<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JfOa8
alp has quit [Ping timeout: 264 seconds]
qrnteen has quit [Quit: Leaving]
shibboleth has quit [Quit: shibboleth]
invra has left #nixos ["WeeChat 2.3"]
tobeportable has quit [Ping timeout: 272 seconds]
butterthebuddha has quit [Remote host closed the connection]
asbachb has quit [Remote host closed the connection]
virus_dave has quit [Quit: virus_dave]
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos
romildo has quit [Quit: Leaving]
ambroisie has joined #nixos
iyzsong has joined #nixos
Kyndig has joined #nixos
<Kyndig>
when a dependency URL in a derivation expires (404 errors at all mirrors) and it is fixed on the latest nixpkgs.unstable, does the fix also get backported to the latest nixpkgs.stable?
sarcasticadmin has quit [Ping timeout: 265 seconds]
invra has joined #nixos
proofofkeags has quit [Remote host closed the connection]
<cole-h>
Kyndig: Assuming this is the avahi patch?
<cole-h>
If so: patch doesn't 404 for me.
proofofkeags has quit [Ping timeout: 256 seconds]
<simpson>
Kyndig: I don't think that there's a formal policy for backports; it seems that it's up to the reviewer or committer to determine whether a backport would be appropriate. I may be totally wrong though.
iyzsong has quit [Ping timeout: 260 seconds]
<cole-h>
Well, 404'ing patches should definitely be fixed. In fact, my fix for the ninja patch 404'ing was backported to 19.09 and 19.03 (which is EOL)
jgeerds_ has quit [Ping timeout: 240 seconds]
<Kyndig>
sorry; wasn't prepared, trying to find referrer
invra has quit [Quit: WeeChat 2.3]
user_0x58 has quit [Read error: Connection reset by peer]
<Kyndig>
i guess i was wondering if this is more likely to happen when the stable version of the package still uses the patch (or whatever URL resource), but the unstable version has moved on and no longer uses the patch
user_0x58 has joined #nixos
<Kyndig>
like the unstable version is being tested more?
<Kyndig>
since i compile from source, maybe I should be following it?
<Kyndig>
seems most ppl use the cache
<Kyndig>
so they aren't bothered
<Kyndig>
can't believe you rememberd ninja cole-h; don't you do like tens of fixes every day?
<{^_^}>
[nixpkgs] @firmero opened pull request #86430 → gnomeExtensions: refactor pnames and derivations headers → https://git.io/JfOw5
jkachmar has joined #nixos
proofofkeags has quit [Ping timeout: 260 seconds]
foineoi has joined #nixos
h0m1 has quit [Ping timeout: 246 seconds]
<foineoi>
I'm trying to use nix-shell to develop a derivation for a package. My postPatch does some substitutions. But when I enter nix-shell and run patchPhase, I'm not able to modify the source because it's a read-only file system. How am I supposed to use nix-shell in this case?
<hyper_ch>
cole-h: dmesg resp. journalctl don't notice anything... accoridng to onscreen notebook reaches poweroff state but it won't get powered off and the fans become really loud
hlisp has quit [Ping timeout: 265 seconds]
Kyndig has quit [Remote host closed the connection]
<bdju>
someone told me a way to find emacs packages in a nix search, but I can't seem to find it in my shell history easily. anyone know what I'm thinking of? it put me into some repl where I could tab-complete stuff, I think, then I'd know the name to put in my home.nix
<cole-h>
hyper_ch: Sorry?
CMCDragonkai1 has joined #nixos
<hyper_ch>
cole-h: the other day you said to check dmes/journalctl if there's anything relevant because my notebook won't power off properly
<bdju>
Ah! It was nix repl I was looking for. precisely this command: nix repl '<nixpkgs>'
shafox has joined #nixos
waleee-cl has quit [Quit: Connection closed for inactivity]
oborot has quit [Ping timeout: 264 seconds]
delroth has quit [Changing host]
delroth has joined #nixos
chagra_ has joined #nixos
chagra has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @kalbasit opened pull request #86433 → asmfmt: do not include gofmt/goimports/goreturns → https://git.io/JfO6M
<cole-h>
hyper_ch: Sorry I can't help more. But I think your next step is to increase the verbosity of your kernel logs, to get more information out of dmesg
<hyper_ch>
cole-h: I'm not sure kernel is even running anymore. last notification i get is powering off point reached.. then screen turns of.. there is no further indication except some notebook leds are running and fan starts to run at full power
<hyper_ch>
cole-h: and don't say sorry :)
<cole-h>
More kernel logs certainly wouldn't hurt... ;^) But I get what you're saying. It could even be a hardware quirk...
proofofkeags has joined #nixos
<Wulfsta>
It's on PyPi, https://pypi.org/project/meshio/, so I figured I would just be able to do `src = fetchPypi ...` but it doesn't seem to work?
<hyper_ch>
it started a few days ago...
<hyper_ch>
I did update the bios also.. nothing changed
<hyper_ch>
it's really weird :)
<Wulfsta>
hyper_ch: I currently have this poweroff bug as well
<hyper_ch>
Wulfsta: notebook? amd? ryzen?
<hyper_ch>
nixos unstable?
<hyper_ch>
it happening on a lenovo notebook for me
<Wulfsta>
`# poweroff -f` gets to poweroff target then sits. unstable channel on Ryzen 3800X
<hyper_ch>
also using nixos unstable and ryzen.. let me check which one
<energizer>
Wulfsta: it has setup.cfg, which is sufficient
<Wulfsta>
I assumed it was a kernel bug that would eventually get resolved ::shrug::
<hyper_ch>
Wulfsta: rsyzen 7 3700u
<energizer>
(or should be)
<hyper_ch>
it has only started maybe a week ago
<Wulfsta>
energizer: I'm still seeing `FileNotFoundError: [Errno 2] No such file or directory: 'setup.py'`
<energizer>
Wulfsta: which pip version are you using?
hlisp has joined #nixos
<Wulfsta>
hyper_ch I assume it's an AMD CPU bug in the kernel then, have you looked around https://bugzilla.kernel.org?
proofofkeags has quit [Ping timeout: 265 seconds]
<hyper_ch>
Wulfsta: nope
<energizer>
Wulfsta: and setuptools version
<hyper_ch>
I'm on 5.4.33
dansho has joined #nixos
<energizer>
Wulfsta: that should work with setuptools v40.9.0 (april 2019)
<hyper_ch>
Wulfsta: I still have previous generations with 5.4.32 and 5.4.28.... I'll try those
<Wulfsta>
energizer: I'm probably doing something wrong, since I don't understand why pip is needed?
<energizer>
Wulfsta: show your code?
<Wulfsta>
hyper_ch: When I noticed this bug I was on that kernel as well, don't know if it's updated since
<Wulfsta>
energizer: sure, let me push it to git first
<Wulfsta>
energizer: so if I wanted to start a shell with this package, how would I do it? `nix-shell -I nixpkgs=./ -p 'python3.withPackages(ps: with ps; [ pygmsh ])' --run python3`?
<energizer>
maybe it's not the same version in bootstrapped-pip
hlisp has joined #nixos
stree has quit [Read error: Connection reset by peer]
stree has joined #nixos
<energizer>
i think so. but if the setuptools version can't handle the pyproject.toml build-system format then it's not gonna work
<Wulfsta>
I am a few days behind, I wonder if I need to pull upstream
<Wulfsta>
energizer I pulled upsteam and merged it so I should have a recent enough setuptools
proofofkeags has quit [Ping timeout: 246 seconds]
<energizer>
Wulfsta: i dont dont understand why it's not working. it seems weird to have setuptools in the propagatedBuildInputs but i dont see that mattering. out of curiosity i might try adding a patch/build step that just adds a setup.py containing "from setuptools import setup; setup()", which is what new setuptools versions do automatically
<energizer>
that's as deep as i can go atm. the nix experts may have better advice
drakonis has quit [Quit: WeeChat 2.8]
<{^_^}>
[nixpkgs] @GTrunSec opened pull request #86435 → Zeek: fix zeekctl and add Zeek module to deploy it → https://git.io/JfOPL
zupo has joined #nixos
Wulfsta has quit [Remote host closed the connection]
wavirc22_ has joined #nixos
wavirc22 has quit [Ping timeout: 260 seconds]
dansho has quit [Quit: Leaving]
morgrimm has joined #nixos
<jluttine>
is there some rule when to use stdenv.lib and when lib when writing package derivations in nixpkgs?
mamo has joined #nixos
morgrimm has quit [Ping timeout: 265 seconds]
palo1 has joined #nixos
chagra_ has quit [Quit: WeeChat 2.7.1]
palo has quit [Ping timeout: 256 seconds]
palo1 is now known as palo
<cole-h>
jluttine: No, but I (now) prefer using `lib` over `stdenv.lib`, because `lib` is its own thing apart from `stdenv`
<jluttine>
cole-h: ok. but there is still some difference or are they the same thing?
<jluttine>
in nix repl, how can i see help for some command? i'd like to read the docstring of builtins.split
<cole-h>
Probably minor ones, but nothing that should matter. e.g. if you're just using it for `lib.optionals`, or `lib.maintainers.cole-h`, either will work
<cole-h>
jluttine: AFAIK there are no docstrings for anything. Your best bet is to read the source over at https://github.com/NixOS/nix
<jluttine>
cole-h: i'm using it, for instance, mapAttrs and attrValues
<cole-h>
If it works, use `lib`. If, for whatever reason, it doesn't work, try `stdenv.lib`.
<cole-h>
That's how I go about it.
<cole-h>
And I haven't had to use `stdenv.lib` since I starting abiding by that
<jluttine>
cole-h: ok, thank
user_0x58 has quit [Quit: Leaving]
cole-h has quit [Quit: Goodbye]
hlisp has joined #nixos
mamo has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @gnidorah opened pull request #86436 → maintainers/scripts/build.nix: script to build all maintainer packages → https://git.io/JfOPz
mamo has joined #nixos
<lovesegfault>
What's the right way to use niv to pin nixpkgs with NixOps?
siel has quit [Ping timeout: 272 seconds]
siel has joined #nixos
alexherbo2 has joined #nixos
hlisp has quit [Ping timeout: 265 seconds]
<mamo>
is enabling a service in nixos diifer from other distros?
<{^_^}>
[nixpkgs] @matthewbauer pushed 36 commits to geoclue-agent-wants-network: https://git.io/JfOPP
<jluttine>
oh, wait, what is niv? maybe i misunderstood..
<{^_^}>
[nixpkgs] @matthewbauer pushed 0 commits to geoclue-agent-wants-network: https://git.io/JfOPX
<jluttine>
lovesegfault: sorry, nevermind me
<mamo>
/
<makefu>
mamo: not sure i understand, but enabling services is fundamentally different. you will not have config files of the service lingering in /etc or /opt or wherever, but you will write asingle file which defines your whole service including service configuraiton
<lovesegfault>
jluttine: niv is really handy, worth checking out
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @Dema to master « emacspeak: init at 51.0 (#73957) »: https://git.io/JfOP9
<makefu>
i checked the nixos options page which describes all available configuration options (https://nixos.org/nixos/options.html#) however it seems there currently is no conman service to configure
alp has joined #nixos
<mamo>
makefu shouldn't this be provided by the package maintainer within the package itself ?
mamo has quit [Remote host closed the connection]
<makefu>
regarding service files, nixos does not normally rely on upstream configuration. in the configuration nixos cobbles together a whole new configuration because normally a lot of assumptions and bias is inside the config
mamo has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<makefu>
mamo: due to different paradigms nixos pursues a lot of things are vastly different from other distros, this solves many problems such as reproducibility (imagine dotfiles but for your whole system) but it also creates some overhead because we need to fix stuff which assumes that specific software resides in pre-defined places (say /usr)
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<otwieracz>
That's when trying to perform `nixos-rebuild switch`
<hyper_ch>
otwieracz: how do you try to migrate?
<otwieracz>
`nixos-rebuild switch --upgrade`
<hyper_ch>
updated channel to 20.03? a nd then nixos-rebuild swithc?
Vikingman has quit [Ping timeout: 276 seconds]
<otwieracz>
More or less, yes. I have rebased my personal nixos tree to `release-20.03`4
<hyper_ch>
try boot instead of switch
<otwieracz>
OK, let's see.
<otwieracz>
And the error is exactly the same.
<otwieracz>
OK, actually, let me check one thing.. that might be PEBKAC!
<hyper_ch>
I love those ones :)
<otwieracz>
I had additional udev-rule to be installed into initrd to keep names persistent.
<otwieracz>
Maybe it has been included in newer release and it's conflicting now.
<hyper_ch>
permission denied on cp sounds like a different issue
<hyper_ch>
actually, check if /nix/store/b3j2m18dg59viaxcmz8fhc3b33kyxv2h-udev-rules/75-net-description.rules exists
linarcx has joined #nixos
<jluttine>
how can i build on the commandline so that i can use overrides? a pseudo example: nix-build '<nixpkgs>' -A "somepackage.override {foo=true;}"
<srk>
yasu: it works, even test passes but not quite there yet due to assets/js compilation
<srk>
also links to older PR with some more details about underlying mini_racer issue
<yasu>
srk, Oh I see, so it might work enough to enable video conference already?
<srk>
yasu: not sure what's needed for that, asset compilation currently SEGFAULTs :(
<yasu>
@srk it would be awesome if I could help debug it but I just need to show it to my cooperative colleagues to see if it works at all or not
<yasu>
I c...
<srk>
I can gist the test I'm using for that so you can try it if you want
hlisp has joined #nixos
<yasu>
@srk ok let me try the docker first and then try the native if you don't mind. You can still send me gist if you wish, yasu@yasuaki.com is my email
<srk>
yasu: overlay/module import is not strictly needed with the PR
<yasu>
@srk btw have you been involved with packaging discourse for sometime? I'm just curious
wavirc22_ has quit [Read error: Connection reset by peer]
avalenn has joined #nixos
<srk>
yasu: not really, I've started working on it because it was listed on NixOS Discourse as nice to have package, someone pointed me to older PR afterwards, was only able to find few references to IRC logs but not the older PR for some reason :)
pjt_014 has quit [Ping timeout: 260 seconds]
<srk>
I should contact upstream and ask them about few things like running from another directory or asset compilation requiring db connection
<yasu>
srk, I see, I have only used NixOS intermittently as and when I needed so I am not even comfortable of basic operations...
<yasu>
got it
<srk>
no problem, we can help if needed :)
<yasu>
@srk Just to confirm, is Discourse the best there is that's open source (or free software)?
<srk>
well, that's quite opinionated, for one I don't like it requires tons of javascript but it is pretty neat to use
aveltras has joined #nixos
<srk>
I don't know many alternatives to it
zebrag has quit [Quit: Konversation terminated!]
<srk>
cool thing is it supports mail so you don't neccessarily have to use web interface
<yasu>
@srk all I need is just a discussion forum that's not Slack or anything proprietary. :-) But I think mailing list is too limited and it seems people don't use NNTP NewsNet anymore??
KindOne has quit [Ping timeout: 260 seconds]
<yasu>
@srk I wonder if there is anything that's very easy to install natively on NixOS
<srk>
phpbb was pretty popular before discourse, never used that tho :) you can also take a look at hyperkitty (aka mailman3)
KindOne has joined #nixos
<niso>
yasu: if you're looking for a slack alternative: how about matrix-synapse?
pingiun has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
domogled has quit [Quit: domogled]
<yasu>
@niso @srk hmm thank you for the pointers - The assumed users are not necessarily very technical and the system really needs to foster discussions to bloom :]
<{^_^}>
#50898 (by whohoho, 1 year ago, open): mididings package
<Avaq>
Good day! I have recently updated to 20.03 from 20.03-pre and I noticed this morning my external monitor is no longer detected. This happened after I disconnected my monitor and my laptop ran out of battery over the night.
<Avaq>
After trying some basic things, and verifying the monitor does work on other devices, I decided to boot an older generation. I tried several generations all the way to a month back, and none detect the screen.
<Avaq>
Does that mean I've reliably excluded the possibility that it's related to the update? Like, the xf86-video-intel driver version is also reverted when I reboot to an older generation, right?
ebopp has joined #nixos
<symphorien>
Yes
ashesham` has quit [Remote host closed the connection]
<srk>
Avaq: yeah, was about to sent it to you, maybe your videoDriver is just filtered in there
choward has quit [Ping timeout: 260 seconds]
linarcx has quit [Quit: WeeChat 2.8]
linarcx has joined #nixos
Desetude has joined #nixos
<srk>
hm, the module should probably warn or even assert when unknown one is use, I guess that's what the FIXME is about :)
<srk>
*used
<Avaq>
Yeah :p
<Avaq>
In here: `if xorg ? ${"xf86video" + name}`? That would mean that the xorg.xf86videoxf86-video-intel attribute doesn't exist. My config has had the value `[ "xf86-video-intel" "intel" ]` for the longest time based on some ancient tutorial I followed like 6 years ago when first installing NixOS.
<Avaq>
It kind of makes sense for xorg.xf86videoxf86-video-intel not to exist, but xorg.xf86videointel to exist.
stree has quit [Read error: Connection reset by peer]
<avalenn>
Is there any licensing checks of what is published on nixos channels ?
<srk>
hydra won't build nonfree stuff
<dramforever>
Context: I maintain the TUNA Nix mirror (more info at [1]), and I'm looking to keep track of how the mirror's keeping up with (basically making a howoldis for the mirror). I found out that the time there can't be UTC, because then it means I have updates from 2 hours in the future
<hyper_ch>
and in the default.nix you have to indicate the license
__monty__ has joined #nixos
fusion809 has quit [Remote host closed the connection]
mallox has quit [Client Quit]
mallox has joined #nixos
<srk>
hyper_ch: that diagonal.. w00t
mallox has quit [Client Quit]
mallox has joined #nixos
<dramforever>
I know mirror-nixos-branch.pl [2] made that timestamp, which should be close to the time the update was actually available for download. I think just need to know the timezone
<{^_^}>
[nixpkgs] @GRBurst opened pull request #86450 → bront_fonts: init at 2015-06-28 → https://git.io/JfO98
smatting_ has joined #nixos
* srk
still loading larger version
<NobbZ[m]>
Something in the tool that generates the page is broken. Sadly. Licenses are specified in the Derivation correctly as far as I know. avalenn
<dramforever>
And it's interesting but not that surprising that [xyxy] = [xy] * 101 is so popular
srl295 has quit [Quit: Connection closed for inactivity]
tobeportable has joined #nixos
<dxtr>
How would I add custom patches to the kernel? I'm trying to do that but the information I find seem pretty outdated and doesn't seem to quite work - and I can't find any good examples of using boot.kernelPatches
jonathan34 has quit [Remote host closed the connection]
<jakobrs>
You just add the patches to boot.kernelPatches?
<jakobrs>
There's a list of packaged patches in pkgs.kernelPatches
<dxtr>
Excuse my ignorance - I started toying with this nixos thing yesterday!
<dxtr>
Haven't quite grasped nix yet
<jakobrs>
It's perfectly alright to ask questions
<dxtr>
Although I am quite satisfied that it only took me a couple of hours from nothing to get it just like I wanted it - with the exception of this one patch
<dxtr>
So I'm gonna try what you just said
chagra has joined #nixos
<jakobrs>
Looking in the nixpkgs source, you might need to package the patch like this:
<jakobrs>
boot.kernelPatches = [
<jakobrs>
{
<jakobrs>
name = "patch-name";
<jakobrs>
patch = ./file.patch;
<jakobrs>
}
<jakobrs>
];
eyenx has joined #nixos
eyenx has joined #nixos
eyenx has quit [Changing host]
<dxtr>
right
<dxtr>
And how would that translate to using pkgs.fetchPatch?
<jakobrs>
Just replace `./file.patch` with `pkgs.fetchpatch { ... }`
<jakobrs>
That's the file where the pre-packaged patches are defined
<dxtr>
Ah cool
zupo has joined #nixos
<dxtr>
Also, second question: What if I want to compile my own kernel and add the patch? Not just change kernel options which I seem to find the most information about
<jakobrs>
Even if they're built for NixOS specifically?
<dxtr>
so if I wanted to use, say, linuxPackages_5_6 I'd do something like boot.kernelPackages = linuxPackages_5_6?
stree has joined #nixos
<jakobrs>
Yes
cr4y1 has joined #nixos
<srk>
jakobrs: other things depend on it, e.g. initrd module collection
Avaq has quit [Quit: Leaving.]
<jakobrs>
Completely irrelevant question, but:
<jakobrs>
I've got this problem here where some containers refuse to die
<srk>
sure there's a way how to do it, not sure if it's worth it tho :) my kernel build for armv7l is like 10h and I can cross compile it in much less time but haven't tried actually using the output of cc
o1lo01ol1o has joined #nixos
<MichaelRaskin>
Some prebuilt kernels are available in the binary cache, most NixOS systems use them
<Setzer22_2>
Hi! I'm finally decided to go out of the VM and try nixOS on the bare metal :) I'd like to set up dual booting with windows, but my current setup is a bit weird, and I was wondering if os-prober would work out of the box. What I currently have is a 128GB ssd with a Windows install, and a 1TB hdd, with a 250GB partition where nixos will go. I'm
<Setzer22_2>
not sure if installing nixOS on this 250GB partition will properly install boot so that I can also boot windows (which is on a different hard drive). Any tips on that?
hlisp has joined #nixos
<clever>
Setzer22_2: is windows booting with uefi or legacy?
<Setzer22_2>
uefi
<jakobrs>
You can run os-prober _before_ installing nixos to make sure it finds windows
<clever>
Setzer22_2: i believe for uefi, you just give linux its own fat32 /boot, and then both OS's will show up in the bios's menu
<jakobrs>
I don't remember the specific command, though
<Setzer22_2>
jakobrs: ok, that makes sense, I've done a similar thing when installing arch..
<clever>
[root@amd-nixos:~]# os-prober
<clever>
/dev/sdc1:Windows 7:Windows:chain
o1lo01ol1o has quit [Remote host closed the connection]
eof has quit [Quit: eof]
<clever>
if i run it with zero args, it does pick up the win7 nearby
<jakobrs>
I did that when there was an issue with os-prober
pingiun has joined #nixos
<jakobrs>
pkgdatadir=$(nix-build '<nixpkgs>' -A grub2 --no-out-link)/share/grub $(nix-build '<nixpkgs>' -A grub2 --no-out-link)/etc/grub.d/30_os-prober
<Setzer22_2>
to clarify, that's so I can run os-prober from the live image, right?
<jakobrs>
Yes
<Setzer22_2>
ok, great!
<jakobrs>
Or just run os-prober to see if it can detect windows
<jophish>
I wish it was possible to do `foo = addExtraLibrary pkgs.myLibrary . doHaddock . disableLibraryProfiling $ bar` instead of `foo = addExtraLibrary (doHaddock (disableLibraryProfiling bar)) pkgs.myLibrary`
<jakobrs>
The longer command will actually generate the grub config, so if something breaks you can just copy-paste it into grub.extraConfig
<jakobrs>
s/Config/Entries
<jophish>
Or even just `foo = addExtraLibrary pkgs.myLibrary (doHaddock (disableLibraryProfiling bar))`
<clever>
jakobrs: i just manually wrote such a grub entry
<jakobrs>
It's just chainloading the windows boot loader, right? I can't check, I don't have windows installed
<jakobrs>
I had the same issue when running one configuration on another machine without changing network interface names
<clever>
jakobrs: NIX_STORE, NIX_LOG_DIR, NIX_CONF_DIR, NIX_STATE_DIR all work together, to put the store at /home/example/nix
Margatroid is now known as Yaniel
<clever>
jakobrs: NIX_REMOTE=local?root=/home/clever/rootfs then prepends that to the paths, so the store ACTUALLY winds up at /home/clever/rootfs/home/example/nix, and the internal strings expect it to be at /home/example/nix
ambroisie has quit [Ping timeout: 240 seconds]
alp has joined #nixos
<clever>
jakobrs: and then the override to nix within the -E, makes the 1st half permanent within the copy of nix its building
<clever>
jakobrs: so you can then copy /home/clever/rootfs/home/example/nix to /home/example/nix of a remote box, and the resulting nix binary will just work
<dxtr>
jakobrs: I guess it's pretty easy to solve that with some nix hackery
<jakobrs>
So it lets you build packages with a different temporary store dir?
hlisp has quit [Ping timeout: 260 seconds]
<jakobrs>
Wouldn't it be easier to use a chroot/proot/whatever?
<clever>
jakobrs: chroot/namespacing would be much simpler
<clever>
jakobrs: NIX_REMOTE=local?root= is if you want to relocate the /nix/store, but you plan to chroot later, and the paths should still think its at /nix/store
<clever>
jakobrs: NIX_STORE, NIX_LOG_DIR, NIX_CONF_DIR, NIX_STATE_DIR is if you want it to actually run from that wonky dir, without any chroot
never_released has quit [Max SendQ exceeded]
<clever>
they can also be combined, if you lack access to that wonky dir, on the build machine
<jakobrs>
Then why did you do this in the first place (without a chroot)?
<clever>
i didnt know how namespacing worked when i wrote that
<jakobrs>
Where are NIX_REMOTE query arguments documented?
<clever>
jakobrs: this will tell the remote nix (after the ssh) to open the store of `local?root=/mnt`
<clever>
and then that causes the remote nix to prefix all things with /mnt/
c382feb92 has joined #nixos
<clever>
jakobrs: ive not found any proper documentation, so i just searched my irc logs for every single use of it, gathered them into one gist (linked above), and then began figuring out what they do
<jakobrs>
Not completely related, but can you set the path with remote builds?
<jakobrs>
Wait that's probably a feature of ssh_config
<clever>
jakobrs: set which path?
<jakobrs>
PATH path
<jakobrs>
used for finding nix-store
<jakobrs>
when remote building
<clever>
jakobrs: ah, dont think so, your better off fixing the .bashrc on the remote box
<jakobrs>
This is a completely unmodified standard multi-user install btw
<clever>
and adding it here, would fix it for good
Setzer22_2 has quit [Remote host closed the connection]
<asbachb>
I see.
<asbachb>
Is someone preparing a fix? Otherwise I'll do.
<thomassgn>
Hi! I'm trying to understand what INSTALL_ROOT is. It is used in several nixpkgs but can't find docs or definition of it by grepping the nix + nixos + nixpkgs sources. Anyone know what this is?
<gchristensen>
thomassgn: it looks like a variable used by various packages' source
<gchristensen>
it isn't anything to Nix/Nixpkgs/NixOS
<pbogdan>
puffnfresh: yeah, that's what I tried but couldn't get it working :-(
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pbogdan>
do you have your config online somewhere so I could compare and see if I missed something?
<rycee>
Anybody know how to fix the Nix in Travis issue? I don't seem to be able to get Nix 2.2 or 2.3 running for darwin so all travis jobs fail to evaluate nixpkgs master.
otwieracz has left #nixos ["WeeChat 2.7"]
<tokudan>
does it make sense to backport a "oh btw, this package is unmainted for 4 years now and could be missing critical security fixes" from master to release-20.03?
<rycee>
Due to sri hashes popping up in Nixpkgs…
<tokudan>
does it make sense to backport a "oh btw, this package is unmainted for 4 years now and could be missing critical security fixes"-warning from master to release-20.03?
<{^_^}>
[nixpkgs] @flokli pushed 3 commits to release-20.03: https://git.io/JfOFC
<tokudan>
emily, my feeling is that the warning in master is warranted and we should just try to fix possible CVEs in p7zip until 20.03 is dead. at least there are no *known* vulnerabilities in p7zip once the two patches are in
lsix1 has quit [Ping timeout: 272 seconds]
<emily>
tokudan: seems reasonable to me, it already has some backported CVE patches after all
<tokudan>
so I'll keep the warning out of 20.03, unless our release managers prefer it otherwise :)
<emily>
incidentally the non-portable 7zip has gotten releases more recently :/
<{^_^}>
[nixpkgs] @flokli merged pull request #86417 → p7zip: fix two CVEs and mark as insecure → https://git.io/JfOcS
<emily>
looks like actually its own implementation too
<{^_^}>
[nixpkgs] @tokudan opened pull request #86467 → p7zip: fix two CVEs → https://git.io/JfOFR
<emily>
tokudan: omg, I hadn't seen "There was another developer that worked for all p7zip related things. And there are no any news or messages from that developer"...
<emily>
I assumed it was just standard FOSS project dormancy
<emily>
tbh the main casualty of investing the 7zip source is that now I have no idea what archiver to use on windows
<tokudan>
emily, nope, sadly the p7zip dev seems to be missing. just the 7zip dev mentioned that in the forum
<tokudan>
emily, 7zip on windows seems fine to me
<emily>
probably fine for being kept relatively up to date, but I saw enough of the code and the existing CVE patches are subtle enough that my trust in it kinda tanked
zupo has joined #nixos
<emily>
it is, after all, usually one of the first things you run on random untrusted files from the internet
<tokudan>
ah, yeah. well, I'm not using windows, so I don't have to worry about that :)
<emily>
maybe I can just use WSL to extract things next time I'm on a windows box
<elux>
im trying to use `docker` command as my wheel user without having to use sudo .. i've added the user to the "docker" group and restarted docker but I still can't `docker ps` as the user, any suggestions? (thank you)
<wagnerf>
I'm trying to have all my packages in 20.03 except this one
<wagnerf>
in 19.09
alp has quit [Ping timeout: 240 seconds]
<wagnerf>
i've read the pinning faq
<wagnerf>
but it's not helping me too much
<wagnerf>
(i'm really a newbie)
<wagnerf>
any idea how to do this kind of things ?
<emily>
wagnerf: it might be easier to just fix your package for the program
<wagnerf>
I can't
<emily>
I don't recommend keeping a package pinned on an old stable version forever
<wagnerf>
the package is fine
pingiun has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
fragamus has quit [Ping timeout: 240 seconds]
lsix1 has joined #nixos
<emily>
well, if it was fine then it would still be working >.>
<wagnerf>
closed source software is a pb
<wagnerf>
how do I link against old qt ?
<emily>
anyway, people can help with both pinning old nixpkgs and updating a package to work with new nixpkgs, just suggesting that one will probably pay off better in the long run
<asbachb>
Around 10W power comsumption :-/
pingiun has joined #nixos
<adisbladis>
wagnerf: Considering Qt has a stable ABI you shouldn't have to link against old Qt
<wagnerf>
well, it stopped working
<emily>
wagnerf: generally you'll want to do something like let nixpkgs_19_09 = import (fetchFromGitHub { ... }) {}; in ... nixpkgs_19_09.callPackage ./my-old-derivation {}
hlisp has joined #nixos
<dxtr>
Hmm, how do I make sure virt-manager can detect that libvirtd is running? I have set up dconf and I am a member of the libvirtd group
<dxtr>
And libvirtd *is* running according to systemctl status
<adisbladis>
wagnerf: Generally speaking the Qt ABI stability guarantee is pretty solid. I'm not convinced your problem is actually Qt related.
<dxtr>
But when I start it it just says "libvirtd is installed but not running. Start the libvirtd service to manage virtualization on this host."
<adisbladis>
Could you paste your expression?
<wagnerf>
adisbladis: to be fair, there is a dependency on libGL
<freeman42x[m]>
in the past there used to be some type of Nix cache provided by IOHK. Does that still need to be configured or is Nix using some cache by default now? I seen cachix references when intalling stuff recently
<wagnerf>
and it tells me it cannot initialize glx
<emily>
wagnerf: that sounds really unlikely to be fixed by pinning nixpkgs
<adisbladis>
Gl is a massive impurity
<emily>
you should really just share your package if you can, it's the only way people will be able to help. don't worry, there's guaranteed to be worse things inside nixpkgs
<{^_^}>
[nixpkgs] @ajs124 opened pull request #86470 → memtest86-efi: replace p7zip with mtools → https://git.io/JfONZ
lsix1 has quit [Ping timeout: 240 seconds]
<b42>
dxtr: try exporting LIBVIRT_DEFAULT_URI=qemu:///system before running virt-manager?
<freeman42x[m]>
actually, not cachix but cache.nixos.org So, is it using a cache by default?
<emily>
wagnerf: I wonder if this could relate to switching to glvnd or something
<emily>
though I think that happened prior to 19.09
<adisbladis>
emily: There were other significant changes around libgl/mesa for 20.03
<emily>
Cadey: I think this means you won NixOS
<Cadey>
oh god
<Cadey>
but yeah do you have any ideas how i can debug this?
<emily>
I'd get into a nix-shell building the derivation so you can look at the lines/variable expansions that are causing an error
<emily>
like cross-reference line 84 of stdenv setup, check what variables are set in the nix-shell, probably one of them has some gigantic pathname
<emily>
I'm guessing the home-manager emacs module is doing some weird stuff
<{^_^}>
[nixpkgs] @bjornfor merged pull request #80922 → programs/chromium: allow extensions outside chrome web store → https://git.io/Jv0ji
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @danielfullmer to master « programs/chromium: allow extensions outside chrome web store »: https://git.io/JfONh
<adisbladis>
Iirc home-manager emacs module is a simple wrapper around emacsWithPackages
hlisp has quit [Ping timeout: 256 seconds]
<emily>
it may be that stdenv has to change to pass less long argument lists, even
<Cadey>
except line 84 does not have a basename call
<emily>
presumably you might be able to work around it if nobody else is having issues though
<adisbladis>
cd $(mktemp -d) && nix-shell /path/to/drv
<adisbladis>
Inside nix-shell: set -x && genericBuild
<emily>
♥ nix: "hm, I should just look at that stdenv myself" → "but I'm not on 20.03" → nix build /nix/store/vilh5ays3ymz3xkwk0fri2a70lha7pfc-stdenv-linux and it materializes
<ajs124>
Cadey: as a shitty hack, you can also set -x in some super early stdenv hook
<emily>
no need, you already shared a globally-reproducible hash cached on hydra :3
<ajs124>
Still >640KB in the environment variable? What would Bill Gates think of you...
wagnerf has joined #nixos
alp has joined #nixos
<clever>
Cadey: `xargs --show-limits`
<clever>
that shows the actual limits, for your system
<Cadey>
POSIX upper limit on argument length (this system): 2088870
oida has joined #nixos
<clever>
of note, the limit is for the sum of both argv and all env vars
<Cadey>
yeah
<{^_^}>
[nixpkgs] @filalex77 opened pull request #86477 → gitui: init at 0.2.0 → https://git.io/JfOpe
<Cadey>
i replaced the auto-detection logic with a static list of emacs packages and i'm still getting the error
<ajs124>
Hm. So I looked at what you pasted again and every path is in there exactly 156 times.
<genevino>
upgrading 'lilyterm-0.9.9.4' to 'lilyterm-2017-01-06'
proofofk_ has joined #nixos
<genevino>
huh, is this normal?
<genevino>
the 2 version things looks extremely different
<LnL>
2017 > 0
dnlkrgr has joined #nixos
<genevino>
LnL: tru
wagnerf has quit [Quit: WeeChat 2.6]
<LnL>
but generally, unless the project switched version scheme using an unreleased version (ie. commit date) should have an -unstable suffix to avoid this
cjpbirkbeck has joined #nixos
corpix has quit [Remote host closed the connection]
<Setzer22_2>
hi! I'm trying to install the nvidia propietary driver following the guide at the wiki (https://nixos.wiki/wiki/Nvidia). Unfortunately, I'm forced to use sync mode due to a hardwired HDMI to the dedicated gpu, so I added hardware.nvidia.prime.sync.enable = true; as suggested, but nixos-install complained about that option not being supported.
<Setzer22_2>
Instead, by looking at the options list I see that there's "hardware.nvidia.optimus_prime.enable", which I could use instead. Does anyone know about this differences?
<Setzer22_2>
If I didn't know any better, I'd think the wiki is just outdated, but the wiki is clearly referencing the new method, with "official" prime offloading being relatively new, and bumblebee deprecated. Instead, the nixos options page lists bumblebee alongside this optimus_prime flag, which suggests the older situation
<{^_^}>
[nixpkgs] @filalex77 opened pull request #86479 → intermodal: init at 0.1.7 → https://git.io/JfOhC
<betaboon>
Florian[m]1: are you FlorianFranzen on github ?
<Setzer22_2>
(brb, If someone has an answer please tell, I'll check the logs ^^')
Setzer22_2 has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @mkg20001 opened pull request #86480 → modules.steam: init module with workarrounds → https://git.io/JfOhB
stepcut has joined #nixos
cole-h has joined #nixos
jakobrs has quit [Quit: WeeChat 2.8]
<stepcut>
I used node2nix to generate some .nix files from my package.json. But everytime I run nix-shell -A shell it seems to rebuild everything.. that can't be right?
<stepcut>
it takes several minutes to compile and gobbles up more disk space
<{^_^}>
[nixpkgs] @peti pushed 196 commits to haskell-updates: https://git.io/JfOhA
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JfOhh
<betaboon>
anyone found a solution how to pass SWAYSOCK to waybar when both are started as systemd-user-unit
<ToxicFrog>
Is there an equivalent to symlinkJoin that uses hardlinks instead? I'm trying to symlinkJoin [read-only program code, generated config file that the program expects to find in the same directory as $0], but it resolves symlinks when finding $0, so it ends up looking in /nix/store/program rather than /nix/store/program-with-config
<ToxicFrog>
(the program in question is Kanboard, written in PHP and using __DIR__ -- I'm trying to write a nixos settings module for it)
<clever>
ToxicFrog: `nix-store --optimize` will create the hardlinks after the build has finished
<clever>
ToxicFrog: so you can just copy normally, then have nix dedup the whole store
<ToxicFrog>
Ok, so is there an equivalent to `symlinkJoin` that uses `cp` instead of `ln -s`? Because if not, and I have to write it myself, I might as well use `ln` and save space during the build.
<turlando>
Another noob question from me: if for example I want to install something for a single user in my system user I should edit .nixpkgs/config.nix and then call nix-build as such user?
<clever>
turlando: yeah, thats one option
<jakobrs>
nix-env -i?
<jakobrs>
Or am i missing something
<turlando>
Thanks clever. Your tips are being very helpful in these days
suchar has joined #nixos
hakujin has joined #nixos
<turlando>
Is there a way to run a nix env inside a linux namespace? (So that it will only see part of the filesystem and I can avoid using docker and such?)
<clever>
turlando: it should work just fine once you map /nix to that namespace
qbit has left #nixos ["WeeChat 2.8"]
<hakujin>
hi folks, we use haskell + hydra on a growing project, with growing compilation times. I'd like to cache as much as possible and ideally allow hydra to reuse unchanged build products between builds. Is there a way to share module-level caching of haskell projects between builds? If not, is my best bet to factor heavy code out into a separate package?
<Yaniel>
separate packages I think
<Yaniel>
look at how rust packages are done
<infinisil>
Yeah Nix can't reasonably do module-level caching
<infinisil>
There is an experimental https://github.com/nmattia/snack which could do module-level caching, but it's not very practical
<infinisil>
Actually it might not be experimental
<infinisil>
Hm, and for a single package this might actually work
<clever>
infinisil: it has several issues, 1st one is that it can only do module level building for the final cabal file
<clever>
and uses plain cabal2nix for all deps
<infinisil>
clever: Yeah, so this might be fine for the CI of a single project
<clever>
so if your project has 2 or more cabal files, things already start to fall apart
<clever>
i wrote a cabal-merger to undo that, and put it all back into 1 cabal file
<clever>
it took over 48 hours just to eval cardano-sl with snack :P
<hakujin>
right now our project has one cabal file
<clever>
just the nix eval alone
<hakujin>
so maybe this is a good fit?
<infinisil>
clever: Why so long though? A ton of modules?
<clever>
in that 48 hours, i was able to rewrite half of snack, to make it eval in 15mins
<clever>
infinisil: lack of memoization
<clever>
infinisil: basically, every time you import Foo.Bar, it re-runs stdenv.mkDerivation for generating Foo.Bar
<infinisil>
I see, yeah that'll be pretty bad
<clever>
infinisil: which has to re-hash the build directions for Foo.Bar, and everythng it depends on, recursively
<turlando>
clever: do I have to run unshare manually? I can't really find any option in Nix. There is no way to automate it through nix?
<clever>
turlando: if you want to sandbox things at runtime, yes
<clever>
turlando: but nix will automaticall sandbox things at build-time for you
<hakujin>
clever: for a monorepo-style project do you recommend snack or just creating a few separate packages for heavy/infrequently changing parts of the codebase?
<infinisil>
clever: cardano-sl seems to have over 1000 Haskell files, so this probably contributed to the 15 minutes eval time :)
<Yaniel>
separate packages is almost definitely going to be more robust
<Yaniel>
and less of a headache to deal with
nschoe has joined #nixos
<clever>
infinisil: yeah, it also depends on the scale of the package
<infinisil>
hakujin: If you're feeling adventurous, I'd give snack a try
<turlando>
clever: yes, I'd like to sandbox running services. For example I have to run Icecast and being a huge pile of old C code I'd rather give it as little privileges as possible. Unless I am really keen to reinvent the wheel I should stick with docker?
<infinisil>
Hm I see
<clever>
and this is the code for cabal-merger (turns many cabal files into one) and also cabal-to-snack (creates a nix file that snack can use)
<jakobrs>
nix build pkgs.Static failed:
<jakobrs>
chmod: missing operand after 'u+w'
<infinisil>
hakujin: Well, maybe don't try snack after all, I doubt it's gonna be an improvement :P
<clever>
turlando: systemd can sandbox services automatically for you
<turlando>
clever I didn't know that. Thanks for pointing out. I will check some documentation about that.
<fresheyeball>
how can I investigate a nginx 500 on nixos?
<clever>
niso: Could not resolve SDK product version for 'macosx'.
<fresheyeball>
logs are not where I expect with googling for this
<clever>
niso: the error is further up
<clever>
fresheyeball: /var/spool/nginx/logs/
<{^_^}>
[nixpkgs] @adisbladis opened pull request #86487 → Python: introduce NIX_PYTHONPREFIX in order to set site.PREFIXES (20.03 backport) → https://git.io/Jf3Uk
<fresheyeball>
clever: thanks
<niso>
clever: oh, thanks just noticed the "view details" link :)
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos
Vikingman has joined #nixos
<niso>
i assume i cannot easily reproduce/fix this issue without owning a machine running OSX right?
<clever>
niso: correct
<niso>
clever: alright, thanks
cp9 has joined #nixos
Ralith_ has quit [Remote host closed the connection]
Ralith_ has joined #nixos
wozeparrot has joined #nixos
dnlkrgr has quit [Ping timeout: 256 seconds]
fps has joined #nixos
johnw has quit [Remote host closed the connection]
<fps>
hi, i have setup a small default.nix for a project i'm working on. i got my build dependencies just fine. but to test it i need to have some lv2 plugins available in the nix-shell
<fps>
the problem is that liblilv doesn't discover the plugins. LV2_PATH isn't set either
<fps>
and since this is my first day of nixos after quite a while i'm not even sure where the packages in the nix-shell environment go. it's a different place from ~/.nix-profile/
<azazel>
fps: if you set an LV2PATH attr on your derivation, it will be available as variable during phase run
<azazel>
fps: If you want to run things when "nix-shell" starts, you can define a "shellHook" attr on your derivation… it will be sourced by nix-shell
shafox has quit [Remote host closed the connection]
<fps>
azazel: hmm. i wonder where the plugins went though on the nix-shell invocation. so i know what to point LV2_PATH to
<fps>
also unrelated: any recommendations for a terminal to use with i3. xterm is kind of dated ;)
<azazel>
fps: I use "sakura"
<cp9>
termite
<cole-h>
Alacritty ;^)
<cp9>
fast + doesn't have any extraneous stuff + supports truecolor
<{^_^}>
[nixpkgs] @FRidh pushed 106 commits to staging-20.03: https://git.io/Jf3Ta
<azazel>
fps: it's nix, so if you have the packages, you have the paths ;-)
<cp9>
I've used kitty and alacritty in the past and liked both, but they both require GPU acceleration that doesn't work easily in VM
<hlolli>
Looking for advice on how to best package a nested directory. In newer graalvm versions, there's included a bundled clang. This is mostly for internal use, but is also needed for graalvm packages which use Sulong. To create a wrapCCWith, I have to refer to a derivation and not store path. Should I call stdenv.mkDerivation and just symlink the the nested directories?
bornjre has joined #nixos
<morgrimm>
Anyone ever had issues with importing a symlink to a package before?
<azazel>
fps: every derivation and package can be converted to string with something like "${pkgs.foo}" to get the base path of it in the nix store
<{^_^}>
[nixpkgs] @jtojnar pushed to master « terminator: clean up »: https://git.io/Jf3kT
butcher has quit [Quit: Leaving]
butcher has joined #nixos
primeos has joined #nixos
proofofkeags has quit [Remote host closed the connection]
splintah has quit [Ping timeout: 265 seconds]
sigmundv_ has quit [Remote host closed the connection]
<morgrimm>
after a few rebuilds and a couple of hard restarts, it's now swapped to an error about missing the `system.build.topLevel` attribute for my host config
dnlkrgr has joined #nixos
<clever>
morgrimm: can you pastebin the full error and the nix files you are using?
pipppluto has joined #nixos
pipppluto has left #nixos [#nixos]
pipppluto has joined #nixos
<morgrimm>
I'm going to take a crack at the flake docs first I think - I guarantee I'm missing some required flake attributes somewhere
<armin>
adisbladis: looks like i'm clearly doing something wrong then
<armin>
error: The option `home.file.git/peyote.__ignoreNulls' defined in `/home/armin/git/nixos-config/phoenix/home-manager/nixpkgs/home.nix' does not exist.
tewfik has quit [Ping timeout: 264 seconds]
chagra has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage2nix: update list of broken builds to fix evaluation »: https://git.io/Jf3It
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jf3IO
<MichaelRaskin>
If upstream does not support Darwin, just set linux-only meta.platforms. If they do but ofborg fails, check if the error is easily searchable, and whether the package is high-impact enough (or already supported on Darwin in the current version) to ping some random Darwin people.
reallymemorable has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed 39 commits to haskell-updates: https://git.io/Jf3Ig
<drakonis>
bqv: say, how do you update the flake lock?
<{^_^}>
[nixpkgs] @peti pushed 19 commits to haskell-updates: https://git.io/Jf3I2
<MichaelRaskin>
niso: if it is hard to fix, it is a reason to wait some time for other people to come up with the fix, then document the problem as well as possible and merge.
butcher_ has joined #nixos
<drakonis>
i'm yet to get it
<niso>
MichaelRaskin: means darwin support is existant, but somewhat treated as a second class citizen?
butcher has quit [Ping timeout: 260 seconds]
<MichaelRaskin>
Tier-2, yes
* niso
could have phrased that a bit better
<niso>
MichaelRaskin: alright, thanks
<{^_^}>
[nixpkgs] @peti pushed 10 commits to haskell-updates: https://git.io/Jf3Iw
reallymemorable has quit [Quit: reallymemorable]
therealsarcastic has joined #nixos
<MichaelRaskin>
Well, we have less Darwin contributors than Linux contributors, and even the Linux work is always more than can be done, so just give people time to react if they want to, and notify about regressions
sarcasticadmin has quit [Read error: Connection reset by peer]
<niso>
MichaelRaskin: i completely understand that, i'm not a darwin user either
mkenigs has quit [Remote host closed the connection]
inkbottle has joined #nixos
<drakonis>
i think this would be overlays maybe?
mkenigs has joined #nixos
zebrag has quit [Ping timeout: 260 seconds]
<drakonis>
or you can build something that uses that on your local copy of the repository
mkenigs has quit [Remote host closed the connection]
<Cadey>
i can use the disk location of nixpkgs as a NIX_PATH entry right?
<drakonis>
yes
<Cadey>
okay cool
<drakonis>
its reasonably easy to test things
endformationage has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-ed25519: jailbreak to fix build with ghc-8.10.1 »: https://git.io/Jf3Lw
inkbottle has quit [Ping timeout: 246 seconds]
zebrag has joined #nixos
detran has quit [Ping timeout: 260 seconds]
zupo has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jf3L5
<Cadey>
i'm getting the same error with a custom NIX_PATH, could my `sudo nixos-rebuild switch` be interfering with this?
Setzer22_2 has quit [Remote host closed the connection]
rickhull has quit [Ping timeout: 240 seconds]
<cole-h>
What is "the same error"?
<colemickens>
Cadey: yes. the `sudo` is losing your env vars
<colemickens>
and thus losing the custom NIX_PATH you're setting
mr_sm1th has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « cabal-install: patch the ghc-8.10.x build to succeed »: https://git.io/Jf3tJ
<colemickens>
cadey: `sudo env NIX_PATH=nixpkgs=/tmp/nixpkgs-pr nixos-rebuild switch` ought to do it
<mr_sm1th>
Hiya. Is the original PhD still worth a read? Or am I better off reading and following along with the Nix Pills?
<srk>
hi, it is!
<srk>
it explains the concepts and motivation quite nicely
<clever>
colemickens: you can also just `-I nixpkgs=/tmp/nixpkgs-pr` instead
butcher_ has quit [Ping timeout: 246 seconds]
o1lo01ol_ has joined #nixos
o1lo01ol1o has quit [Ping timeout: 265 seconds]
<cole-h>
mr_sm1th: IMO those are two different things. Pills is for basic use of the language (and relatively outdated, at that), while the PhD thesis is for Nix's rationale in its entirety, AFAIK
<{^_^}>
[nixpkgs] @prusnak opened pull request #86493 → gzip: add setup-hook to stop gzip from adding timestamps to headers → https://git.io/Jf3tE
<{^_^}>
[nixpkgs] @GTrunSec opened pull request #86494 → zeek: add zeek module and fix zeekctl to deploy it. → https://git.io/Jf3tu
<{^_^}>
[nixpkgs] @GTrunSec closed pull request #86435 → zeek: fix zeekctl and add Zeek module to deploy it → https://git.io/JfOPL
<Fare>
and how do I learn how to answer such a question?
zebrag has joined #nixos
<simpson>
Fare: Depends on exactly what you want to do in the container. If you don't need Nix in the container, then maybe dockerTools is a good starting point: https://nixos.org/nixpkgs/manual/#sec-pkgs-dockerTools
<simpson>
If you *do* need Nix in the container, maybe just start with a Dockerfile with `FROM nixos/nix`
<simpson>
I got the former from reading the manual, and the latter from happening to see somebody else do it.
<manveru>
i thought guix had an unfree repo somewhere
<drakonis>
but then you dont get to enjoy the niceties of having working hardware across the board
<drakonis>
it does, yes.
<drakonis>
but it isn't backed by a hydra instance
<drakonis>
it has a very limited range of packages
<manveru>
given that i use nvidia, it's a no-go for me anyway...
<drakonis>
i tried getting nvidia to work, but as it turns out, nix has a lot of scaffolding to enable it
<drakonis>
so i'd have to replicate all of that and then maintain it
<manveru>
yeah
<drakonis>
it already has issues on nix
<manveru>
i'm really thankful to whoever does that, it never failed me once :)
<drakonis>
the extra repository covers packages that don't match policy
<drakonis>
ie: the other useful bits of clojure live here
MightyJoe has joined #nixos
<drakonis>
the annoying thing is having to manually build the kernel every point release, so pinning ahoy
alp has joined #nixos
<drakonis>
there seems to be some measure of patchelf scaffolding
<drakonis>
there's also no nix shell equivalent
cyraxjoe has quit [Ping timeout: 256 seconds]
<drakonis>
but that's a derail
fragamus has quit [Ping timeout: 240 seconds]
<Fare>
OK, so I built my image with dockertools, and I am surprised to find that it's > 1.1GB when the "equivalent" on top of Ubuntu is < 400MB.
<manveru>
that usually means that we can improve the closure size of the derivations :)
<adisbladis>
Fare: What are you building?
<clever>
Fare: use pkgs.writeText to create a file that contains every storepath you told nix to put into the docker image (you can just "${toString [ the list ]}")
<clever>
Fare: then run `du -hc --max=0 $(nix-store -qR result) | sort -h` on the result of building that writeText
<clever>
that should show everything that gets included into the docker image, sorted by size
<Fare>
zsh: command not found: nix-store
<Fare>
the image doesn't even contain nix itself!
<manveru>
why would it?
<clever>
Fare: you have to include nix in the list if you want nix in the docker image
<adisbladis>
dockerTools is only including the _runtime closure_ of a package
<Fare>
doing just that... it's a development / test image, after all (that I want to use in github / gitlab CI)
gustavderdrache has quit [Quit: Leaving.]
butcher_ has joined #nixos
cyraxjoe has joined #nixos
inkbottle has joined #nixos
<adisbladis>
Fare: So what exactly do you want to end up with?
MightyJoe has quit [Ping timeout: 240 seconds]
zebrag has quit [Ping timeout: 265 seconds]
mallox has quit [Quit: WeeChat 2.8]
<Fare>
adisbladis, a replacement for the current ubuntu-based image used in CI
<Fare>
something I can run build and test commands on top of
<Fare>
not a production image (yet, that will come later)
<adisbladis>
Fare: I mean very specifically
<adisbladis>
A toolchain to build some package presumably. Which one?
<codygman__>
hm, I have another complication. home-manager is still alive and can't be deleted and my environment is what's keeping it alive. There is no previous generation to roll back to: http://dpaste.com/03JJNXT
<adisbladis>
It seems to be equivalent to the "official" docker image
<adisbladis>
But with buildLayeredImage awesomeness :)
splintah has joined #nixos
__monty__ has quit [Quit: leaving]
knupfer has quit [Ping timeout: 244 seconds]
<manveru>
adisbladis: could use some tidying, but pretty awesome :)
reallymemorable has quit [Quit: reallymemorable]
<manveru>
really would like to make a nix-built docker image our official one...
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
splintah has quit [Ping timeout: 265 seconds]
<adisbladis>
manveru: Yeah :) I just hacked it up as quick as possible
griff__ has joined #nixos
jgeerds__ has quit [Remote host closed the connection]
sgo__ is now known as stigo
domogled has quit [Quit: domogled]
<Fare>
adisbladis, how in this script you wrote do I create a home directory owned by some user?
<codygman__>
clever I forgot to tag earlier, but I don't have a previous generation to roll back to and home-manager gc root is kept alive by my environment it seems: http://dpaste.com/03JJNXT
<clever>
codygman__: is there only one generation of nix-env's profile?
<{^_^}>
[nix] @Ma27 opened pull request #3557 → Improve help-message for nix-repl → https://git.io/Jf3OR
stigo has quit [Remote host closed the connection]
dnlkrgr has joined #nixos
stigo has joined #nixos
<adisbladis>
Fare: I guess you can add a user in the users attrset and then create a home directory and chown appropriately in extraCommands
<fresheyeball>
ok I am just going to ask here
<fresheyeball>
I have nginx virtual hosts with locations
<fresheyeball>
I want one location to serve one root
dnlkrgr has quit [Ping timeout: 246 seconds]
<fresheyeball>
and another location to serve another
<fresheyeball>
that's it
morgrimm has quit [Ping timeout: 260 seconds]
<HedgeMage>
exit
<{^_^}>
Let me do that for you
<fresheyeball>
locations."/static/" = {
<fresheyeball>
root = "${chatstrike-webchat}";
<fresheyeball>
};
<HedgeMage>
oops, wrong window :o
<fresheyeball>
I don't get why setting root at that location, seems to do nothing
<fresheyeball>
I just get 404s
fragamus has quit [Ping timeout: 240 seconds]
<Fare>
adisbladis, somehow nix-build tells me error: attempt to call something which is not a function but a set, at (the line where I try to use buildLayeredImageWithNixDb)
fendor has quit [Read error: Connection reset by peer]
__Sander__ has quit [Ping timeout: 258 seconds]
<Fare>
adisbladis, I copypasted what you had in http://ix.io/2knJ/nix for the definition of buildLayeredImageWithNixDb
thelounge79 has joined #nixos
<nh2>
fresheyeball: `builtins.trace` the path, check on the server that it actually exists, check that nginx can read it (e.g. error logs), if still in doubt, use `strace` to check that it does try to ready it
zarel_ has joined #nixos
<fresheyeball>
nh2: nginx is running fine
<fresheyeball>
I cated the conf file nix builds
zarel has quit [Ping timeout: 265 seconds]
<fresheyeball>
and am not seeing the problem
<codygman__>
clever Yes, this is a fresh vm
<nh2>
fresheyeball: then strace is probably the fastest way, check that it tries to read it when making the request
<clever>
codygman__: if the only root is your nix-env profile, then you can use `--delete --force` and it will just blow away the entire nix-env profile
<fresheyeball>
nh2: what is your suspicion?
<clever>
codygman__: then you can just reinstall home-manager from scratch
<nh2>
fresheyeball: usually editing some part of the config that is accidentally not in effect
<nh2>
fresheyeball: in any case, strace will give you good hints
o1lo01ol1o has joined #nixos
<fresheyeball>
nh2: ahh, I know it's having effect
<fresheyeball>
because if I use proxy_pass it works
ninegrid has joined #nixos
<nh2>
fresheyeball: `strace -fyp "$(pidof nginx)" -e open,openat` on your server, then make the request
Henson has joined #nixos
<Henson>
hi everyone, I have a system that has some caches defined in nix.binaryCaches but they are no longer accessible. I am unable to use nixos-rebuild to rebuild the configuration to remove them, because it's trying to access them to do the rebuild. How do I get out of this problem?
<Fare>
aha I had a ''; leftover at the end of the invocation... so I presume it was trying to call the *result of calling the function* with an empty string.
<fresheyeball>
[pid 4947] openat(AT_FDCWD, "/var/spool/nginx/html/static", O_RDONLY|O_NONBLOCK) = -1 ENOENT (No such file or directory)
<armin>
ok weird, i just built a custom iso, everything is working nicely so far, but i only have dynamic width fonts under both konsole and xfce4-terminal
<nh2>
fresheyeball: sorry, I meant to say "does NOT act"
chloekek has quit [Ping timeout: 258 seconds]
<fresheyeball>
nh2: so I tried that
Thra11 has quit [Quit: WeeChat 2.8]
<fresheyeball>
[pid 6255] openat(AT_FDCWD, "/nix/store/nwfp685pldpsdln8qcv806n8y0v1gsv2-ui-strikechat-staging/static", O_RDONLY|O_NONBLOCK) = -1 ENOENT (No such file or directory)
<fresheyeball>
which makes sense
<fresheyeball>
since that nix-store path is for the top level root directive
<fresheyeball>
not the one I set for static
<nh2>
fresheyeball: that looks better
<Fare>
adisbladis, interestingly, it looks like the ln succeeds in many iterations before failing for the gambit path (!)
stigo has quit [*.net *.split]
Ralith_ has quit [*.net *.split]
eoli3n_ has quit [*.net *.split]
butcher_ has quit [*.net *.split]
KindOne has quit [*.net *.split]
epta has quit [*.net *.split]
cross has quit [*.net *.split]
c_wraith has quit [*.net *.split]
sveitser2 has quit [*.net *.split]
lzmartinico has quit [*.net *.split]
lorimer has quit [*.net *.split]
bukkitgerman8608 has quit [*.net *.split]
plutes has quit [*.net *.split]
neobit has quit [*.net *.split]
Scriptkiddi has quit [*.net *.split]
ajs124 has quit [*.net *.split]
kapilp has quit [*.net *.split]
MarcWeber has quit [*.net *.split]
avn has quit [*.net *.split]
niksnut has quit [*.net *.split]
itorres has quit [*.net *.split]
mupf has quit [*.net *.split]
seanparsons has quit [*.net *.split]
inf has quit [*.net *.split]
minicom has quit [*.net *.split]
meck has quit [*.net *.split]
mvnetbiz_ has quit [*.net *.split]
Heirlung has quit [*.net *.split]
Gohla has quit [*.net *.split]
acowley has quit [*.net *.split]
lukash_|away has quit [*.net *.split]
ornxka has quit [*.net *.split]
gentauro has quit [*.net *.split]
luigy has quit [*.net *.split]
otti0815 has quit [*.net *.split]
meatcar has quit [*.net *.split]
adamtkh6 has quit [*.net *.split]
Nazral has quit [*.net *.split]
hyper_ch has quit [*.net *.split]
vyorkin` has quit [*.net *.split]
primeos has quit [*.net *.split]
hlolli has quit [*.net *.split]
user_0x58 has quit [*.net *.split]
lsix1 has quit [*.net *.split]
bqv has quit [*.net *.split]
Desetude has quit [*.net *.split]
m0rphism has quit [*.net *.split]
thc202 has quit [*.net *.split]
aswanson has quit [*.net *.split]
andi- has quit [*.net *.split]
mumuluxi has quit [*.net *.split]
xO1 has quit [*.net *.split]
siel has quit [*.net *.split]
dingenskirchen has quit [*.net *.split]
NinjaTrappeur has quit [*.net *.split]
snupples[m] has quit [*.net *.split]
tbenst[m] has quit [*.net *.split]
dckc has quit [*.net *.split]
konobi has quit [*.net *.split]
dhess has quit [*.net *.split]
xantoz has quit [*.net *.split]
aminechikhaoui has quit [*.net *.split]
plp_ has quit [*.net *.split]
qyliss has quit [*.net *.split]
U-armin has quit [*.net *.split]
klys has quit [*.net *.split]
kgz has quit [*.net *.split]
immae has quit [*.net *.split]
aranea has quit [*.net *.split]
{^_^} has quit [*.net *.split]
rail_ has quit [*.net *.split]
cyphase has quit [*.net *.split]
gerschtli has quit [*.net *.split]
woffs has quit [*.net *.split]
mog has quit [*.net *.split]
fionera has quit [*.net *.split]
lopsided98 has quit [*.net *.split]
greizgh has quit [*.net *.split]
drewr has quit [*.net *.split]
infty has quit [*.net *.split]
Reisen has quit [*.net *.split]
thonkpod has quit [*.net *.split]
plp__ has joined #nixos
neobit_ has joined #nixos
Ralith_ has joined #nixos
siel_ has joined #nixos
dckc has joined #nixos
dhess has joined #nixos
xO1 has joined #nixos
Nazral has joined #nixos
niksnut has joined #nixos
snupples[m] has joined #nixos
tbenst[m] has joined #nixos
U-armin has joined #nixos
user_0x58 has joined #nixos
butcher_ has joined #nixos
Desetude has joined #nixos
avn has joined #nixos
<genevino>
\m/
NinjaTrappeur has joined #nixos
luigy has joined #nixos
epta has joined #nixos
bqv has joined #nixos
gerschtli has joined #nixos
aswanson has joined #nixos
m0rphism has joined #nixos
dingenskirchen has joined #nixos
itorres has joined #nixos
sveitser2 has joined #nixos
andi- has joined #nixos
tdeo has joined #nixos
eoli3n_ has joined #nixos
thonkpod has joined #nixos
vidbina has quit [Ping timeout: 265 seconds]
adamtkh6 has joined #nixos
lzmartinico has joined #nixos
Desetude has quit [Changing host]
Desetude has joined #nixos
xantoz has joined #nixos
primeos has joined #nixos
Scriptkiddi has joined #nixos
cross has joined #nixos
inf has joined #nixos
Reisen has joined #nixos
kapil_ has joined #nixos
mupf has joined #nixos
eoli3n_ has quit [Client Quit]
meatcar has joined #nixos
fionera has joined #nixos
eoli3n_ has joined #nixos
lukash_|away has joined #nixos
immae has joined #nixos
m0rphism is now known as Guest35509
tdeo has joined #nixos
tdeo has quit [Changing host]
meatcar has joined #nixos
meatcar has quit [Changing host]
primeos is now known as Guest90765
inf is now known as Guest84206
drewc has quit [Ping timeout: 244 seconds]
hlolli has joined #nixos
hlolli has quit [Remote host closed the connection]
Heirlung has joined #nixos
acowley has joined #nixos
seanparsons has joined #nixos
cyphase has joined #nixos
plutes has joined #nixos
<Fare>
adisbladis, aha, somehow I need to remove duplicates
c_wraith has joined #nixos
vyorkin` has joined #nixos
hlolli has joined #nixos
qyliss has joined #nixos
ornxka has joined #nixos
minicom has joined #nixos
Gohla has joined #nixos
greizgh has joined #nixos
mog has joined #nixos
mumuluxi has joined #nixos
stigo has joined #nixos
KindOne has joined #nixos
kgz has joined #nixos
lopsided98 has joined #nixos
MarcWeber has joined #nixos
klys has joined #nixos
gentauro has joined #nixos
infty has joined #nixos
rail_ has joined #nixos
hyper_ch has joined #nixos
otti0815 has joined #nixos
Avaq has joined #nixos
* Fare
greps his way to lib.unique
drewc has joined #nixos
<Raito_Bezarius>
Do someone know how to set up IPv6 /64 on Scaleway under NixOS (unstable) ?
<Raito_Bezarius>
I have enableIPv6 = true; and set statically the IPv6
thelounge792 has joined #nixos
amir has quit [Ping timeout: 256 seconds]
<Raito_Bezarius>
but "curl api6.ipify.org" tell me that I'm still on IPv4
amir has joined #nixos
<Raito_Bezarius>
(I used the data provided by Scaleway at the <magic-ip>/conf)
<Fare>
are you on *both* IPv4 and IPv6 ?
siel_ has joined #nixos
siel_ has quit [Changing host]
meck has joined #nixos
<Raito_Bezarius>
Yes
<Fare>
can you "just" disable IPv4?
<Raito_Bezarius>
Well, it's a server, so if I disable it and cannot SSH back, it'll be a bit annoying
<qyliss>
Have you set networking.defaultGateway6 and networking.interfaces.<name>.ipv6.addresses?
thelounge79 has quit [Ping timeout: 256 seconds]
thelounge792 is now known as thelounge79
<Raito_Bezarius>
Yes
Guest90765 has quit [Quit: WeeChat 2.8]
primeos_ has joined #nixos
morgrimm has joined #nixos
<Raito_Bezarius>
and I can see them in effect in "route -6 -n" and "ip addr"
<armin>
so i'm building custom iso's, everything works like expected, i even get a fully working kde desktop and whatnot, but the fonts in my terminal windows (both konsole and xfce4-terminal) look weird (no monospace fonts available there, too). i have zero idea how to debug this.
<Raito_Bezarius>
armin: fc-list ?
<armin>
Raito_Bezarius: well those fonts are available even in kde's system settings
<armin>
Raito_Bezarius: so they must be present in some way or the other
<armin>
Raito_Bezarius: will definitely continue to test and try :)
covidian has joined #nixos
<Raito_Bezarius>
armin: one hardcore way is to strace your terminal binary
<Raito_Bezarius>
take a look if it's trying to look for the fonts properly
<Raito_Bezarius>
you can filter out based on the filesystem syscalls I believe
smatting_ has joined #nixos
<armin>
Raito_Bezarius: i'll have a look, thank you :)
<adisbladis>
Fare: The / FHS is constructed from nix store paths
primeos_ has quit [Quit: WeeChat 2.8]
<adisbladis>
So you can just create whatever directory structure you want in a derivation and add it to contents :)
hyper_ch has quit [Remote host closed the connection]
hyper_ch has joined #nixos
Ralith_ has quit [Remote host closed the connection]
<covidian>
’’ is not ''
<covidian>
btw.
primeos_ has joined #nixos
<covidian>
gotta know ure '’´`
<Avaq>
Hi! Is it possible to "query" my nixo(p)s configuration for info on a derivation it *would* use? Eg. `nix show-derivation nixpkgs.hello` but using the nixpkgs as they are after evaluating a configuration.nix (applying overlays etc). The command `nix show-derivation -f ./configuration.nix pkgs.hello` produces "error: cannot auto-call a function .." of course.
arjen-jonathan has quit [Ping timeout: 246 seconds]
<lovesegfault>
my blog has a dark theme?
<lovesegfault>
it looks white to me
pjt_014 has quit [Ping timeout: 260 seconds]
<Avaq>
infinisil: hm, maybe take a step back. Is it possible for `nix-instantiate --eval` to show be the info I want from a regular /etc/nixos/configuration.nix file? I'm working towards just finding what package version would be installed by a specific configuration file.
<Avaq>
So I tried `nix-instantiate --eval -A 'pkgs.hello' ./configuration.nix` but that also tells me it can't call the function it encounters automatically.
<clever>
Avaq: configuration.nix doesnt return a set containing pkgs
<energizer>
eoli3n_: some browsers respect your desktop prefer-dark-theme config
<infinisil>
Avaq: Or `nix-instantiate --eval '<nixpkgs/nixos>' -A pkgs.hello.name --arg configuration /etc/nixos/configuration.nix` if you want to be explicit about the file
covidian has quit [Quit: Konversation terminated!]
<Avaq>
infinisil: --arg configuration was what I was missing! :)
<eoli3n_>
lovesegfault: i mean theme of the comment widget
<lovesegfault>
eoli3n_: Fascinating, I guess it picks the dark theme depending on ????
<Fare>
(of course, the right thing is to tell gambit about the "right" path to sed)
<eoli3n_>
as energizer said
<infinisil>
,runtimeDeps Fare
<infinisil>
Oh
<cole-h>
,runtimedeps Fare
<codygman__>
clever `nix-store --delete --force` says `error: unknown flag '--force'`
<lovesegfault>
eoli3n_: Which browser are you using?
<cole-h>
Hello
<infinisil>
{^_^} not here
<cole-h>
Robot
<eoli3n_>
lovesegfault: firefox
<cole-h>
Fare: "In order of preference: Patch source OR ((if it uses PATH -> wrap with new $PATH) AND (if it uses dlopen, (patchelf --set-rpath in postFixup OR wrap with new LD_LIBRARY_PATH)))"
<eoli3n_>
latest on archlinux without nothing particular
<Fare>
ok, next nix task is finding the "right" way to build and expose gerbil libraries
<Fare>
each library will create its own set of compiled files. What's the right way to make the collection of them plus whatever the user is currently building available to the user?
<Fare>
a wrapper that exports some variables and executes gerbil ?
<Fare>
something that adds all the stuff you want in ~/.nix-profile/lib/gerbil ?
quinn has joined #nixos
<Fare>
(or some other such semi-fixed path?)
Avaq has quit [Quit: Leaving.]
Avaq has joined #nixos
<clever>
Fare: development stuff is only meant to work inside nix-shell
<clever>
Fare: enless you want to re-engineer a lot of stuff, its recommended you just run nix-shell as normal
dnlkrgr has quit [Quit: WeeChat 2.6]
<energizer>
or lorri
<clever>
energizer: that just extracts the env vars from nix-shell, and sets them in your shell
<Fare>
clever, I'll still have to engineer the nix-shell for Gerbil --- same difficulty, really
<energizer>
clever: what is the alternative?
<energizer>
iow, in what ways is lorri's environment incomplete relative to nix-shell's
xkapastel has joined #nixos
<clever>
Fare: even basic gcc use wont work right if your outside of nix-shell
<Fare>
plus the whole point of Gerbil is interactive development. I *want* this stuff available at runtime.
<clever>
Fare: then run the IDE under nix-shell
<Fare>
clever, I have a script to reproduce what nix-shell does, to work with GCC
<Fare>
I don't want to reproduce my entire IDE under the nix-shell.
<Fare>
I want my IDE to commandeer my nix-shell's, not the other way around
<clever>
Fare: why do you have to reproduce the IDE under nix-shell?
<Fare>
uh, how else to "run the IDE under nix-shell" ?
<clever>
Fare: you could even make it 100% automatic
<clever>
Fare: modify the Makefile to just do `-DFOO_PATH=$(which foo)`
<clever>
Fare: then it will just find it from $PATH, and hard-code that into things
<Fare>
instead of replacing the literal "gsc" by the path to GSC
<clever>
then if its in buildInputs, it gets found
<Fare>
ok, what about the part where I use ${coreutils}/bin/env instead of /usr/bin/env ? Just use /usr/bin/env ?
<clever>
Fare: what are you doing with env?, a #! line? running it? at build or runtime?
<Fare>
that was a #! line
{^_^} has joined #nixos
<clever>
Fare: use #!/usr/bin/env foo, but then run patchShebangs on the directory if its ran at build-time
<Fare>
I also modify shell scripts to call the "right" gsc, gsi, gxc, gsi.
<Fare>
these scripts are part of the source
<clever>
Fare: you should just run "gsc" directly, let $PATH do its job
<clever>
Fare: `#!/usr/bin/env foo` will work just fine outside of nixos as well
<clever>
it will look in $PATH
stree has quit [Excess Flood]
<johnjay>
clever: i tried running a python script earlier today and got an error from Imagemagick
<johnjay>
went away when I added #!/usr/bin/python3. XD
<Fare>
clever, but to make gerbil available outside nix-shell, don't I need see the path to gsc ?
alp has quit [Ping timeout: 244 seconds]
<clever>
Fare: thats what wrapProgram is for, to prepend things to PATH before the thing gets ran
stree has joined #nixos
<Fare>
actually, on a mac, the path to gcc is almost certainly wrong for gambit and gerbil: gcc on a mac is usually an alias to clang, but gambit will be confused by that, and even if you unconfuse it will run 10x slower with clang than with gcc.
<clever>
Fare: use $CC and $CXX to get the compiler
<clever>
Fare: or use gccStdenv to get a real gcc
<clever>
> gccStdenv.mkDerivation
<{^_^}>
<LAMBDA>
<Fare>
clever, what I did was patch gambit to replace its calls to nake "gcc" by calls to the full path.
<clever>
Fare: it should just respect $CC, and then the makefile can provide a default when its missing
<Fare>
not sure how good an idea it is: if you use a CC that is different from the one used to compile gambit, you may have "surprises".
<Fare>
(especially if you use any C++, where the ABI often changes)
<Fare>
(gambit compiles to C)
alp has joined #nixos
* Fare
uses docker push... those hundred layers take forever to push
justanotheruser has quit [Ping timeout: 260 seconds]