kraem: got the same issue
<kraem>
hm ok. when doing `find /nix/store | grep vulkan-headers` i can see that my vulkan-headers is a newer version than the assert
polman has quit [Read error: Connection reset by peer]
Fendor has joined #nixos
chloekek has quit [Ping timeout: 244 seconds]
<kraem>
oh sorry it's an older version
polman has joined #nixos
polman has quit [Read error: Connection reset by peer]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
polman has joined #nixos
ned64 has quit [Remote host closed the connection]
<Taneb>
Is there a way to do a postFetch-type thing for a Hydra git input like you can do with pkgs.fetchGit?
<kraem>
teto: okok. a fix to bump the version was merged into master 19 hours ago. any idea why updating the channel may not be including the fix? just curious, seems like i'm missing something
polman has quit [Read error: Connection reset by peer]
tom__ has quit [Ping timeout: 244 seconds]
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<teto>
kraem: cool, do you have the commit id ? eventually master will become nixos-unstable, just wait the next channel bump
<chiefgoat>
I am receiving an assertion failed at `vulkan-loader`. Is there a way to see the package depending on this library so I can remove it?
<kraem>
chiefgoat: it seems to be mpv when looking at the dependency tree of nix
chloekek has joined #nixos
jgeerds has joined #nixos
<kraem>
for me at least
<kraem>
chiefgoat: you can run `nix-store -q --tree /nix/var/nix/profiles/system` to find out i think
<kraem>
if you are managing your packages declaratively
<teto>
kraem: ty that fixed it
polman has joined #nixos
<teto>
arf vdirsyncer and flask-caching fail as well :/
drewr has quit [Ping timeout: 264 seconds]
inkOne has joined #nixos
polman has quit [Read error: Connection reset by peer]
zupo has quit [Remote host closed the connection]
polman has joined #nixos
zupo has joined #nixos
jluttine has quit [Ping timeout: 245 seconds]
turion has joined #nixos
<{^_^}>
[nixpkgs] @basvandijk opened pull request #67930 → trivial-builders: add the applyPatches function → https://git.io/fjpah
<kraem>
teto: how did you fix it? did you somehow make sure that commit was included when building your system? if so, how? or did you remove packages depending on vulkan?
jluttine has joined #nixos
<{^_^}>
[nixpkgs] @Izorkin opened pull request #67931 → fail2ban: init at 0.11.dev3-2019-08-22 → https://git.io/fjpVe
polman has quit [Read error: Connection reset by peer]
_zincy has joined #nixos
Chiliparrot has joined #nixos
Makaveli7 has quit [Quit: WeeChat 2.5]
sigmundv has joined #nixos
domogled has quit [Ping timeout: 246 seconds]
drewr has joined #nixos
<teto>
kraem: jsut cherry picked the commit in my local nixpkgs
bakakuna has joined #nixos
Makaveli7 has joined #nixos
joshmeredith has joined #nixos
polman has joined #nixos
polman has quit [Ping timeout: 245 seconds]
jgeerds has quit [Ping timeout: 246 seconds]
polman has joined #nixos
<chiefgoat>
kraem: Thank you! I am managing my packages declaratively.
__monty__ has joined #nixos
<turion>
Any HLedger users here? What's your setup? I'm using nginx as a reverse proxy and I'm considering putting this setup on nixpkgs. Would be interested in your feedback/guidance/recommendations
<Ashy>
is there an env var i should be setting that points to my ~/.ssh/config ?
<immae>
Ashy: it says that you need to set NIX_PATH=ssh-config=/path/to/some/ssh/config that contains the necessary configuration to access the private git repository
<immae>
(it assumes that you access that repository via ssh)
<pbb>
Is it possible to get suid binaries from nix on non-NixOS systems?
<Ashy>
immae: ah ok thanks, where's the manual section on NIX_PATH?
<Ashy>
i wasnt having much luck searching within the nix manual
<immae>
Ashy: I don’t know if there is such thing, but basically NIX_PATH is a list of key=value, which makes <key> accessible in your nix files
<immae>
pbb: for sure not in the /nix/store, so you would need to write a script that copies some outputs to another directory and sets uids there (that’s what nixos do)
<Ashy>
which doesnt mention the key=value syntax at all
<immae>
it does actually
<immae>
it only mentions "nixpkgs" as key
<immae>
but nothing prevents you from putting anything else
<Ashy>
oh yeah nice, thanks
<immae>
(like ssh-config in your case)
<Ashy>
so does setting it append it to the existing value or do i need to append it similar to `export PATH=$PATH:/foo/bar/bin`?
<immae>
it’s an environment variable, so it’s like PATH but for nix
<das_j>
this may sound weird, but can I substitute against a local store? Like /mnt/nix/store?
mexisme has quit [Ping timeout: 258 seconds]
<hyperfekt>
das_j: try --substituters /mnt
<das_j>
hyperfekt: Nice, thanks!
<hyperfekt>
don't thank me before you've tried it :b
<das_j>
what would be even better: is it possible to nix-copy-closure into a store that is not at /nix?
<das_j>
because right now, my local machine pushes to the store of the iso from which deivations are realised and pushed to the other store at /mnt/nix
<{^_^}>
[nix] @edolstra pushed to flakes « nix dev-shell --profile: Support relative path »: https://git.io/fjp6r
<eyJhb>
aanderse: ohh then get ready! :D Would love to get them to use NixOS, but no clue if possible, but 100% not possible if it doesn't support multiple sites :(
<aanderse>
eyJhb: them?
<eyJhb>
aanderse: yes, the organisation I will be spending the next semester helping :)
<aanderse>
ah
<aanderse>
how do you feel about declarative containers? :-)
<eyJhb>
aanderse: Depends on which kind of containers, but I think I would quite like it :p WHy?
<aanderse>
have you worked with containers on nixos at all?
<averell>
i use them to give them their own network via wireguard (they can claim their own interface and move it into the container). pretty great.
Parabox has joined #nixos
<averell>
and isolate closed-source crap, like plex.
<Parabox>
anyone online that has a few minutes they wouldnt mind helping me with an lvm issue?
endformationage has joined #nixos
bakakuna has joined #nixos
<srhb>
Parabox: You're best off stating your problem and seeing if someone who knows how to help answers :)
chloekek has joined #nixos
<Parabox>
lol well its a really long story but short I was making backups of a server with dd piped into gunzip to create an image. I just needed to recover and I extracted that image to another drive and when I try to moung the lvm I get cant read superblock
<Parabox>
So I have done xfs_repair and still cant get it to mount
<clever>
Parabox: does blkid identify the fs correctly?
<eyJhb>
averell: not trusting the closed source? ;)
<Parabox>
well it sees it as an lvm member
<averell>
not more than i have to at least
<kaliumxyz>
firefox is unkillable
<clever>
Parabox: what does `pvscan` say?
<kaliumxyz>
my system has a process on it called firefox-wrapper which cannot be killed :d
<clever>
kaliumxyz: does the cpu usage overall agree?, zombie processes cant consume cpu
<kaliumxyz>
I have no way of testing that beyond looking at what tools report.
<clever>
kaliumxyz: `ps -eH x`, what is the parent of firefox-wrapped?
<Parabox>
If you want to get into the server Its a rebuild so Im not opposed to giving you acess if you dont mind leting me watch you work because Im kind of at a loss on this and Im desperate
<kaliumxyz>
Systemd.
<Parabox>
you can use screen to let me attach to your tty session.
<clever>
Parabox: did you run blkid on the lv or the pv?
<Parabox>
it sais that a bunch of times then runs for hours then it sais
<Parabox>
.Sorry, could not find valid secondary superblock
<clever>
Parabox: how was the backup made? how was it restored?, do you have the exact commands saved?
ee1943 has quit [Read error: Connection reset by peer]
Ariakenom has quit [Read error: Connection reset by peer]
<kaliumxyz>
Thanks for the help clever c:
ee1943 has joined #nixos
bakakuna has quit [Ping timeout: 245 seconds]
<clever>
kaliumxyz: yep
<boxdot>
Do I understand correctly that nix would rebuild my packages from the same repo if something changes in the repo?
<boxdot>
(using fetchGit)
<srhb>
boxdot: What do you mean "same repo" ?
<boxdot>
I have a mono repo with several packages
bakakuna has joined #nixos
<eyJhb>
boxdot: it does not check if the repo changes
<srhb>
boxdot: Normally you would fetchGit using a permanent immutable reference.
<clever>
bakakuna: pkgs.fetchgit or builtins.fetchGit?
<srhb>
boxdot: In which case there's no rebuild (because nothing changes ever)
<boxdot>
ah, there are two. Right now I am using builtins.fetchGit
<clever>
bakakuna: with what arguments?
<Parabox>
yeah I had dd if=/dev/sda conv=sync,noerror bs=64K | gzip -c > /backup/jackbox.tar.gz running as a cron job once a week. after the crash I rebuilt the server and moved the tar ball to the new server then I ran gunzip -c jackbox.tar.gz | dd of=/dev/sda (esentially laying the image back down on the same drive it came from)
Ariakenom has joined #nixos
<clever>
Parabox: that backup wont be atomic, so the data at the start of the disk wont match up with the data at the end of the disk, which can lead to the entire backup being unusable
<Parabox>
oh man please dont tell me that
<Parabox>
like this is about 10 years of important documents pictures like everything
<clever>
Parabox: you must run `zfs_freeze` before doing that, and then un-freeze it afterwards
<clever>
Parabox: to ensure nothing changes while making the copy
<Parabox>
so what can I do now?
waleee-cl has joined #nixos
<Parabox>
will a data recovery company be able to get this data back?
<clever>
Parabox: possibly, i would also recomend just switching to zfs, since it has backup stuff built into it, that wont have these issues (snapshots, send/recv)
<Parabox>
well I think Im going to be moving this shit to an s3 bucket if I can recover it
<Parabox>
is there nothing I can do at home first before calling a data recovery compoany?
<kaliumxyz>
how can systemd be the cause of this tho clever?
<kaliumxyz>
Or rather, be involved in this
<clever>
kaliumxyz: systemd didnt call waitpid() on its children
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<kaliumxyz>
oh okay.
<clever>
Parabox: you would need somebody that knows the xfs internals heavily, to be able to recover something like this
<kaliumxyz>
thanks
<Parabox>
any clue where I might be able to find this person?
chloekek has quit [Ping timeout: 244 seconds]
_zincy has quit [Ping timeout: 244 seconds]
<clever>
Parabox: my only guess would be the xfs irc channel
<Parabox>
its not comming up on the channels list in freenode, do they have their own server?
<ddima>
(not with XFS though)
<Parabox>
thank you ddima its wortha shot for me
<qyliss>
Lots of channels are +s which means they don't show up in /list
<qyliss>
Makes it harder for spammers to find you, etc.
bakakuna has quit [Remote host closed the connection]
<ddima>
Parabox: as a small side note, possibly for the future - LVM supports snapshots, which can be used to obtain a frozen state to backup from. Bun in general, dont consider a backup existing before you have performed tests of restoring it.
philr has quit [Ping timeout: 245 seconds]
<Parabox>
well ddima I did something even stupider than what I have shared, I wa also keeping file level backups, I wrote over them to add that drive to an arrey on the new rebuilt server because I was going to recover from this image then use s3 to start running file level backups
<Parabox>
so I should absolutly not be in this situation but I was really really stupid
<ddima>
happens
drakonis1 has joined #nixos
t58 has joined #nixos
<Parabox>
shouldnt to me lol
schjetne has quit [Ping timeout: 245 seconds]
<Parabox>
I should have known better and I cant even start to tell you how important this data is
lsyoyom has quit [Ping timeout: 245 seconds]
<Parabox>
the tax papers and that kind of shit I can get back but this is just about every picture and video I have taken my entire life. From vacations to my daughter who is now 13
<Parabox>
this is just going to be devistating to me and my wife if I cant fix this
<Parabox>
and Im a linux admin for a living :-O
<ddima>
Well, make sure to keep an unmodified copy of the image before you try anything. Usually its possible to recover quite a bunch if just small parts of the FS metadata are corrupted. Best of luck :)
<ddima>
always takes a small reminder every few years
bakakuna has joined #nixos
<Parabox>
thank you
fusion809 has quit [Read error: Connection reset by peer]
<simpson>
Parabox: You are right to think of a data recovery company. The overall goal is to use a tool which can statistically examine your backup files to try to guess them. The techniques aren't super-hard but AFAIK there's no good Free Software tool for doing it.
<simpson>
Out of curiosity, have you tried doing a loopback mount of a disk image yet? They *might* be mountable that way.
<Parabox>
you know what im thinking happened? Because dd was overwriting the image if the image got smaller then it wouldnt have overwrote the entire image right?
incognito999 has joined #nixos
<simpson>
Sure.
fusion809 has joined #nixos
<Parabox>
I havnt, I will try that. Im re extracting it to the drive as we speek, I wanted to try again with a clean image I havnt altered in anyway
<Parabox>
so if my theory is correct I could use a tool to try cutting some bits off the end of the image and if I cut the from the right bit then itll be a perfict image right?
<Parabox>
I know thats not realistic because we are talking about billions of bits potentially but its possible right?
kaliumxyz has quit [Remote host closed the connection]
cjpbirkbeck has joined #nixos
<superbaloo>
I run 2.2.2 though
<Ariakenom>
me noob. I tried to add displaylink to configuration.nix but I get stopped by "[...]linux/evdi/default.nix:28 is marked as broken, refusing to evaluate."
<superbaloo>
should I run master?
<Ariakenom>
allowBroken yes or no?
<clever>
superbaloo: i believe the platform thing is in master now
Tucky has quit [Quit: WeeChat 2.5]
<srhb>
Ariakenom: It's unlikely to help you. Things are usually marked broken because, well, they are. Either at run- or, more often, build time.
<Ariakenom>
right
<superbaloo>
clever: alright, thanks
<Ariakenom>
thanks
hmpffff has quit [Quit: nchrrrr…]
<srhb>
Ariakenom: fwiw I was able to get it running some months back, using a lot of hackery with various ubuntu packages for displaylink. When it finally worked, it was a decidedly poor and buggy experience, and I decided that DisplayLink is the devil.
<srhb>
Ariakenom: (in other words I suggest you try and see if there's another, less evil method you an use)
<Ariakenom>
srhb: yeah I've had colleagues fail to get this dock to work with linux so I wasn't too hopeful from the start
<Ariakenom>
srhb: I'll have to live without my 2nd monitor for a bit
<Ariakenom>
srhb: Actually I had already disconnected one monitor when on windows. It was noticably laggy.
Okinan has joined #nixos
<Ariakenom>
(and connected it straight to computer)
kaliumxyz has joined #nixos
orivej has quit [Ping timeout: 245 seconds]
<hyperfekt>
i still have an encrypted luks partition on this hard drive that doesn't have a partition entry (thanks to dd lol)
<clever>
hyperfekt: there are some utils to search for fs headers, and discover where partitions should begin
mawis has joined #nixos
<hyperfekt>
clever: Yeah I tried the usual utils and they found some stuff but it turned out to not be that
<mawis>
Has someone here GitLab installed on NixOS 19.03? While I could call the gitlab:backup:create job on 2018-08-11 it fails today with „NameError: uninitialized constant Gitlab::Settings”. – I wonder if someone else experiences the same problem.
boxscape has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 26.2)]
<Stuck_>
Hi, scrub here, I'm trying to install haskell packages by nix and I'm lost now, I'vs installed the random package, but when I try to import it (System.Random) and doesn't find it, when I do "nix-shell -p 'haskellPackages.ghcWithPackages (p: [p.random])'" it surprisingly works, how can I import it in my env?
<clever>
Stuck_: you must use the ghc provided by haskellPackages.ghcWithPackages, installing the library wont work, installing a normal ghc wont work
<evanjs>
Is there anything extra that needs to be done to enable DBUS services in NixOS with i3? Helping ashkitten debug openrazer support, which works fine for me under xmonad, but when running their config under nixops, I can reproduce the issue (
<Stuck_>
ok, I've removed ghc by nix-env -e ghc and introduced in my config in the environment.systemPackage (haskellPackages.ghcWithPackages (p : [ p.ghc p.random ])) and still random doesn't import
casaca has quit [Ping timeout: 246 seconds]
<clever>
Stuck_: if you remove it from systemPackages, does `type ghc` say its gone?
casaca has joined #nixos
<mikky>
hey, got trouble with nixops - one package in `environment.systemPackages = with pkgs; [ ... ]` is seemingly ignored while other package gets installed correctly. Specifically `[ rainloop-community htop ]`. `htop` gets installed while `rainloop-community` does not make any difference at all. The definition is in an included module and is host-specific.
<Stuck_>
I don't have ghc in my systemPackages
<clever>
Stuck_: does `type ghc` agree with you? what does it say?
<mikky>
clever: don't know, I was able to install it on my local machine without any trouble
<clever>
mikky: ls -l /run/current-system/sw/index.php
<jonge[m]>
corpix: i don't know how to solve this, but once you got it i would be very very interested in seeing this in the nixos wiki or somewhere.
<mikky>
clever: `no such file or directory`
<clever>
mikky: run that nix-store command on the machine with nixops, not the target
<clever>
mikky: what files do you expect rainloop-community to provide once installed?
drakonis_ has joined #nixos
<mikky>
clever: something like `/nix/store/<hash>-rainloop-community-1.x.x/` where the full source from the tarball would be. iirc, this is what it does on the local machine.. let me check
<clever>
mikky: rainloop-community isnt meant to go in systemPackages, your meant to use it in an expression that configures the http server
<clever>
mikky: which will insert the proper path into the config files
<clever>
mikky: i think the problem, is that rainloop-community doesnt contain any bin directories, so systemPackages only depends on it at buildtime, not runtime
<clever>
mikky: and nixops is copying the runtime closure, so it ignores rainloop-community
drakonis has quit [Ping timeout: 252 seconds]
<mikky>
clever: that was my intention in the end - put the path using something like "${pkgs.rainloop}/index.php" to nginx config
<mikky>
clever: but the absence of the package had me apparently chasing ghosts
ryantrinkle has quit [Remote host closed the connection]
<evanjs>
does DBUS work on i3 under NixOS without any extra work? Debugging a dbus service and d-feet cant event connect to it, but both it and dbus seem normal per systemctl status
Soo_Slow has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @jtojnar opened pull request #67957 → glib-networking: clean up & add installed tests → https://git.io/fjp1m
Soo_Slow has joined #nixos
<mawis>
After downgrading and upgrading NixOS/gitlab combined with restarting everything, the backup is now working again. (No configuration changes.) … Strange.
<Phillemann>
Hmmm, assertion failed at /nix/store/q9zzj3jmghbix58968vcf1jn9dsw8wmz-nixexprs.tar.xz/pkgs/development/libraries/vulkan-loader/default.nix:8:1
<Phillemann>
Latest nixos-unstable
<pbb>
that's fixed on master
<Phillemann>
Ah, okay. I'll postpone my update then, thanks ;)
o1lo01ol1o has quit [Ping timeout: 245 seconds]
mikky has quit [Ping timeout: 276 seconds]
<Phillemann>
(How did that get past the unstable tests though?)
<pbb>
I guess no test evaluates that part
<pbb>
you'll have to either wait a while for the unstable channel to update, or do "git checkout 7d5375ebf4c; git cherry-pick e54007370fafe4b0d50ce4f0b5dcfab34e80042d; sudo nixos-rebuild switch -I nixpkgs=$PWD"
<clever>
Phillemann: so the error will appear in this page
<pbb>
clever: I don't think so, because it's not part of the packages that get built by hydra
<clever>
one glance at that will tell you why nobody noticed an error
<clever>
pbb: if its in the pkgs tree, then the nixpkgs jobset will get it
<clever>
and this is a combined jobset, nixpkgs + nixos
<srhb>
pbb: It is, but it's not part of the blocking set..
<Phillemann>
clever: Oh, damn.
<Phillemann>
_Everything_ is broken according to that tab :D
<clever>
Phillemann: feel free to file PR's to get those errors down
<srhb>
Phillemann: That will always be true for that tab :-)
<clever>
Phillemann: for unfree packages, you want to set hydraPlatforms, so its still marked as "working" but hydra wont try to build it
<Phillemann>
I can hardly tell the authors of the packages that ther should change their licenses though.
<srhb>
heh
<srhb>
You *can* :P
<Phillemann>
Yeah...that'd be the _ideal_ solution anyway
<Phillemann>
I'll start with Nvidia. Damn closed source blobs in my kernel...
<clever>
Phillemann: my system locks up solid if i unplug a monitor before turning it off
<Phillemann>
clever: Any more luck with nouveau though?
<Phillemann>
Anyhow, seems silly to report all those errors every time.
<clever>
Phillemann: amd card
blitzclone[m] is now known as blitzclone_
bgamari has quit [Ping timeout: 245 seconds]
casaca has quit [Ping timeout: 258 seconds]
bgamari has joined #nixos
schjetne has joined #nixos
casaca has joined #nixos
<fresheyeball>
anyone running selenium tests in a nix-build?
<{^_^}>
[nixpkgs] @ambrop72 opened pull request #67958 → houdini: Add the GPU driver to LD_LIBRARY_PATH. → https://git.io/fjpMJ
schjetne has quit [Ping timeout: 264 seconds]
polman has quit [Ping timeout: 245 seconds]
polman has joined #nixos
Ariakenom has joined #nixos
<fresheyeball>
oh wait
<fresheyeball>
I think I know the problem
<fresheyeball>
I need XAUTHORITY
<fresheyeball>
and I don't have it in the nix build
<clever>
fresheyeball: you probably need to launch an entire X server
<fresheyeball>
clever: I am trying to run chrome --headless
<clever>
fresheyeball: the nixos test cases can do that
<clever>
if your dealing with XAUTHORITY, then your not headless
<fresheyeball>
but I think it still needs x
<clever>
the nixos tests let you just enable xorg as normal
vyorkin has joined #nixos
<mightybyte>
Is specifying `--option trusted-public-keys` on the command line not allowed?
<fresheyeball>
hey mightybyte
<clever>
mightybyte: it only works if you are a trusted user in the main nix.conf file
<clever>
mightybyte: root is trusted by default
<fresheyeball>
guess we both looking for the clever gold
mikky has joined #nixos
aither has joined #nixos
<mightybyte>
Hmmm, let me see...
knupfer has quit [Ping timeout: 264 seconds]
<m1cr0man>
Hey folks does anyone use the weechat service? How do I connect to the screen session it starts? It seems to use a setuid wrapped call to screen in the systemd service.
<ivan>
can I replace Linux machines by somehow getting Linux to stop all reads from the block device and then writing over /dev/sda
<Ralith>
is there a specific reason you need to reformat?
<Ralith>
usually easier to leave the filesystem intact
<clever>
ivan: zfs and xfs support some of that, so it heavily depends on what fs your using
<ivan>
yeah, I prefer to have non-LVM XFS instead of the LVM ext4 I get from IONOS
<clever>
ivan: ah, do you want to wipe the disk and install nixos on it?
<ivan>
yes
<clever>
ivan: then you want my kexec stuff
<vyorkin>
Have anyone tried making derivations for OCaml packages? I want to make one for the ocaml-rpc, but the ocaml-rpc
<vyorkin>
package contains smth around 5 libraries and I'm not sure how to
<vyorkin>
organize such derivations
<fresheyeball>
clever: ok I am now sure that I need a x server even for headless chrome, because its needed for keycodes
<ivan>
clever: I've looked at that and it looks like something that could work outside your setup unless I rewrite stuff
<evanjs>
m1cr0man: wait there’s a weechat service 😯
<clever>
ivan: ?
Jackneill has quit [Remote host closed the connection]
<pbb>
ivan: it works perfectly, just look at the stuff in session.md
<ivan>
ok, I'll give it a try largely unchanged
glittershark has joined #nixos
<pbb>
only change you have to do is add ip config (if not dhcp) and ssh key to configuration.nix
zupo has joined #nixos
<glittershark>
Hey all - I'm getting "ghc: can't find a package database at /home/griffin/.cabal/store/ghc-8.6.5/package.db" when trying to nix-build a Haskell package with Nix
<evanjs>
m1cr0man: looks like `services.weechat.sessionName` defaults to “weechat-screen”. Can you just `screen -r weechat-screen`?
Guest81054 is now known as benny
<clever>
glittershark: you have some new-build junk in the dist dir, `rm -rf dist` and try again with `cabal configure` and `cabal build`
<glittershark>
so when developing I'm running `cabal new-build` etc. inside of the nix-shell
<glittershark>
should I be using v1-build instead?
<clever>
glittershark: yeah
schjetne has joined #nixos
<glittershark>
oh ok
<glittershark>
rm -rf dist dist-newstyle && nix-build is giving the same error as before
<clever>
glittershark: is there any hidden files like `.ghc*` ?
<glittershark>
ah, yes, I removed that and it's working now
<glittershark>
thanks! I assume that's what was pointing to the package db in my ~ ?
<clever>
new-build also makes that file, which breaks old-build
<fresheyeball>
I think this should work
<fresheyeball>
Xvfb :10 -ac &
<fresheyeball>
DISPLAY=:10
chloekek has joined #nixos
<clever>
fresheyeball: Xvnc is another hack ive used before
<clever>
fresheyeball: its a headless X server, with a vnc server baked into it
<fresheyeball>
clever: do you think that should work? do far no change
<glittershark>
hmm is there anyway I can configure nix-build to use new-build, or is that just not supported
<clever>
fresheyeball: both Xvfg and Xvnc should work for your issue
zupo_ has joined #nixos
<clever>
glittershark: nixpkgs only supports the old build
<glittershark>
k
<glittershark>
I guess that's just my happy path for now
<glittershark>
np, thanks for your help clever
fusion809 has quit [Remote host closed the connection]
zupo has quit [Ping timeout: 244 seconds]
o1lo01ol1o has joined #nixos
Melkor333 has joined #nixos
schjetne has quit [Ping timeout: 264 seconds]
<fresheyeball>
clever: I have not used Xvfg before
<fresheyeball>
so I need to worry about XAUTHORITY?
DariusTh` has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<clever>
fresheyeball: you have to point XAUTHORITY to a file, before you run the X server
<clever>
fresheyeball: i believe the X server will then generate the cookie and write it to that file for you
<clever>
fresheyeball: though you may need to run the x server with -auth pointing to the file
<Melkor333>
Is it possible to set an option within a package derivation, if the package is installed on NixOS? I would need to set `programs.dconf.enable = true` for a package I maintain and it would be nice if I don't have to do it manually..
<clever>
Melkor333: not really, you just have to tell users to do that
<clever>
Melkor333: you could also make your own nixos option, that both sets settings, and puts your package into systemPackages
DariusTh` has quit [Ping timeout: 245 seconds]
<Melkor333>
Oh okay thank you! :) Yeah the other way around it wouldn't really make much sense!
casaca has quit [Ping timeout: 258 seconds]
jtojnar has joined #nixos
casaca has joined #nixos
<m1cr0man>
evanjs, Sorry went afk there. Yeah I tried that, I also tried "sudo -u weechat screen -list" and screen -list as root but no dice. My bet is that weechat is failing to start?
<clever>
m1cr0man: check `systemctl status` on that service
<m1cr0man>
clever, says it's active/rc 0. it's running screen -DmS so it's forking the process and it looks like it's not set up to track any other pid
<clever>
m1cr0man: what user is screen running as, `ps aux | grep -i screen` ?
<clever>
-D -m This also starts screen in "detached" mode, but doesn't fork a new process. The command exits if the session terminates.
<m1cr0man>
hm, no one apparently. There is no screen process
<m1cr0man>
oh right
<clever>
-S sessionname
<m1cr0man>
right so weechat is broken
sm has left #nixos ["ERC (IRC client for Emacs 27.0.50)"]
<m1cr0man>
let me see if I can run it without screen a second
<clever>
m1cr0man: also, `sudo -u weechat screen -x` will fail, because screen cant open /dev/tty, which is owned by the wrong user
<clever>
m1cr0man: you need to change the owner of whatever `tty` prints out, before you can successfully do it
<clever>
m1cr0man: but screen does have a multi-user attach mode ive never used, that can smooth over those issues
<m1cr0man>
clever, Yeah I'm aware of the permissions error but nevertheless -list returned nothing
<m1cr0man>
hm multi user attach might be useful
growpotkin has joined #nixos
<m1cr0man>
ugh got it. "Error: cannot create directory "/.../weechat/.weechat"
<m1cr0man>
Now interestingly the folder is owned by weechat
<clever>
m1cr0man: i also have my own weechat service i wrote a while back
<esclear>
It seems to download the extensions, but when I start visual studio, they aren't installed.
<esclear>
s/visual studio/(vs) code/
<sphalerite>
systemd doesn't want to go into normal mode because of a missing filesystem. How can I convince it to proceed anyway?
<{^_^}>
[nixpkgs] @gnidorah opened pull request #67959 → qmmp, dfilemanager: use qt5's mkDerivation → https://git.io/fjpDt
<sphalerite>
(missing filesystem being /boot, which is annoying but not critical for system startup at the point it's currently reached…)
<clever>
sphalerite: you need to add a nofail thing to the fstab options
<sphalerite>
clever: yes, but I can't currently rebuild my system because of this
<sphalerite>
aah but I can add it into fstab in the ugly way
qqlq has quit [Ping timeout: 245 seconds]
<clever>
sphalerite: yeah, delete the /etc/fstab symlink, replace it with a file containing the same contents, then edit-away
<clever>
but nixos will undo that at the next boot or switch
<clever>
i also discovered, that if you add invalid entries to fstab (i had escaping trouble with a space in a mount point), systemd will happily kill the entire machine, the instant you rebuild-switch
<m1cr0man>
^ been there too ;)
tokudan1 has joined #nixos
<m1cr0man>
was fun when I was setting up gluster
<sphalerite>
clever: sed -i is nice enough to create a temporary file then rename() it into place, sed -i /boot/d /etc/fstab :D
<m1cr0man>
hm, that's really cool
tokudan1 is now known as tokudan[m]
<m1cr0man>
Oh jeeez actually I don't know if that is cool XD I'll have to remember that sed -i on a symlink replaces the symlink and not the original file
<{^_^}>
[nixos-hardware] @exi opened pull request #118 → Add better pulseaudio support for thinkpad t490 → https://git.io/fjpDY
<clever>
oh, nice
emptyflask has joined #nixos
andgate has joined #nixos
evanjs| has quit [Ping timeout: 245 seconds]
<fresheyeball>
ok
<fresheyeball>
so Xvfb dies wit this
<fresheyeball>
_XSERVTransmkdir: ERROR: euid != 0,directory /tmp/.X11-unix will not be created.
<fresheyeball>
I can get rid of the error by passing -nolisten unix
infinisil has quit [Read error: Connection reset by peer]
nikivi has joined #nixos
<m1cr0man>
clever, evanjs turns out the issue with the weechat service was that the parent directory of its data directory did not have sufficient permissions. Found by `sudo -u weechat bash` then just cd home
Phillemann has quit [Ping timeout: 264 seconds]
andi- has quit [Ping timeout: 264 seconds]
Myhlamaeus4 has joined #nixos
aither has joined #nixos
Phillemann has joined #nixos
<pbb>
is there some utility function in nixpkgs to merge two derivations?
<pbb>
(or actually more than two)
<clever>
pbb: nixos option definitons or plain sets?
martyet-o has joined #nixos
<pbb>
derivations actually
ehmry has joined #nixos
jtojnar has quit [Read error: Connection reset by peer]
o1lo01ol1o has quit [Remote host closed the connection]
<clever>
pbb: why do you want to merge 2 together?
Lem_ has joined #nixos
andi- has joined #nixos
<thoughtpolice>
pbb: I'm not sure exactly what you mean, but if you just want an output directory in the store that contains all the artifacts of two (or more) derivations, you can use pkgs.buildEnv, which is perhaps what you want?
<pbb>
yup thanks
<pbb>
that's what I want
<thoughtpolice>
OK cool. The two derivations will need to output a directory of files (some only output a single file but that's rare) but other than that you should be good to go
<pbb>
clever: coreboot has quite some utility programs, which are usually packaged in one coreboot-utils package. In this case it's much easier to build each one seperately as an own derivation and merge them together to become coreboot-utils in the end.
<pbb>
(because each one has their own makefile and quirks)
<clever>
pbb: ah, yeah, definitely a use for buildEnv
<thoughtpolice>
pbb: Ah fun. Are you building the cross compilers in a separate derivation, by chance?
<pbb>
no it's just some utils like ifdtool, cbfstool, cbmem and so on
<thoughtpolice>
I was actually fiddling with a coreboot cross build the other day and was wondering if I could separate the two but didn't try hard. In the end I used my own and set CONFIG_ANY_TOOLCHAIN=y
<thoughtpolice>
ah ok
<pbb>
to build coreboot I usually use the coreboot ci docker image, because we don't have a working ada compiler in nixpkgs :(
<clever>
pbb: cant grub already do that with its normal background image stuff?
<pbb>
yeah but at that stage there is no graphics if you can not build the graphics init code
<pbb>
so you would not see anything until the kernel starts
zupo_ has quit [Ping timeout: 245 seconds]
<clever>
pbb: x86 or arm?
<pbb>
x86
<clever>
pbb: i would expect most x86 bios to provide working graphics
<pbb>
yeah but I'm building my own bios, that's what coreboot is about
<clever>
enless that coreboot was very minimal?
levdub has joined #nixos
o1lo01ol1o has joined #nixos
<pbb>
clever: coreboot includes code for graphics initialization, but that's writteen in ada, so it can not be compiled on NixOS
<clever>
ahhh
<pbb>
or I could rely the proprietary vgabios blob but meeeeh
dasj19 has quit [Quit: dasj19]
o1lo01ol1o has quit [Remote host closed the connection]
<thoughtpolice>
clever: Yeah it's just coreboot's way of doing it AFAIK. I'm guessing Grub can do the same thing if you booted with UEFI but you can use Grub directly as a coreboot payload too
<thoughtpolice>
Or you could chain load a bunch I guess.
<clever>
thoughtpolice: yeah, ive heard about the whole payload stuff, and i think there was something about how you need the seabios payload if you want the legacy bios api available to grub
<pbb>
even if you wanted to do "coreboot -> tianocore -> grub", you would need some kind of graphics init in coreboot, since tianocore and grub rely on the graphics already being initialized
<thoughtpolice>
pbb: ah that's a good point, you're right.
kp__ has quit [Quit: WeeChat 2.5]
<pbb>
using grub as primary payload and loading the NixOS grub config from the boot partition is very awesome
<clever>
pbb: that reminds me of a recent bug i discovered with systemd-boot, lol
<thoughtpolice>
clever: Yeah I'm not using it for x86 at all at the moment, and this is my first foray with coreboot so I'm not too familiar with payload specifics on x86
<clever>
pbb: nixos grub uses /boot/grub/state to detect when the version in the boot partition is out of sync with nixos, and auto-reinstalls the bootloader
<pbb>
it's a bit hard to read, but on the picture you can see the first boot entry is "Load config from (ahci0,gpt1)", which is my NixOS boot partition
<clever>
pbb: but systemd-boot does have anything like that, at all
o1lo01ol1o has joined #nixos
<clever>
so if you switch from grub to systemd-boot, it doesnt even install the systemd-boot binaries!
<pbb>
oops :D
<clever>
so you only get the config file, with no binary
<thoughtpolice>
pbb: If you use grub2 as the coreboot payload directly, do you configure nixos to just write the grub config? And then upgrade grub/coreboot at once? Or does the payload you use load grub2 from a partition?
<clever>
pbb: which is actually what you want, when grub is the payload, the cfg, without the loader!
psyanticy has quit [Quit: Connection closed for inactivity]
<clever>
thoughtpolice: ive yet to find a motherboard in my collection that is supported
<pbb>
thoughtpolice: I do boot.loader.grub.device = "nodev";
<clever>
that will instruct grub to not update the stub in the MBR
<pbb>
yeah
<clever>
but it will still install legacy-bios modules to /boot/grub/
<clever>
which costs you a whole 12mb!!
<thoughtpolice>
clever: In theory the RISC-V device I have can run Coreboot on, but it's pretty spotty right now. The best bet at the moment is probably a chromebook honestly.
<pbb>
yeah I don't care if they're there
zupo has joined #nixos
<pbb>
thoughtpolice: that reminds me of the boot chain of the nintendo switch
<fresheyeball>
clever: so if I run Xvnc in a nix-shell
<fresheyeball>
its seems to turn on and be happy
<fresheyeball>
using Xvnc :10 -listen unix
<fresheyeball>
but if I do it in a nix-build
reallymemorable has joined #nixos
<fresheyeball>
I get _XSERVTransmkdir: ERROR: euid != 0,directory /tmp/.X11-unix will not be created.
<clever>
fresheyeball: same problem then, its upset about the uid's of /tmp things, try `--option sandbox false` ?
<pbb>
nintendo switch be like: BootROM Exploit → coreboot loader → coreboot → ARM trusted firmware → coreboot → u-boot → Linux
<thoughtpolice>
pbb: I have a Tegra myself on my desk, actually. I assume it's uh, very special, like most ARM boot procedures.
<thoughtpolice>
Yeah I saw that lol.
superbaloo has quit [Ping timeout: 245 seconds]
<pbb>
I still want to write a NixOS module for booting on the Nintendo Switch
<thoughtpolice>
There were some reports of NixOS working fine on the X1 already with "little" special help, so that's promising.
isHavvy has joined #nixos
<pbb>
but I partially lost motivation because this tegra graphics stuff is so meh
<thoughtpolice>
I have the X2 which was a bit spottier, I have no idea if anyone has tried anything on Xavier yet.
dansho has quit [Remote host closed the connection]
fpob has quit [Ping timeout: 250 seconds]
<pbb>
thoughtpolice: I just replaced the ubuntu root with a NixOS, and it booted up fine, but graphics are a bit harder and I really don't want to mess around with this proprietary nvidia shit too much
dansho has joined #nixos
<thoughtpolice>
The graphics are nice and I wish they would just give you a simple driver package like on the desktop!
<pbb>
yes
polman has quit [Ping timeout: 245 seconds]
<pbb>
I think they only provide a complete Ubuntu 18.04 root image with the drivers installed (L4T/Linux 4 Tegra), right?
polman has joined #nixos
<thoughtpolice>
The NixOS shit for Nvidia is horrid already but if they did that we could in theory just reuse most of it. They're allegedly bringing CUDA to ARM but I doubt they care about GeForce that much.
<thoughtpolice>
Basically, yeah. For the X1 you need the kernel packages at least for NixOS, which you can download separately (somewhere after clicking 50 links on nvidia dot com)
Haskellfant has joined #nixos
<thoughtpolice>
The graphics would actually require unpacking/scraping stuff from the BSP images
<pbb>
meeeeeh
<pbb>
can nouveau run tegra?
o1lo01ol1o has quit [Remote host closed the connection]
Havvy has quit [Ping timeout: 268 seconds]
<thoughtpolice>
No, I think all Tegra chips you can buy have signed firmware preventing nouveau from doing much. The situation is probably even worse than on desktop, if I had to guess.
feep has joined #nixos
<pbb>
why did I buy this device and support nvidia with their proprietary bullshit >.>
<pbb>
Linus was right
<feep>
yyep
<pbb>
we are talking about tegra stuff which appearantly is even shittier than the desktop parts
mac10688 has joined #nixos
<pbb>
well, I'll get back to coreboot-utils
<Yaniel>
I thought it was the other way around
fpob has joined #nixos
<thoughtpolice>
I do run Nvidia and personally on NixOS, it works... surprisingly well. For pure mobile options it might be best to wait and see how some of the new Mali drivers in Mesa (Panfrost, etc) works out over the next year though for FOSS options.
<thoughtpolice>
Yaniel: There are upstream drivers for things like the actual chipsets that the Tegra includes, which Linux has to know about. But the actual "graphics stack" is still just as closed source as the desktop.
ehmry has quit [Ping timeout: 245 seconds]
Thra11 has quit [Quit: WeeChat 2.5]
y0x3y[m] has quit [*.net *.split]
mariusz-kowalski has quit [*.net *.split]
jonreeve[m] has quit [*.net *.split]
kyleondy[m] has quit [*.net *.split]
AntonSchirg[m] has quit [*.net *.split]
infinisil-m has quit [*.net *.split]
Fluffy[m] has quit [*.net *.split]
corpix[m] has quit [*.net *.split]
das-g[m] has quit [*.net *.split]
marius851000[m] has quit [*.net *.split]
alexarice[m] has quit [*.net *.split]
frislie[m] has quit [*.net *.split]
bkl[m] has quit [*.net *.split]
Nyanloutre[m] has quit [*.net *.split]
zmlww[m] has quit [*.net *.split]
is0x3F has quit [*.net *.split]
dwagenk has quit [*.net *.split]
rihardsk[m] has quit [*.net *.split]
otini has quit [*.net *.split]
BinkyTheClown has quit [*.net *.split]
wak-work has quit [*.net *.split]
bikki[m] has quit [*.net *.split]
simbergm has quit [*.net *.split]
nh2[m] has quit [*.net *.split]
musicmatze has quit [*.net *.split]
kaychaks[m] has quit [*.net *.split]
sylvie[m] has quit [*.net *.split]
suvash_away has quit [*.net *.split]
cocreature has quit [*.net *.split]
Haskellfant is now known as cocreature
<Yaniel>
spoeaking of graphics stack fuckery, what is the current correct way to package anything that needs graphics drivers?
<Yaniel>
lutris complains about missing libgl and libvulkan and renderdoc flat out crashes
<Yaniel>
*due to not finding libnvidia-spirv
<pbb>
I think libgl is impure, it's always in /run/opengl-driver
otini has joined #nixos
nh2[m] has joined #nixos
<thoughtpolice>
Yes, if you're on NixOS, things in most cases will work fine. If you're using Nix elsewhere, you'll need something like nixGL so you can "side load" the mesa drivers from your linux distro into things that live within /nix/store
<Yaniel>
and there is no way to package things to Just Work?
<pbb>
what's the shortest way to convert an attribute set to a list (just drop the keys)?
simbergm has joined #nixos
infinisil-m has joined #nixos
kaychaks[m] has joined #nixos
<thoughtpolice>
nixGL works alright admittedly, IME, but if that's a bridge too far, then yes. But it's not easy to fix in general.
y0x3y[m] has joined #nixos
<thoughtpolice>
There were some recent discussions about the future of nixGL
<Yaniel>
well it used to Just Work until a few weeks ago
musicmatze has joined #nixos
wak-work has joined #nixos
<Yaniel>
this is on nixos unstable fwiw
BinkyTheClown has joined #nixos
superbaloo has joined #nixos
<thoughtpolice>
Oh, that's significantly different then.
<thoughtpolice>
You shouldn't need nixGL there
<Yaniel>
yes, but what should the packages be doing?
is0x3F has joined #nixos
sylvie[m] has joined #nixos
Fluffy[m] has joined #nixos
sylvie[m] has joined #nixos
sylvie[m] has quit [Changing host]
is0x3F has joined #nixos
is0x3F has quit [Changing host]
Fluffy[m] has joined #nixos
Fluffy[m] has quit [Changing host]
<pbb>
got it: attrValues
<{^_^}>
[nixpkgs] @dkabot opened pull request #67964 → gnomeExtensions.arc-menu: init at 31 → https://git.io/fjpyS
<thoughtpolice>
Yaniel: I'm afraid I don't know OTTOMH -- as said, it should "just" find your graphics/libgl drivers in /run/opengl-drivers as usual. there may have simply been a regression that caused it to e.g. to miss some files, it depends on lot on your graphics stack and versions of everything.
<thoughtpolice>
If you have a revision of nixpkgs that worked before, filing a bug would be very useful.
<thoughtpolice>
(Or even if you don't)
<Yaniel>
I don't
<thoughtpolice>
Admittedly the graphics stack stuff isn't perfect on NixOS either in some cases. But just, like, "boot up machine and log in and run opengl app" should work fine
<fresheyeball>
clever: with sandbox false, it dies even sooner, because our tests exit even sooner now, because it has expectations about tmp
<thoughtpolice>
And if not that's probably a bug
dansho has quit [Ping timeout: 244 seconds]
<clever>
fresheyeball: have a look at the socat man page, you should be able to make socat listen on a unix socket, and forward everything to a tcp port
<clever>
fresheyeball: then just tell X to not listen on unix
<fresheyeball>
clever: I don't understand the suggestion
<fresheyeball>
how does proxying the unix socket to a tcp port helpful?
<clever>
fresheyeball: most X programs expect a unix socket in /tmp/.X11-unix/
<clever>
fresheyeball: which routes to the X server
<fresheyeball>
right
<fresheyeball>
with you so far
<clever>
you can use socat to create such a socket, and proxy it to the tcp port X is listening on
<clever>
to get around X being upset about /tmp having the wrong permissions
<Yaniel>
actually now it's borkign already on libgl/libegl
<{^_^}>
[nixpkgs] @mstarzyk opened pull request #67965 → tab: init at 7.1 → https://git.io/fjpyA
<fresheyeball>
,locate socat
<{^_^}>
Found in packages: cheat, socat, socat2pre
glittershark has quit [Ping timeout: 246 seconds]
<Yaniel>
interestingly running the target application without renderdoc works fine
<Yaniel>
no wait it doesn't
o1lo01ol1o has joined #nixos
domogled has joined #nixos
m1cr0man has quit [Quit: "\\//," - Spock]
knupfer has joined #nixos
<Shyim[m]>
Hey, i have executed nixos-rebuild switch --upgrade. Since that the channels are broken. Someone an idea how to fix that?
<Shyim[m]>
The error is
<Shyim[m]>
error: syntax error, unexpected $end, at /nix/store/y1gvp3bp13ygi05d5jsvfibha516hc4g-env-manifest.nix:1:1
m1cr0man has joined #nixos
drakonis_ has joined #nixos
<clever>
Shyim[m]: ls -lL /nix/var/nix/profiles/per-user/root/channels/manifest.nix
zupo has quit [Read error: Connection reset by peer]
<clever>
fresheyeball: also, that is not the right path to the socket
<clever>
$ ls -lh /tmp/.X11-unix/
<clever>
srwxrwxrwx 1 root root 0 Sep 1 15:01 X0
<clever>
total 512
<clever>
fresheyeball: its a directory, with sockets for each server
<Shyim[m]>
clever: nope. But maybe its something with pantheon as desktop. I am getting some notifications, but i don't know how to solve it. But last times it worked also with that notice :)
<clever>
and X wants it to have the same sticky bit as /tmp, so anybody can create a socket, and own it properly
<clever>
Shyim[m]: the assertion means that the headers and main package for vulkan are of different versions
<Shyim[m]>
Maybe i should stop using unstable :D
<clever>
Shyim[m]: or just wait a day for somebody else to fix it
<fresheyeball>
clever: so I updated it to /tmp/.X11-unix/X10
spacekookie has quit [Read error: Connection reset by peer]
xkapastel has quit [Quit: Connection closed for inactivity]
zupo has joined #nixos
<Shyim[m]>
Ahh okay :)
<Shyim[m]>
I have now just removed the package which requires vulkan :)
<Shyim[m]>
Thanks <3
<pbb>
Shyim[m]: you can also do this in your nixpkgs directory: "git checkout 7d5375ebf4c; git cherry-pick e54007370fafe4b0d50ce4f0b5dcfab34e80042d; sudo nixos-rebuild switch -I nixpkgs=$PWD"
mexisme has joined #nixos
wildtrees has quit [Quit: Leaving]
<kraem>
pbb: ah nice! i'm currently researching on how to do that :)
<{^_^}>
[nix] @edolstra pushed to flakes « nix repl: Run in impure mode »: https://git.io/fjpSa
pie_ has joined #nixos
<kraem>
how would one do that declaratively if one would like to cherry pick two fixes?
<pbb>
that's a bit of a meta problem
<clever>
kraem: might be simpler to just read that git rev, and then write an overlay that changes the versions of both vulkan packages
<pie_>
does anyone know why my home manager package specifications are being ignored? almost minimal repro at https://bpaste.net/show/z5e-
<pie_>
im probably doing somethign dumb beause with all the googling ive done, this seems to work fine for everyone else, and ive been staring at this for days
<pbb>
I noticed that quite a few builds on hydra are failing because one of the packet.net workers ran out of disk space. Who would be the person to talk to and are they aware of the problem?
<kraem>
clever: Ok so the solution with cherry picking wouldn't count as an overlay? Currently watching Nicolas B Pierron talk about overlays.. So creating an overlay would basically be copy pasting solutions found in the commits?
<clever>
kraem: yeah
<kraem>
thanks! that saved me alot of trial and error i guess :)
<pie_>
or can someone at least check if they can reproduce my issue?
<pbb>
kraem: you can do both, they apply for different cases. the ones in ~/.config/nixpkgs/overlays are used by user commands like nix-env and nix-shell, while the ones in /etc/nixos/configurations.nix are used when building the NixOS system with nixos-rebuild
m1cr0man has quit [Quit: '\\//,' - Spock]
<Shyim[m]>
Has someone an idea how to fix ``trace: Default desktop manager (pantheon) not found at evaluation time.`` ?
<Shyim[m]>
this notice comes on every rebuild
<kraem>
thanks again!
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
vyorkin has quit [Remote host closed the connection]
evanjs| has joined #nixos
qubasa has joined #nixos
<kraem>
clever: i think you lost me with the src= part :/ do i need to change the sha256? shouldn't tofu work in that case?
<clever>
kraem: you need to override src= also, because the rec { means the version= alone wont do it
knupfer has quit [Ping timeout: 264 seconds]
__monty__ has quit [Quit: leaving]
abrar_ has joined #nixos
<fresheyeball>
this is some painful pinoeer tax mang
abrar has quit [Ping timeout: 268 seconds]
qubasa has quit [Ping timeout: 245 seconds]
<kraem>
clever: alright. can't find fetchFromGithub if i leave it there and can't find version if I remove it. am I on the right path with this? https://bpaste.net/show/hbsu
<clever>
kraem: you want self.fetchFromGithub
qubasa has joined #nixos
<clever>
kraem: and you need to call fetchFromGithub when overriding src, your overlay just sets src to a set with just one attr
<corpix[m]>
jonge: I found the reason why nixpkgs fetchurl with `curlOpts` was not working for me... well, there was two reasons. First: I have blindly copied the sha256 sum from some random package which exists in my /nix/store at the time of derivation evaluation(this is why I have got some strange contents inside). Second: there was anoher problem with escaping, `-H 'PRIVATE-TOKEN:${token}'` will become `-H
<corpix[m]>
''\''PRIVATE-TOKEN:XXXXXXXXXXXXXXXXX'\'''` and curl will send HTTP header with quotes `'PRIVATE-TOKEN:XXXXXXXXXXXXXXXXX'`. Not sure, but looks like this escaping stuff applied somwhere in nix itself, thinking of opening an issue about this...
<danderson>
so if I `nixos-rebuild switch` and a systemd unit fails to start during activation, how do I retry that? Would another `nixos-rebuild switch` try to activate the unit again?
<danderson>
I'm trying to figure out secret key management, and there's a circular dependency between users getting created during activation, and me setting the correct permissions on a secret. Trying to figure out how to resolve that, and one option would be to just switch, push keys, and switch again to retry anything that failed due to missing keys.
<ivan>
systemctl start something
<danderson>
right, but I'm trying to generically retry anything that failed during activation, rather than manually patch up specific instances.
chloekek has quit [Ping timeout: 246 seconds]
Rusty1 has joined #nixos
<pie_>
danderso: well, ideally...you would express dependencies, as dependencies, and not leave it up to chance