<NixOS_GitHub>
[nixpkgs] oxij opened pull request #24733: firefox: refactor, add options; tor-browser: package a version directly usable via nix (master...pkg/firefoxPackages) https://git.io/vSrSs
stasku has left #nixos []
bhipple has joined #nixos
<bhipple>
Is anyone else having issues with Docker on NixOS? I have the virtualization option enabled and added my user to the docker group, but it seems the docker process fails to start with systemd
Guest60_ has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
<bhipple>
I'll try upgrading everything to NixOS 17 and see if that helps
Dr8128 has quit [(Ping timeout: 264 seconds)]
cryptsetup has quit [(Ping timeout: 260 seconds)]
bhipple has quit [(Ping timeout: 268 seconds)]
Wizek has quit [(Ping timeout: 246 seconds)]
Wizek_ has quit [(Ping timeout: 260 seconds)]
drp has quit [(Quit: Page closed)]
hamishmack has quit [(Quit: hamishmack)]
hamishmack has joined #nixos
hamishmack has quit [(Client Quit)]
Dr8128 has joined #nixos
jrolfs has joined #nixos
bhipple has joined #nixos
<disasm>
bhipple: on boot?
jrolfs has quit [(Ping timeout: 264 seconds)]
<disasm>
if so, it's a known issue. 17.03 adds a virtualisation.docker.enableOnBoot = true
<gchristensen>
disasm: that is a known issue, or does docker just not start until the socket is used by default?
<gchristensen>
can you link me to an issue?
<bhipple>
I didn't reboot or change anything after upgrading from nixos 16 to 17; the nixos-rebuild switch restarted the docker process as soon as it finished, and the restart "just worked"
filterfish has quit [(Ping timeout: 240 seconds)]
<bhipple>
haven't rebooted yet; I guess I should add the enableOnBoot flag to my configuration.nix
<disasm>
gchristensen: the latter, with 16 you have to do a docker ps for containers to launch on boot. with 17 if you have enableOnBoot it works as expected.
takle has joined #nixos
<gchristensen>
seems very strange
isidore has joined #nixos
<gchristensen>
is there a changelog item?
<gchristensen>
ohhh enableOnBoot is fiing the issue
stepcut has quit [(Remote host closed the connection)]
rardiol has joined #nixos
<disasm>
gchristensen: exactly :)
takle has quit [(Ping timeout: 260 seconds)]
byteflame has joined #nixos
<gchristensen>
thanks, disasm :)
byteflame has quit [(Ping timeout: 256 seconds)]
stepcut has joined #nixos
slack1256 has joined #nixos
ryanartecona has joined #nixos
<disasm>
gchristensen: question for ya... right now I'm just using pypi2nix for a dev env for some flask restless stuff I'm doing, but would there be any value in making a pull request to add the library to nixpkgs?
ryanartecona has quit [(Remote host closed the connection)]
ryanartecona has joined #nixos
<gchristensen>
what library?
endformationage has quit [(Quit: WeeChat 1.7)]
dpren has quit [(Quit: Connection closed for inactivity)]
mbrgm has quit [(Ping timeout: 260 seconds)]
filterfish has joined #nixos
slack1256 has quit [(Remote host closed the connection)]
filterfish_ has joined #nixos
filterfish has quit [(Read error: Connection reset by peer)]
mbrgm has joined #nixos
jrolfs has joined #nixos
Wizek_ has joined #nixos
jrolfs has quit [(Ping timeout: 260 seconds)]
Supersonic112 has quit [(Disconnected by services)]
Supersonic112_ has joined #nixos
zeus_ has quit [(Remote host closed the connection)]
Supersonic112_ is now known as Supersonic112
ertes has joined #nixos
sellout- has quit [(Quit: Leaving.)]
Guest60_ has joined #nixos
hexagoxel has quit [(Ping timeout: 256 seconds)]
gbbrt_ has joined #nixos
isidore has quit [(Quit: WeeChat 1.7)]
hexagoxel has joined #nixos
stepcut has quit [(Remote host closed the connection)]
gbbrt has quit [(Ping timeout: 260 seconds)]
sellout- has joined #nixos
Guest62308 has quit [(Ping timeout: 264 seconds)]
stepcut has joined #nixos
zraexy has quit [(Ping timeout: 256 seconds)]
bhipple has quit [(Ping timeout: 260 seconds)]
stepcut has quit [(Ping timeout: 240 seconds)]
ryanartecona has quit [(Quit: ryanartecona)]
Guest60_ has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
bhipple has joined #nixos
athan has joined #nixos
ebzzry has joined #nixos
jrolfs has joined #nixos
Guest60_ has joined #nixos
YellowOnion has joined #nixos
<YellowOnion>
whats the best way to get a cidr file in to the firewall?
<gchristensen>
what is a cidr file?
<gchristensen>
clever: can our grub boot zfs? :)
<gchristensen>
or do I need to make a /mnt/boot partition?
<YellowOnion>
gchristensen, it's a list of ip zones in the cidr format.
<c74d>
there's a boot.loader.grub.zfsSupport option, but then the only system I've tried to use it with never successfully booted
<gchristensen>
c74d: annoying :)
<c74d>
(that option wasn't necessarily at fault)
<YellowOnion>
gchristensen, yeah, seems like I could use fetchurl, and map.
vaibhavsagar has joined #nixos
<ToxicFrog>
gchristensen: yes it can; I have /boot on ZFS on a few machines here
<ToxicFrog>
That said, on modern machines you're going to need a vfat ESP anyways
<gchristensen>
oh? any installation suggestions?
<ToxicFrog>
On EFI systems I just make /boot the ESP and make sure it has enough room for a few kernels/initrds
<gchristensen>
ah
<c74d>
my successful NixOS installations are all (I think) using systemd-boot with the FAT ESP at the start of the disk, ZFS in the middle, and swap at the end
<ToxicFrog>
On my EFI system I have separate pools for /boot and /, but this is more for historical reasons involving lots of drive shuffling and deficiencies in what the platform can boot from
<ToxicFrog>
Er
<ToxicFrog>
On my legacy systems
mguentner has quit [(Quit: WeeChat 1.7)]
davidak has quit [(Quit: Leaving.)]
<gchristensen>
this is a bit annoying, the server takes several minutes to reboot so testing is slightly hell
<gchristensen>
I started down this road thinking it might be easier than waiting for mdadm to assemble a raid5 of 6 x 480GB SSDs
<c74d>
my /boot (= EFI system partition, FAT) on this system is 1 GiB, which has filled up with NixOS kernels and initramfses once so far
<ToxicFrog>
gchristensen: is it legacy or EFI?
mguentner has joined #nixos
<gchristensen>
ToxicFrog: I suspect it is EFI but the host defaults to legacy
byteflame has joined #nixos
Guest60_ has quit [(Quit: My Mac Pro has gone to sleep. ZZZzzz…)]
filterfish_ has quit [(Ping timeout: 240 seconds)]
<peterhoeg>
I'm trying to build the hyperv daemons that are distributed as part of the linux kernel source. So I need to depend on the current kernel.
<peterhoeg>
I see that the virtualbox-guest-additions derivation takes a "kernel" parameter, but when I try doing that in mine, I just get 'anonoymous function called without required argument ‘kernel’'
<peterhoeg>
Does anybody have any bright ideas?
schoppenhauer has quit [(Ping timeout: 256 seconds)]
jmeredith has quit [(Quit: Connection closed for inactivity)]
schoppenhauer has joined #nixos
matthewbauer has joined #nixos
colescott has left #nixos []
<rardiol>
Why does the font for java programs look so awful and where do i change it?
AllanEspinosa has quit [(Ping timeout: 240 seconds)]
<peterhoeg>
rardiol: set the _JAVA_OPTIONS environment variable as follows:
pmn has quit [(Remote host closed the connection)]
mkoenig has quit [(Ping timeout: 246 seconds)]
<c74d>
YellowOnion: I'm guessing there's a \n at the end of the source file, which is getting included in the last element of the list returned by splitString
byteflame has joined #nixos
<YellowOnion>
c74d, well that was obvious, now that I think about it :-|
<c74d>
but no, testing in nix-repl shows `lib.splitString "," "a,b,c,"` is ["a" "b" "c" ""]
jrolfs has quit [(Quit: leaving)]
<c74d>
I'd recommend passing each line through a trim function, but afaict there isn't one
byteflame has quit [(Ping timeout: 260 seconds)]
pmn has joined #nixos
<YellowOnion>
c74d, each line? I could just remove the last line?
colescott has joined #nixos
colescott has left #nixos []
<c74d>
I guess, if it's not important
<YellowOnion>
is there a function application operator or do I just nest all these brackets?
<c74d>
if you mean like Haskell's ($), then no
<NixOS_GitHub>
[nixpkgs] vmchale opened pull request #24736: tweet-hs: init at 0.5.3.22 (master...master) https://git.io/vSrAc
<YellowOnion>
c74d, yeah that is what I was thinking.
<YellowOnion>
how do you find the correct store associated with the current configuration.nix
stepcut has quit [(Remote host closed the connection)]
<c74d>
I've never heard of having multiple stores at once
cmr has quit [(Quit: FIN)]
stepcut has joined #nixos
<YellowOnion>
c74d, I have about 10 different versions of firewall-start
<c74d>
oh, I guess you mean where in the store things have been put to realize the configuration.nix
<YellowOnion>
yeah
cmr has joined #nixos
cmr has quit [(Changing host)]
cmr has joined #nixos
<YellowOnion>
well listing file dates doesn't help, they all have jan 1970 :\
<c74d>
YellowOnion: `nix-store --query --requisites /run/current-system`, then search for "firewall-start"
<c74d>
`--query --requisites` can be abbreviated as `-qR`
stepcut has quit [(Ping timeout: 260 seconds)]
<c74d>
the command should open a list of store paths in your $PAGER
matthewbauer has quit [(Ping timeout: 246 seconds)]
<YellowOnion>
wtf, so its adding a phantom new line...
hamishmack has joined #nixos
<c74d>
(those store paths are the dependencies of the current system generation)
<YellowOnion>
sorry I mean the splitSTring command is wonky.
<YellowOnion>
lib.splitString "\n" "a \n b \n c \n"
<YellowOnion>
[ "a " " b " " c \n" ]
<c74d>
oh huh
colescott has joined #nixos
<c74d>
(I added that about the store paths listed by `nix-store -qR /run/current-system` because I realized I'd forgotten to explain what the command was showing you)
<c74d>
so I was right about lib.splitString; I just didn't test it right
scott has joined #nixos
<c74d>
YellowOnion: this seems like a bug to me; I suggest that you file an issue report for it
<YellowOnion>
c74d, I'll try to remember to, I'll write a work around for now.
cmr has quit [(Quit: FIN)]
cmr has joined #nixos
cmr has quit [(Changing host)]
cmr has joined #nixos
[0x4A6F] has quit [(Ping timeout: 240 seconds)]
cmr has quit [(Remote host closed the connection)]
[0x4A6F] has joined #nixos
ogkloo has quit [(Ping timeout: 256 seconds)]
cmr has joined #nixos
cmr has quit [(Changing host)]
cmr has joined #nixos
cmr has quit [(Client Quit)]
blahdodo has quit [(Remote host closed the connection)]
cmr has joined #nixos
cmr has quit [(Changing host)]
cmr has joined #nixos
blahdodo has joined #nixos
colescott has quit [(Ping timeout: 240 seconds)]
<YellowOnion>
yay looks like it's working now.
rardiol has quit [(Ping timeout: 240 seconds)]
<YellowOnion>
only 5000~ rules to whitelist NZ and AU zones for my server ha.
<YellowOnion>
c74d, thanks for your help.
moet has joined #nixos
<YellowOnion>
actually gah I make a typo, and now I'm getting a stack overflow.
<YellowOnion>
my string is literally too long to process with map?
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 3 new commits to master: https://git.io/vSrxp
<NixOS_GitHub>
nixpkgs/master 47e943e Vanessa McHale: init tweet-hs at 0.5.3.2
<NixOS_GitHub>
nixpkgs/master 8a78fc5 Vanessa McHale: Merge branch 'master' of github.com:vmchale/nixpkgs
<CMCDragonkai>
My kernel logs keep showing a rejected packet from the router IP to my IP source port 67 to destination port 68, basically one of the DHCP messages, either DHCP offer or DHCP acknowledge, should I be allowing port 68 in `allowedUDPPorts`?
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vSoeL
<NixOS_GitHub>
nixpkgs/master 26766f0 Vladimír Čunát: Merge branch 'staging'...
hyphon81 has quit [(Read error: Connection reset by peer)]
<NixOS_GitHub>
[nixpkgs] aristidb closed pull request #24692: haskell: add the linkWithGold function to link packages with ld.gold (master...haskell-link-with-gold) https://git.io/vSVEL
<NixOS_GitHub>
[nixpkgs] aristidb pushed 2 new commits to master: https://git.io/vSoJJ
<NixOS_GitHub>
nixpkgs/master 60f18cf Petr Mikusek: advancecomp: fix license
takle has quit [(Remote host closed the connection)]
<simukis_first_ni>
yeah, my usecase is not exactly hacking but rather inserting a module that's not in linux yet, nix would preferably figure out that the module needs to be recompiled and everything on update etc
takle has joined #nixos
<unlmtd[m]>
Im guessing you would modify that in your nixpkgs then
<unlmtd[m]>
override something in the kernel expression
<simukis_first_ni>
(another unrelated question: are there mirrors of some sort? the default mirrors seem to be extremely slow (~600KB/s on 100mbps fibre))
<unlmtd[m]>
'Its all code man!'
<simukis_first_ni>
just starting with nix so having trouble figuring out stuff. Spent whole day yesterday figuring out how to get `cc` binary happen. Turned out I needed gcc-wrappers package
<unlmtd[m]>
ya im a few months into it and still wrapping my brain around it. but its the right way of doing things, thats for sure
wkennington has quit [(Quit: Leaving)]
takle has quit [(Ping timeout: 240 seconds)]
Dr8128 has joined #nixos
mudri has quit [(Ping timeout: 246 seconds)]
iyzsong has quit [(Ping timeout: 260 seconds)]
Wizek_ has joined #nixos
iyzsong has joined #nixos
bennofs has joined #nixos
<unlmtd[m]>
ok I seem to have figured it out ```packageOverriges.pkgs.nodePackages = pkgs.nodePackages // pkgs.callPackages ./node-packages```
jrolfs has quit [(Ping timeout: 256 seconds)]
<unlmtd[m]>
that was packageOverrides - colemak still messes with me
<unlmtd[m]>
but worth it
mudri has joined #nixos
<unlmtd[m]>
feels like I used to type with one hand
<unlmtd[m]>
but qwerty makes you look like a hard worker
DutchWolfie has quit [(Quit: Konversation terminated!)]
frigate_freedom has left #nixos []
<simukis_first_ni>
> At the moment it’s not possible to use fish as login shell. But you can always start it as default shell of your terminal emulator.
<simukis_first_ni>
aww, seems like nixos is still not ready for general use :(
<steveeJ>
garbas we need binary caches for mozilla-nixpkgs :-D
<bennofs>
simukis_first_ni: you can put exec fish in your bashrc
<bennofs>
(with a check for interactive terminals)
ebzzry has quit [(Ping timeout: 240 seconds)]
kier has quit [(Ping timeout: 240 seconds)]
ksf has joined #nixos
<ksf>
I am having the strangest of problems: A particular build script is using fuse to create a filesystem image. fusermount -u (unmounting) works perfectly fine when invoked normally, but fails when run via nix-shell
mudri has quit [(Ping timeout: 240 seconds)]
lorilan has joined #nixos
YellowOnion has quit [(Read error: Connection reset by peer)]
zero812n has quit [(Quit: Bye)]
<pem__>
Hi! I'm trying to compute the sha256 expected by fetchurl (to package rust crates automatically)
<pem__>
How does nix compute that?
<ksf>
use nix-prefetch-*
<pem__>
And… It's working! Awesome, thanks ksf!
<ksf>
another option would be to write down a bogus hash, run the derivation and have nix download the file, complain about the wrong hash, correct the hash, and re-download :)
pie_ has joined #nixos
<Heffalump>
that's what I always do, I often have trouble quickly converting the hash formats other commands output
<pie_>
so i installed samba but i dont have any smbfs mounting commands :( wht?
<pie_>
*why?
<simukis_first_ni>
would making fish put `/run/current-system/sw/bin` into path make it work as a login shell?
froglegs1 has quit [(Remote host closed the connection)]
ixxie has quit [(Ping timeout: 246 seconds)]
froglegs1 has joined #nixos
bohan_ is now known as bohan
<ksf>
aaaah. there's a setuid wrapper for fusermount, which I circumvented by adding fuse to buildInputs.
jrolfs has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vSoIz
<NixOS_GitHub>
nixpkgs/master 016a319 Tim Steinbach: linux: 4.9.20 -> 4.9.21
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.03: https://git.io/vSoIg
<NixOS_GitHub>
nixpkgs/release-17.03 3f4d221 Tim Steinbach: linux: 4.9.20 -> 4.9.21...
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vSoIr
<NixOS_GitHub>
nixpkgs/master 1988c1f Tim Steinbach: linux: 4.10.8 -> 4.10.9
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.03: https://git.io/vSoIo
<NixOS_GitHub>
nixpkgs/release-17.03 d098c20 Tim Steinbach: linux: 4.10.8 -> 4.10.9...
pie_ has quit [(Changing host)]
pie_ has joined #nixos
jrolfs has quit [(Ping timeout: 260 seconds)]
mudri has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/vSoIy
<NixOS_GitHub>
nixpkgs/master 79f9544 Tim Steinbach: linux: 4.4.59 -> 4.4.60
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.03: https://git.io/vSoIS
<NixOS_GitHub>
nixpkgs/release-17.03 cae3e55 Tim Steinbach: linux: 4.4.59 -> 4.4.60...
elninja44 has quit [(Remote host closed the connection)]
aminechikhaoui has quit [(Ping timeout: 256 seconds)]
aminechikhaoui has joined #nixos
<NixOS_GitHub>
[nixpkgs] volth opened pull request #24742: reason: init at 1.13.3 (master...reason-1.13.3-master) https://git.io/vSoLZ
<simukis_first_ni>
cc -ltinfo complains about /nix/store/p6wcdqa2ks82xq7d4d5019drrjxlqrdj-binutils-2.27/bin/ld: cannot find -ltinfo, despite libtinfo.so{,.6} being present in /run/current-system/sw/lib/, why?
ebzzry has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 5 new commits to release-17.03: https://git.io/vSoL5
<NixOS_GitHub>
nixpkgs/release-17.03 565dc28 Linus Heckemann: Add removeReferencesTo for removing specific refs...
<NixOS_GitHub>
nixpkgs/release-17.03 b348e75 Joachim Fasting: docker: use removeReferencesTo...
<NixOS_GitHub>
nixpkgs/release-17.03 b37957a Tim Steinbach: docker: 1.13.1 -> 17.03.0-ce...
mudri has quit [(Ping timeout: 264 seconds)]
drasich has quit [(Ping timeout: 246 seconds)]
<NixOS_GitHub>
[nixpkgs] ljli opened pull request #24743: haskell: Correct dependencies for haskell-tools-*_0_6_0_0 (master...hs-haskell-tools) https://git.io/vSotL
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 2 new commits to master: https://git.io/vSotm
<NixOS_GitHub>
nixpkgs/master 775531c Periklis Tsirakidis: qt58: extend darwin compatibility
<NixOS_GitHub>
nixpkgs/master e0b1288 Thomas Tuegel: Merge pull request #24340 from periklis/topic_qt58_modules...
mizu_no_oto has joined #nixos
kier has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 2 new commits to master: https://git.io/vSot4
<NixOS_GitHub>
nixpkgs/master 4ceeac7 Tim Steinbach: skopeo: default policy
<NixOS_GitHub>
nixpkgs/master 184e323 Tim Steinbach: Merge pull request #24544 from NeQuissimus/skopeo_policy...
Supersonic112 has quit [(Ping timeout: 260 seconds)]
Supersonic112 has joined #nixos
cpennington has joined #nixos
mood__ has quit [(Quit: Gone.)]
<jophish>
Hi all
<jophish>
I'm writing a tool which calls out to nix-build and friends, I'd like to write some tests for it, however they will require modifying the nix store
<jophish>
I think it would be best to do this either in a small VM or with a local nix store
<jophish>
Does anyone have any examples or experience in doing this which they can share?
jrolfs has joined #nixos
jrolfs has quit [(Ping timeout: 260 seconds)]
<gchristensen>
mbrgm: ack I forgot to commit an broke it, heh
<gchristensen>
ping again in like 8 hours? :P
<jophish>
hey gchristensen, how's things?
<jophish>
I'm working on the haskell nix-build interface today (finally)
<bennofs>
jophish: oh, for hercules?
zarel has joined #nixos
<jophish>
bennofs: yeah, I'm trying to use NIX_BUILD_HOOK to get the list of packages required by nix-store --realise (and nix-instantiate in the case of IFD)
estewei has quit [(Quit: Quit)]
<bennofs>
jophish: hmm. does this handle 'if predicate (import ${someDrv}) { ... } else { ... }' dependencies?
<jophish>
bennofs: it should do
<bennofs>
jophish: I don't know of any nix command that can handle that. But perhaps NIX_BUILD_HOOK is the missing one?
<bennofs>
hmm yeah NIX_BUILD_HOOK may workl
<jophish>
I call nix-instantiate on the file, and that (tells the binary at NIX_BUILD_HOOK that it needs someDrv) if predicate is true
<jophish>
(parentheses added to disambiguate)
<bennofs>
jophish: that only works if you're not using nix-daemon, right? or does NIX_BUILD_HOOK get somehow magically transfered to the daemon?
<bennofs>
hmm well it may work
<jophish>
I think it works with the daemon, gchristensen came up with the idea
<bennofs>
that's a nice one
<jophish>
yeah, it's fiddly to use and the documentation for what should live at NIX_BUILD_HOOK is "read the source"
<jophish>
I'm thinking at the moment about the best way to test this
<jophish>
A local nix store in /tmp would be nicest, however it might be harder to set up than just instantiating a VM with nix on
<bennofs>
jophish: a docker container with /nix?
<jophish>
that might be best, better on OSX than the stuff in nixos/tests
<jophish>
heh, I'd really like: runInContainer :: ContainerDesc -> IO a -> IO a
<jophish>
There's a lot of fancy VM machinery in the nix ecosystem, it seems a shame to not be using it
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 1 new commit to master: https://git.io/vSoqz
<NixOS_GitHub>
nixpkgs/master d0954b5 Thomas Tuegel: nixos/fontconfig-ultimate: Restore presets...
<jophish>
bennofs: I'll give that a try as soon as I've got this up and running!
[0x4A6F] has quit [(Ping timeout: 260 seconds)]
<Heffalump>
I have nixdb that I need to dump and restore because I accidentally updated the sqlite schema with a newer nix. But I also have a single corrupted store path, which means that nix-store --load-db is refusing to restore it. Is the format of the dump documented/can I manually remove the broken path somehow?
<DeaDSouL>
hi, how can I force some user to logout... I used to use `fuser -k /home/username` which will kill everything used by that user.... I didn't find the fuser package in the repo... and the plasma/system is frozen I logged in remotely via ssh... so, how can I force that user to logout
niklob has joined #nixos
davidak has joined #nixos
<tilpner>
DeaDSouL - You might want nixpkgs.psmisc
<DeaDSouL>
tilpner: thanks :)
ebzzry has quit [(Ping timeout: 240 seconds)]
RchrdB has joined #nixos
<DeaDSouL>
tilpner: how did you find it ? because I need the `skill` package too
<DeaDSouL>
tilpner: `nix-env -iA nixpkgs.psmisc` .... gave me .... error: attribute ‘nixpkgs’ in selection path ‘nixpkgs.psmisc’ not found
<tilpner>
I had "fuser" in my path, so I knew it existed, then searched nixpkgs for it, but I could also have used nox/nq or the web search
<DeaDSouL>
just wondering... since nixos structures are different from a traditional linux distros... could it be the reason why everything died when I used `fuser -k /home/user1` ? it got me logged out via ssh, and disconnect from wireless, although I was logging-in from user2 not user1
<DeaDSouL>
even if user1 and user2 are logged it... I can't see anyone but the user I'm logged-in with.. even if I'm root!!
<shlevy>
pierron: Also, are these comonads? :o
sellout has quit [(Quit: Leaving.)]
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 2 new commits to master: https://git.io/vSoYA
<NixOS_GitHub>
nixpkgs/master 2214b63 Thomas Tuegel: nixos/fonts: install gyre-fonts by default...
<NixOS_GitHub>
nixpkgs/master a40600b Thomas Tuegel: fontconfig-penultimate: 0.3.2 -> 0.3.3
<jophish>
bennofs: test.nix works without a problem
<bennofs>
jophish: nice!
<jophish>
One downside of this is that if a derivation required for evaluation is already present in the store then we don't hear about it
<shlevy>
Hmm, is there any modification we could make to get something like http://sprunge.us/OXXH to evaluate to true? I don't think it's actually what we want with respect to overrides, but "super represents the whole set as if self had been fixed at that point" is not obviously wrong semantics either.
<shlevy>
pierron: ^
<bennofs>
shlevy: "super represents the whole set as if self had been fixed at that point" is not correct for nixpkgs btw
<shlevy>
Right, the expr I pasted currently evaluates to false
<bennofs>
shlevy: in nixpkgs, super.callPackage looks at 'self' (the real self, not the self "at that point")
<shlevy>
I'm saying would it make sense to have a version of extensions that behaves this way
<niklob>
Hello, does the nixos kernel have EFI stub support enabled?
<shlevy>
niklob: Yes
nh2 has joined #nixos
<shlevy>
bennofs: Hmm, I don't quite understand pierron's security thing. Why would self.callPackage be problematic?
<bennofs>
shlevy: i don't understand it 100% either. I opted to wait till a more details of that get known for now
<shlevy>
bennofs: My intuition aligns with your initial position, that super.callPackage would be the callPackage you would have gotten if you'd stopped just before the current layer, and self.callPackage is the final callPackage
<niklob>
shlevy: Ty for your quick response. Is it possible to edit the Built-in kernel command line setting?
<bennofs>
shlevy: it seems to me like pierron wants to make a distinction between static linking and dynamic linking by whether you use self.bla or directly refer to bla
<shlevy>
bennofs: I think to get the desired behavior you'd just need to change "let super = rattrs self;" to "let super = fix' rattrs", right?
<shlevy>
I'll test
<bennofs>
shlevy: that would not work, as you want deps in super to have access to self though. that's the whole point
<shlevy>
Oh, right
<shlevy>
No, wait
zarel has quit [(Quit: Leaving)]
<bennofs>
shlevy: or maybe not... this is confusing :)
<shlevy>
rattrs is the full set from before
<shlevy>
OK I think I know what it should look like to behave the way I want
<shlevy>
Let me test :)
CMCDragonkai has quit [(Ping timeout: 260 seconds)]
<bennofs>
right, super should not contain overriden stuff "from the future" :)
<bennofs>
shlevy: problem with that: we currently hack around the missing self: super: in nixpkgs config by making self an attribute of super (the pkgs attribute)
<bennofs>
shlevy: that does make sense, but I fear it may be very inefficient to compute the fixpoint at every stage
<shlevy>
bennofs: Why? You only have to resolve it deep enough to get the attribute you care about
<bennofs>
shlevy: hmm wait, how would I ever override a dep of a package in that fixpoint?
<shlevy>
bennofs: Refer to self
<bennofs>
shlevy: say foo depends on bar
<bennofs>
shlevy: the final set has super.foo (due to //)
<shlevy>
Note super vs super'
<shlevy>
The naming isn't right :D
<bennofs>
shlevy: ah
<bennofs>
yes that makes more sense
<shlevy>
pierron: Would you object to something like this, assuming we figured out compatibility with existing usage? http://sprunge.us/HBdf
jrolfs has joined #nixos
<shlevy>
pierron: It seems more semantically correct IMO. The 'super' argument cannot change based on what comes after it in the stack
<bennofs>
shlevy: note that we currently do something like this for the "local" all-packages fixpoint i noted in my email
<bennofs>
shlevy: perhaps we can get of rid of that oddity then as well
ixxie has quit [(Ping timeout: 240 seconds)]
<shlevy>
Hmm
<shlevy>
I'm not sure whether this is what we want actually :/
<bennofs>
shlevy: it does require you to use self.callPackage throughout
<shlevy>
If I say foo = super.foo.override { bar = bar_1_2; }; and then later do baz = bazFull;, if foo depends on baz it won't pick it up
<shlevy>
Right?
<pie_>
anyone know how i can mount a windows smb share on the filesystem?
<shlevy>
Even if you do self.callPackage, because foo was fixed to super.foo at that layer
<gchristensen>
jophish: yeah, I abandoned that idea FWIW :P
<gchristensen>
jophish: better to just go around nix-build
<bennofs>
shlevy: hmm right
<shlevy>
Damn, so I don't think this is workable
<bennofs>
shlevy: seems like what you actually want here is to take the recipe for foo from super but the deps from self
<jophish>
gchristensen: what do you mean, my program doens't use nix build already
<shlevy>
Even though it's itchy that super can change based on the rest of the stack
jrolfs has quit [(Ping timeout: 240 seconds)]
<shlevy>
bennofs: Yeah, maybe the right answer is a better way to do overrides that knows about self vs super
<bennofs>
shlevy: which could be made work if you took super.foo.__the_function_passed_to_callPackage and re-callPackage it with self as scope
<bennofs>
(the new "self")
<shlevy>
Right
<bennofs>
the problem really seems to be that we tie where the recipes for packages come from together with where we resolve deps from
<shlevy>
Yeah
<shlevy>
Maybe we shouldn't be calling callPackage until the end anyway...
<bennofs>
shlevy: I wonder how this used to work, where `super` was even more magic :) (I think nixpkgs overrides used to eval each attribute with a different super?)
<bennofs>
did anyone understand how it worked back then? :D
<shlevy>
Hmm, I remember how things worked before we ever had super, and self was lazier so you could override stuff if you were careful
iyzsong has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
colescott has joined #nixos
johnsonav has joined #nixos
<jophish>
bennofs: Ah you were right about NIX_BUILD_HOOK not being used when the daemon is active. In fact, it was me who filed an issue about this and forgot!
<bennofs>
jophish: ugh :/. perhaps just patch nix-instantiate to call a script before it checks if a store path is already build?
<bennofs>
or write it to a file
<bennofs>
or something
<shlevy>
It would be unsafe for a user's NIX_BUILD_HOOK to affect the daemon!
<shlevy>
You need to set the env for the daemon
<bennofs>
jophish: like the .d files created by some compilers with preprocessor deps info
<jophish>
shlevy: I'd like the opposite, I'd like for the daemon to not be called at all when NIX_BUILD_HOOK is set
<shlevy>
jophish: Ah, you probably want to look into the new store mechanism in 1.12
<jophish>
(obviously the daemon will be called somewhere down the pipe, but only after my program has had a peek at the traffic)
<jophish>
shlevy: for some context, this is for hercules, I'd like to get some precise information about what's needed for evaluation and building, as well as deciding which machines to perform these builds on
<jophish>
shlevy: I'll take a look, is it documented?
<shlevy>
jophish: You'd basically write a nix::Store implementation that does your stuff and calls out to the daemon when it's done, and set NIX_REMOTE to the uri of the implementation
<jophish>
bennofs: that's not a bad idea. Eventually I want to rewrite the nix evaluator to allow for better parallel instantiation
<shlevy>
But currently you''ll need to LD_PRELOAD it, we should have proper plugins
<jophish>
wouldn't it be great to evalutate nixpkgs across multiple machines!
<bennofs>
if nixpkgs eval takes so long that you have to split it across multiple machines, something should be changed
<jophish>
shlevy: hmm, I wonder how Haskelly this can be, I've not done any c++/Haskell interfacing before
<bennofs>
jophish: don't write the LD_PRELOAD in haskell.
<jophish>
yeah, A c++ program which communicates with haskell-land might be easier
<jophish>
bennofs: how long does hydra take to eval nixpkgs at the moment?
<bennofs>
jophish: hmm, that might be an idea: run a "proxy store" that tracks the deps
<bennofs>
jophish: if hercules implements the store api, you could use a hercules endpoint tagged with an id to collect the deps of a build
derjohn_mob has quit [(Ping timeout: 256 seconds)]
Gravious has joined #nixos
davidak has quit [(Quit: Leaving.)]
<bennofs>
is there a "isPrefixOf" builtin/lib function?
<bennofs>
hasPrefix...
mudri has joined #nixos
ixxie has joined #nixos
<Gravious>
hi, i believe there's a nix command for generating the sha256 of a tarball url, but i can't find reference to it anymore?
alx741 has joined #nixos
<bennofs>
Gravious: nix-prefetch-url url prints it
<Gravious>
bennofs, ah nice thanks :)
<jophish>
bennofs: what do you mean by collect the deps of a build?
<bennofs>
jophish: well, I suppose you wanted this to find the dependencies of evaluation?
<bennofs>
jophish: if evaluation uses your binary-cache proxy, you can see what stuff it probes
eacameron has joined #nixos
<jophish>
bennofs: ah, that's what I thought. Where does the "id" fit into this?
<jophish>
shlevy: I'm not sure I *entirely* understand how the store api fits into nix. I see it's instantiates by a local store
<jophish>
oops, hit enter too early
<bennofs>
jophish: if you have multiple builds running at the same time, you have to make sure that each get their "own" binary cache proxy hence the id
<jophish>
I see that it's instantitated by a s3 store and a local store, the s3 store is not capable of building things. I think for what I want to do (capturing log output too) I'd need to implement a `Worker` too, is this correct?
<jophish>
bennofs: ah, of course!
alx741 has quit [(Read error: Connection timed out)]
<bennofs>
jophish: can't you capture log output from nix-instantiate output?
alx741 has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 21 new commits to master: https://git.io/vSosk
<NixOS_GitHub>
nixpkgs/master 22d9f53 Philipp Hausmann: Init Perl AduioScan
<NixOS_GitHub>
nixpkgs/master 718f3ba Philipp Hausmann: Init Perl ImageScale
<NixOS_GitHub>
nixpkgs/master ffa3dbe Philipp Hausmann: init slimserver
<jophish>
bennofs: I can, it's what I'm doing at the moment. I'm not sure how this will fit into the libstore approach though
<NixOS_GitHub>
[nixpkgs] bennofs pushed 1 new commit to master: https://git.io/vSosO
takle has quit [(Remote host closed the connection)]
<jophish>
I suppose that this can't just be a logging passthrough straight to a local store if I want to direct which machines perform the build myself
davidak has joined #nixos
<bennofs>
jophish: for a hydra-like thing, don't you want to link against the nix libraries directly instead of relying on tools like nix-build?
<jophish>
bennofs: There are downsides to that too, for example replicating the steps locally becomes harder
<jophish>
The end goal is to be able to give a .nix file to a function, have it evaluate and build the derivations in that expression, while piping the logs somewhere and keeping a record of what was required for that build
<jophish>
it seems as though most of that can be done with NIX_BUILD_HOOK, although the libstore stuff does seem to have some tempting advantages
<gchristensen>
ok first reboot trying to install nixos + zfs. let's play, Will It Boot
<clever>
gchristensen: ive done 5 systems with that, and one of them is zfs on lvm on luks
<gchristensen>
:o :o :o
<gchristensen>
if this worked on my first try, I might just die
coltfred has joined #nixos
<bennofs>
jophish: the only way to get accurate requirements info seems to be a custom nix store
<NixOS_GitHub>
[nixpkgs] edanaher opened pull request #24745: acme: Add "domain" option to separate domain from name (master...add-acme-domain) https://git.io/vSosP
<jophish>
bennofs: Sounds worth a try. Getting the live logs of the builds without reimplementing build.cc might require some careful thinking
ksf has quit [(Ping timeout: 260 seconds)]
jrolfs has joined #nixos
takle has joined #nixos
<clever>
bbl
clever has quit [(Quit: leaving)]
<bennofs>
jophish: hmm, or perhaps you could use both NIX_BUILD_HOOK and a custom store
<gchristensen>
it worked :O
<bennofs>
jophish: NIX_BUILD_HOOK for the log reporting, custom store for the dependency tracking
<gchristensen>
ok now to try it automatically on pxe boot...
<gchristensen>
copumpkin: ^ if this works, it shaves 40 minutes off the T2 startup time
<bennofs>
jophish: like nix-build with remote store = hercules ---> hercules tracks build request --> hercules calls nix-build, knowing the exact dependencies and knowing that no other deps need to be built
<bennofs>
or even just, nix-store --realise
ixxie has quit [(Ping timeout: 268 seconds)]
takle has quit [(Remote host closed the connection)]
<bennofs>
jophish: yeah, that's just the nix-daemon "API"
coltfred has quit [(Ping timeout: 260 seconds)]
<jophish>
bennofs: nix-store --realise on a remote machine might be more tricky
markus1199 has joined #nixos
<bennofs>
jophish: btw, even with all this stuff, I think there's still a way to create a network dependency that is not tracked: builtins.fetchurl
hamishmack has quit [(Quit: hamishmack)]
takle has quit [(Ping timeout: 260 seconds)]
<jophish>
That might have to wait until I write my own nix evaluator!
mkoenig has joined #nixos
hyper__ch has joined #nixos
<hyper__ch>
hi there, having issues installing nixos on a larger ssd.... I formated the sdd accordingly, mounted it in the live environment to /mnt.... edited the hardwre configuration file to represent the new uuid... but when I run nixos-install it tells me: error: opening file '/etc/nix/nix.conf': No such file or directory
<hyper__ch>
but it does exist
<hyper__ch>
also restored all stuff from the old install excpet /nix
niklob has quit [(Ping timeout: 260 seconds)]
marsel has joined #nixos
NickHu[w] has joined #nixos
NickHu has left #nixos ["Kicked by @appservice-irc:matrix.org"]
Mercuria1Alchemi has joined #nixos
calvertvl has joined #nixos
hyper__ch has quit [(Quit: Page closed)]
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 1 new commit to master: https://git.io/vSoZC
<NixOS_GitHub>
nixpkgs/master 16e12a6 Michael Raskin: glusterfs: 3.10.0 -> 3.10.1
eacameron has quit [(Remote host closed the connection)]
eacameron has joined #nixos
xadi has joined #nixos
eacameron has quit [(Ping timeout: 246 seconds)]
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 1 new commit to release-17.03: https://git.io/vSonJ
<NixOS_GitHub>
nixpkgs/release-17.03 f3a0c16 Michael Raskin: glusterfs: 3.9.1 -> 3.10.1
<jophish>
$ nix-build: error: secret key is corrupt
<jophish>
much better
eacameron has joined #nixos
<jophish>
ah, if I run this nix version will it helpfully upgrade my db
<jophish>
perhaps best to do this in a VM
c0ff33 has joined #nixos
<c0ff33>
Howdy
<dtzWill>
hmm, what's the versioning scheme used in nixUnstable? the suffix in particular, it's "pre${MONOTONICALLY_INCREASING_SOMETHING}_${GIT_SHORT_REV}" AFAICT but dunno what the first part corresponds to. Build somewhere? svn rev? lol
<shlevy>
jophish: nix 1.11.7 or so is forwards compatible to nix 1.12's db
<c0ff33>
Anyone doing some Python development on NixOS? I have this strange issue where most packages basically don't install. My workflow goes like that - spawn nix-shell with basic Python tooling (python3, pip, virtualenv), create virtualenv for my project, pip install gmpy. Sadly it crashes because of lack of gmp.h. gmp library is obviously installed.
<dtzWill>
oh, it is?? oh yay lol
eacameron has quit [(Ping timeout: 260 seconds)]
<bennofs>
jophish: hmm, do you really need to implement a new store in C++ though? I wonder if it would be possible to just implement a HTTP store endpoint
<bennofs>
s/store/binary cache
<shlevy>
bennofs: You can't do nix-store -r on an http store
<shlevy>
You can query it and fetch from it but not realize stuff
<jophish>
shlevy: could it be some kind of lazy store?
mudri has joined #nixos
<bennofs>
shlevy: oh? but the store could "magically" have everything that you query for
<jophish>
^
<bennofs>
shlevy: by building it in the background
<shlevy>
Sure
<jophish>
bennof can type faster than I can
<bennofs>
:)
ksf has joined #nixos
eacameron has joined #nixos
<bennofs>
shlevy: hmm, is a http binary cache told which drv the request is for?
<bennofs>
shlevy: or is it only told the output path
<shlevy>
What command are you running specifically?
jrolfs has joined #nixos
<bennofs>
shlevy: if you run nix-instantiate for example
<bennofs>
shlevy: (with IFD)
<shlevy>
bennofs: I don't think that will work
<shlevy>
It *might*
<shlevy>
and if it does it will know which drv you want
<shlevy>
But I don't think it will work
<bennofs>
shlevy: why would it not work?
<shlevy>
Because it still tries to call the buildPaths store operation
<shlevy>
buildPaths just does no work if the path is already there, but I don't think HttpStore implements it at all
<bennofs>
shlevy: ah hmm yeah timeouts could be a problem. I thought I could perhaps "lie" and say I have the path and build it, but I guess nix won't wait till I have finished building it before fetching
<shlevy>
Not because of timeouts
eacameron has quit [(Ping timeout: 240 seconds)]
<bennofs>
shlevy: well if I act like a binary cache and also serve the .nar (build it on the server) then nix won't even have to call buildPaths
<bennofs>
shlevy: oh why would it call buildPaths if it substitutes from the bin cache?
<shlevy>
Because it's buildPaths that *does* susbtitution
<bennofs>
oh wut? then what is http binary cache store used for?
Gravious has quit [(Ping timeout: 260 seconds)]
eacameron has joined #nixos
<bennofs>
i thought that's what regular nix uses for cache.nixos.org
aminechikhaoui has quit [(Ping timeout: 240 seconds)]
<shlevy>
if you have it in your substituors
<shlevy>
Then it's used at substitution phase by the daemon
aminechikhaoui has joined #nixos
<bennofs>
shlevy: but shouldn't I be able to fake a binary cache that lazily builds everything that is requested?
<shlevy>
Sure, you need to make it a substitutor for the daemon then
<bennofs>
shlevy: like a cache.nixos.org that would just have a narinfo for every hash that triggers a build of it
vaibhavsagar has quit [(Ping timeout: 240 seconds)]
<shlevy>
Yeah, that's fine
<shlevy>
I thought you were trying to avoid the daemon
<bennofs>
shlevy: but that might run into timeouts, right? because I would have to finish the build in the time nix tries to fetch it
<jophish>
ah, that was earlier when I was doing the NIX_BUILD_HOOK trick
<shlevy>
I don't know if a timeout is applied to subsitutors actually
<jophish>
bennofs: how would that work in the case that the paths were already present in the store
<jophish>
presumably the substitutor wouldn't be called
eacameron has quit [(Ping timeout: 256 seconds)]
alfie has joined #nixos
<shlevy>
Correct
<bennofs>
hmm right. you'd need to make it take priority over the local file store
<shlevy>
Then you can't use the daemon
<shlevy>
You'll need to write some C++ :P
<bennofs>
ok, so any way, seems like you need to write C++. So you might as well go the correct way and hook the store
NickHu[w] has quit [(Quit: WeeChat 1.7)]
<bennofs>
shlevy: hmm, what's the difference between local-fs-store and local-binary-cache-store?
<jophish>
While we're at it, thanks shlevy and bennofs for the help and advice today!
<bennofs>
hmm local-{,fs}-store seems to be what is used to drive lookup from /nix/store?
NickHu_ has joined #nixos
NickHu_ has quit [(Client Quit)]
NickHu_ has joined #nixos
c0ff33 has quit [(Quit: Lost terminal)]
nckx has quit [(Quit: restarting my ZNC)]
nckx has joined #nixos
<jophish>
hmm, nix-build (Nix) 1.12pre5122_c60715e fails with: error: getting attributes of path ‘Versions/Current/CoreFoundation’: No such file or directory
Dr8128 has joined #nixos
<jophish>
1.11.8 works without issue
bennofs has quit [(Ping timeout: 264 seconds)]
<NixOS_GitHub>
[nixpkgs] vmchale opened pull request #24748: init madlang at 2.1.0.1 (master...master) https://git.io/vSoCF
mkoenig has quit [(Read error: Connection reset by peer)]
<jophish>
shlevy: I've got nix starting with LD_PRELOAD, however "Registering Dummy Store" is never printed. Do I need to add an unhandled "hercules://foo" URI somewhere?
<ksf>
is there a proper way to get at the path "/var/setuid-wrappers/fusermount" from inside nixpkgs?
<ksf>
I can depend on fuse, but that doesn't seem to imply that the necessary setuid wrapper is actually installed
<ksf>
(or it might, I have no idea)
<ksf>
passing pkgs.fuse into buildInputs at least puts the exactly wrong (non-setuid) thing in path which is what won't work as the build script actually uses it.
<hyper_ch>
dman, nixos is completed f**** up again.... virtual box 17.03 stable won't build
<hyper_ch>
neither stable nor unstable are working for me :(
<jophish>
shlevy: good idea, thatnks
<jophish>
thanks*
<Mateon1>
Hi, why does patchelf rip out $out/lib and $out/lib64 from the rpath? lib is certainly used, and in fact, I get a crash when running the executable because it can't find an .so file, unless I completely override the fixupPhase to do nothing.
<Dezgeg>
is the build system perhaps adding a rpath to the build directory? (/tmp/nix-build-foo/foo-1.2.3/lib or so)?
<bennofs>
Mateon1: it rips it out if it is not required by any NEEDED entries afik?
<Mateon1>
"NEEDED" entries?
erictapen has joined #nixos
<erictapen>
good eveing everyone, did somebody manage to run php on nginx in nixos?
<Mateon1>
Well, true. The .so file is not marked (NEEDED) in readelf -d output
<Mateon1>
How can I fix that?
<bennofs>
Mateon1: I think patchelf has an --add-needed switch?
<Mateon1>
bennofs: Should I try that in the preFixup hook?
marusich has joined #nixos
endformationage has joined #nixos
NickHu has joined #nixos
pie_ has quit [(Ping timeout: 256 seconds)]
<Mateon1>
Nice! that fixed it, thanks bennofs!
<bennofs>
Mateon1: yeah sounds good. or just do it after the line where you patch it
<Mateon1>
I don't manually patch it, I let the default fixup phase run.
Gravious has joined #nixos
<bennofs>
oh that's weird then. whatever, if it fixed it, nice
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 1 new commit to master: https://git.io/vSoBL
<NixOS_GitHub>
nixpkgs/master 333923c Thomas Tuegel: dropbox: 22.4.24 -> 23.4.17
<NixOS_GitHub>
[nixpkgs] hrdinka pushed 1 new commit to master: https://git.io/vSoBt
<NixOS_GitHub>
nixpkgs/master 6016023 Christoph Hrdinka: nsd: 4.1.14 -> 4.1.15
yorick has quit [(Remote host closed the connection)]
<NixOS_GitHub>
[nixpkgs] hrdinka pushed 1 new commit to release-17.03: https://git.io/vSoBY
<NixOS_GitHub>
nixpkgs/release-17.03 a00c4ae Christoph Hrdinka: nsd: 4.1.14 -> 4.1.15
<NixOS_GitHub>
[nixops] ljli opened pull request #644: GCE backend: Add preemptible option for instances (master...gce-preemptible) https://git.io/vSoB3
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 6 new commits to master: https://git.io/vSoBs
<DeaDSouL>
how can I point to the current used bash in my scripts? I used to use `#!/bin/bash`
eacameron has joined #nixos
<srhb>
DeaDSouL: Scripts as in things you've written derivations for? If so, you can use patchShebangs for get the store path of the interpreter. Otherwise, /usr/bin/env bash will probably suffice.
erictapen has quit [(Ping timeout: 268 seconds)]
<DeaDSouL>
srhb: thanks :)
<DeaDSouL>
is there a global bin directory where I can save my scripts to be available to all users ? like `/usr/bin/` and `/usr/sbin/` ?
<DeaDSouL>
that it's being included by default in $PATH ?
Itkovian has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
<copumpkin>
gchristensen: how was it taking 40 minutes in the first place? :O
<toppler>
Is anyone set up with postfix here?
<ronny>
DeaDSouL: wha the heck do you intend to do? without context it sounds like you want to go all that nix/nixos is intended to do
<ronny>
*against
<NixOS_GitHub>
[nixpkgs] copumpkin pushed 1 new commit to master: https://git.io/vSoRp
<NixOS_GitHub>
nixpkgs/master 72d9016 Dan Peebles: darwin.make-bootstrap-tools: fix to use LLVM 4...
<ronny>
DeaDSouL: typically if youwant to install scripts/make them availiable you package them as nix package, declare them as part of the system package set and after the next switch its availiable to everyone
<DeaDSouL>
ronny: sometimes you wrote a useful script that you'd like to let it be available to all users.. where would you store it?
<ronny>
DeaDSouL: i typically package it as python application for python, then aadd a system level package to nix
<ronny>
you know, instead of jsut smearing it somewhere, i make correct packagesin the language ecosystem and then use the correct nix tools to make them availiable to nix users
<ronny>
correct and repeatable package setup takes a bit longer than just dumping random files somewhere, but it means it is repackageable, reusableand testable
<DeaDSouL>
wow, that would take a lot of time, to re-do all my scripts :D
<DeaDSouL>
ronny: can you point me where to start learning about packaging a small scripts?
erictapen has joined #nixos
justin_ has quit [(Ping timeout: 240 seconds)]
justin_ has joined #nixos
erictapen has quit [(Ping timeout: 260 seconds)]
DeaDSouL has quit [(Remote host closed the connection)]
justin__ has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<yorick>
how do I output "${a}" in a '' block?
<yorick>
''\${a}'' gives \\a
<yorick>
oh, two single quotes
<yorick>
ew
justin_ has quit [(Ping timeout: 256 seconds)]
johnsonav has quit [(Ping timeout: 246 seconds)]
Gravious has quit [(Remote host closed the connection)]
<NixOS_GitHub>
[nixpkgs] copumpkin pushed 1 new commit to master: https://git.io/vSo0K
<NixOS_GitHub>
nixpkgs/master fbdc461 Dan Peebles: nix: fix on LLVM 4.0
<dtzWill>
\o/
vandenoever has quit [(Ping timeout: 260 seconds)]
jrolfs has quit [(Read error: Connection reset by peer)]
acarrico has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
Gravious has joined #nixos
jrolfs has joined #nixos
<LnL>
oh! llvm-4 in master already?
<copumpkin>
yes
<copumpkin>
I think vcunat merged staging -> master
<copumpkin>
I think nixpkgs-unstable is also unblocked
<copumpkin>
so we might even get a channel release with it
acarrico has quit [(Ping timeout: 268 seconds)]
eacameron has joined #nixos
jrolfs has quit [(Ping timeout: 240 seconds)]
jrolfs has joined #nixos
NickHu_ has quit [(Quit: WeeChat 1.7)]
Filystyn has joined #nixos
<Ralith>
huh? wasn't llvm 4 in 17.03?
batdog_ has quit [(Ping timeout: 260 seconds)]
<LnL>
yes, but it's the default for darwin now
Gravious has quit [(Ping timeout: 240 seconds)]
eacameron has quit [(Ping timeout: 240 seconds)]
<copumpkin>
(our stdenv uses llvm so it changes everything, unlike on linux)
reinzelmann has joined #nixos
systemfault has quit [(Quit: Bye!)]
justin__ has quit [(Ping timeout: 240 seconds)]
schoppenhauer has quit [(Quit: Adé)]
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 2 new commits to master: https://git.io/vSouB
<NixOS_GitHub>
nixpkgs/master 9d94e10 Tim Steinbach: coursier: 1.0.0-M15-5 -> 1.0.0-RC1
<NixOS_GitHub>
nixpkgs/master 33b4be3 Tim Steinbach: Merge pull request #24750 from NeQuissimus/coursier_1_0_0_rc1...
oever has joined #nixos
vandenoever has quit [(Ping timeout: 240 seconds)]
oever is now known as vandenoever
reinzelmann has quit [(Quit: Leaving)]
Ivanych has quit [(Ping timeout: 260 seconds)]
vandenoever has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
hiratara has quit [(Ping timeout: 246 seconds)]
hiratara has joined #nixos
eacameron has quit [(Ping timeout: 252 seconds)]
marsel has quit [(Ping timeout: 260 seconds)]
bennofs has quit [(Ping timeout: 256 seconds)]
eacameron has joined #nixos
Filystyn has quit [(Read error: Connection reset by peer)]
Filystyn has joined #nixos
Filystyn has quit [(Changing host)]
Filystyn has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
justin__ has joined #nixos
johnsonav has joined #nixos
acarrico has joined #nixos
eacameron has quit [(Ping timeout: 252 seconds)]
Filystyn has quit [(Read error: Connection timed out)]
Filystyn has joined #nixos
eacameron has joined #nixos
aminechikhaoui has quit [(Ping timeout: 260 seconds)]
eacameron has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
justanotheruser has quit [(Ping timeout: 264 seconds)]
eacameron has quit [(Remote host closed the connection)]
justin__ has quit [(Ping timeout: 252 seconds)]
Khorne[m] has joined #nixos
erictapen has joined #nixos
derjohn_mob has joined #nixos
schoppenhauer has quit [(Ping timeout: 240 seconds)]
sigmundv has quit [(Ping timeout: 240 seconds)]
schoppenhauer has joined #nixos
<NixOS_GitHub>
[nixpkgs] corngood opened pull request #24757: mesa: patch vulkan manifests when any driver is enabled (master...mesa-manifest-fix) https://git.io/vSoak
eacameron has joined #nixos
AllanEspinosa has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
acarrico has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 5 new commits to master: https://git.io/vSoaC