<iqubic>
Why is it that home-manager has options for setting the pointerCursor theme, but standard NixOS doesn't?
<iqubic>
How am I supposed to change the theme of my mouse cursor on NixOS?
<samueldr>
because they accomplish their work in different contexts
SrPx has joined #nixos
<samueldr>
NixOS' concerns stop mostly at where the user's configs concern starts, by design
<SrPx>
I'm trying to install nixos on my dell xps 13 9300, but it doesn't find the wifi. I managed to share my macbook internet with a thunderbolt cable. I followed the instructions on the manual, but when I typed nixos-config, I got a lot of "unable to download" ... "couldn't resolve host name" and it failed to build. What do I do?
ingenier1ariel has quit [Ping timeout: 260 seconds]
<samueldr>
SrPx: have you confirmed you have shared the internet, and not only made a local network with the two computers?
<samueldr>
(sorry, can't help much more because to me networking is hell)
<iqubic>
samueldr: I see. So, if I'm not using home-manager, how can I change the look of the mouse cursor on X11?
<samueldr>
through whatever means the desktop environment provides, which would be the same means as on other distros
<SrPx>
samueldr: nope, i dont know much about networking
<iqubic>
samueldr: I'm using i3, and it doesn't provide that utility.
<samueldr>
SrPx: might want to ping a well known hostname that responds to pings, like hydra.nixos.org was just verified to
<praduca>
clever Ok, will try this tks but i think this is a bug... the option is for the soft to use to ask password for ssh... I used both last year and have no problem...
<SrPx>
i've tried everything I could at this point then. no way to get an internet connection from the installer
<samueldr>
iqubic: you should always check beforehand that it will respond to ping
<samueldr>
google could stop responding at any point
philr_ has joined #nixos
<iqubic>
samueldr: isn't that true of all websites?
<samueldr>
no
shibboleth has quit [Quit: shibboleth]
<samueldr>
that's not even related to websites, that's at a lower level
<excelsiora>
I might have all I need with a shell.nix and elmPackages.elm as a buildInput, but I want to see other ways of doing it. Anybody got an elm project that uses nix?
<samueldr>
ping in itself is a protocol, not like, but not unlike HTTP is
<samueldr>
though it's common courtesy to respond to pings
<iqubic>
Well, now I have to relog into i3 to see if this change I made will change my cursor.
iqubic has quit [Remote host closed the connection]
<samueldr>
(for ipv4)
<{^_^}>
[nixpkgs] @loafofpiecrust opened pull request #93793 → farge: init at 1.0.8 → https://git.io/JJ49V
<SrPx>
is there anything I can do? any ideas? i guess i'll give up :(
<samueldr>
SrPx: new wifi hardware sometimes is painful, other than getting an alternative connection, it's going to be harder
<samueldr>
it could also be an older wifi usb dongle, or a usb ethernet dongle
<samueldr>
for your current situation, you would need to figure out the right networking incantations to get the mac's internet sharing working with the nixos iso
<SrPx>
i don't even have an ethernet cable tbh, and xps 13 only has usb3 ports. all wifi here
<samueldr>
the good thing is that it's likely any old linux commands to do so
<SrPx>
hm
<samueldr>
(yeah, I assumed so for usb, an usb ethernet dongle is mightily handy to have laying around :))
<SrPx>
can't i just make an usb installer that doesn't need internet though?
<evanjs>
in this case, the CMake file references /nix/store/kmry1p38h1xzics8yqzy3854npkcbji8-exiv2-0.27.2-dev/lib/libexiv2-xmp.a but cannot find it
<samueldr>
it's possible, though that would be harder than to make an usb installer with the right drivers
<SrPx>
i see
<samueldr>
SrPx: do you know what's required for wifi to work on your laptop?
<SrPx>
i have no idea at all
<samueldr>
I have an idea
<samueldr>
I'm waiting on a couple pages to load
<samueldr>
but we build an unpublished iso with the latest kernel
<samueldr>
maybe you need a newer kernel than the LTS
<samueldr>
that should be using kernel 5.7, compared to 5.4
<samueldr>
energizer: overrideAttrs takes a function as an argument, a function which receives the old arguments as its arg
<samueldr>
the old attrs as its arg*
<iqubic>
So my WM doesn't provide me with a convient way of setting the look of the X11 mouse cursor. I'm not sure what I can do to fix that.
<energizer>
samueldr: oops, ty
<samueldr>
energizer: n/p, was the explanation on point for you? :) I seem to recall you're relatively fine with nix overall so I went explaining the exact "high level" issue
<energizer>
samueldr: yep
lafrenierejm has joined #nixos
<lafrenierejm>
Whenever I attempt `nixos-rebuild switch` I get an error "file 'nixos-config' was not found in the Nix search path". Any pointers what that means or how to resolve?
<iqubic>
samueldr: I assume following the steps listed under XDG specification here will work on NixOS?
<samueldr>
I don't know for sure, but it's likely to
<iqubic>
Cool
<iqubic>
I'm not going to modify the system-wide config, because A) This is a single user system, and B) I don't know how to set things like that on NixOS.
<iqubic>
And actually, looking at home-manager's source right now tells me that this is exactly how the xsession.pointerCursor settings work there.
kim0 has quit [Quit: Connection closed for inactivity]
<SrPx>
samueldr: hmm, thanks
<SrPx>
will try
redj has joined #nixos
catid has joined #nixos
<catid>
Anyone know what causes the /run/opengl-driver/lib/dri/i965_drv_video.so file to show up in NixOS?
<samueldr>
this is a known impurity to handle the different switchable opengl backends
iqubic has quit [Remote host closed the connection]
Emantor has joined #nixos
jumper149 has quit [Quit: WeeChat 2.9]
<catid>
Using the vaapiIntel extraPackages seems to work on an i5 but the i7 computer doesn't get the driver to show up
<catid>
Both CPUs support iGPU
<{^_^}>
[nixpkgs] @tnias opened pull request #93794 → nixos/chromium: update link in docs → https://git.io/JJ4HK
<samueldr>
if links under there are not the same, for the same nixos release, it must be because *something* isn't setting hardware.opengl.package the same way in the two configs
<clever>
../config.h:169:16: error: 'rpl_malloc' was not declared in this scope 169 | #define malloc rpl_malloc
<energizer>
i'm trying to write a systemd service for airmon-ng. afaict the only way it will write structured output is if i pass it `--write filename` and then it will write to `filename-01.csv`. but i want the output to go into system journal instead of that file. what can i do?
<clever>
config.h is trying to remal malloc to rpl_malloc, but rpl_malloc doesnt exist
<hazel[m]>
apparently theres some autoconf wizardry you can do with environment variables
<clever>
hazel[m]: its definitely configure thats doing it, maybe it has an off swich?
<clever>
autom4te.cache/traces.1:m4trace:configure.ac:153: -1- AH_OUTPUT([malloc], [/* Define to rpl_malloc if the replacement function should be used. */
<hazel[m]>
got it
codygman has joined #nixos
<clever>
checking for GNU libc compatible malloc... no
proofofkeags has quit [Remote host closed the connection]
<energizer>
i want to give this package runtime access to `ip` so i tried adding iproute to the propagatedBuildInputs but i still get failure "You have neither ip (iproute2) nor ifconfig installed." https://bpa.st/S5EQ how to fix?
proofofkeags has quit [Remote host closed the connection]
proofofkeags has joined #nixos
<clever>
energizer: propagatedBuildInputs is only available at build time, for other things that depend on that package
<clever>
energizer: you want wrapProgram and --prefix PATH
luna has quit [Ping timeout: 240 seconds]
justache has quit [Remote host closed the connection]
<lafrenierejm>
infinisil: Thank you for the pointer. I'll see if I can find docs on the format of the environment variable's value.
lafrenierejm has left #nixos ["Fixing NIX_ENV"]
sigmundv_ has quit [Ping timeout: 240 seconds]
<energizer>
clever: https://bpa.st/775Q it complains that /nix/store/7xyx5wl6ybm77q4wygdm4r2x7h8j4iw4-aircrack-ng-1.6/bin/airodump-ng is not executable and when i look manually that file doesnt even exist
ylixir has joined #nixos
<energizer>
clever: i added the --prefix in the preFixup. should i have put it somewhere else?
<clever>
energizer: when does it complain? when building aircrack-ng?
<ylixir>
hey all, i just got a new machine (amd ryzen 4700u) and tried to boot it up with the plasma 5 nixos image. couldn't get X to start though. i wasn't surprised when the 20.03 image didn't work, but i was when the nightly didn't. anyone get one of the new ryzen chips working yet?
<sb0>
if I set up hydra to place its outputs in a different directory than the system nix store, how do the builds work? do they first occur in the system nix store and then they are copied to the store_uri location?
<sb0>
also, how do I remove old builds (gc) the custom store_uri location?
marsh has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @jtojnar pushed commit from @sbourdeauducq to master « pam_p11: add »: https://git.io/JJBvH
h0m1 has quit [Quit: WeeChat 2.9]
h0m1 has joined #nixos
<srhb>
energizer: At a guess, a derived dependency is lacking some information in the toml for naersk to generate the requisite keys
<srhb>
energizer: Specifically, it looks like it's leveldb
<energizer>
srhb: how did you figure that out?
<srhb>
energizer: I patched naersk to throw ${e.url} in that error message in build.nix, because I figured since it was in a toml conversion it'd be something like that. :)
<energizer>
srhb: nice
<energizer>
srhb: where is the rev missing from so i can patch it in?
h0m1 has quit [Client Quit]
maier has joined #nixos
h0m1 has joined #nixos
<srhb>
energizer: cargo.toml, I believe.
ylixir has joined #nixos
<srhb>
energizer: Or rather, it's probably missing a version
maier has quit [Ping timeout: 240 seconds]
mounty has joined #nixos
<bqv>
energizer: same problem it always is with rust
<bqv>
energizer: there's a git dependency
<bqv>
energizer: and *none* of nix's rust tooling supports those well
<bqv>
Iirc your best bet is attempting cargo2nix and manually fixing the result, but even that will be arduous in many cases
malik has joined #nixos
justache has quit [Remote host closed the connection]
pmy has quit [Ping timeout: 240 seconds]
justache has joined #nixos
pmy has joined #nixos
afreak has joined #nixos
alp_ has joined #nixos
<energizer>
i tried just cloning the `vector` package, editing the Cargo.toml manually and pointing src=./vector but i get the same error
lambda-11235 has quit [Quit: Bye]
<bqv>
energizer: Ah, then yeah it's that your filepath has no hashes to put in the lock. The rust infra relies on those
<bqv>
It's brittle as hell
<bqv>
Probably better to patch the src after gen, not before
<{^_^}>
[nixpkgs] @cdepillabout merged pull request #93799 → harfbuzz: enable support for gobject-introspection → https://git.io/JJ4jG
<{^_^}>
[nixpkgs] @cdepillabout pushed 2 commits to staging: https://git.io/JJBJt
<{^_^}>
[nixpkgs] @Lassulus merged pull request #93758 → writers.writePython2 and writePython3: use "bare" python if no deps are needed → https://git.io/JJ4IO
<{^_^}>
[nixpkgs] @Lassulus pushed commit from @adrian-gierakowski to master « writers.makeScriptWriter: fix on Darwin\MacOS (#93757) »: https://git.io/JJBTd
maier has joined #nixos
WilliamHamilton has joined #nixos
<WilliamHamilton>
gst-plugins-base-0.10.36 is broken in current master. How can I find out which one of the packages in my configuration.nix is trying to use that?
<WilliamHamilton>
by bisecting my package list, I know that the problematic ones were steam and appimage_run. Could I have determined that in a more principled way?
<Reventlov>
aside from networking.firewall.allowedTCPPorts, do I need something else to allow http connections to reach a nixos VM on port 8000?
WilliamHamilton has quit [Remote host closed the connection]
<hexa->
Reventlov: bind to an externally reachable IP address
<hexa->
bind your service*
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
<Reventlov>
ah
<Reventlov>
I thought binding to 127.0.0.1 would work, but no, thanks
<{^_^}>
[nixpkgs] @Lassulus pushed to master « tor-browser: disable hardening by default (#93154) »: https://git.io/JJBIX
<Arahael>
SomeDude: A great excuse to upgrade! :)
<SomeDude>
plugged everything in. monitor goes into a constatnt bootloop. nixos dmesg log "amdgpu no edid read". try different connections(displayport, hdmi) different power cords. )
<SomeDude>
Now not even the power led lights up when power is connected
<SomeDude>
Arahael sadly this was a new monitor...
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos
<Arahael>
SomeDude: Bah, that's annoying. At least it should be under waranty.
Guest60997 is now known as JJJollyjim
<{^_^}>
[nixpkgs] @peti pushed to master « neuron-notes: create a top-level attribute for haskellPackages.neuron »: https://git.io/JJBIp
<{^_^}>
[nixpkgs] @jtojnar pushed to master « ostree: enable ed25519 support »: https://git.io/JJBIh
octe has joined #nixos
afreak has quit [Quit: leaving]
szicari has quit [Ping timeout: 246 seconds]
justache has quit [Remote host closed the connection]
pmy has quit [Ping timeout: 264 seconds]
justache has joined #nixos
pmy has joined #nixos
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
<prusnak>
how do I test whether a nix list contains an item?
<prusnak>
e.g. "3 in [1 2 3]"
Pwnna has quit [Quit: Bye]
Pwnna has joined #nixos
<b42>
> builtins.elem 3 [ 1 2 3 ]
<{^_^}>
true
<prusnak>
thx!
orivej has quit [Ping timeout: 240 seconds]
orivej has joined #nixos
freezeboy has joined #nixos
<freezeboy>
hello, on my local overlay I added some overrides for vscode-utils, vscode-with-extensions and vscode-extensions for vscodium, i saw that master has a new vscodeEnv that does improve the experience when providing custom extension sets for vscode so I am not sure my changes are very useful now.. has anyone any experience and ideas about that ?
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
Fare has joined #nixos
reanimus has quit [Quit: Logging off...]
nisstyre has quit [Changing host]
nisstyre has joined #nixos
maier has joined #nixos
<buffet>
hey, im trying to pin my home-manager version using niv, so i tried a few versions of `imports = [ sources.something/nixos ];`, but i get `value is a path while a set was expected`. how can i make that work?
o1lo01ol1o has joined #nixos
maier has quit [Ping timeout: 246 seconds]
Fare has quit [Ping timeout: 260 seconds]
arjen-jonathan has joined #nixos
nature has joined #nixos
<nature>
Anybody has experience with nixos on aarch64 ? More specifically on rpi4 ?
<nature>
I am about to flash a sd card and wonder how I can "preconfigure" it so I can just plug it in and then ssh in
<nature>
I don't have acced to an external monitor
<nature>
*access
romildo has joined #nixos
justache has quit [Remote host closed the connection]
justache has joined #nixos
romildo has left #nixos [#nixos]
orivej has quit [Ping timeout: 260 seconds]
orivej_ has joined #nixos
<nature>
Is hydra.nixos.org down ?
m0rphism has quit [Quit: WeeChat 2.7.1]
<hexa->
DBIx::Class::Storage::DBI::catch {...} (): DBI Connection failed: DBI connect('dbname=hydra;host=10.254.1.9;user=hydra;','',...) failed: could not connect to server: Connection refused
<hexa->
looks like it
m0rphism has joined #nixos
<nature>
ok thanks, I'll have to wait then :)
<{^_^}>
[nixpkgs] @bobismijnnaam opened pull request #93813 → Ensure wpa_supplicant.conf is written when userControlled and extraConfig are used → https://git.io/JJBqi
Neo--- has quit [Ping timeout: 272 seconds]
knupfer has joined #nixos
Rusty1 has joined #nixos
<gchristensen>
huh
<gchristensen>
fixed
<gchristensen>
postgres isn't ordered after the wireguard interface...
asbachb has joined #nixos
<asbachb>
Hi. I'm looking for a way to ensure that letsencrypt certs are generated initially when nixos-rebuild switch is run.
<{^_^}>
[nixpkgs] @vbgl opened pull request #93814 → coqPackages.bignums: init at 8.12.0 for Coq 8.12 → https://git.io/JJBq9
<immae>
Hello there, I’m trying to use the latest (master branch) nixops but I cannot find how to use a plugin with it, is there any documentation (or even just a hint) somewhere about how to do that?
<{^_^}>
[nixpkgs] @mweinelt opened pull request #93821 → python3Packages.pyfritzhome: init at 0.4.2 → https://git.io/JJBO5
Matt125 has quit [Remote host closed the connection]
<dhess>
gchristensen: have you used GitHub Actions with Nix projects much?
<adisbladis>
dhess: I'll answer for gchristensen :P Yes.
<dhess>
adisbladis: have you as well then?
<adisbladis>
dhess: Yes, for NixOps for example.
<dhess>
adisbladis: ahh cool, thanks, I found the ci.yml in the NixOps repo and that's the kind of thing I was looking for. Got time for a few questions?
<adisbladis>
Shoot =)
knupfer has quit [Ping timeout: 240 seconds]
<dhess>
thanks :)
alp_ has quit [Ping timeout: 246 seconds]
<dhess>
adisbladis: as background, I'm considering switching from a Hydra-based CI to a GitHub Actions-based CI, so I've been spelunking a bit to figure out how I could do that.
<dhess>
adisbladis: one thing I like about the Hydra approach is that Hydra automatically figures out (based on meta.platforms) which packages should build on which platforms. It looks like I'd have to re-do that work manually in a GitHub Actions-based workflow, yes?
<dhess>
as in, create a package set for macOS, and another for NixOS, and then use 'runs-on' in the ci.yml to build only the applicable attribute/package set.
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
<adisbladis>
Yes, you'd have to figure out what to build "manually"
<LnL>
if you put them in set or list and null out unsupported packages in release.nix you don't need to list attributes explicitly
<dhess>
LnL: so I have working release.nix files that I use for Hydra CI. The Hydra (or the "release-lib.nix" functionality of nixpkgs) does all that for me automatically. I would need to come up with equivalent machinery to get this to work "automatically" with GitHub Actions, is my point. Right?
<LnL>
all that does is group packages with their respective platforms
<LnL>
which also means nix-build won't recurse far enough anymore unless you specify attributes
<LnL>
yes, because hydra-eval-jobs evaluates differently
<hexa->
github time outs :<
<dhess>
whatever attributes are defined in my packageSet (which comes from the project's default.nix's attributes) are automatically discovered and built on the appropriate platform.
<dhess>
LnL: my point is, if I want equivalent functionality on GitHub Actions, I'm going to have to write that myself.
<Zer000>
So for some weird reason this shell script works fine if I use nix-shell shell.nix, but llvmPackages_6 doesn't exist anymore, and my shell doesn't have the "opt" command that comes with llvm. Also if I remove the "_6" part from the script, it still doesn't seem to install it. https://dpaste.com/8XSTCCWZ4.txt
<LnL>
release-lib does't really work for nix-build evaluation
<dhess>
adisbladis: OK next question -- I assume GitHub's runners don't have any KVM functionality so NixOS tests will not run in GitHub Actions right?
<adisbladis>
Yep, no KVM :/
<Athas>
Does steam work for the rest of you with current nixpkgs-unstable? I get a complaint that 32-bit libraries cannot be found.
<dhess>
LnL: I can also run `nix-build release.nix -A native.hydra-unstable` and it builds fine, even on macOS.
<dhess>
as it knows not to build it on macOS. Just as one example.
<dhess>
adisbladis: ok that's what I figured. I can work around that with my own hosted runners, but not for public repos.
<dhess>
adisbladis: do you know of any GitHub Action examples that filter PRs and only build them when the requestor is included in a list of trusted folks?
<LnL>
while hydra will recurse down without having to specify each attribute in the job config
<dhess>
LnL: That was my point. Moving from a Hydra-style release.nix to GitHub Actions is going to require either a) listing out each attribute for each platform by hand, or b) reproducing Hydra's machinery for traversing the attrset and filtering based on meta.platforms
iceypoi has joined #nixos
<simpson>
Why does meta.platforms matter at the top? Aren't individual packages using .isDarwin and friends?
<LnL>
right, but that structure doesn't really make sense outside of hydra
orivej has quit [Quit: No Ping reply in 180 seconds.]
<dhess>
LnL: I understand that, but it would be nice to be able to re-use it outside of Hydra, since it's already there in nixpkgs.
<LnL>
hydra doesn't need it either, it's just so you get the package.platform as job name
dbmikus has quit [Ping timeout: 264 seconds]
orivej has joined #nixos
icey_ has quit [Ping timeout: 246 seconds]
<dhess>
simpson: that doesn't work for a large package set that has packages that build on some platforms and not others. It's fine for simple packages I'm sure.
<dhess>
in fact that action is only for ubuntu-latest so it doesn't even cover macOS, for example.
<dhess>
to be clear I'm not complaining about the state of things, I'm just trying to figure out how we can go from a Hydra-based CI to GitHub Actions-based CI with Nix, and how much work it'll be.
<simpson>
I suppose I'd answer that they do quite different things! In particular, one runs on your machines and one runs on somebody else's machiens.
<dhess>
simpson: even if we use hosted runners, there are issues.
<LnL>
nix flake check will hopefully make this nicer in the future
<simpson>
I know that there are issues. I'm suggesting that the solution isn't a matter of porting, but a matter of building a totally new CI setup.
<dhess>
which we'll have to do if we want to run NixOS tests, for example.
knupfer has joined #nixos
tru_tru has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed commit from @r-ryantm to release-20.03 « tcpreplay: 4.3.2 -> 4.3.3 (security) »: https://git.io/JJBsJ
<dhess>
Here's another question: I understand that GitHub Actions does things to filter your secrets so they don't show up in logs. Let's say someone makes a PR that changes your .github/workflows/ci.yml and adds a job that echoes '${{secrets.CACHIX_SIGNING_KEY}}' to stdout. So GitHub Actions will filter that, right?
<infinisil>
Even if it filters stdout, somebody could just curl PUT the secret to their own url
<TakeWill>
is there anyway to hide message of this bot {^_^}?
<TakeWill>
I'm new to this IRC thing
<infinisil>
TakeWill: Most IRC clients allow you to install filters. Maybe check your clients manual
<dhess>
infinisil: right, so I don't follow how triggering build actions automatically on PRs from randos prevents secret leakage. What's the story on that?
<infinisil>
dhess: I honestly have no idea either and am also interested to know
<TakeWill>
I'm using rcirc (an Emacs package)
<TakeWill>
anyone using the same?
<dhess>
infinisil: another scenario is that they just rot13 the secret before echo'ing it to stdout and then unrot13 the logs
<dhess>
etc.
<simpson>
dhess: I'd suggest not doing Cachix on PRs. But yeah, this is a good question and I think that GH will have to answer it.
<infinisil>
TakeWill: This isn't an IRC support channel though, I suggest asking for help on that somewhere more appropriate :)
Rusty1 has quit [Read error: Connection reset by peer]
<simpson>
TakeWill, infinisil: The bot uses the /notice mechanism, and many IRC clients have ways to filter notices.
<simpson>
dhess, infinisil: I just checked a build log. The URL's got a funky signature to prevent sharing (yay?) but I can say that the signing key is not echoed by the Cachix CLI tool into the log.
<TakeWill>
infinisil sorry, I'll ask in other places
<simpson>
`signingKey: ***` literally what it emits. Very polite.
<TakeWill>
simpson thanks for your info
<JJJollyjim>
like it only does periodic evaluations
<JJJollyjim>
am i right in understanding that hydra can't really link things to git commits?
<JJJollyjim>
(i think those two messages sent out-of-order)
<dhess>
simpson: I'm sure that domenkozar has handled the cases where cachix itself might output the key. To be clear, I'm worried about the scenario where the attacker's PR modifies your ci.yml file and includes a script that has access to the key via the GitHub Actions secret templating.
<simpson>
JJJollyjim: Right, while Hydra usually can show you which commit is to blame, sometimes it shows you a range of commits instead. And in those cases, you'll have to bisect, if you can't manually see at a glance which commit is the likely offener.
<JJJollyjim>
TakeWill: yeah you can /ignore a username in rcirc
<dhess>
simpson: ok cool, just wondered if anyone in the channel might have already looked into this and knows how it's prevented. It seems to obvious not to have been addressed.
<dhess>
too obvious
<JJJollyjim>
imo if a community is making people to learn to use irc, it's not unreasonable for them to ask questions about it
<simpson>
I agree that they could do better, and explicitly say that enabling secrets for PR workflows is dangerous.
<infinisil>
JJJollyjim: I guess, but I think it's an important part of open source to know where to get support for things
<dhess>
simpson: yeah I've read that, but I don't see how that prevents a PR (which, as I understand it, automatically kicks off a build if made on a public repo where GitHub Actions are enabled) from doing it.
<TakeWill>
JJJollyjim many thanks man. This is really helpful
<JJJollyjim>
infinisil: sorry, i'm bitter cos i work at a company that has irc as the main form of communication and doesn't help people with it, so the non-technical staff just give up and are cut off from everyone else -_-
<dhess>
doesn't that say right there at the top that it runs on pull_request ?
<dhess>
or am I misinterpreting that
<simpson>
dhess: Yes, because I want it to run on PRs. I'll probably change that. Don't worry, my fuckups aren't indicative of GH's poor documentation~
<dhess>
simpson: heh ok
<dhess>
simpson: that seems like a major problem though! Someone as experienced as you can easily make that mistake! Which makes me think there must be something else to this that we're missing.
mbr has joined #nixos
<{^_^}>
[nixpkgs] @blitz opened pull request #93824 → nixos/lib/make-ext4-fs: fix after mkfs.ext4 refactoring → https://git.io/JJBs0
niso_ has joined #nixos
work_ has quit [Quit: Connection closed for inactivity]
<simpson>
dhess: Or maybe I copied it from the neighboring YAML for the IRC action.
<Mic92>
bqv: actually the mkdir in use_flakes is not necessary
<simpson>
And yes, GH is shit, including GH Actions, but it's also free and what the community uses. To revive a 2010s meme, "Actions suck, use Actions".
<Mic92>
simpson: buildkit works great for me.
meh` has joined #nixos
<infinisil>
simpson: What's bad about github actions? (Other than the secret problem)
<dhess>
simpson: this would be easy to fix if either a) you could filter PRs by submitter or b) you could make build logs visible only to trusted team members.
niso has quit [Ping timeout: 260 seconds]
<dhess>
although I guess b) doesn't fix the scenario that infinisil pointed out
<mbr>
i'm looking for a way to specify `--pure` in my `shell.nix` to avoid having to pass it on the cmdline; is that possible?
<simpson>
infinisil: GH is closed-source and unauditable; it's the sort of thing that, in the long term, it would be nice to have a decentralized community solution for managing and publishing. Every fresh dependency on GH's new products puts that goal further out of reach.
<dhess>
Mic92: yeah I've been considering using BuildKite as well but now there's so much momentum behind GitHub Actions, I'm a bit worried about their future.
<Mic92>
dhess: it is, but buildkit with nix is a lot faster if you don't have to redownload everything all the time. I mean a lot faster.
<dhess>
simpson: I can see the page anyway. What do you mean by "raw log"?
<Mic92>
I use it in one project where I need nixos tests
<dhess>
Mic92: yes if we use GitHub Actions, it'll be with hosted runners.
<simpson>
dhess: To the right of the header-bar "Push to Cachix / tests", past the search controls, click the ... and choose "View raw logs"
<Mic92>
dhess: yeah, it's just still a pain to set up with NixOS.
<dhess>
simpson: got it. yes I can see the raw log
user_0x58 has joined #nixos
<dhess>
Mic92: GitHub Action hosted runners, you mean, or BuildKite?
<simpson>
So that rules out (b). (a) is kind of possible, I think, in that it's possible to set up an organization so that only some people have PR permission, but I'm not 100% sure. (I could swear that I'm in an org with a setup like that...)
<Mic92>
dhess: hosted github action runners.
<dhess>
Mic92: got it. It looks like BuildKite is pretty easy due to the support in nixpkgs, yes?
<dhess>
simpson: so an interesting experiment would be for me to make a PR against your cachix.yml which includes a script that echos the secret using a rot13 or something like that. But I'm not going to do that :)
<Mic92>
Yes. Github action might get their one day too
<dhess>
Mic92: right. looks quite nice. Have you gotten it to run in a container?
<simpson>
dhess: In #monte I've already started the discussion to fix this. If there's no opposition, I'll push a fix later.
kriztw has quit [Ping timeout: 260 seconds]
Rusty1 has joined #nixos
SomeDude has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @Lassulus pushed commit from @deliciouslytyped to master « qtile: 0.13.0 -> 0.16.0 (#93705) »: https://git.io/JJBnX
cjpbirkbeck has joined #nixos
orivej_ has quit [Ping timeout: 246 seconds]
dbmikus has joined #nixos
orivej has joined #nixos
foxmitmox has joined #nixos
boothead has joined #nixos
<boothead>
Hey folks, I'm trying to get the v4l2loopback module working. I've added it to boot.kernelModules but that doesn't seem to have done the trick. Essentially I want the equivalent of sudo modprobe v4l2loopback exclusive_caps=1 max_buffers=2
dbmikus has quit [Ping timeout: 256 seconds]
<lassulus>
maybe boot.extraModprobeConfig ?
<Zer000>
Can anyone please tell me how to make a haskell derivation that installs llvm? I adding passing pkgs.llvmPackages.llvm to the executableHaskellDepends list and it doesn't work.
<boothead>
thanks lassulus - do you know what the syntax for the options/flags would be offhand?
<Zer000>
(I'm calling the derivation using haskellPackages.callPackage
<lassulus>
from the example: ''options parport_pc io=0x378 irq=7 dma=1''
orivej has quit [Quit: No Ping reply in 180 seconds.]
<boothead>
Yup, just found that... This is my first time fiddling with kernel options - would you expect the modprobe command to work? I'm getting "Module v4l2loopback not found in directory /run/current-system/kernel-modules/lib/modules/5.3.0-rc5"...
arjen-jonathan has quit [Ping timeout: 244 seconds]
proofofkeags has joined #nixos
orivej has joined #nixos
<Zer000>
nix-shell -iA nixos.llvmPackages.llvm will give me what I want, it's just when I make a shell from a derivation that it can't find it. This used to work before too...
<Reventlov>
I'd like to use django 3.0 ( https://github.com/NixOS/nixpkgs/pull/92848 ). Can I somehow use packageOverrides to say django is actually "another thing" defined in another directory ?
<{^_^}>
[nixpkgs] @pulsation opened pull request #93833 → quisk: init at 4.1.66 → https://git.io/JJBWo
eyenx|m is now known as eyenx
<Mic92>
dhess: it was actually not necessary to mount /dev/kvm for nixosTest since it runs in the host nix builder. However I still need to bind mount /nix/store because channels/tarballs are imported there: https://github.com/Mic92/sops-nix/pull/25
<{^_^}>
Mic92/sops-nix#25 (by Mic92, 23 minutes ago, open): add TODO
<Mic92>
dhess: I think I can also make it safe for public use if I define the pipeline steps in the webinterface instead of the repo and enable restrict-eval in nix
dbmikus has joined #nixos
LinusSexTips[m] has joined #nixos
orivej_ has quit [Ping timeout: 240 seconds]
<dhess>
Mic92: I haven't used BuildKite yet. Why would it matter whether you define the pipeline steps in the web interface? (I assume you mean the BuildKite web interface.)
<Mic92>
dhess: So people in a pr cannot change the code that is executed in the CI. Even with a sandbox they could still read the buildkite api token + ssh key
moet has joined #nixos
<Mic92>
my sandbox now only executes nix-build with restrict eval.
<Mic92>
You have the same issue with self-hosted github actions btw.
<Church->
Hmm, anyone have thoughts on what's the best thing to do for a deriv upgrade that will cause db upgrades on install? Making rollback infeasible unless you have a cp/dump made before hand?
<Church->
I'm debating on just adding an implicit dump step and tar up the sqlite db before the service starts. Add a blurb to the release notes about it, but even that's not the best way to go about things I think.
<atemu12[m]>
srid: Hm, not familiar with systemd-boot but AFAIK it should detect Windowsy stuff automagically
<mwelt>
Hi. Short question. Is there a way to get show the "symbol" for a package. `nix-env -qa "*.prolog.*" returns package names, which are not of much use for creating a shell `nix-shell '<nixpkgs>' -p ...`
<atemu12[m]>
srid: If that doesn't work you'll have to jump in an manually clean up a few files that definitely aren't needed anymore. `ncdu` is very useful for this.
<srid>
when I explicitly specify the device, I get ^
<TheSirC[m]>
I was trying to see if everything was ok in the file
<boothead>
thanks! What's unstable there? Something in lib?
<mwelt>
atemu12[m]: if the windows partition is efi, than it should work by removing the boot flag from the linux efi partition and set it back onto the windows recovery partition...
<boothead>
Also - to you run gphoto and ffmpeg somehow?
<atemu12[m]>
mwelt: boot flag isn't a thing with GPT
<boothead>
Sorry - I meant this bit: boot.extraModulePackages = with unstable; [
<boothead>
TheSirC[m] how do you run the gphoto | ffmpeg bit? Do you have a service or something?
<TheSirC[m]>
Oh! I think that it is because the version on the unstable channel is the latest to circumvent the issue cited in the first comment
moet has quit [Ping timeout: 240 seconds]
<TheSirC[m]>
boothead: Yes I do. I use a bash script which is not really great to connect the camera to the device it is not amazing but it works at least
MarcWeber has quit [Ping timeout: 240 seconds]
<boothead>
Cool, I'm going to try it your way, cos rebuilding my whole system to unstable just now failed
<TheSirC[m]>
boothead: I updated the gist with the shell script if you are interested
<TheSirC[m]>
Tell me if you find any bugs with the script <_<
<notgne2>
is there some basic guide on how to use flakes and what might need to be changed when switching? also does anybody know if there is a stateless nixops alternative that supports flakes
<{^_^}>
#92261 (by FRidh, 3 weeks ago, merged): gst-plugins-base: mark as broken
fresheyeball has quit [Quit: WeeChat 2.7.1]
<boothead>
So how can I get that fix if 20.03 is broken for me? Do I have to switch to unstable?
gentauro has quit [Read error: Connection reset by peer]
juliosueiras has joined #nixos
gentauro has joined #nixos
<juliosueiras>
is creating non-sandbox derivation consider an anti-pattern?
<infinisil>
juliosueiras: The sandbox is what makes builds pure, so yeah, requiring the sandbox to be turned off for a derivation is not good
<juliosueiras>
emm... k, then probably need to create a fetcher for lein project then
evanjs has quit [Read error: Connection reset by peer]
<juliosueiras>
lein-based project*
maier has joined #nixos
<Reventlov>
So i'd like to create an overlay "overriding" pkgs.python3Packages.django, what should I do ? Having « python3Packages.django = super.callPackage ./django/3_0.nix { }; » is not really working, as I get an error message "attribute 'buildPythonApplication' missing"
<{^_^}>
[nixpkgs] @lucperkins opened pull request #93839 → humioctl: init at 0.25.0 → https://git.io/JJB1L
<boothead>
{^_^} - should that gstreamer thing effect 20.03? I'm thinking I've got something screwy in my config if I'm getting bitten by it...
evanjs has joined #nixos
<energizer>
Reventlov: you're trying to override django systemwide?
<gchristensen>
I'm looking to cross compile to thumbv7em-none-eabi -- anyone know how I configure nixpkgs to target this? I have no idea what thumbv7em even means, and can't really find much about it :/
<energizer>
Reventlov: my preference these days is to use poetry2nix
CelliCel317 has quit [Ping timeout: 245 seconds]
<energizer>
just `poetry add 'django>=3.0'` and poetry2nix takes care of the rest
<boothead>
OK - I'm stumped now... How can I figure out where the dependency on gst-plugins-base is coming from so I have remove it (or bring in the fix from master)? I don't even really understand why I'm seeing it at all if I'm buiding against 20.03?
<yilin>
Mic92 If you have a simple `mkShell` derivation with a single buildInput dependency, re-evaluate the functions (and make sure buildInputs is in nix-shell-inputs) and start a shell using (nix-eshell "my-file") and see if the eshell buffer created has the binaries that you expect
<boothead>
Thanks symphorien - I'll give that a try. I had a look through the bug and is seems like there are some pretty fundamental libs that are effected... fingers crossed!
<boothead>
symphorien how come the build is working but the switch fails straight away? (I would have expected build to be the step in switch that fails)
<symphorien>
if build worked you can do result/bin/switch-to/configuration switch
orivej has quit [Ping timeout: 240 seconds]
<boothead>
It didn't work - but it also didn't fail as fast as switch...
orivej has joined #nixos
<symphorien>
ah that can be because many derivations can be built in parallel and the order can vary
<boothead>
Ah i see - well this gst thing turns up A LOT... Don't know that I'm going to be able to get rid of it... Do you know how I can get the fix?
malook has joined #nixos
malook has quit [Client Quit]
justanotheruser has quit [Ping timeout: 272 seconds]
<symphorien>
what fix? I didn't read the scrollback
<symphorien>
looks like the cause of the problem, not the fix
hyper_ch has quit [Read error: Connection reset by peer]
<boothead>
Yeah - but this is since 20.03 right (I mean this doesn't affect everone who has 20.03 as their nixos channel)? TheSirC[m] you're on 20.03 aren't you?
hyper_ch has joined #nixos
<{^_^}>
[nixpkgs] @lucperkins opened pull request #93845 → Add build version to open-policy-agent package → https://git.io/JJBxH
zeta_0 has joined #nixos
<symphorien>
no this shouldn't affect 20.03
<symphorien>
in nix-store -q --tree look for the exact full path of the failing drv
<TheSirC[m]>
boothead: I tried all the configuration on markhor, 20.03 but right now I am on nixos-20.09pre235279.5717d9d2f7c
hyper_ch has quit [Read error: Connection reset by peer]
<TheSirC[m]>
It is not going well but I am on nixos-20.09pre235279.5717d9d2f7c
<symphorien>
there might be several versions of gst-base in your system, some from unstable if you have overlays
<symphorien>
also if you changed it recently, it only takes effect after nix-channel --update
<boothead>
Aaah ok - I thought rebuild was taking care of that!
<boothead>
Thanks, I think I'm moving again :-)
maier has quit [Ping timeout: 240 seconds]
<freezeboy>
hopefully nix-channel is handling channel update and nix-rebuild is only working on current channel :D
<ar>
hm. i have a weird issue with booting and some "nixos-rebuild switch"-es. /nix/store/66lvb19j4ag603363qd0lhp8vvia5mqg-update-users-groups.pl gets run, and ends up in an apparent infinte loop (full commandline: https://dpaste.com/HFWAGQ75T.txt, contents of users-groups.json: https://dpaste.com/7SBMD7WXH#)
<freezeboy>
@ar dpaste is sending not found
wnklmnn has joined #nixos
<bqv>
Mic92: o? I had issues with it before I thought
<zeta_0>
is there another #channel that i can ask about the lorri nix tool, or is this the only one?
<ar>
here both work
<freezeboy>
oh there is a dangling comma in your first message ...
<ar>
it exits the infinite loop by running out of memory
buckley310 has quit [Quit: Connection closed for inactivity]
<samueldr>
ar off of stable or unstable?
<samueldr>
(or master?
<ar>
samueldr: unstable
<ar>
but i've had this issue for a while now, i just don't reboot very often
<Orbstheorem>
Hi o/ I created a derivation that uses home-manager using `nixos-rebuild build-vm-with-bootloader`. Such derivation is supposed to run on an airgapped system. The system boots okay, but home-manager fails: during activation it calls `nix-env -i /nix/store/...-home-manager-path`, which in turn connects to the nix-daemon and invokes `wopQueryMissing` which tried to reach the binary cache. Since I just built the
<Orbstheorem>
derivation all necessary paths should be in the nix-store. How can I prevent nix-env from calling wopQueryMissing?
<samueldr>
ar: a while, like months and months that it could also happen on 20.03?
<zeta_0>
do any of you know a nix lorri expert that i can ask for help? my problem is probably simple, but i'm new to lorri, so i'm not sure how to do it
<yilin>
What issues do you have with lorri @zeta_0 ?
<eyJhb>
Any way I can get commands such as `nix-prefetch-git --fetch-submodules --url https://gitlab.com/cznic/xc --rev v1.0.0` to NOT use /run/user/1000 ? And use /tmp
bruce is now known as Guest88361
<Guest88361>
Hello,
<Guest88361>
I am trying to install nix for the first time can I ask you a few questions?
<{^_^}>
[nix-mode] @matthewbauer pushed 2 commits to master: https://git.io/JJRtk
<yilin>
zeta_0 Were the type errors in the rust process (as in rust daemon or elsewhere?)
<Guest88361>
I spend some time on my configuration.nix
<Guest88361>
and still have to
<Guest88361>
but when I run nixos-install
<Guest88361>
it says
<Guest88361>
error: The option `powerManagement.cpuFreqGovernor` is defined both null and not null, in `/mnt/etc/nixos/configuration.nix' and `/nix/var/nix/profiles/per-user/root/channels/nixos/nixos/modules/services/hardware/tlp.nix'.(use '--show-trace' to show detailed location information)
orivej has quit [Ping timeout: 240 seconds]
boothead has quit [Remote host closed the connection]
<Guest88361>
I did define cpuFreqGovernor = "ondemand";
orivej has joined #nixos
<symphorien>
tlp has another opinion
<symphorien>
you can override it with `= lib.mkForce "ondemand"`
<Guest88361>
why would tlp put null?
<ar>
i've ran the script with -d:Trace and it appears to get stuck in File/Slurp.pm. https://dpaste.com/5S5DLF4F4
<symphorien>
no idea
<Guest88361>
jolie pseudo symphorien
<yilin>
*lorri daemon
maier has joined #nixos
asdfasdfasdf has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<zeta_0>
yilin: no, those were nix type errors, but i'm starting from scratch now
<asdfasdfasdf>
halp
<yilin>
Hmm, does that mean your typical nix-build isn't working?
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
<yilin>
zeta_0 sorry, let me backtrack a little. 1) can you run the relevant nix-build/nix-shell. 2) otherwise, is it that the environment isn't getting pushed into emacs?
aetherolio has quit [Remote host closed the connection]
growpotkin has joined #nixos
<{^_^}>
[nix-mode] @matthewbauer pushed 3 commits to master: https://git.io/JJR3P
orivej has quit [Ping timeout: 256 seconds]
orivej_ has joined #nixos
yilin has quit [Remote host closed the connection]
maier has quit [Ping timeout: 264 seconds]
nature has quit [Ping timeout: 272 seconds]
nature has joined #nixos
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
arjen-jonathan has quit [Ping timeout: 272 seconds]
<duairc>
From the NixOS configuration, is there a way I can determine what the architecture of the system I'm running on is? I'd like to get back a string like "x86_64"
<duairc>
Or maybe there's something in pkgs that has that
<cole-h>
> builtins.currentSystem
<{^_^}>
"x86_64-linux"
<duairc>
cole-h: Lol, that was easy! Thanks :)
o1lo01ol1o has joined #nixos
kreyren has quit [Remote host closed the connection]
<infinisil>
duairc: cole-h: Note that builtins.currentSystem is the system you're *evaluating* on, which doesn't have to be the same as the system you're deploying to
<infinisil>
For getting the string for the system you're deploying to you can use
<infinisil>
> pkgs.stdenv.system
<{^_^}>
"x86_64-linux"
<cole-h>
>> is there a way I can determine what the architecture of the system I'm running on is?
<cole-h>
I was answering that part :)
edio has joined #nixos
<duairc>
Yeah, my question was ambiguous! pkgs.stdenv.system is probably what I actually want
<duairc>
infinisil: Thanks!
<cole-h>
Darn, infinisil always stealing my thunder.
<cole-h>
:P
<infinisil>
:P
<edio>
Hi! I'm trying to boot NixOS plasma on Lenovo X13 AMD and I can't start sddm there. I tried both 20.03 and unstable 20.09pre. Is there somethign I'm missing?
<Guest88361>
I got this warning warning: The following options are deprecated: - services.xserver.desktopManager.defaultPlease use services.xserver.displayManager.defaultSession = "xfce";instead.
shibboleth has quit [Quit: shibboleth]
raju[m] has joined #nixos
<Guest88361>
but when I try with displayManager it doesn't work
<Guest88361>
saying error: The option `services.xserver.displayManager.default' defined in `/mnt/etc/nixos/configuration.nix' does not exist.
<Guest88361>
is the wiki out of date?
<cole-h>
Because it should be `.defaultSession`, no?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Guest88361>
thanks
smatting_ has joined #nixos
<energizer>
i do too. it has advantages and disadvantages
<notgne2>
is there any workaround for https://github.com/NixOS/nixpkgs/issues/93704 ? I'm a little confused on where the issue is coming from but it's preventing me from rebuilding on unstable
<{^_^}>
[nixpkgs] @philandstuff opened pull request #93850 → yubikey-agent: fix on darwin → https://git.io/JJRZG
nature has joined #nixos
<notgne2>
or at least, I'm getting the error `ln: failed to create symbolic link '/nix/store/v1b8jrszwf53xf342pfnijd80ws614pz-etc/etc/tmpfiles.d/lvm2.conf': Permission denied` when rebuilding
cole-h has quit [Quit: Goodbye]
juliosueiras has quit [Ping timeout: 245 seconds]
nckx is now known as AMDinc
AMDinc is now known as nckx
<eyJhb>
Okay so I have this - https://termbin.com/rx4x and I get output - https://termbin.com/esnc , how con I specify that it should not look at is as a archive, but add it to the store or.. Idk? How should I go about this?
o1lo01ol1o has quit [Remote host closed the connection]
<Guest88361>
how long does it take in general to build a system?
<symphorien>
depends on your internet connection
<eyJhb>
And which packages you have
<notgne2>
and how many are already in a locally available store
<Guest88361>
it seems to stay for ages on 1/1/76 built] building source: 100 127 100 127 0 0 769
<symphorien>
if you suspect it's a hiccup during a download, you can safely interrupt it with ^C and start again
<__monty__>
eyJhb: Have you tried using a path literal rather than a string? I'm not sure why it's expecting an archive.
<Guest88361>
I don't know my connection is fine
<eyJhb>
Worked
<eyJhb>
Damn it __monty__
<eyJhb>
:P
<Guest88361>
I tried to interrupt and start again but it just says [1 build] just before interrupting
<notgne2>
Guest88361: some things have to be built locally too, afaik usually for license reasons
<notgne2>
nvidia drivers for instance build locally
<__monty__>
eyJhb: Yeah I run into that too sometimes. It's a little hard to tell when paths are ok/expected imo.
<Guest88361>
I assume I must not interrup all my pa download this way?
Fare has joined #nixos
<eyJhb>
Got another problem now, with fetchpatch stripping things away I guess
<notgne2>
Guest88361: if any derivations are downloading or building, cancelling could interrupt those, but everything else will remain and won't need re-downloading/building
<buffet>
hey, say im packaging a script, how should i handle binary dependencies? atm im using sed to find and replace all occurences, which works, but isnt pretty
Guest88361 has quit [Remote host closed the connection]
szicari has quit [Ping timeout: 244 seconds]
dbmikus has joined #nixos
proofofkeags has quit [Remote host closed the connection]
cr4y1 has quit [Remote host closed the connection]
szicari has joined #nixos
<__monty__>
buffet: resholved sounds like it might become what you're looking for.
<eyJhb>
Got it working.
<eyJhb>
But I hate finding the actual build drv
orivej_ has quit [Ping timeout: 246 seconds]
orivej has joined #nixos
mwelt has quit [Remote host closed the connection]
<energizer>
Guest88909: sometimes if i download too much stuff my modem gets bogged down
<energizer>
that happens when building a new system occasionally
<installingNIX>
I don't understand I download jdk from oracle
<energizer>
erm guest left
<installingNIX>
by the way just reboot I am guest
<energizer>
ok
<installingNIX>
so I did sudo nix-prefetch-url --type sha256 file:///home/nixos/Downloads/jdk-8u261-linux-x64.tar.gz as nix recommanded me
<installingNIX>
but I still have
<installingNIX>
builder for '/nix/store/dc5i1hlph07ni9lwnq1a9hpzi8fhw187-jdk-8u261-linux-x64.tar.gz.drv' failed with exit code 1; last 10 log lines: *** Unfortunately, we cannot download file jdk-8u261-linux-x64.tar.gz automatically. Please go to http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html to download it yourself, and add
<installingNIX>
it to the Nix store
<clever>
installingNIX: the .tar.gz you download, must be binary identical to the one nix is expecting
<installingNIX>
the sha hash must be the same?
<clever>
yeah
<installingNIX>
can I change the expected sha hash?
<installingNIX>
would that be a solution?
<clever>
you can do a package override against it, as with any other package
<adyatlov>
Can I see build output from a running nix-daemon with journalctl ?
<clever>
adyatlov: nope, the logs always go to /nix/var/log/
<clever>
adyatlov: `nix-store -l` and `nix log` can read them
<installingNIX>
where can I see the exected sha during the install?
<clever>
installingNIX: run `nix show-derivation` on the failing .drv file
<clever>
installingNIX: its the outputHash
<adyatlov>
clever: Is that a real time log?
<adyatlov>
Lorri fired off a long compilation, can't get to it
<clever>
adyatlov: only the process triggering the build can get realtime logs
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93856 → knetwalk: init at 20.04.3 → https://git.io/JJRnh
smatting_ has quit [Ping timeout: 240 seconds]
jumper149 has joined #nixos
<installingNIX>
editing the nix/store/...drv
<installingNIX>
is a bad idea?
<installingNIX>
I assume
<eyJhb>
installingNIX: /nix/store is readonly
<eyJhb>
What are you trying to do
<clever>
installingNIX: very, thats why you do an override in the expression
<installingNIX>
smart
<installingNIX>
:)
<installingNIX>
(the read-only)
<AmandaC>
,locate xkbcommon.pc
<{^_^}>
Found in packages: libxkbcommon.dev, libxkbcommon_7.dev
<installingNIX>
the documentation says to override to as in this exemple
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93857 → knights: init at 20.04.3 → https://git.io/JJRc3
<installingNIX>
environment.systemPackages = with pkgs; [ (python35.withPackages(ps: with ps; [ numpy toolz ])) ];
<installingNIX>
I should do
<clever>
installingNIX: you want something like oraclejdk.overrideAttrs (old: { src = ./path/to/file.tar.gz; })
<AmandaC>
,locate gles.pc
<installingNIX>
environment.systemPackages = with pkgs; [ (oraclejdk8.withPackages(outputHash: with outputHash; [ newSha ])) ];
<{^_^}>
Couldn't find in any packages
<colemickens>
Where can I see the website portion of how AWS images are shown on nixos.org?
<clever>
installingNIX: .withPackages is a python only thing
<installingNIX>
@clever so I don't override the sha but the path?
<clever>
installingNIX: outputHash isnt a function
davidv7 has joined #nixos
xelxebar has quit [Remote host closed the connection]
<installingNIX>
I must say I find the syntax really hard to read
<clever>
installingNIX: it would be more like oraclejdk8.overrideAttrs (old: { src = requireFile { hash = "..."; something = "something"; })
xelxebar has joined #nixos
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93858 → knavalbattle: init at 20.04.3 → https://git.io/JJRc8
<clever>
installingNIX: oraclejdk8.overrideAttrs says to look at the attribute on oraclejdk8 that is called overrideAttrs, exactly the same as object.attr would in js or c++
<clever>
installingNIX: and then it is followed by an argument in (), so treat overrideAttrs as a function, and call it
<clever>
installingNIX: old: { ... } is then another function, that is given the old set, and returns a set of things to change
<adyatlov>
Can I just do nixos.jetbrains.clion.override { version = "2020.2" } in my overlay?
<{^_^}>
[nixpkgs] @veprbl pushed commit from @oxalica to master « osu-lazer: 2020.714.0 -> 2020.725.0 (#93726) »: https://git.io/JJRcp
<clever>
it will insert the current host platform (32bit or 64bit) into that, and do something like sha256.x86_64-linux
<installingNIX>
wait when I do nix edit …
<installingNIX>
I can edit the local defination of the package?
<clever>
the file it opens is read only
<installingNIX>
or I can't since the store is read only
<installingNIX>
okay
<clever>
2020-07-25 19:38:53 < clever> installingNIX: you need to read the old expression, and then copy/paste the requireFile part over, changing the hash
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93862 → kteatime: init at 20.04.3 → https://git.io/JJRCf
<installingNIX>
I am afraid I don't undenstand
<installingNIX>
old is the name of a expersion in the nix file?
<clever>
adyatlov: fetchTarball returns the path to nixpkgs
<clever>
adyatlov: you can also newPkgs.path to get the path
<installingNIX>
but my modem is working fine otherwise we would not be chating
<clever>
installingNIX: requireFile never downloads anything
<installingNIX>
no it's my sudo nixos-install command
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93868 → katomic: init at 20.04.3 → https://git.io/JJRCQ
<installingNIX>
runned with -vv
<clever>
installingNIX: what is it downloading?
<installingNIX>
last thing is building '/nix/store/6ybb0ysfk33zf4fr1h35g1w0v42nf033-source.drv'..
<clever>
installingNIX: run `nix show-derivation` on that drv, what is the srcs?
<installingNIX>
it seems this drv isn't in my /nix/store
<clever>
installingNIX: its in /mnt/nix/store/
<installingNIX>
error: path '/mnt/nix/store/6ybb0ysfk33zf4fr1h35g1w0v42nf033-source.drv' is not in the Nix store
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93869 → kblocks: init at 20.04.3 → https://git.io/JJRCd
work_ has joined #nixos
<kekw736[m]>
anyone having problems with custom css for firefox ?
zeta_0 has joined #nixos
<installingNIX>
is there a way to clean/reset my nix-store?
<infinisil>
,ask kekw736[m]
<{^_^}>
kekw736[m]: Just ask your question. It's the best way to know if anybody can help. Or look at other ways of getting in touch with the Nix community here: https://nixos.wiki/wiki/Get_In_Touch
<clever>
installingNIX: you dont need to
rixily has quit [Quit: Leaving]
<installingNIX>
what should I do than?
<installingNIX>
error: path '/mnt/nix/store/6ybb0ysfk33zf4fr1h35g1w0v42nf033-source.drv' is not in the Nix store
<installingNIX>
I assume this mean I have a problem with my drv file?
<clever>
installingNIX: no, the check is checking if the path begins with /nix/store/
<clever>
installingNIX: just run cat on it then, what is the srcs ?
tilcreator has quit [Ping timeout: 258 seconds]
<kekw736[m]>
<infinisil ",ask kekw736"> I am trying to customize firefox with userChrome.css and it doesn't work. I'm curious if this is a nixos thing
<zeta_0>
do you mind taking a look at the `paste: end paste` section: https://dpaste.org/1MMt
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93870 → kalzium: init at 20.04.3 → https://git.io/JJRWv
<clever>
installingNIX: its downloading nerd-fonts, which is a fairly big file
<clever>
installingNIX: github also doesnt report the size of the file, so you have no progress information
<zeta_0>
the `?` underneath `paste:` is throwing the `unexpected symbol ?` syntax error, so i'm guessing i'm going to need to move all of those?
tilcreator has joined #nixos
<clever>
installingNIX: all you can do is wait
<clever>
installingNIX: or not install nerd-fonts
<installingNIX>
okay
<installingNIX>
do I nead them?
<clever>
installingNIX: only if you want them
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93872 → kturtle: init at 20.04.3 → https://git.io/JJRWt
<installingNIX>
I read someone advising to use theç
<zeta_0>
i'm basically trying to wrap all the nix code in the shell.nix of the ghc.nix repo with code generated by `lorri init`
<installingNIX>
anyway
<clever>
zeta_0: line 23-37 is defining a function, but you jammed it into the middle of buildInputs, which wants a derivation
<installingNIX>
atleast I know why I don't have any progress bar :)
<clever>
zeta_0: you want to put that function into its own file, and then do (pkgs.callPackage ./ownfile.nix {})
<clever>
installingNIX: github is also rather slow at generating those files, i have fiber internet, and its still not done
<kekw736[m]>
<infinisil "kekw736: Since I just added a us"> yep, that was definitely on me. I thought I had that enabled through user.js but it was mispelled :). Well... this made me check thanks!
<infinisil>
:)
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93873 → klettres: init at 20.04.3 → https://git.io/JJRWs
<clever>
installingNIX: download finished, its 2.1gig
<installingNIX>
okay
<installingNIX>
well for now I will comment this part of my configuration
<zeta_0>
clever: thanks, i'll try implementing your solution now
orivej has quit [Quit: No Ping reply in 180 seconds.]
<installingNIX>
in the service do I need to add it as a package?
orivej has joined #nixos
<clever>
installingNIX: if you enable xfce, it installs it for you
<installingNIX>
does it set all the classic desktop thingy
<installingNIX>
like automount?
<installingNIX>
of usb key
<clever>
not sure
mbrgm_ has joined #nixos
mbrgm has quit [Ping timeout: 260 seconds]
mbrgm_ is now known as mbrgm
dsiypl4_ has joined #nixos
orivej has quit [Ping timeout: 240 seconds]
orivej has joined #nixos
<zeta_0>
clever: would the best approach be to run `lorr init`, then in lorri's generated shell.nix file, use the callPackage function to call ghc.nix's default.nix file and run it's code(i'll rename default.nix so lorri does not collide with shell.nix)?
<zeta_0>
clever: what do you think is the best approach for my issue?
<clever>
zeta_0: maybe
dsiypl4 has quit [Ping timeout: 265 seconds]
<adyatlov>
clever: Ok, this is the solution I made. I copied the files like so http://ix.io/2slR Any remarks?
<clever>
adyatlov: if your not changing common.nix, you can just do (pkgs.path + "/nixos/pkgs/applications/editors/jetbrains/common.nix") for its path
<adyatlov>
clever: I don't think path is in scope from jetbrains.nix
Zer000 has quit [Quit: Leaving]
<clever>
adyatlov: then add pkgs to the args where it gets every other package
<installingNIX>
so I can't both use networkmanager and preconfigure my wifi?
<installingNIX>
the only way would be to go throw wpa supplicant?
<clever>
installingNIX: personally, i use naked wpa_supplicant, with /etc/wpa_supplicant.conf not managed by nix
<clever>
installingNIX: so i can add a pw, without having to nixos-rebuild, which can be a problem, when you dont have internet
<installingNIX>
pw?
<installingNIX>
password?
<clever>
password
afreak has quit [Ping timeout: 264 seconds]
<installingNIX>
beceause if it was managed by nix you could not add one both manually and in the nix conf file?
<clever>
installingNIX: yeah, if nix manages the wifi config, you have to nixos-rebuild to change the wifi config
<clever>
and then you get into problems where you need wifi to fix the wifi
<installingNIX>
ouch
<installingNIX>
I naively thought nix would create the conf file for wpa_supplicant but than you colud manually edit them and it would be erase at the next rebuild
<clever>
everything nix creates is read-only
<installingNIX>
I am attracted to nix cause I've been using either debian/ubuntu/arch
<installingNIX>
I always ended up in some upgrade creating bugs
<installingNIX>
or small regression
<clever>
same, ran into that on both ubuntu and gentoo
<installingNIX>
would you say nix fix that issue with the possibility to roll-back and upgrade?
<clever>
yeah
<clever>
and also how each version is self contained and pure
<clever>
it doesnt depend on the mess of config from past versions