day|flip has quit [Remote host closed the connection]
noudle has quit []
npmccallum has joined #nixos
matt`` has joined #nixos
matt` has quit [Ping timeout: 268 seconds]
matt`` has quit [Remote host closed the connection]
justanotheruser has joined #nixos
justan0theruser has joined #nixos
justanotheruser has quit [Ping timeout: 244 seconds]
jluttine has joined #nixos
lsyoyom has joined #nixos
<hpfr[m]>
Compton isn’t working for me
<hpfr[m]>
I ran it in a terminal emulator and started opening windows, and I get a badwindow error every time. I’m on i3
<hpfr[m]>
It worked fine on arch
<hpfr[m]>
Never mind. I needed a config file. Is it normal for nixos to just not make default config files when you enable packages in configuration.nix?
xkapastel has quit [Quit: Connection closed for inactivity]
<mddawley>
It's a WiFi only Chromebook. I've already gotten it to the point of booting from a USB stick but I can't get WiFi working so I can download packages
abathur has joined #nixos
<mexisme>
hi there, I'm also struggling with Go on MacOS (as per @psalm + @LnL conversation in 12th April)
<mexisme>
I understand the underlying problem with not having access to include files
<mexisme>
but my difficulty is that I use VSCode, and any of the IDE tooling will always complain about being unable to find CoreFoundation, et al
<mexisme>
anyone has any thoughts / pointers for dealing with this?
<mexisme>
it's sounding a bit like the only easy way forward might be to write a bunch of shell-level wrappers around go, golang-ci, et al
<mexisme>
(btw: apologies if this is the wrong channel to ask)
<mexisme>
as an aside: I've been kind-of wondering if this would be a problem even in NixOS (non-Darwin) as well?
mexisme has quit [Quit: WeeChat 2.5]
<abathur>
heh
mexisme has joined #nixos
<abathur>
not sure about the acute question, does VSCode provide a way to introspect/debug where it's looking for files?
<abathur>
the nix-darwin channel might be a worthwhile resource
<abathur>
erg; might be a good resource if just because it has a higher density of macos+nix users
<hpfr[m]>
I’m confused with some of the modules for configuration as well. For example, Compton has a fair number of options available, but there’s plenty more that can only be changed in a config file. Is the split just arbitrary? Seems like it adds needless complexity. Furthermore, if I have something like services.compton.enable set to true, why don’t I need to include Compton in my pkgs? Now half my packages are enabled throug
<hpfr[m]>
module options and half are in the actual list
<mexisme>
aha, yes that's more appropriate for my Mac question
<mexisme>
but, i did wonder whether this might probably affect any IDE that does introspection
<abathur>
mexisme: it's specifically for the nix-darwin package/project--I don't think you're really *off topic* here; just saying that you might get higher quality (if fewer) answers there
<mexisme>
VSCode uses Language Servers (like gopls) a lot
<mexisme>
i'll re-ask the Darwin portion over there
<abathur>
I do use nix on darwin, but use neither go nor vscode, so I'm not certain specifically
<mexisme>
am i right to think that any IDE that wants to do introspection, therefore needs to find include libs, etc, might have a problem?
<mexisme>
Go might be slightly unusual in that it makes it quite easy to do reflection, and some types of coding requires it
<abathur>
maybe? it probably depends on how they try to look, and whether the projects you're building/installing with nix are in your profile
<teto>
mexisme: you can run the IDE in a nix-shell with the appropriate config. Alternatively, I have some code that generates a per-project neovim configuration so that it can find what it needs (and only that).
<mexisme>
right, so that's helpful.
<mexisme>
i couldn't quite tell if i was bashing my head against the problem (and being dumb)
<mexisme>
i think maybe i need to teach VSCode how to run exactly the commands I need.
<mexisme>
teto: i'd love to see your neovim generation code, btw
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<monokrome>
,locate libpg-fe.h
<monokrome>
,locate libpq-fe.h
<{^_^}>
Found in packages: postgresql94, postgresql95, postgresql96, postgresql_10, postgresql_11
<{^_^}>
[nixpkgs] @worldofpeace pushed 3 commits to master: https://git.io/fjPHv
<AnswerGuy>
Under NixOS is every user given an isolated environment? I just added a new (headless) user to my /etc/nixos/configuration.nix, did a nixos-rebuild switch, and go this error: gpg-connect-agent: failed to create temporary file '/home/ops/.gnupg/.#lk0x000 ... No such file; but also trying to use a simple git pull I got "file not found"
<AnswerGuy>
I'm sure I have it for my other account.
* AnswerGuy
checks
<AnswerGuy>
Yep.
Shem[m] is now known as Catalyst[m]
<AnswerGuy>
(I'm setting up gitea with a repo for operations, and testing a read-only deployment key).
<bob_twinkles>
is there a good guide for integrating "one off" libraries? I'm trying to build some software that relies on forks of upstream libraries and it would be helpful to not have to try and shove them into a global nixpkgs fork
<teto>
simpson: yep for sure it will bring benefits. It will replace NUR, make the transition of projects to nix easier, also potentially allow to "purge" nixpkgs of some packages that could be moved to flakes
Okinan has joined #nixos
abathur has quit [Ping timeout: 248 seconds]
mexisme has joined #nixos
adamantium has joined #nixos
ensyd_ has quit [Quit: Leaving]
adamantium has quit [Remote host closed the connection]
stepcut has quit [Read error: Connection reset by peer]
andersk has quit [Read error: Connection reset by peer]
spacekookie has quit [Quit: No Ping reply in 60 seconds.]
stepcut has joined #nixos
spacekookie has joined #nixos
raoul has quit [Quit: Ping timeout (120 seconds)]
raoul has joined #nixos
rembo10 has quit [Remote host closed the connection]
trfl has quit [Remote host closed the connection]
duckfullstop has quit [Quit: Why do you need...?]
trfl has joined #nixos
duckfullstop has joined #nixos
andersk has joined #nixos
rembo10 has joined #nixos
<bob_twinkles>
can I use an overlay just for a specific nix-shell? It seems like they must be configured globally
<slack1256>
you can reuse the pattern of overlays at different levels
mexisme has quit [Ping timeout: 248 seconds]
<slack1256>
For example pkgs.haskellPackages.overrides expects a function that works as a overlay but just for that subtree
<teto>
bob_twinkles: you can add include paths to NIX_PATH via -I nixpkgs-overlays=/toto
<bob_twinkles>
so if I have a directory structure with shell.nix at the root, and then derivations/default.nix (overlay defn. containing libmyfork = super.callPackage ./libmyfork {}; ) and a derivation for libmyfork at derivations/libmyfork/default.nix
<bob_twinkles>
and invoke nix-shell with -I nixpkgs-overlays=$(pwd)/derivations, that should pick up libmyfork?
<bob_twinkles>
I should probably pastebin some of this stuff
<nefix>
Hello! I'm trying to pack chronograf. The thing is that before building, it needs to have some "go generate" things and have some JavaScript code built. How is the JavaScript suposed to be built? It needs to be embedded after
endformationage has quit [Quit: WeeChat 2.5]
<bob_twinkles>
oh, you mean I should just write a package and then invoke a shell there?
<bob_twinkles>
that's probably a good idea
<teto>
bob_twinkles: yep
<bob_twinkles>
do I need an overlay in that case?
gyroninja___ has quit [Ping timeout: 244 seconds]
<bob_twinkles>
no, I can just use the same trick above right
<teto>
bob_twinkles: yep nix-shell by default looks into default.nix then shell.nix
<teto>
so you can write your package in a default.nix and then just run "nix-shell"
<bob_twinkles>
does it run default.nix in a callPackage context? I think I need things like the autoconf hook
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<slack1256>
Where are usually the ssl certificates in nixos?
<slack1256>
Oh, there are symlinks on /etc
<slack1256>
nvm
<slack1256>
ls
markus1189 has quit [Ping timeout: 245 seconds]
iqubic has joined #nixos
mexisme has joined #nixos
<{^_^}>
[nixpkgs] @danbst merged pull request #64267 → system.autoUpgrade: optionally allow rebooting the system on kernel change → https://git.io/fj6sm
<{^_^}>
[nixpkgs] @danbst pushed commit from @tokudan to master « system.autoUpgrade: optionally allow rebooting the system on kernel change (#64267) »: https://git.io/fjPHj
sushi_ has quit [Ping timeout: 258 seconds]
kvda has joined #nixos
kvda has quit [Client Quit]
iqubic has left #nixos ["ERC (IRC client for Emacs 26.2)"]
ashesham` has joined #nixos
iqubic has joined #nixos
<teto>
bob_twinkles: I never remember, worst case you add a "with import <nixpkgs> {};" on the first line
kvda has joined #nixos
<iqubic>
Anyone know how to change the look of the X11 cursor. Also commonly known as the mouse.
<teto>
I am getting "HTTP error 200 (curl error: Failure when receiving data from the peer) error 7 while decompressing xz file" from cachix.org. anyone ?
hyper_ch2 has quit [Remote host closed the connection]
jasongrossman has quit [Ping timeout: 248 seconds]
<abbec>
hi, quick question, I am using nix copy to create a local binary cache for a rust project but some derivations seem to not get copied to the local directory
<abbec>
any hints on what to look for?
hyper_ch2 has quit [Remote host closed the connection]
<abbec>
for example "rust-src" derivation seems to be copied and cached but not "clippy-preview" and "rustfmt-preview"
hyper_ch2 has joined #nixos
<clever>
abbec: nix copy will only get the runtime deps, but not all of the build time deps
<abbec>
but I tell it to copy buildInputs
<clever>
but not all of the inputs are in buildInputs
jgt has joined #nixos
<abbec>
hmm ok but "rust-src" is specified at the exact same place as "clippy-preview" for example
<domenkozar[m]>
most probably for backwards compatibility
knupfer has quit [Ping timeout: 244 seconds]
stepcut_ has joined #nixos
<teto>
even if I set it as a list of string, mkMerge doesn't seem to work, options are still overriden instead of merge hum..
stepcut has quit [Ping timeout: 245 seconds]
justanotheruser has quit [Ping timeout: 246 seconds]
ThatDocsLady has joined #nixos
no-n has joined #nixos
<{^_^}>
[nixpkgs] @mrVanDalo opened pull request #64653 → bitwig-studio3: add version 3 for bitwig → https://git.io/fjP5l
<abbec>
hmm, after running nix-instatiate, garbage collection does not clean up my shell packages anymore?
karetsu has quit [Quit: WeeChat 2.4]
<abbec>
hmm now it did
<abbec>
strange
<colemickens>
worldofpeace: I'm not sure. When I was carrying commits, the mesa definition looked a bit different. I guess if mesa builds iris by default ("auto") then we are. I can try in a bit, need to rebuild chromium again remotely first.
<colemickens>
since others were interested in iris in 19.1 as well, I'm inclined to think we probably are building it...
kvda has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
<__monty__>
Just ran into: "error: Package ‘xvfb-run’ in /nix/store/.../xvfb-run/default.nix:10 is not supported on ‘x86_64-darwin’, refusing to evaluate." How do I find out why this is being built? Since I'm not requesting xvfb-run explicitly in my config.
<clever>
__monty__: --show-trace?
hyper_ch2 has quit [Remote host closed the connection]
<__monty__>
jgt: I think it's because hashable depends on quickcheck, which depends on hashable...
<__monty__>
I'm not sure how to break the loop properly though. Easiest way might be to dontCheck hashable.
<__monty__>
Maybe you shouldn't be using hpNew.callpackage?
<jgt>
__monty__: that's strange; this is QuickCheck.nix http://ix.io/1OeP#
<__monty__>
Ok, so at the bottom of your error, there's some blurb about "infinite recursion," then I looked at where it occurred "while evaluating blah blah 'hashable'," so I looked up hashable in nixpkgs (I walk the github tree down to hackage-packages.nix then search that with C-f, I'm sure there's a better way.) then I look at what hashable depends on and I see Quickcheck and remembered that from further up in
<__monty__>
your error.
<__monty__>
Same thing basically. Depends on Quickcheck.
<jgt>
__monty__: same thing, but this time with hspec-meta?
<__monty__>
I think the real problem is you're overriding Quickcheck improperly. But other than vaguely hinting that overlays are clearer than overrides I can't really help with that.
<__monty__>
clever++ --show-trace hit the spot.
<{^_^}>
clever's karma got increased to 152
___laika has joined #nixos
<tdeo>
anyone know how to switch a PR from master to staging without pinging a bunch of people? no matter what order I do it in it still does it
<tdeo>
i should just check rebuilds before making one, i guess
<etu>
tdeo: Checkout staging locally, make a new branch based on that. Cherry-pick over your commits from your branch based on master.
___laika has quit [Ping timeout: 245 seconds]
<tdeo>
yep, i did that this time, github still notifies the code owners though
bvdw has quit [Read error: Connection reset by peer]
pie_ has joined #nixos
abathur has joined #nixos
___laika has joined #nixos
Mateon2 has joined #nixos
Mateon1 has quit [Ping timeout: 248 seconds]
Mateon2 is now known as Mateon1
npmccallum has joined #nixos
<abathur>
odd question I guess, but I've been fiddling around with some shell profile/rc scripts that print the time (i.e., printf "%(%T %Z)T") at some points and noticed some indeterminate behavior from how my timezone resolves within a nix-shell as it starts up
<abathur>
so there'll be a point during that startup where a single one fails to resolve and print UTC; I checked the /etc/timezone link both before and after the UTC printf
<abathur>
and it's fine at both points
is_null has quit [Ping timeout: 246 seconds]
<abathur>
but it works if I check the link first; so either looking before I leap fixes it, or there's some sort of race-condition?
gagbo has quit [Ping timeout: 244 seconds]
<abathur>
also, it seems to work fine if I explicitly set TZ in the outer environment
<clever>
abathur: the TZ env var can override /etc/localtime
<clever>
TZ='America/Moncton' date
<clever>
that will always print it in my zone, and ignore whatever zone your in
knupfer has quit [Quit: knupfer]
gagbo has joined #nixos
knupfer has joined #nixos
<abathur>
yeah; in the case described above it worked when I set "export TZ=:/etc/timezone"
bakakuna has quit [Remote host closed the connection]
bakakuna has joined #nixos
Guest30053 has quit [Quit: WeeChat 2.4]
Makaveli7 has quit [Quit: WeeChat 2.5]
grabb0id has joined #nixos
avalan has joined #nixos
justanotheruser has joined #nixos
srid6390803 has quit [Ping timeout: 248 seconds]
wrl_ has joined #nixos
dweller has quit [Disconnected by services]
avalan is now known as dweller
wrl has quit [Read error: Connection reset by peer]
_deepfire has quit [Ping timeout: 245 seconds]
JadoJodo has quit [Quit: Ping timeout (120 seconds)]
MasterdonY has quit [Ping timeout: 245 seconds]
<grabb0id>
Hi all, I'm trying to add network interfaces for VLAN tagged traffic. Messing around with vconfig and ip tools it seems to work, however is there also a way to declaratively specify this stuff in the config somewhere?
<{^_^}>
[nixpkgs] @luispedro opened pull request #64669 → sortmerna: init at 2.1 → https://git.io/fjPxW
Diagon has quit [Quit: Leaving]
<sauyon>
what package is lib32-gcc in?
M-Gregoire49 has joined #nixos
M-Gregoire4 has quit [Ping timeout: 258 seconds]
gagbo has quit [Ping timeout: 246 seconds]
<clever>
sauyon: pkgs.pkgsi686Linux is the 32bit version of nixpkgs, so pkgs.pkgsi686Linux.callPackage and pkgs.pkgsi686Linux.stdenv will just give you 32bit everything
<sauyon>
ah, awesome, thanks!
cfricke has joined #nixos
<sauyon>
I'm currently building an internal repository that requires both 32bit and 64bit everything and is a bit of a cluster...
<nakkle>
Hi all, I have a bit of a problem understanding, where exactly nix-builds are reproducible, in contrast to managing my environment with, e.g., apt. If I do not pin my build to a specific nixpkgs snapshot, then (as far as I understand it) the reproducibility is completely dependent on me never running nix-channel --update. There's no guarantee that someone else building my package will get the same result on their machine. It seems to me
<nakkle>
hat as long as I don't pin my package to a specific version, a build environment managed by nix is as reproducible as a build environment managed manually by apt. Is my understanding of this issue correct, or did I miss something?
<sauyon>
hmmm apparently adding pkgs.i686Linux.libgcc didn't help, I guess that's because of the default GCC environment?
<xcthulhu>
nakkle: That’s not how you pin to a version snapshot
<clever>
sauyon: i'm not sure what the right way is to get both 32 and 64 at once
<xcthulhu>
nakkle: The whole idea is that you have `{pkgs ? import <nixpkgs> {}, …}:` in all of your `.nix` files, and if you want to version pin you force `pkgs` to point to a particular git revision.
<xcthulhu>
(if pkgs isn’t pegged to some particular revision then you have to be careful about doing nix-channel --update)
acarrico has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
___laika has joined #nixos
grabb0id has joined #nixos
alex`` has quit [Quit: WeeChat 2.5]
<nakkle>
xcthulhu: yeah, that's what I meant by pinning. so basically, as long as I don't pin nixpkgs in my files, there's no guarantee that what I'm building is reproducible, right?
__monty__ has quit [Quit: leaving]
<clever>
nakkle: i usually start with pinning nixpkgs to whatever i'm currently on
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<nakkle>
are there any best practices on how I get important security updates when I pin my dependencies? Just don't forget to regularly update the pinned nixpkgs version?
Baughn has joined #nixos
xcthulhu has quit [Quit: xcthulhu]
FRidh has quit [Quit: Konversation terminated!]
is_null has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed 4 commits to release-19.03: https://git.io/fjPxh
___laika has quit [Ping timeout: 245 seconds]
grabb0id has quit [Ping timeout: 258 seconds]
shomodj has joined #nixos
<sauyon>
I can't tell if using gcc-unwrapped has solved my problem or not because now I'm getting locale errors T.T
emptyflask has joined #nixos
grabb0id has joined #nixos
grabb0id has quit [Read error: Connection reset by peer]
<clever>
arcnmx: try nix-build pkgs/top-level/release.nix -A perl528Packages.StringShellQuote.x86_64-darwin
<clever>
in a checkout of nixpkgs master
romildo has joined #nixos
<arcnmx>
clever: as in, it might be fixed already?
virus_dave has joined #nixos
<clever>
arcnmx: dont think so yet, that will just confirm it
pie__ has joined #nixos
<arcnmx>
sure, I was building against nixpkgs-unstable so not too far off and am getting the same log as hydra (which despite VERBOSE doesn't seem to be very...)
pie_ has quit [Remote host closed the connection]
pie__ has quit [Remote host closed the connection]
pie__ has joined #nixos
<arcnmx>
3 commits behind :p
romildo has quit [Client Quit]
pie__ has quit [Remote host closed the connection]
pie__ has joined #nixos
pie__ has quit [Remote host closed the connection]
pie_ has joined #nixos
<clever>
arcnmx: i can reproduce the failure on master
<clever>
arcnmx: and with the nixpkgs rev on the green build, its in the binary cache
<clever>
Bisecting: 574 revisions left to test after this (roughly 9 steps)
<arcnmx>
yeah was trying to see if I could bisect by finding the first failure point on hydra but I'm not sure how to navigate it to find out what dependency caused the rebuild
<clever>
for both the failing and passing one, click the inputs tab
<arcnmx>
Yeah, there's that... a few weeks between them though, I was hoping hydra could tell me which dependencies triggered the failing build to begin with, but the Build Dependencies list is linking to builds that failed so that doesn't make sense to me...
domogled has quit [Remote host closed the connection]
domogled has joined #nixos
<arcnmx>
mostly just because I wanted to understand the interface, thanks for running a bisect!
xkapastel has quit [Quit: Connection closed for inactivity]
<ajs124>
You're also having problems with perlPackages? Maybe you're also hitting this:
o1lo01ol1o has quit [Remote host closed the connection]
is_null has quit [Ping timeout: 248 seconds]
bakakuna has quit [Ping timeout: 246 seconds]
<arcnmx>
ajs124: oh I was really confused in trying to see if that was the cause before I realised it was merged into staging first... but yeah, seems likely to be the problem
siel has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @veprbl merged pull request #64460 → cdrtools: fix for darwin → https://git.io/fjiig
<{^_^}>
[nixpkgs] @veprbl pushed commit from @Amar1729 to master « cdrtools: fix for darwin (#64460) »: https://git.io/fjPjL
Jackneill has quit [Remote host closed the connection]
Drakonis has joined #nixos
shomodj has joined #nixos
gagbo has quit [Ping timeout: 268 seconds]
gagbo has joined #nixos
gxt has quit [Ping timeout: 260 seconds]
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
<toppler>
I'd like to package an open source computer game (Keeper RL), an ASCII rogue-likish which has optional licensed graphical assets. By default, if the game's binary has these in /data_contrib of its working directory, they will be used. I obtained mine from Steam, and play in a nix-shell, but I'd like to package it properly. Are there any examples of similar packages I can copy where the user can say that they have obtained the assets and
<clever>
toppler: you could have one derivation called keeper-unwrapped, that contains just the raw game, then something like keeper-assets or keeper-map, that uses requireFile to download it
<clever>
toppler: then have a keeper package using runCommand or buildEnv, to merge the 2 together
<timokau[m]>
How can it happen that instantiating the same derivation on the same nixpkgs derivation gives two different results on two different machines?
<stites>
timokau[m] well, one could be OSX and the other Linux
<stites>
the toolchains for the foundational builds are all totally different
<timokau[m]>
stites: No, both linux. I noticed that I wasn't getting any substitutes for that package recently, so I compared with a different linux machine and got different derivations
orivej has quit [Ping timeout: 246 seconds]
<samueldr>
timokau[m]: overlays on the system?
<samueldr>
(well, in NIX_PATH might be more appropriate to say)
<stites>
: ) out of my depth then. I imagine there are different tricks required to work around different operating systems (even, perhaps, different versions of operating systems), but I can't really say speak to that
<samueldr>
and the "did you try turning it off and on again" type question: for the same revision of nixpkgs? :)
<infinisil>
timokau[m]: Yeah overlays and config are the other impure pieces
<timokau[m]>
samueldr: at least nothing in `~/.config/nixpkgs/overlays`, could there be other ways to add them?
<timokau[m]>
samueldr: Yeah same rev, I double and triple checked. Same nix-path too.
<timokau[m]>
Are there any other config values that could have an effect besides overlays?
<infinisil>
timokau[m]: See the link I posted..
noudle has quit [Read error: Connection reset by peer]
<infinisil>
You can eliminate the overlays and config impurities by passing overlays = [] and config = {}
noudle has joined #nixos
Xal has joined #nixos
Shoubit has quit [Ping timeout: 246 seconds]
noudle_ has joined #nixos
<stites>
(last call for the linode referral!)
noudle_ has quit [Client Quit]
<gchristensen>
stites: maybe try #nixos-chat :)
<stites>
oh! okay! sorry
<gchristensen>
no worries, not a problem, but youmight get some bites there :)
noudle has quit [Ping timeout: 252 seconds]
<timokau[m]>
infinisil: ah, sorry missed that. But `system` is also the same on both systems and even eliminating all overlays and config still results in different drvs
<timokau[m]>
Okay there's something fishy with my nixpkgs checkout. Its on the right revision and reports a clean working tree, yet I get a different result with a fresh checokut
<clever>
timokau[m]: git diff --staged ?
<{^_^}>
[nixpkgs] @costrouc opened pull request #64677 → qt-box-editor: init at unstable-2019-07-12 → https://git.io/fjXf7
<timokau[m]>
clever: Yeah that gives the same as the fresh checkout and the other machine
<timokau[m]>
clever: Empty result
<clever>
git status?
dashkal has quit [Ping timeout: 246 seconds]
<timokau[m]>
working tree clean. Also empty `git diff 95395fbf549908e8fe59922653e9dadd4da4792e`
<timokau[m]>
But that still doesn't explain why I also don't get the substitute when I tried to update my system using channels o.O
<clever>
timokau[m]: what is the storepath for the thing its trying to build?
<timokau[m]>
clever: any reason why some invalid store path could mess with this on one repo but not the other? Or just grasping at straws? (its running)
Soo_Slow has quit [Remote host closed the connection]
<timokau[m]>
And where should I put the --pure? I though that's only a nix-shell arg
<clever>
timokau[m]: its also a nix-instantiate arg i believe
<clever>
it disables using NIX_PATH and other things
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<timokau[m]>
clever: `diff` reports a bunch of differences in .giff of course, otherwise nothing interesting (a `.version-suffix` file, nix-reviews `.review`, an empty folder)
<timokau[m]>
nxi-store --verify is still working, maybe my store is too big
<timokau[m]>
A fresh worktree of the original repo gives the same result as the new checkout
shibboleth has joined #nixos
<timokau[m]>
Very curious
<clever>
timokau[m]: so, uhhh, what was the original problem? lol, i didnt see it
o1lo01ol1o has quit [Ping timeout: 244 seconds]
tokudan has joined #nixos
<timokau[m]>
clever: Well the original original problem is that I keep rebuilding sage on channel updates. So I checked hydra, and noticed that I even have to rebuild it when I manually check out the latest version hydra built
virus_dave has quit [Quit: virus_dave]
<timokau[m]>
So then I noticed that the nixpkgs checkout I usually use gets a different result for `nix-instantiate` than a clean checkout (the clean checkout also gets the hydra substitute)
<timokau[m]>
But to make everything even weirder, my system uses channels so the difference in my local checkout doesn't really explain why I have to rebuild on channel updates
<clever>
timokau[m]: if you are getting 2 different .drv files for the same nixpkgs rev, run nix-diff on both drv files
<clever>
that should answer things
<timokau[m]>
I'm sure there is a perfectly simple and monumentously stupid reason for all this
<timokau[m]>
--verify --check-content finished without any output by the way
<timokau[m]>
Good idea
gagbo has quit [Ping timeout: 268 seconds]
gagbo has joined #nixos
<timokau[m]>
clever: Can you help me interpreting the output? I have no experience with nix-diff and have a hard time understanding it
<clever>
Bisecting: 13 revisions left to test after this (roughly 4 steps)
<clever>
and this is the 2nd python thing i'm bisecting now, the other one is still stuck in mass-rebuild territory :P
<clever>
it has had 13 revs left for hours
<clever>
eval-time problems are trivial compare to darwin-only build problems
<timokau[m]>
clever: You want to do it? I should really be working, this was just too juice an opportunity to procrastinate
<clever>
timokau[m]: same :P
<timokau[m]>
clever: Okay then I can do it, it was my problem after all and you already helped :)
o1lo01ol1o has quit [Ping timeout: 268 seconds]
<{^_^}>
[nixpkgs] @timokau opened pull request #64678 → python.pkgs.scipy: inherit the correct version → https://git.io/fjXJ2
<timokau[m]>
clever: ^. Thanks for your help :)
<clever>
timokau[m]: your welcome
myskran has joined #nixos
<timokau[m]>
There are many welcomes like this, but this one is mine
<timokau[m]>
(First time I'm seeing your github handle, although I see you in irc all the time)
<adamantium>
Hi-- Home-Manager Question: home.StateVersion ... do I need this? I was browsing https://rycee.gitlab.io/home-manager/options.html and saw this and realized it is an option that may be declared
Myhlamaeus has joined #nixos
abathur has quit [Ping timeout: 248 seconds]
johnny147 has joined #nixos
johnny101m has quit [Ping timeout: 246 seconds]
domogled has joined #nixos
<{^_^}>
[nixpkgs] @ambrop72 closed pull request #60365 → buildEnv: Don't warn about collisions if contents match when ignoreCollisions=true. → https://git.io/fjGnA
<{^_^}>
[nixpkgs] @ambrop72 closed pull request #64427 → chromium: Strip with --strip-unneeded to reduce binary size. → https://git.io/fjild
<nh2>
when I use `NIX_PATH=nixpkgs=https://github.com/nh2/nixpkgs/archive/a2d7e9b875e8ba7fd15b989cf2d80be4e183dc72.tar.gz`, should it use the ETag to avoid the download even beyond `tarball-ttl`?
<gchristensen>
no, because github doesn't support etags
<gchristensen>
oh it redirects to a URL which does offer an etag. hmm it should probably follow the redirect, and then indeed use the etag
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
Thra11 has quit [Ping timeout: 268 seconds]
matt` has joined #nixos
MmeQuignon has quit [Quit: WeeChat 2.4]
MmeQuignon has joined #nixos
wildtrees has joined #nixos
<matt`>
i'm building a package using cmake and even though the build completes without error, the install phase doesn't create a bin dir in the store even though the cmake specifies it should be created. is there some special configuration i need to do for cmake files?
<matt`>
s/cmake files/cmake packages/
<matt`>
the cmake install command is INSTALL(FILES ... DESTINATION bin PERMISSIONS ...)
<matt`>
does nix know how to translate this into making the bin directory and populating it with files? or do i need create a custom install phase that does it?
levdub has joined #nixos
<wildtrees>
ok I am trying perl6 on nixos, I have openssl installed globally via nix-env, how do I get the perl6 package manager to see the .so's for openssl? or is this broken, I am pretty new to nix/os
<alexarice[m]>
anyone know if evtest-qt is packaged?
<__monty__>
,locate evtest-qt
<{^_^}>
Couldn't find in any packages
<alexarice[m]>
guess not
<__monty__>
wildtrees: What do you mean? If it needs the openssl libs then it should be packaged with those.
rotaerk has quit [Ping timeout: 250 seconds]
Myhlamaeus has quit [Read error: Connection reset by peer]
<wildtrees>
mmm the perl6 on nixos stable channel I use, was from like 2017 , so I used rakudobrew to manually install perl6 and its zef package manager, the zef package manager was missing from the official perl6 build on nixos
levdub has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @aske opened pull request #64682 → buildPythonPackage: add support for setupPyDistFlags → https://git.io/fjXTm
<__monty__>
wildtrees: Ah, then the easiest answer is you can't : )
o1lo01ol1o has joined #nixos
<wildtrees>
do I have to make a .nix for it and run it from nix-shell?
<__monty__>
If you can specify a library path you can pull it off. Or you can try an FHSenv.
<wildtrees>
FHSenv?
levdub has joined #nixos
<__monty__>
What I would do is fork nixpkgs, update the expression for perl6 and submit a PR. That way everyone benefits : )
<__monty__>
wildtrees: It's a nix function that mimics the standard linux file hierarchy. It pretends the dynamic libs are where programs expect them to be. Not generally a good idea though.
<wildtrees>
perl6 seems to work fine both ways, its just the zef package manager and libraries that are a problem (:
<__monty__>
More like an escape hatch.
<wildtrees>
how bad of an idea is it?
<__monty__>
It's a way worse idea than submitting a PR. You wouldn't save much work.
<kumikumi>
I'm stuck with 1.14 and can't connect because my client is at 1.14.3 (newest stable version)
lambda-11235 has quit [Ping timeout: 245 seconds]
<__monty__>
kumikumi: Either fork nixpkgs, update the version and the hash then submit a PR. Or do the same in an overlay.
v88m has quit [Ping timeout: 258 seconds]
<kumikumi>
__monty__: I think I'll try the overlay approach. But the package _should_ be designed in a way that allows the user to choose the desired version (the official minecraft launcher allows downloading and managing multiple versions). I'm not skilled enough to make such modification to the package anyway
<__monty__>
The overlay approach isn't any easier. You need to make the same changes.
<matt`>
is it possible to access the ${out} path in a mkDerivation?
<__monty__>
kumikumi: It's not a common thing afaik. Normally you'd want to run the most recent version of things.
<__monty__>
kumikumi: Looks like that package *does* have an option to specify the version though.
<kumikumi>
__monty__: expect that we can't expect nixpkgs to keep up with what the most recent version is. but yeah I noticed it (however it refers to the package I pointed to earlier)
<kumikumi>
but I'll see if I can figure out a solution
<infinisil>
matt`: Yeah the $out environment variable is set to it
<__monty__>
kumikumi: Ah, if the newer version isn't in nixpkgs yet you can't refer to it of course. Yeah, only way forward is to write an expression for it.
<infinisil>
matt`: Since Nix 2.0 also accessible through ${placeholder "out"}
<__monty__>
kumikumi: And if enough people pitch in nixpkgs *can* keep up : )
<matt`>
thanks infinisil
<__monty__>
kumikumi: If you can write a configuration.nix you know enough to update a package btw.
<{^_^}>
[nixpkgs] @alexarice opened pull request #64684 → evtest-qt: init at 0.2.0 → https://git.io/fjXT6
gyroninja___ has quit [Quit: WeeChat 2.5]
gyroninja has joined #nixos
MmeQuignon has joined #nixos
<levdub>
I'm trying to debug why my remote build doesn't work.
<levdub>
I have `ssh://builder-node x86_64-linux` in /etc/nix/machines, and the proper ssh config (`ssh builder-node` works fine). When I run a nix-build I do see `connecting to ssh://builder-node`, but immediately after I see `cannot build on 'ssh://builder-node': cannot connect to 'builder-node': Received disconnect from 192.168.56.115 port 22:2: Too many authentication failures`
<levdub>
Is it possible to tell nix-build to run ssh in verbose mode, so I can see what the ssh problem is?
MmeQuignon has quit [Client Quit]
<gchristensen>
levdub: can you paste the output from nix-info ?
NightTrain has joined #nixos
v88m has joined #nixos
<levdub>
I'm on darwin, so nix-info doesn't print anything, but nix-build should still work outside of a nix shell.
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Neo-- has quit [Ping timeout: 272 seconds]
jgt has quit [Ping timeout: 252 seconds]
civodul has quit [Quit: ERC (IRC client for Emacs 26.2)]
nefix has joined #nixos
ng0 has quit [Quit: Alexa, when is the end of world?]
<gchristensen>
nh2: I have poked them in the only way I know how
day|flip has joined #nixos
<{^_^}>
[nixpkgs] @pacien opened pull request #64687 → riot-web: fallback to example config → https://git.io/fjXk4
drakonis_ has joined #nixos
ddellacosta has joined #nixos
jackdk has joined #nixos
<wildtrees>
there isn't -dev packages for libraries in nixos is there?
sigmundv has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
<day|flip>
wildtrees: don't take my word at truth. but from what I understand. *.nix file is handle everything when it come to building your own packages. so you have to include all the package that it depends one.