ryantrinkle has quit [(Ping timeout: 248 seconds)]
zennist has quit [(Remote host closed the connection)]
zennist has joined #nixos
oida_ has joined #nixos
griff_ has joined #nixos
oida has quit [(Ping timeout: 240 seconds)]
_ris has quit [(Ping timeout: 260 seconds)]
vaibhavsagar_ has joined #nixos
epta has quit [(Remote host closed the connection)]
rmarko is now known as srk
slack1256 has quit [(Remote host closed the connection)]
zennist has quit [(Remote host closed the connection)]
zennist has joined #nixos
m0rphism has quit [(Quit: WeeChat 1.9)]
nh2 has quit [(Ping timeout: 240 seconds)]
zennist has quit [(Ping timeout: 252 seconds)]
eacameron has joined #nixos
bfrog has joined #nixos
eacameron has quit [(Ping timeout: 260 seconds)]
ison111 has joined #nixos
zennist has joined #nixos
nh2 has joined #nixos
zennist has quit [(Ping timeout: 240 seconds)]
rotaerk has joined #nixos
Rotaerk2 has quit [(Quit: Leaving)]
bfrog has quit [(Ping timeout: 246 seconds)]
colescott has joined #nixos
<colescott>
Hey guys, I kinda broke my nix-env and need some help
<colescott>
I get this when I run nix-env: "nix-env: symbol lookup error: /nix/store/6qdj0jlgqan9g54l20jxrm4dx4g2ibw2-curl-7.53.1/lib/libcurl.so.4: undefined symbol: SSL_CTX_set_alpn_protos"
<colescott>
Something happened possibly involving the installation of Enpass i think
<clever>
colescott: nixos or another distro?
vaibhavsagar_ has quit [(Remote host closed the connection)]
<colescott>
nixos
vaibhavsagar_ has joined #nixos
<clever>
colescott: env | grep LD_LIBRARY gives what output?
<clever>
hmmm, and if you "unset LD_LIBRARY_PATH" does that fix nix-env?
<colescott>
yep!!!
<clever>
/home/cole/.nix-profile/lib probably broke it
<colescott>
hmm
<clever>
thats not normal, and it causes the libraries you nix-env to have priority
<clever>
you probably have something odd in your .bashrc or .bash_profile
<colescott>
let me check my zshrc
<colescott>
yeah idk why that's there
<colescott>
thanks!
<colescott>
works perfectly in a new shell :-)
m0rphism has joined #nixos
zennist has joined #nixos
bfrog has joined #nixos
thc202 has quit [(Ping timeout: 240 seconds)]
zennist has quit [(Ping timeout: 260 seconds)]
colescott has quit [(Ping timeout: 260 seconds)]
jsgrant has joined #nixos
bfrog has quit [(Ping timeout: 246 seconds)]
bfrog has joined #nixos
nh2 has quit [(Ping timeout: 252 seconds)]
python476 has quit [(Ping timeout: 248 seconds)]
Supersonic112 has quit [(Disconnected by services)]
Supersonic112_ has joined #nixos
zennist has joined #nixos
Supersonic112_ is now known as Supersonic112
nh2 has joined #nixos
griff_ has quit [(Quit: griff_)]
zennist has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jerith666 opened pull request #29043: amarok: fixup by building with gcc-5 (master...amarok-gcc-5) https://git.io/v52V6
NixOS_GitHub has left #nixos []
mbrgm has quit [(Ping timeout: 248 seconds)]
mbrgm has joined #nixos
Kingsquee has joined #nixos
beemo has quit [(Ping timeout: 260 seconds)]
bfrog has quit [(Ping timeout: 240 seconds)]
bfrog has joined #nixos
jellowj has quit [(Ping timeout: 260 seconds)]
elninja44 has quit [(Remote host closed the connection)]
zennist has joined #nixos
scribble_ has quit [()]
bfrog has quit [(Ping timeout: 252 seconds)]
glenn has joined #nixos
zennist has quit [(Ping timeout: 248 seconds)]
ee1943 has joined #nixos
<ee1943>
So there's been a new version of a package released that fixes critical bugs. It's only a few hours old so Hydra hasn't built it yet. Is there any simple way I can build it locally to test in the meantime?
glenn is now known as scr1bler
<tilpner>
Which one is it?
<tilpner>
Do you want to install or just try it?
scr1bler has left #nixos []
<ee1943>
irony-server
scribbler has joined #nixos
<ee1943>
install I suppose
<tilpner>
The last commit was on Aug 3. I feel like Hydra had enough time, but I don't know how long a full build typically takes
<tilpner>
That version is in nixos-unstable. What channel are you currently on?
<mpcsh>
god dammit I've been without command-not-found for a fucking month
<mpcsh>
thank you :)
<tilpner>
:)
nh2 has quit [(Ping timeout: 264 seconds)]
<tnks>
I could manually put a symlink in gcroots/per-user, but is there a switch for nix-build that will do that for me with less hassle?
<tnks>
I know there's a symlink put into "auto".
<clever>
tnks: nix-build -o foo
<tnks>
I guess in a way that's similar. I just never use gcroots/per-user.
vaibhavsagar_ has quit [(Ping timeout: 252 seconds)]
<tnks>
clever: does the "-o" switch do anything other than change the name of the "result" symlink?
<clever>
thats all it does
<tnks>
I think even without "-o", nix-build creates a symlink in gcroots/auto.
<clever>
--no-outlink stops that i believe
<tnks>
yeah.
<tnks>
just curious what command worked well with gcroots/per-user.
<tnks>
(without me managing it by hand).
<clever>
only nix-env as far as i know
<tnks>
nix-env definitely manages profiles/per-user... but also gcroots/per-user?
<tnks>
maybe gcroots/per-user was a feature implemented, but not really used by much.
<clever>
ah, maybe
nslqqq has joined #nixos
scribbler has joined #nixos
zennist has joined #nixos
nh2 has joined #nixos
<tnks>
clever: I guess it's easy enough to use the path with nix-store --realize --add-root /nix/var/nix/gcroots/per-user/$USER/some_name some_output_link
<tnks>
you have to spell out "/nix/var/nix/gcroots/per-user" but... at least I guess that's in line with the rest of the bumpiness of Nix's CLI UX.
vaibhavsagar_ has joined #nixos
zennist has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] matthewbauer opened pull request #29045: pinentry-mac: no ‘fill plist’ stuff (master...pinentry-no-fill-plist) https://git.io/v52PC
NixOS_GitHub has left #nixos []
scribbler has quit [(Remote host closed the connection)]
scribbler has joined #nixos
scribbler has quit [(Ping timeout: 248 seconds)]
hiratara has quit [(Ping timeout: 246 seconds)]
hiratara has joined #nixos
<clever>
tnks: you will also want to nix-store --query --roots on the storepath, to confirm its detecting the root
oida_ has joined #nixos
scribbler has joined #nixos
oida has quit [(Ping timeout: 248 seconds)]
scribbler has quit [(Remote host closed the connection)]
scribbler has joined #nixos
<tnks>
clever: yup, it works fine.
<tnks>
(I'd be surprised if it didn't).
nh2 has quit [(Ping timeout: 255 seconds)]
bgamari has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
bgamari has joined #nixos
marusich has quit [(Quit: Leaving)]
eacameron has quit [(Ping timeout: 240 seconds)]
pxc has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 1 new commit to master: https://git.io/v5214
<NixOS_GitHub>
nixpkgs/master 28e4975 John Ericson: zile: disable help2man for cross builds...
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 1 new commit to release-17.09: https://git.io/v5210
<NixOS_GitHub>
nixpkgs/release-17.09 601e989 John Ericson: zile: disable help2man for cross builds...
NixOS_GitHub has left #nixos []
fresheyeball has joined #nixos
<fresheyeball>
hey
<fresheyeball>
I'm building a docker image containing a single haskell binary
<fresheyeball>
how the hell is it 2GBs?
<clever>
fresheyeball: run a command like this on the storepath for that haskell binary: du -ch $(nix-store -qR /run/current-system) --max=0 | sort -h
pxc has quit [(Ping timeout: 248 seconds)]
<fresheyeball>
clever: what do you mean?
<clever>
fresheyeball: what is the path to the haskell binary?
scribbler has quit [(Remote host closed the connection)]
<clever>
fresheyeball: line 10, your build directly depends on ghc at runtime, likely dynamic linking, what does "grep -r --color mqcmpgd6i0hggx7mvlvx7gvi6cx124c4 /nix/store/f3zyazp6ddx3bcqmi7j6vm8k0xccl2wc-mysymptoms-web-dashboard-0.1.0.0" say?
<fresheyeball>
clever: it says my binary file matches
<clever>
yeah, thats a problem with the dynamic linking
<fresheyeball>
how can I build without dynamic linking?
<clever>
run haskell.lib.enableStaticLibraries over the derivation i believe
<fresheyeball>
clever: trying
<fresheyeball>
mySymptoms = pkgs.haskell.lib.enableStaticLibraries (haskellPackages.callPackage f {});
<clever>
fresheyeball: ah, i was thinking of haskell.lib.justStaticExecutables
<fresheyeball>
I tried adding -static to my cabal file
<fresheyeball>
under ghc-options
<fresheyeball>
docker image is no2 31MB
<fresheyeball>
is now*
<clever>
sounds good, and you could probably further optimize it, re-do the previous du to see what the fattest package is now
<fresheyeball>
clever: fattest is gcc
<rotaerk>
I just use that so that I never have to run cabal2nix manually, and never have to store the resulting files
<clever>
rotaerk: haskellPackages.callCabal2nix
<fresheyeball>
clever: that was a HUGE help
<fresheyeball>
thank you
<rotaerk>
ah right, forgot about that; there was some reason I used mine instead...
<fresheyeball>
how can I get the file name only for a nix package
<fresheyeball>
clever: now I just need to be able to string interpolate the binary name instead of the full path
<clever>
"${mySymptoms}/bin/foo
<clever>
"
<fresheyeball>
that will get me the path to the binary
<fresheyeball>
but not the binary name
<clever>
ah
<fresheyeball>
I am making a docker image
<clever>
the issue, is that the nix expressions run before the build has even started
<fresheyeball>
and it's a nastly long file name ending in tar.gz
<clever>
so you need to know that name in advance
<clever>
ah, not sure there
<fresheyeball>
is there a way to just get whatever is after the last "/" in the path?
<clever>
use wildcards in the bash that happens in the next derivation
Ivanych has joined #nixos
<rotaerk>
I think my runCabal2Nix.forLocalPath is basically the same as callCabal2Nix, but the forHackagePackages one is different from the one in nixpkgs
<fresheyeball>
clever: I don't understand
scribbler has joined #nixos
<rotaerk>
the one in nixpkgs uses only the snapshot of hackage that it has within than nixpkgs version, but mine uses the latest
<rotaerk>
though that might be a problem for reproducibility
<clever>
rotaerk: stack2nix is also an option, but that generates a lot of stuff, so its not much different from cabal2nix
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to release-17.09: https://git.io/v52MX
<NixOS_GitHub>
nixpkgs/release-17.09 9da9004 Joachim Fasting: tor-browser-bundle-bin: 7.0.4 -> 7.0.5...
<NixOS_GitHub>
nixpkgs/release-17.09 20f4bfd Joachim Fasting: tinycc: 0.9.27pre-20170710 -> 0.9.27pre-20170821...
NixOS_GitHub has left #nixos []
<clever>
its getting late here, i should head to bed
scribbler has quit [(Remote host closed the connection)]
<rotaerk>
night
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to master: https://git.io/v52Mb
<sphalerite>
hefnurg: by the looks of it udisks2 is enabled by default but there aren't any further options for it. It's probably something polkit-related, I'll see what I can find out about it (I'd like to be able to mount stuff through my file manager too)
<hefnurg>
sphalerite, <(I'd like to be able to mount stuff through my file manager too)> good, I don't want to owe you too much :D
<jluttine>
sphalerite: `nix-env -qaP somepackage` still didn't find anything.
<sphalerite>
hefnurg: I haven't made much progress so far, it won't even show my USB stick for me, but I'm still on it
eacameron has joined #nixos
MP2E has quit [(Quit: reboot)]
<hefnurg>
sphalerite, what it auto-mounted for me was secondary disk drive partitions - haven't tried with USB or any external
<sphalerite>
ah right, I don't think I have any non-mounted disk partitions
<hefnurg>
I think it might not have been automounting on Plasma, but it did automount on xfce
<hefnurg>
from what they say Thunar is doing the auto-mounting
<hefnurg>
my settings are just the base xfce settings in the manual for enabling it
MP2E has joined #nixos
<sphalerite>
well no, by the looks of it it isn't mounted at all
nh2 has quit [(Ping timeout: 252 seconds)]
eacameron has quit [(Ping timeout: 246 seconds)]
<hefnurg>
sphalerite, I didn't do anything special and they got mounted for me... maybe because of the NixOS generated hardware file?
<sphalerite>
but they don't seem to be mounted
hotfuzz_ has joined #nixos
<sphalerite>
It says it failed to mount them in the screenshot you sent
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/v52HI
<NixOS_GitHub>
nixpkgs/master fefdc00 Peter Hoeg: arc-theme: Fix build with gnome 3.24
NixOS_GitHub has left #nixos []
<hyper_ch>
nioxs-unstabled moved to 18.03 now?
hotfuzz has quit [(Ping timeout: 264 seconds)]
<hyper_ch>
also what meaning does system.stateVersion = "17.09"; have in the configuration.nix?
nslqqq has quit [(Ping timeout: 248 seconds)]
<sphalerite>
if you change that and rebuild, you won't be able to go back
nslqqq has joined #nixos
<hyper_ch>
can't even boot into an old one?
<sphalerite>
or you can, but it may have made incompatible state changes
<sphalerite>
so stuff is likely to break
<sphalerite>
you can just leave it at 17.09
<hyper_ch>
I thought the whole point was to have atomic upgrades that can be reverted
proteusguy has joined #nixos
<sphalerite>
yes, but that's quite incompatible with state
<vaibhavsagar_>
hi, how do I run tests for a haskell package in nix-shell?
<hyper_ch>
so once I confirmed that 18.03 runs well, then it's ok to alter that?
<sphalerite>
e.g. if you upgrade postgres, it will want to migrate the database to a new format, which can't be reversed
<sphalerite>
and the database data itself isn't nixos's responsibility
<hyper_ch>
postgres... I still am on warpath with it :)
nh2 has joined #nixos
<sphalerite>
so postgres will only be upgraded if the stateVersion has been bumped
<hyper_ch>
good to know
Isorkin has joined #nixos
<hyper_ch>
one should always have backups anyway :)
<sphalerite>
^
<hyper_ch>
I'm starting to use sqlite more and more instead of creating databases in MariaDB
<sphalerite>
sqlite is awesome
jmeredith has quit [(Quit: Connection closed for inactivity)]
stanibanani has quit [(Ping timeout: 240 seconds)]
<hyper_ch>
e.g. I fetch publicated court decision by the Swiss Federal Courts and some cantonal ones.... I used to use MySQL/MariaDB for it, now I just use SQLite :)
<sphalerite>
yeah as long as I don't need shared R/W access I always go with sqlite
<hyper_ch>
nah, it's one per day checking if there's new stuff :)
<hefnurg>
sphalerite, ah yes, they are not mounted, they just show up for mounting in Thunar
MP2E has quit [(Quit: reboot)]
nschoe has quit [(Quit: Program. Terminated.)]
freeman42x[nix] has quit [(Remote host closed the connection)]
<sphalerite>
jluttine: aah right, you're not setting nixos. So you could also have done nix-env -iA nixpkgs.foo
<jluttine>
sphalerite: ok! thanks
pie_ has joined #nixos
nikola_i has joined #nixos
zennist has joined #nixos
zennist has quit [(Ping timeout: 240 seconds)]
civodul has joined #nixos
<sphalerite>
hefnurg: got it working!
m has joined #nixos
m is now known as epta
epta has left #nixos []
eacameron has joined #nixos
zraexy has quit [(Quit: Leaving.)]
<sphalerite>
hefnurg: so it needs gnome3.gvfs to be installed (to provide some DBus services) and thunar to be invoked with the GVFS Gio module added
zennist has joined #nixos
<sphalerite>
hefnurg: try `nix-env -iA nixos.gnome3.gvfs`, log out and back in (so that the dbus session daemon gets restarted and discovers the new services), and then `GIO_EXTRA_MODULES=$(nix-build '<nixpkgs>' -A gnome3.gvfs --no-out-link) thunar`
<sphalerite>
I know next to nothing about the haskell ecosystem though, so I can't help you any further than that — sorry!
<Geraldus>
Ok. `stack unpack yesod-bin & cd yesod-bin-xxx & stack init`. Then manually edited `stack.yaml` appending `nix: enable: true packages: [ zlib ]` do the trick
<Geraldus>
After that `stack install .` succeeded
goibhniu has joined #nixos
<Geraldus>
sphalerite: Look like exactly what I've looked for! Many thanks
<sphalerite>
great
<vaibhavsagar_>
Geraldus, another thing you might like to try is `nix-shell -p haskellPackages.yesod-bin`
<sphalerite>
hefnurg: any luck?
DrWaste has joined #nixos
FRidh has joined #nixos
DrWaste has left #nixos []
nh2 has joined #nixos
silver_hook has joined #nixos
silver_hook has quit [(Changing host)]
silver_hook has joined #nixos
<hefnurg>
sphalerite, is this a command? `GIO_EXTRA_MODULES=$(nix-build '<nixpkgs>' -A gnome3.gvfs --no-out-link) thunar`
Neo-- has joined #nixos
_ris has joined #nixos
<hefnurg>
sphalerite, I'm at work now so I have to work on work :D but thank you, I'll try it
thc202 has joined #nixos
<sphalerite>
yes, it is
hke has quit [(Ping timeout: 246 seconds)]
<sphalerite>
and no problem, just wanted to be sure you'd seen it :)
<hefnurg>
sphalerite, I wasn't actually assuming a reply this late, just happened to remote into home PC a bit and seen your message :) thank you! much obliged
<sphalerite>
np, ping me if you have any more issues
<Geraldus>
vaibhavsagar_: Indeed, `nix-shell -p haskellPackages.yesod-bin` allowed me to run `yesod devel` directly
<Geraldus>
However to be consistent I want to run it with `stack exec yesod devel`
<Geraldus>
Is it possible to select which shell to use as `nix-shell`?
takle has quit [(Remote host closed the connection)]
<vaibhavsagar_>
but you can do `nix-shell --run 'exec zsh; return'`
<vaibhavsagar_>
or similar
<sphalerite>
the thing with changing the shell for nix-shell is that it exposes some shell functions, which won't generally work in shells other than bash
<taktoa>
weird question, but does anyone know the terminology that distinguishes an IP address (1.2.3.4) from an IP-address-with-port (1.2.3.4:8080)?
leat has joined #nixos
takle has quit [(Ping timeout: 248 seconds)]
<sphalerite>
taktoa: I'd call the latter a TCP address (assuming it is TCP)
<taktoa>
hmm, that's not bad
<sphalerite>
not sure if that's any sort of standard terminology but I think it makes what I mean fairly obvious
ertes-w has quit [(Ping timeout: 248 seconds)]
aloiscochard has joined #nixos
ertes-w has joined #nixos
<Geraldus>
Ok, I've simply added `haskellPackages.yesod-bin` and `haskellPackages.stack` to `nix.packages` section in my project's `stack.yaml`.
<Geraldus>
Now I'm able to run `stack exec yesod devel`
<hefnurg>
sphalerite, will do, cheers
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 4 new commits to master: https://git.io/v52bw
<NixOS_GitHub>
nixpkgs/master b2bff7e Peter Simons: LTS Haskell 9.3
<NixOS_GitHub>
nixpkgs/master e1a6791 Peter Simons: haskell-happy: use 1.19.7 version (despite lts-9) to fix test suite failure
<NixOS_GitHub>
nixpkgs/master f2a76b6 Peter Simons: hackage-packages.nix: automatic Haskell package set update...
<sphalerite>
fearlessKim: where did you put the file and is that the full contents of it?
<taktoa>
then in the repl do offlineimap.propagatedBuildInputs
<taktoa>
err
<taktoa>
propagatedNativeBuildInputs
<taktoa>
make sure that keyring is in there
<fearlessKim[m]>
keyring is in there let me double check the wrapper
<fearlessKim[m]>
which offlineimap returns /run/current-system/sw/bin/offlineimap, is that the offlineimap installed via nixos-rebuild ? I have offlinemap installed via /etc/nixos/configuration.nix, and also via nix-env -i offlinemap so that it picks up the overlay for my user (and override the system one).
<taktoa>
yes, that is the global offlineimap
<taktoa>
~/.nix-profile/bin/offlineimap is the one with the overlay
<fearlessKim[m]>
ha ok
<fearlessKim[m]>
this one has the keyring fine perfect ! Thanks taktoa
jellowj has joined #nixos
<fearlessKim[m]>
learning curve is tough but nixos deserves it. Without this channel and the helpful community, would have given up so please keep it up :)
<taktoa>
:-)
nikola_i has quit [(Quit: Leaving)]
ison111 has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 1 new commit to master: https://git.io/v5aeE
<NixOS_GitHub>
nix/master 1277aab Eelco Dolstra: Fix abort when the GC thread gets an exception
NixOS_GitHub has left #nixos []
hke has quit [(Ping timeout: 240 seconds)]
atondwal has quit [(Quit: leaving)]
yee has joined #nixos
eacameron has joined #nixos
nh2 has quit [(Ping timeout: 255 seconds)]
<sphalerite>
Does anyone know why everything graphical in QEMU with EFI seems to run very slowly for me?
kuznero has joined #nixos
<kuznero>
Hi All!
<kuznero>
Is there any way to install additional fonts (I have ttf files) without creating a derivation in NixOS? Or is it still the preferred way?
<goibhniu>
sphalerite: not related to EFI, but I found `-vga vmware` was quite good
<sphalerite>
goibhniu: alright, will have to try that. I've tried -vga virtio which doesn't seem to have helped
<chreekat>
I'm on nixos, stable channel. I'd expect the version of postgresql that I get with 'nix-env -i postgresql' to be the same version I get with 'nix-shell -p postgresql', but it's not: 9.6.3 in the former case, and 9.5.7 in the latter. What's the underlying reason for that?
felixsch__ has joined #nixos
jellowj has quit [(Ping timeout: 246 seconds)]
<sphalerite>
chreekat: as a general rule, don't use nix-env -i without -A
<seequ>
Oh?
<sphalerite>
chreekat: nix-env -i searches through all available packages by name and then somehow chooses one that matches the name (I'm not sure how it chooses)
<sphalerite>
with -A you specify more precisely, it selects a package by its attribute path (which nix-shell does anyway)
<sphalerite>
so if you do nix-env -iA nixos.postgresql you should get the same version as with nix-shell -p postgresql
<chreekat>
"all available packages"... ahh because there's more than one, even in a single channel
<sphalerite>
Yes. You can use nix-env -qaP to show all the available things, then search through it with grep, or you can use nox for a more convenient search interface
hke has joined #nixos
<chreekat>
Does "/etc/nixos/configuration.nix".environment.systemPackages have similar ambiguity?
<sphalerite>
no, that also selects by attribute path
<sphalerite>
nix-env is basically the only thing that gets it wrong :p
<chreekat>
ha
<chreekat>
thx
eacameron has joined #nixos
<sphalerite>
the other disadvantage of searching by name is that it's much slower, but I don't think that's as much of an issue as not quite knowing what you're going to get
vaibhavsagar_ has quit [(Ping timeout: 248 seconds)]
<sphalerite>
chreekat: what led you to use nix-env -i? It's a fairly frequent problem and we should probably track down the documentation that mentions it and fix it
jellowj has joined #nixos
<chreekat>
Most of my current understanding and habits come from reading the nix manual and the nixos manual
<chreekat>
But also examples from a former coworker, so maybe let's blame him
<sphalerite>
haha
<chreekat>
I was planning on rereading the manuals anyway... I'll keep an eye out
<sphalerite>
yeah the nix manual should really not recommend using nix-env -i :(
eacameron has quit [(Ping timeout: 246 seconds)]
ShalokShalom_ has joined #nixos
<chreekat>
I think the problem is more fundamental. If you read the nix-env man page, the impression you get is that -A is Just Another Option
<chreekat>
Plus, people have clearly put work into this 'priority' business, so you might find people who disagree with your recommendation to ignore it and always use attributes...
pvgoran has joined #nixos
<pvgoran>
Hello. Is it a good place for questions about NixOps, or I should rather go to #nixops or something?
ShalokShalom has quit [(Ping timeout: 260 seconds)]
Mateon3 has joined #nixos
<taktoa>
sphalerite: `nix-env -i <regex>` evaluates every derivation in nixpkgs, reads the "name" attribute, and matches it against the provided regex
<taktoa>
oh you meant that you weren't sure how it chooses if there's a conflict
<sphalerite>
taktoa: yeah
<taktoa>
I think it might just install _all_ the derivations matching the regex
<taktoa>
if that's true, then what gets installed with `nix-env -i '.*'`
<taktoa>
the initial object in the priority poset :^)
<sphalerite>
"If there are multiple derivations matching a name in args that have the same name (e.g., gcc-3.3.6 and gcc-4.1.1), then the derivation with the highest priority is used. A derivation can define a priority by declaring the meta.priority attribute. This attribute should be a number, with a higher value denoting a lower priority. The default priority is 0."
<sphalerite>
it installs one package for every name for which there are packages that match
<taktoa>
oh, does it only use priority in the case where only the version number is different?
<sphalerite>
"If there are multiple matching derivations with the same priority, then the derivation with the highest version will be installed."
<pvgoran>
Suppose I have a nixops network defined in two files: mynetwork.nix and mynetwork-physical.nix. Is it possible to define some variable/attribute in mynetwork-physical.nix and then use its value in mynetwork.nix?
<sphalerite>
taktoa: yeah no I do think the name attribute is useful
<michaelpj>
pvgoran: I don't think so - what are you trying to do?
<pvgoran>
michaelpj: I'm trying to define some location in the machine's filesystem, (potentially) dependent on the way how the machine is provisioned, and then use it as a data path for one of the services.
eacameron has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh force-pushed python-unstable from b26ea27 to 454b0c0: https://git.io/v5aIf
<NixOS_GitHub>
nixpkgs/python-unstable 2f3b691 wisut hantanong: pythonPackages.pyshp : move to separate expression...
<NixOS_GitHub>
nixpkgs/python-unstable d8d13de wisut hantanong: pythonPackages.tempora : move to separate expression...
<pvgoran>
michaelpj: It's not. :) This way, the physical specification needs to be aware of the logical specification internals, what services it declares, etc.
eacameron has quit [(Ping timeout: 240 seconds)]
jellowj has quit [(Ping timeout: 246 seconds)]
<michaelpj>
but that is going to be true for anything where you need to establish certain resources to match what the logical spec expects. It seems sane for information to go logical -> physical but not vice versa
chreekat has quit [(Quit: quitting)]
<pvgoran>
I can't agree with that. Rather, the physical part needs to meet certain expectations (performance) and provide certain resources (RAM, storage) that the logical part will then use.
<pvgoran>
Like, the physical part says "this is the fast disk for frequent access, and this is the large disk for big data", and the logical part uses these resources to do its things.
odi has quit [(Ping timeout: 255 seconds)]
<michaelpj>
so you want a way for the logical part to say "the data for X should go on a big disk" and the physical part to say "here is a big disk", or something like that?
Lisanna has quit [(Ping timeout: 240 seconds)]
<michaelpj>
you could define your own module that has options for "big filesystem" "fast filesystem" etc, and use them in the logical part while filling them in in the physical part
JagaJaga has joined #nixos
nh2 has joined #nixos
<sphalerite>
DiThi: it depends on what you want to do really, if you're actually developing the application it's more likely that you want to use nix-shell to build it incrementally rather than getting a clean build every time as you do by overriding the expression
<DiThi>
sphalerite: yes, that's what I want to do, since compile times are very long if not done incrementally
<sphalerite>
DiThi: in that case, use `nix-shell '<nixpkgs>' -A nameOfThePackage` to get a shell suitable for compiling it
<sphalerite>
so with all the dependencies available and stuff
<DiThi>
oh, I see
odi has joined #nixos
frnwtr has joined #nixos
mudri has quit [(Ping timeout: 240 seconds)]
<DiThi>
awesome
<DiThi>
sphalerite: thanks
<DiThi>
once I have it compiled and I've made a package in a git clone of nixpkgs, how do I test it?
jaga1 has joined #nixos
JagaJaga has quit [(Read error: Connection reset by peer)]
<pvgoran>
michaelpj: Yes, this is what I would want to do. I can do it on filesystem level by using bind mounts or something, it's straightforward but I'd rather want a solution on configuration level. (Besides, resources can be other things, like network interfaces or whatever.)
<pvgoran>
How do I define this kind of a module? Any example or documentation? Maybe someone else already implemented a module like this that I could just use? :)
<michaelpj>
well, what options make sense depends on your use case, no? you're trying to abstract over the kinds of deployment machines that you expect to deploy on
<michaelpj>
"fast drive" vs "big drive" might make sense for you, but not for other people
<michaelpj>
I think you can just define another nixos module - try copying one of the existing ones
jbgi has joined #nixos
<michaelpj>
there's probably something you have to do to make it be picked up...
<michaelpj>
sadly, there isn't much documentation on how the module system actually works
<sphalerite>
DiThi: `nix-build path/to/nixpkgs/clone -A packageName` will build it and generate one or more result symlinks containing the build output
eacameron has joined #nixos
<DiThi>
sphalerite: thank you
<DiThi>
so the first argument of both nix-shell and nix-build in the examples above is either '<nixpkgs>' or the local path to the repo?
<symphorien>
yes. if you don't specify anything, it mean "."
<symphorien>
+s
Lisanna has joined #nixos
<sphalerite>
Yes. <nixpkgs> will be resolved against NIX_PATH
<NixOS_GitHub>
nixpkgs/python-unstable d0d861b wisut hantanong: pythonPackages.pyshp : move to separate expression...
NixOS_GitHub has left #nixos []
<michaelpj>
pvgoran: actually, I'm not really sure how you'd do this in the nixops setting. Conceptually it makes sense to me, but I'd have to go source spelunking to figure it out :/
<michaelpj>
bah
<michaelpj>
there might also be an easier way to do this
pxc has quit [(Ping timeout: 248 seconds)]
chpatrick has quit [(Quit: Connection closed for inactivity)]
justbeingglad has joined #nixos
earldouglas has joined #nixos
takle_ has joined #nixos
gnuhurd has joined #nixos
proteusguy has quit [(Remote host closed the connection)]
<pvgoran>
michaelpj: I see. I'm considering doing something like: with import ./mymachine-physical.nix; { <logical part> }
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 6 new commits to master: https://git.io/v5aOg
<NixOS_GitHub>
nixpkgs/master ad44a76 Michael Raskin: pythonPackages.browsermob-proxy: make python2-only
<NixOS_GitHub>
nixpkgs/master cca09d8 Michael Raskin: pythonPackages.manifestparser: make python2-only
<NixOS_GitHub>
nixpkgs/master f20ccd4 Michael Raskin: pythonPackages.wptserve: make python2-only (upstream uses relative import in the latest release)
NixOS_GitHub has left #nixos []
takle has quit [(Read error: No route to host)]
<silver_hook>
For an Epson all-in-one to scan via the net, do I have to do anything else than have services.printing.drivers [ pkgs.epson-escpr ] and hardware.sane.netConf set up?
takle has joined #nixos
<pvgoran>
Not the most elegant way, because if I do this, I hard-code the physical file name na lose the ability to use the same logical file with a different physical file.
<michaelpj>
right, that seems to not really get you anything over just hardcoding the value
<michaelpj>
then at least you could set the "data goes here" argument for a given deployment depending on what physical config you instantiated it with
<pvgoran>
Well, the import approach still lets me modify and replace mynetwork-physical.nix as I see fit; I lose the freedom to use a different name for the physical part, but not the freedom to use different contents.
ShalokShalom_ is now known as ShalokShalom
takle_ has joined #nixos
<pvgoran>
Network arguments are detached from the physical part, this approach looks too easy to get wrong.
takle_ has quit [(Read error: No route to host)]
takle_ has joined #nixos
jellowj has joined #nixos
takle__ has joined #nixos
michiel_l has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
takle_ has quit [(Ping timeout: 240 seconds)]
<michaelpj>
well, when you create a deployment you pass the physical part as an argument - but your logical part will still import "mynetwork-physical.nix" whatever you pass at creation time
nh2 has quit [(Ping timeout: 255 seconds)]
jaga1 has quit [(Ping timeout: 260 seconds)]
<pvgoran>
I can just pass the logical part, and let the physical part be imported. :)
erictapen has quit [(Ping timeout: 240 seconds)]
<michaelpj>
huh, does that work?
<pvgoran>
Why wouldn't it? I think I can pass any number of file arguments to nixops create, not necessarily two.
<michaelpj>
apparently so. I have no idea what that will do :D
cpennington has quit [(Remote host closed the connection)]
<sphalerite>
it merges all the sets together AFAIU
<pvgoran>
That's what I think too.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/v5aGP
<NixOS_GitHub>
nixpkgs/master 67651d8 Frederik Rietdijk: Merge pull request #28884 from FRidh/python-fixes...
NixOS_GitHub has left #nixos []
<michaelpj>
cute
<michaelpj>
would be nice if it said that somewhere ;)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to staging: https://git.io/v5aG9
<NixOS_GitHub>
nixpkgs/staging 38dc9a8 Frederik Rietdijk: Merge remote-tracking branch 'upstream/master' into HEAD
NixOS_GitHub has left #nixos []
a6a3uh has joined #nixos
takle__ has quit [(Remote host closed the connection)]
civodul has quit [(Read error: Connection reset by peer)]
oahong has quit [(Ping timeout: 248 seconds)]
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
eacameron has quit [(Ping timeout: 248 seconds)]
hypersw_ is now known as hypersw
oahong has joined #nixos
nikola has joined #nixos
python476 has joined #nixos
peacememories has joined #nixos
nh2 has quit [(Ping timeout: 264 seconds)]
nikola has quit [(Client Quit)]
nikola has joined #nixos
nikola has quit [(Client Quit)]
Arcaelyx has joined #nixos
nikola_i has joined #nixos
ona has quit [(Quit: ...)]
oida has joined #nixos
FRidh has joined #nixos
oida_ has quit [(Ping timeout: 240 seconds)]
thblt has quit [(Remote host closed the connection)]
<pvgoran>
michaelpj: Looks like I can access the other parts of machine configuration using the "config" parameter of the machine configuration function.
cpennington has joined #nixos
<pvgoran>
So I do { machine = { pkgs, ...}: { passthru.mypath = "/path"; }; } in network-physical.nix, and { machine = { pkgs, config, ... }: { service.dataPath = config.passthru.mypath; }; } in network.nix.
<pvgoran>
("services.someservice.dataPath" instead of "service.dataPath" would be a better example.)
nh2 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 3 new commits to staging: https://git.io/v5a4D
<NixOS_GitHub>
[nixpkgs] FRidh closed pull request #28916: secretstorage init at version 2.3.1 (master...secretstorage) https://git.io/v502s
NixOS_GitHub has left #nixos []
ThatDocsLady has quit [(Ping timeout: 264 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh closed pull request #28861: A Collection of new python packages (master...pr-betaboon-pythonpackages) https://git.io/v5B5m
NixOS_GitHub has left #nixos []
<Unode>
any suggestion on how to address "output path ‘/share/nix/store/1xn822y87l9l3ymv2zjrwksglpfmmyg2-named.root’ has sha256 hash ‘01n4bqf95kbvig1hahqzmmdkpn4v7mzfc1p944gq922i5j3fjr92’ when ‘0qsyxpj5b3i7n162qfyv76ljqbvnwjii7jk8mpfinklx0sk01473’ was expected" ?
<symphorien>
the sha256 in a fetch* function is wrong
<Unode>
sphalerite: it's part of the dependency tree. Not a direct dependency.
freeman42x[w] has quit [(Read error: Connection reset by peer)]
nh2 has joined #nixos
<sphalerite>
Unode: I just built R and I don't have any copies of named.root in my store
<dash>
Getting a very odd issue with pypy packages: "nix-shell '<nixpkgs>' -A pypyPackages.flake8" gives the error "configparser-3.5.0 not supported for interpreter pypy"
<Unode>
the failing tree now includes, openjdk -> cups -> gnutls -> dns-root-data -> named.root
<Unode>
sphalerite: in case you didn't notice, not running on /nix but /share/nix. So everything is from sources.
<sphalerite>
oooh right
<sphalerite>
but yeah that must have been updated or you're getting MITMed very nastily :p
<Unode>
that would be interesting :)
<Unode>
sphalerite: any suggestions on which file I need to poke to workaround this?
<shlevy>
Hi all (esp. grahamc ): Trying to install nix on darwin, the store path theinstaller downloaded only has nix-shell in the bin/ dir and so the installer fails :p
<NixOS_GitHub>
nixpkgs/master 9405bc2 Joachim F: Merge pull request #28965 from jlesquembre/mcomix...
NixOS_GitHub has left #nixos []
<rolf1>
Is there anyone here who uses the xfce4-power-manager? Whenever I try to change something in xfce4-power-manager-settings I get a critical error saying the program cannot set the value for the property.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to python-unstable: https://git.io/v5ayk
<NixOS_GitHub>
nixpkgs/python-unstable 2f09b4f Frederik Rietdijk: mycli: move out of python-packages.nix and fix build
NixOS_GitHub has left #nixos []
<Denommus>
hm, I need xvfb (not xvfb_run) to be able to use the Python package xvfbwrapper
<Denommus>
(just venting, not actually a doubt)
ison111 has joined #nixos
tmaekawa has joined #nixos
<hyper_ch>
anyone using KDE here?
<sphalerite>
hyper_ch: how many times have you asked that question now? :p
<hyper_ch>
sphalerite: not enough
<hyper_ch>
I just noticed the systemsettings entry is gone from kicker's computer submenu, still can open it manually
<hyper_ch>
and when I do and go to audio, no devices are listed
acowley_away is now known as acowley
<goibhniu1>
hyper_ch: FWIW, I don't have those problems here (on unstable) ... did you try with a fresh user account?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 2 new commits to python-unstable: https://git.io/v5a9T
<NixOS_GitHub>
nixpkgs/python-unstable 4fe9e3f Frederik Rietdijk: python.pkgs.jsonnet: disable for Python 3
<NixOS_GitHub>
nixpkgs/python-unstable 6a365aa Frederik Rietdijk: python.pkgs.koji: fix build
NixOS_GitHub has left #nixos []
goibhniu1 is now known as goibhniu
mrkgnao has joined #nixos
<maurer>
Just to check before I go copy-pasting code - we still can't use nixos modules to create user-systemd services as a user, right?
<clever>
only nixos modules can be used to create systemd-user services
<maurer>
Context: I'm trying to develop a postgresql extension. PostgreSQL requires extensions to live in $SHAREDIR set at configure time. This is obviously not a writable location
<clever>
so the user services are still global, and available to every user
<maurer>
The way PostgreSQL modules are instantiated in normal usage is via a variable on the nixos module to specify additional plugins
rolf1 has quit [(Ping timeout: 240 seconds)]
<maurer>
I'm basically wondering whether, in my dev environment, there's a reasonable way for me to say "bring up another copy of postgres, with this fresh extension added"
<sphalerite>
hyper_ch: if nobody knows anything, you won't get a response to "Does anyone use KDE?" nor to your description of the problem. If anybody does know anything, you'll get a response more rapidly by going into specifics. If somebody does use KDE but doesn't know anything about your problem, you're wasting their time in the former case
<sphalerite>
The description of the problem is a far superior thing to put :)
<clever>
maurer: i think this line is key, it copies the main binary into the buildenv
<clever>
so when the binary tries to open something relative to itself, it goes into the new share dir
<maurer>
Yeah, I suppose it's possible that SHAREDIR is relative
<maurer>
and LIBDIR wasn't, so they had to override it
leat has quit [(Quit: leat)]
rolf1 has quit [(Ping timeout: 240 seconds)]
<hyper_ch>
sphalerite: not really... since this seems to be kde specific, I first have to find someone who also uses kde and then check if it works for them or not... only afterwards it's of interest what the actuall problem is..... how would someone using xfce, lxde, gnome know about kde issues?
odi has quit [(Ping timeout: 248 seconds)]
<dash>
It seems to me that there's no good way to override 'disabled' settings on Python derivations since an error gets thrown very early in the evaluation of mk-python-derivation.nix -- anybody have experience otherwise?
dieggsy has joined #nixos
<dash>
I'm trying to turn a package back on for pypy that was disabled in the past
<chreekat>
hyper_ch: It is standard IRC etiquette to just ask one's question directly, and as specifically as possible.
justbeingglad has left #nixos []
<hyper_ch>
chreekat: it would make sense in #kde where everyone uses kde... but the number of users using kde in #nixos would be only a fraction
<goibhniu>
hyper_ch: but people don't know if they can help you until you actually ask your question
wes has joined #nixos
wes is now known as Guest8969
Guest8969 has left #nixos []
bsoudan has joined #nixos
<bsoudan>
hey all... trying to install nix but i need to delete a mirror from fetchurl/mirrors.nix since easyname.at seems to be blocked at work. but, it seems like it's changing all the package hashes and causing a world rebuild
<bsoudan>
any ideas on how to eliminate one and still be able to use the binary cache?
julm has joined #nixos
ertes-w has quit [(Ping timeout: 240 seconds)]
simukis has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to master: https://git.io/v5abg
<NixOS_GitHub>
nixpkgs/master ca54a86 makefu: dnscrypt-wrapper module: fix permissions and options...
<NixOS_GitHub>
nixpkgs/master 9bb4005 Joachim F: Merge pull request #29050 from makefu/module/dnscrypt-wrapper/fix...
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 5 new commits to python-unstable: https://git.io/v5aNr
<NixOS_GitHub>
nixpkgs/python-unstable 7f7776a Frederik Rietdijk: python.pkgs.astropy: fix build
<NixOS_GitHub>
nixpkgs/python-unstable f7a9e6c Frederik Rietdijk: python.pkgs.pylint: use fetchPypi
<NixOS_GitHub>
nixpkgs/python-unstable aa65095 Frederik Rietdijk: conan: 0.21.2 -> 0.26.2, and fix build
NixOS_GitHub has left #nixos []
julm has left #nixos []
<bsoudan>
or maybe some way where i don't change it, but when it fails i manually retrieve the package from a different mirror, and then resume nixos-install? manually installing the file using curl doesn't seem to work, nixos-install tries to redownload it
remexre has left #nixos ["Leaving"]
<clever>
bsoudan: run "nix-store -r /nix/store/foo" to download the path from a binary cache
Denommus has quit [(Ping timeout: 252 seconds)]
<dieggsy>
...if rev and sha are excluded from fetchgit, does it just take HEAD from master?
<clever>
dieggsy: it needs a sha256, it will throw an error without it
<dieggsy>
clever: I just commented out the sha256 and it didn't seem to
<dieggsy>
clever: well it's building
<bsoudan>
clever: it looks like i already have the binary (*-acl-2.2.52) but nixos-install is trying to download the source for some reason (*-acl-2.2.52.src.tar.gz). maybe nix-prefetch-url? trying that now
<clever>
dieggsy: ah, looks like it doesnt throw an assert, but the sandbox will break that
<dieggsy>
clever: ?
<clever>
bsoudan: you can also use -r on the storepath of the source
<bsoudan>
ok, i'll try that now
<clever>
dieggsy: when the nix sandbox is enabled, it will block all network access
<dieggsy>
clever: and then why is the default to the rev argument HEAD? isn't the sha256 for HEAD not a constant?
<clever>
nix expects the expression to give the same source code every time
<clever>
doing otherwise will cause problems down the road
<dieggsy>
clever: can you describe "down the road" - and what's the nix sandbox
<dieggsy>
My point is, if it's problematic, it seems like what i just did (no rev, no sha, and it's building) should just not be allowed
<clever>
dieggsy: nix has no idea when it should rebuild the package because the nix expression hasnt changed
<clever>
dieggsy: if the sandbox was enabled, it wouldnt allow this
<clever>
dieggsy: what OS are you on?
<dieggsy>
clever: Kubuntu
<bsoudan>
nix-store -r on the source fails too. but it's also having trouble downloading nix-cache-info i just realized. maybe that's the root of my problems
<clever>
bsoudan: yeah, sounds like the binary cache is blocked
<clever>
dieggsy: ah, no simple way to turn the sandbox on in that
<dieggsy>
clever: so you're saying this is a bug, more or less
<bsoudan>
we have an irritating https proxy here. self signed certificate which breaks lots of things. lemme hack on that for a bit, i fixed this before on regular nix
<dieggsy>
clever: my question still stands, why would they make the rev arg not mandatory, in that case?
<dieggsy>
I suppose you could have a repo that just doesn't change, but seems like a bit of a gamble
<clever>
dieggsy: not sure why its not setup like that
<mog>
i have a channel called unstable on my nixos so i can install a few packages from unstable easily. how do i tell a service to use those packages or how do i use the unstable version of the service
mrkgnao has quit [(Quit: WeeChat 1.7.1)]
<bsoudan>
niksnut: yeah, i was mistaken, it's our work https proxy that's causing all the pain
<bsoudan>
clever: making progress.. i can download files with nix-store if i specify --option verify-https-binary-caches false, but the same option doesn't seem to help nixos-install
__Sander__ has quit [(Quit: Konversation terminated!)]
<bsoudan>
nor does setting it in various combinations of ~/.config/nixpkgs/config.nix, ~/.nixpkgs/config.nix, or /mnt/etc/nixos/configuration.nix
<bsoudan>
so i'm hacking around trying to discover why it's not honoring my setting during nixos-install. i don't want to manually download every package using nix-store :)
<tilpner>
mog - Sometimes the service has an option services.name.package. Does yours?
<clever>
bsoudan: one minute
<bsoudan>
i also tried nixos-install --option nixpkgs.verify-https-binary-caches false as well. currently i'm trying to find the place where it's actually invoking curl but not having any luck yet
<aminechikhaoui>
bsoudan: if you do nix-env -v --version shouldn't that give the conf file to use for the nix daemon ?
<clever>
bsoudan: this is the general area where nixos-install builds nix
<viaken>
bsoudan: Don't know if it would help, but to deal with my work's proxy, I'm running squid as a user service with work set as upstream. When I leave the network, I remove the cache_peer and reload.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 4 new commits to python-unstable: https://git.io/v5VJq
<NixOS_GitHub>
nixpkgs/python-unstable b43c416 Frederik Rietdijk: python.pkgs.ftfy: disable Python 2
<NixOS_GitHub>
nixpkgs/python-unstable 03b0796 Frederik Rietdijk: python.pkgs.pybfd: fix build
<NixOS_GitHub>
nixpkgs/python-unstable 9ea934b Frederik Rietdijk: python.pkgs.yamllint: fix build
NixOS_GitHub has left #nixos []
<mog>
tilpner, its firefox syncserver
<mog>
ill check
<bsoudan>
viaken: i have thought about setting something that up too... do they have an https proxy too though? that's the problem here, it's injecting a self-signed certificate which messes up tons of utilities
<bsoudan>
to be clear i've been using plain nix here after solving all the issues, but i'd really like to actualy get nixos itself installed now. it's just that nixos-install seems to not be honoring the verify-https-binary-caches option. i need to disable it
<ixxie>
mog: essentially, you can give the channel an alias in the config, and the prefix the packages you wanna upgrade like unstable.packagename
goibhniu has quit [(Ping timeout: 248 seconds)]
<mog>
right ixxie i do the same thing for packages, but i want to do this for a service
cfricke has quit [(Quit: WeeChat 1.9)]
<tilpner>
mog - You could try to globally override pythonPackages.syncserver, probably by globally overriding pythonPackages
<tilpner>
Or add the appropriate "package" option to the sync-server module and PR that :P
cricriiiiii has joined #nixos
<mog>
i put in a bug for the stable version
<mog>
i see where i can just override the env
<mog>
i will just do that i think
<ixxie>
hmm not sure mog, sorry I misunderstood the question
<bsoudan>
ah. you can't override the NIX_PATH during an install. that's disappointing
<viaken>
bsoudan: They don't MITM me, no.
<bsoudan>
viaken: it's really irritating.
<bsoudan>
viaken: i hope the amount of money they've paid me to deal with all the issues surrounding it is worth the extra "security"
georges-duperon has quit [(Ping timeout: 252 seconds)]
ambro718 has joined #nixos
<mpcsh>
hey - how can I run nix-shell with a package name instead of an attribute name?
<clever>
nix-env is the only tool that accepts package names, every other tool in the system uses attribute names
<viaken>
What's the difference?
<mpcsh>
clever: maybe I have it backwards then. how can I run, say, `nix-shell -p python36Packages.fontforge`?
<clever>
package names are just the value of the .name attribute on a derivation
<clever>
mpcsh: that is an attribute path, it should just work like that
<clever>
mpcsh: and when i run what you pasted, it does just work
<mpcsh>
clever: `error: undefined variable ‘python36packages’ at (string):1:66`
<clever>
does python 3.6 exist in your copy of nixpkgs?
<clever>
mpcsh: what if you run nix-repl '<nixpkgs>' and then try to tab-complete pythonPackages ?
<Denommus>
I want to create a package that doesn't include all of the files in the current directory in it, how do I do that?
<michaelpj>
I would give so much to have nix-repl's tab completion when using nix commands
<clever>
Denommus: builtins.filterSource
nckx has joined #nixos
<mpcsh>
clever: it works
<bsoudan>
looks like i had to set NIX_CONF_DIR manually for it to read my custom config. i didn't realize it the install environment already had explicit NIX_CONF_DIR
<clever>
mpcsh: what is the output of nix-channel --list and sudo nix-channel --list?
<bsoudan>
not sure why --option verify-https-binary-caches doesn't work for nixos-install (which invokes nix-env) but it does for nix-store though
<bsoudan>
i had to remove a few mirrors probably blocked by our proxy (easyname, gnu ftpmirror) but now it seems to be progressing, thanks for the help all
<clever>
2017-09-06 13:57:41 < mpcsh> clever: maybe I have it backwards then. how can I run, say, `nix-shell -p python36Packages.fontforge`?
<clever>
mpcsh: packages vs Packages
<mpcsh>
oh man
<mpcsh>
facepalm
<mpcsh>
uuuuuuuuugh lmao thank you clever
<Denommus>
clever: thanks
<clever>
your welcome * 2
pxc has quit [(Client Quit)]
<Denommus>
clever: is there a function to convert a dictionary of derivations into a list of derivations? That is, instead of I having to do `buildInputs = [ something.pkgs.one something.pkgs.two ]`, I do something like `buildInputs = someFunction something.pkgs`?
<pop>
How do I delete old generations / cleanup my root partition?
<clever>
Denommus: thats not a directory, thats an attribute set, and you probably want builtins.attrValues
<clever>
pkgs/build-support/fetchurl/boot.nix:let mirrors = import ./mirrors.nix; in
<clever>
that is only loaded in 2 places
mudri has joined #nixos
<clever>
bsoudan: oh right, there is also a weird glitch in nixos-install
<clever>
bsoudan: if it cant contact cache.nixos.org because of network problems, it thinks there is zero binary cache coverage
<clever>
bsoudan: so it tries to rebuild everything from scratch
rolf1 has quit [(Ping timeout: 240 seconds)]
<bsoudan>
hmm
mkoenig has joined #nixos
<bsoudan>
it appears to be trying to rebuild the world again. at least it's rebuilding openssl which is where it started last time
<clever>
it computed things as being identical to the old config and reused them
<clever>
it might be the cache cache
<clever>
bsoudan: try deleting /nix/var/nix/binary-cache-v3.sqlite and its journal, on both / and /mnt
<bsoudan>
how can i tell if that's getting triggered? i don't see any errors with cache.nixos.org or whatever the correct address is like i did before now that i've figured out how to properly set the nix configuration
<bsoudan>
ok trying
<bsoudan>
no luck. i'm going to completely restart, reformat, and retry from scratch
Filystyn has joined #nixos
Filystyn has quit [(Changing host)]
Filystyn has joined #nixos
<catern>
hey #nixos
<catern>
is there a way to create all the low-level derivations for a given derivations closure?
hotfuzz_ has joined #nixos
takle has quit [(Remote host closed the connection)]
<catern>
the low-level store derivations corresponding to the high-level derivations in the closure of a given high-level Nix derivation, I mean
<catern>
like a command line thing or something
<catern>
ah
<catern>
nix-instantiate
hotfuzz has quit [(Ping timeout: 260 seconds)]
jasom has quit [(Quit: WeeChat 1.7.1)]
iyzsong has quit [(Ping timeout: 246 seconds)]
ertes has joined #nixos
ThatDocsLady has joined #nixos
ThatDocsLady has quit [(Client Quit)]
ThatDocsLady has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti closed pull request #28918: haskell/natural-transformation: add dependency on semigroups for GHC 7.10 (master...natural-transformations) https://git.io/v50rM
NixOS_GitHub has left #nixos []
ThatDocsLady has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] rycee pushed 1 new commit to release-17.09: https://git.io/v5Vmb
<NixOS_GitHub>
nixpkgs/release-17.09 58ed8fa Markus Hauck: rawtherapee: 5.0-r1 -> 5.2...
NixOS_GitHub has left #nixos []
ThatDocsLady has joined #nixos
nh2 has joined #nixos
<bsoudan>
i'm going to give up for now. after i restarted and retraced all my steps, now it's having trouble downloading from the binary cache again even though i've set verify-https-binary-caches in 4 different places
<bsoudan>
i'm not sure if it had just stopped trying before or not
<bsoudan>
nix-store works fine. nixos-install just doesn't seem to want to honor the verify-https-binary-caches flag. any troubleshooting suggestions would be appreciated, i'll probably pick it up later today or tomorrow
<bsoudan>
option definitely does not work for nixos-install, but it does for nix-store. plain curl works fine if i pass the '--insecure' flag, which is what that option does if you look at download-from-binary-cache.pl, but the perl version
<clever>
what error does curl give?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 1 new commit to master: https://git.io/v5VYR
<NixOS_GitHub>
nixpkgs/master 57ef1a3 Daiderd Jordan: go: fix darwin build
NixOS_GitHub has left #nixos []
<bsoudan>
can't copy paste but basically 'self-signed certificate in certificate chain. blah blah if you'd like to turn off curl's verification, use the -k or --insecure option'
<clever>
ah
<clever>
so its not a real proxy, but a transparent proxy
<bsoudan>
right
<bsoudan>
plain nix, everything works fine for me once i set that option. i've been using it for a few months here on ubuntu
<bsoudan>
but i'd like to get nixos installed finally. in any case gotta run for a bit, i'll be back to try more. i'll give /mnt/etc/nix/nix.conf a shot too
<clever>
and then upload the logfile to a pastebin
hamishmack_ has joined #nixos
hamishmack has quit [(Ping timeout: 240 seconds)]
hamishmack_ is now known as hamishmack
* viaken
needs to figure out why CIFS won't load.
acowley_away is now known as acowley
<mpcsh>
is anyone having their nixos-rebuilds fail somewhere in an autoconf with `configure: error: Invalid GNOME version: 3.24`?
<mpcsh>
*anyone else
<clever>
mpcsh: what derivation gave that error?
simukis has joined #nixos
odi has joined #nixos
<catern>
is there a pretty-printer for low-level derivations somewhere?
<catern>
i.e. an aterm pretty-printer?
hotfuzz has joined #nixos
hotfuzz_ has quit [(Ping timeout: 240 seconds)]
<mpcsh>
clever: I can't tell :/
nh2 has quit [(Quit: Leaving.)]
<clever>
mpcsh: it should say what failed, right after that error
<mpcsh>
clever: ah, it's arc-theme
suvash_away has joined #nixos
<Denommus>
clever: can I get the relative name of a file somehow?
<clever>
Denommus: got an example of what your doing?
<Denommus>
that is, a function f where (f path ./.) gives me the name of path relative to ./. or something
freeman42x[nix] has joined #nixos
rolf1 has joined #nixos
<clever>
Denommus: if you do ./. in nix, it will automatically become absolute
<ee1943>
the irony-server package inherits it's version from the autogenerated emacs irony package. when a new version is released, how long does it typically take for hydra to package it on unstable?
<Denommus>
clever: I understand that. The thing is that I need to filter off some of my directories called "assets", but not EVERY directory called asset
<Denommus>
Ankhers: yes, that's exactly what I was looking for, thanks
<Ankhers>
No problem :)
<Ankhers>
If I am developing a library for $LANGUAGE, would it be possible to run the test suite against a matrix of langauge and database versions using nix?
jensens has quit [(Ping timeout: 264 seconds)]
<Denommus>
Ankhers: maybe you want something like Hydra?
<Ankhers>
I was hoping to not have something so heavy handed.
<LnL>
sure
<LnL>
you can callPackage the expression with a bunch of different versions of the dependencies
<Denommus>
ah, yes. Most .nix files are functions, you can just parameterize whatever you want
<LnL>
and if you put all the variants in an attrset you can even build them all with a single nix-build command
<Ankhers>
I will have to give that a go once we have finished getting the BEAM fixed up.
<Ankhers>
Thanks for that. I will need to look into that for sure.
jensens has joined #nixos
gnuhurd has quit [(Remote host closed the connection)]
ison111 has quit [(Ping timeout: 240 seconds)]
<catern>
hey #nixos
<katyucha>
hi
<sphalerite>
grahamc: finally got the nvidia working \*_\*
<katyucha>
ok, so finally, problem on consul services ... hard to find
<catern>
is there any kind of protection in nix/nixpkgs against, passing two different versions of the same library into the same derivation where they are linked into the same process?
python47` has quit [(Remote host closed the connection)]
<pshendry>
Noob packaging question here: I'm using fetchurl on a tarball that contains a single directory; how can I extract the directory contents instead of the archive root? I figure this must happen all the time but I haven't stumbled upon an example of it.
python476 has joined #nixos
<pshendry>
Ah, I think I'm onto an answer, looking at the unpackPhase in nixpkgs docs
<fresheyeball>
I tried adding the version bounds for servant and servant-server to my cabal file
<fresheyeball>
and re-running cabal2nix
<fresheyeball>
I updated my cabal2nix
<fresheyeball>
I did a channel update
<fresheyeball>
a cabal clean
<fresheyeball>
but so far it always fails with the same error
<fresheyeball>
which is very annoying
<rycee>
Anybody know how to reduce the verbosity to "errors only" in `nix-build`? It defaults to "informational" and adding `-v`, `-vv`, etc. increases verbosity. But I have no luck using, e.g. `-v0` to reduce verbosity…
<rycee>
Is it the `--no-build-output, -Q` that corresponds to "errors only"?
FRidh has quit [(Quit: Konversation terminated!)]
<LnL>
there's only -Q AFAIK
pxc has joined #nixos
<rycee>
LnL: Ok, I'll use that. Thanks.
revtintin has quit [(Ping timeout: 252 seconds)]
<LnL>
rycee: you could use nix-store --read-log foo.drv afterwards if you know what failed
<fresheyeball>
I downgraded the compiler to 8.0.1, not what I would have liked
<rycee>
LnL: Cool! Didn't know about that. Thanks :-)
<fresheyeball>
but it configures now
<LnL>
with nix 1.12 you can even get the hydra logs from the cache :D
<nliadm>
is there a timeline for the new nix commands?
<nliadm>
like, are they planned to land for 18.03?
<sphalerite>
nliadm: you can get them now with nix-env -iA nixos.nixUnstable
<nliadm>
yeah, I just... don't want the unstable ones ;)
<sphalerite>
as for them becoming standard, I don't think anyone knows yet
<sphalerite>
At least I've asked the same question a few times and not got a satisfying answer either :p
<nliadm>
ah, okay
Synthetica has joined #nixos
goibhniu has joined #nixos
<viaken>
I'm getting "mount error(6): No such device or address" trying to mount a CIFS share. This exact command works fine in my previous distro. Is there a daemon I need running or something?
civodul` has joined #nixos
<Denommus>
... ok, I need to come up with a way of making my settings.py pure somehow
civodul has quit [(Ping timeout: 255 seconds)]
civodul` is now known as civodul
<rycee>
Hmm, and is there any way to suppress the "these derivations will be built …" and "building path(s) …" stuff that nix-build prints?
<Denommus>
rycee: do you mean completely stopping the output from nix-build, or just specific parts?
<rycee>
Basically I want nix-build to be completely silent except if an evaluation failed, i.e., it couldn't produce an output path.
<Synthetica>
Okay, so I tried installing NixOS in a vm, however, /etc/nixos/configuration.nix seems to be ignored, I tried replacing the lines in the default config with these:
<Synthetica>
However, in the default display manager (which is still the one from KDE), xfce doesn't show up. And no new entries are added in the GRUB menu. Yes, I did a nixos-rebuild switch
<Synthetica>
What am I doing wrong?
<rycee>
I'm thinking that I may have to redirect stderr to a file and cat that to the terminal if nix-build exits with an error.
<LnL>
your best bet is probably to call nix-instantiate and nix-store -r separatly
<LnL>
or something like that
<Denommus>
rycee: you just made me notice that nix-build always output to stderr. I didn't know that :-/
<LnL>
it only sends the output path to stdout so you can use it with $() etc.
the-kenny has quit [(Ping timeout: 240 seconds)]
ToxicFrog has quit [(Ping timeout: 240 seconds)]
marusich has joined #nixos
the-kenny has joined #nixos
ToxicFrog has joined #nixos
<fresheyeball>
erp nope
<fresheyeball>
ghc801 did not help
<fresheyeball>
I'm out of ideas
python476 has quit [(Remote host closed the connection)]
python476 has joined #nixos
python476 has quit [(Remote host closed the connection)]
Arcaelyx has quit [(Ping timeout: 246 seconds)]
<LnL>
are you sure the versions of servant, etc. match what it's looking for?
Arcaelyx has joined #nixos
georges-duperon has quit [(Ping timeout: 240 seconds)]
Arcaelyx_ has joined #nixos
stanibanani1 has quit [(Ping timeout: 248 seconds)]
Arcaelyx has quit [(Ping timeout: 260 seconds)]
ambro718 has quit [(Ping timeout: 248 seconds)]
thblt has quit [(Remote host closed the connection)]
erictapen has quit [(Ping timeout: 260 seconds)]
<rycee>
Ah, found that nix-build recognizes `--quiet` but it isn't in the man page :-)
Wizek has quit [(Quit: Connection closed for inactivity)]
katyucha has quit [(Ping timeout: 240 seconds)]
Itkovian_ has joined #nixos
Itkovian has quit [(Ping timeout: 260 seconds)]
Itkovian_ has quit [(Client Quit)]
testuser has joined #nixos
hiratara has quit [(Ping timeout: 240 seconds)]
<testuser>
hi, I've added a new NVM disk into my laptop, which is the third disk inside the laptop and is LUKS-encrypted. The first two are labeled /dev/sda and /dev/sdb, while the new disk is labeled as /dev/nvme0. However, the disk is not detected at boot, since adding: "{ name = "nvmssd"; device = "/dev/nvme0n1"; preLVM = true; allowDiscards = true; }" to boot.initrd.luks.devices provides the following error during boot: "stage-1-init: Device
<testuser>
/dev/nvme0n1 doesn't exist or access denied.". Therefore, the /dev/nvme0n1 is not available at the time the LUKS tries to decrypt the disk, is there any reason why this is so and what to do to make the disk available at that boot stage?
zennist` has quit [(Remote host closed the connection)]
stanibanani has quit [(Ping timeout: 248 seconds)]
<testuser>
clever, thanks, let me try this.
<testuser>
clever, btw: what's the difference between boot.initrd.availableKernelModules and boot.initrd.kernelModules ?
the-kenny has quit [(Quit: WeeChat 1.7.1)]
<clever>
availableKernelModules are present in the initrd, but not forcibly loaded, the udev layer will just auto-load things as needed
<clever>
kernelModules makes it present, and forcibly loads it
the-kenny has joined #nixos
hiratara has joined #nixos
<testuser>
clever, thanks, let me reboot to see the change.
pxc has quit [(Ping timeout: 240 seconds)]
testuser has quit [(Quit: Leaving)]
bfrog has quit [(Ping timeout: 252 seconds)]
jsgrant has joined #nixos
jammerful has joined #nixos
bfrog has joined #nixos
dbmikus has quit [(Quit: WeeChat 1.9)]
goibhniu has quit [(Ping timeout: 252 seconds)]
jammerful has quit [(Ping timeout: 260 seconds)]
_ris has joined #nixos
jellowj has quit [(Ping timeout: 255 seconds)]
slyfox has quit [(Quit: no need to cry)]
ona has quit [(Ping timeout: 248 seconds)]
ona has joined #nixos
<bsoudan>
clever: ok, found the problem. download-from-binary-cache.pl has a system() callout to curl that does not set --insecure if the verify-https-binary-caches flag is set to false
<bsoudan>
hacked up a local version, fixed it, and install is progressing normally now finally. i'll submit a patch
<bsoudan>
all done already. that was much better :)
<bsoudan>
thanks again for the help today
pxc has joined #nixos
erictapen has quit [(Ping timeout: 240 seconds)]
jellowj has joined #nixos
Sonarpulse has quit [(Ping timeout: 260 seconds)]
civodul has quit [(Quit: ERC (IRC client for Emacs 25.2.1))]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
<sphalerite>
Synthetica: what do `readlink /run/booted-system` and `nix-env -p /nix/var/nix/profiles/system --list-generations` say?
slyfox has joined #nixos
slyfox has quit [(Client Quit)]
slyfox has joined #nixos
pxc has joined #nixos
griff_ has quit [(Quit: griff_)]
<sphalerite>
Synthetica: I suspect the boot entries aren't being generated because of a bootloader config issue or because /boot isn't mounted or similar
<sphalerite>
Synthetica: in the meantime, see if your changes get applied if you nixos-rebuild switch then systemctl restart display-manager
supremacsy has quit [(Quit: Page closed)]
digitus has quit [(Quit: digitus)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Profpatsch opened pull request #29061: CODEOWNERS: add Profpatsch to various paths (master...codeowners) https://git.io/v5VdY
NixOS_GitHub has left #nixos []
nh2 has joined #nixos
nh2 is now known as nh2-web
<nh2-web>
given that there exists a release-17.09 branch now, do I need to explicitly pull-request changes to master also for that branch if I want them in?
<tk_>
Hi all, how do I specify which compiler nix uses on macOS? I need a compiler with openmp
<nh2-web>
tk_: IIRC clang >= 3.8 has openmp supported by default
<tk_>
nh2-web: Sure, but nix seems to be using Apple's clang
<tk_>
nh2-web: It actually appears to be a newer clang, but I still am wondering how to switch the compiler to gcc
oida_ has joined #nixos
<nh2-web>
I don't know that but I'd imagine copumpkin would know
<copumpkin>
Nix doesn't use Apple's clang at all
<nh2-web>
tk_: ^
<tk_>
copumpkin: Yes, I was wrong about it using Apple's clang. It reports version 4.0.1. However, it seems to lack proper openmp support. It can't find omp.h
<copumpkin>
acowley stuff with openmp iirc, but I don't know, sorry