<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « openfst: 1.6.9 -> 1.7.1 (#55245) »: https://git.io/fhbO3
<ajs124>
I'm trying to "resurrect" a kernel that was dropped from nixpkgs upstream through an overlay, but copy&pasting the "linux_4_18 = " and "linuxPackages_4_18 = " lines from all-packages.nix fails with undefined variable 'kernelPatches'
<ajs124>
And yes, I _could_ just fork nixpkgs and revert that patch, but I'd really prefer not to do that
<{^_^}>
[nixpkgs] @xeji pushed commit from @dtzWill to master « caddy: 0.11.1 -> 0.11.4 (#56214) »: https://git.io/fhbOP
rcshm has quit [Remote host closed the connection]
rcshm has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
rcshm has quit [Remote host closed the connection]
rcshm has joined #nixos
<{^_^}>
[nixpkgs] @matthewbauer closed pull request #56116 → libgcc: fix build on darwin → https://git.io/fhFIA
<ddellacosta>
iqubic: you may find more help in #haskell if you haven't already asked there, I've found it's a tossup whether I get help here or in #haskell when I have a haskell+nix-related question
<ddellacosta>
as far as hoogle, I think you'd have to run an instance locally to get that working, but I'm not sure how you'd do that in nix
<{^_^}>
[nixpkgs] @xeji pushed commit from @marsam to master « mblaze: 0.4 -> 0.5 (#55526) »: https://git.io/fhbOF
dermetfan has quit [Ping timeout: 246 seconds]
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
vidbina has quit [Ping timeout: 246 seconds]
hedning has quit [Quit: hedning]
wavewave has quit [Quit: Connection closed for inactivity]
dramforever has quit [Ping timeout: 250 seconds]
<bpye>
Hey, when I'm writing a script am I better using say "${pkgs.coreutils}/bin/mktemp" or just "mktemp" as mktemp is in coreutils which I'm pretty sure is always in teh system path?
<bpye>
Or even "echo" vs "${pkgs.coreutils}/bin/echo"
<clever>
bpye: no real way to access stuff hidden in a let block, your only choice is to copy the code out
o1lo01ol1o has quit [Ping timeout: 244 seconds]
<bpye>
Ah okay
<bpye>
I guess I'll probably end up splitting these out into more general modules anyway, but to get it working it's easier to do it inline like this it seems
<bpye>
Is there a plan for what NixOS is going to do when we run out of service UIDs?
<bpye>
I guess that's the way to go for services that don't need to store persistent data?
xeji has quit [Quit: WeeChat 2.2]
<ottidmes>
bpye: persistent data is OK too, its more about whether other things need to access that persistent data too, like other services, then it becomes a problem
<infinisil>
bpye: We don't need to allocate static uids, just creating a user without a static uid works too and is perfectly fine
<bpye>
ottidmes: Sweet, that sounds like a good fit
<infinisil>
DynamicUser if it works is still pretty good though
<bpye>
infinisil: Sure, but then don't we end up in a weird state if you deploy a config with userA, write some data, deploy a new config with userA and userB, and then a config with just userB?
<infinisil>
bpye: NixOS remembers which users got which uids
<infinisil>
bpye: In /var/lib/nixos/gid-map
<infinisil>
bpye: And /var/lib/nixos/uid-map
doyougnu has quit [Ping timeout: 245 seconds]
<Izorkin>
infinisil: checked PR to nginx?
<bpye>
ottidmes: Is DynamicUser in 18.09?
<bpye>
Hm yeah, looks like it is!
<{^_^}>
[nixpkgs] @nomeata opened pull request #56272 → Avoid problematic test case of llvm6 on Android → https://git.io/fhb3B
<ottidmes>
bpye: yep, I have used it in 18.09 myself
doyougnu has joined #nixos
georges has quit [Ping timeout: 244 seconds]
<lopsided98_>
ryantm: That looks like an ofborg bug to me, there is no log for the test
rcshm has quit [Read error: Connection reset by peer]
rcshm_ has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
ottidmes has quit [Ping timeout: 246 seconds]
o1lo01ol1o has quit [Ping timeout: 258 seconds]
<Arahael>
bpye: I've seen that before with an odd openssl or certificate issue, I forget the details entirely, though.
rcshm has quit [Read error: Connection reset by peer]
<Arahael>
(And that was on a mac, too)
<bpye>
That seems possible, it's a https request. To make wget happy I had to make it store the hsts list to tmp
rcshm has joined #nixos
<bpye>
Hm, I wonder, with DynamicUser can I set two services to the same StateDirectory or is that a crazy idea?
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
o1lo01ol1o has joined #nixos
<infinisil>
bpye: Can't do that
eadwu has quit [Quit: WeeChat 2.4]
<infinisil>
bpye: Every dynamicUser service runs as a different user, and StateDirectory sets up a private directory for the services user
<infinisil>
So either systemd complains that they use the same state directory, or it just fails for one of them, because the directory got created with permissions for the other
<infinisil>
I just recently realized that while DynamicUser works well when your service is self-contained, it quickly falls apart when you need interaction between services
<infinisil>
systemd has a lot of other sandboxy features you can enable for your service to make it more secure though
<bpye>
It seems that setting User=xyz might make it work? Shouldn't they both end up as the same user? Obviously that has to be acceptable for your specific service...
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<bpye>
In my case I'm testing setting up a DNS server and auto updating a blacklist. The former is long running but the latter needs to be able to change files they both read
<infinisil>
bpye: If you use an already existing user in the User= setting it won't use dynamic users
<infinisil>
If the user doesn't exist yet it must throw some error, because it would try to create two users with the same name
<bpye>
That doesnt seem to be the behaviour
<infinisil>
Oh?
<infinisil>
What is the behavior?
<bpye>
"When used in conjunction with DynamicUser= the user/group name specified is dynamically allocated at the time the service is started, and released at the time the service is stopped — unless it is already allocated statically (see below). If DynamicUser= is not used the specified user and group must have been created statically in the user database no later than the moment the service is started, for example using the
<bpye>
sysusers.d(5) facility, which is applied at boot or package install time." It seems the first (in my case the long running service) with allocate the user dynamically, the second seems to run as the same user as it does suceed
<{^_^}>
[nixpkgs] @matix2267 opened pull request #56277 → vim_customizable: Include manpages in the output → https://git.io/fhbsd
<infinisil>
With BindReadOnlyPaths the permissions of the files still need to permit the other user to access it
<infinisil>
The DynamicUser+User way doesn't seem half bad though
<infinisil>
But, then again, it doesn't really get the benefit of DynamicUser (it being isolated from other services), if any service can just get access to the services state directory too if it has the correct settings..
<bpye>
I guess one hopes that you only configure your system that way IF you know you need it, BindReadOnlyPaths would work for me as one service only needs read access, the other needs rw, so i guess the service with rw would "own" the directory, and the read only service would bind it
<{^_^}>
[nixpkgs] @matthewbauer pushed to master « harfbuzz-icu: disable error for cast-align »: https://git.io/fhbGU
yokyolu has joined #nixos
vidbina has joined #nixos
<yokyolu>
i'm trying to add a haskell package from a pinned nixpkgs to an overlay
<yokyolu>
but once in shell cabal doesn't see it there
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
<infinisil>
yokyolu: (Nobody can help you if you don't post the code)
<kyren>
I'm experiencing a very strange behavior, and I'm not sure how to track this down
<kyren>
after a recent update, if I run, just as an example `nix-shell -p xorg.xeyes`, I get "error: getting status of '/<hash>-<hash>-<hash>.tar.gz': No such file or directory
<kyren>
if I revert to an earlier nixpkgs, the error goes away
<kyren>
this happens with *any* nix-shell invocation
<kyren>
aha, okay it looks like that path exists in the nix store, where it just the contents of nixpkgs
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
<kyren>
okay, this is starting to make sense, but I'm still not sure why it's broken... my NIX_PATH is basically just '/etc/nixpkgs', and this path is valid
<clever>
kyren: NIX_PATH=/etc/nixpkgs means to look for <nixpkgs> at /etc/nixpkgs/nixpkgs
<clever>
kyren: just like -I/usr/include makes #include <foo.h> look at /usr/include/foo.h
<kyren>
well, no it's NIXPKGS='nixpkgs=/etc/nixpkgs'
<kyren>
sorry
<clever>
you probably want NIX_PATH=nixpkgs=/etc/nixpkgs
<clever>
kyren: you can also test it with nix-instantiate --find-file nixpkgs
<kyren>
that's what it is, sorry I wasn't being precise
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
<kyren>
I haven't bisected it or anything, but this behavior started recently, like within the last few days
<clever>
kyren: how did you configure /etc/nixpkgs ?
<kyren>
so that link exists in the nixpkgs store, and it's valid for where the final target of the link is
<ibrahims>
Hello people, i am having difficulty installing ghc844 since yesterday, it tries to build from source and fails miserably, any idea why i might be missing the cache?
<kyren>
er, sorry it's valid for where /etc/static/nixpkgs points to
<clever>
kyren: it sounds like a bug in nix, when following multiple relative symlinks
<justanotheruser>
I'm trying to figure out what to search here. Is there a term for the configuration.nix equivalent which is specific to a user, rather than the system?
<ibrahims>
clever: thanks, any idea when it would reflect on 'nix-channel --update'?
<clever>
kyren: looks like something worth filing a bug against nixpkgs, but you can use my example as a work-around
lsyoyom has quit [Ping timeout: 246 seconds]
<clever>
justanotheruser: config.nix
<justanotheruser>
thanks
<clever>
ibrahims: i think its already in the latest channel, but i havent confirmed that
<kyren>
clever: so you think this is a bug in nix or nixpkgs? Should I file a bug for nixpkgs if it's unclear?
<ibrahims>
clever: thank you!
<clever>
kyren: hmmm, it could be either, it could be setup-etc.pl breaking things when handling relative links, or it could be a bug in nix, following relative links
<clever>
kyren: setup-etc.pl is part of nixpkgs
fusion809 has joined #nixos
schjetne has joined #nixos
sigmundv has quit [Ping timeout: 255 seconds]
o1lo01ol1o has joined #nixos
ircuser5678 has joined #nixos
ircuser5678 has quit [Quit: Going offline, see ya! (www.adiirc.com)]
<kyren>
clever: okay, thanks a lot for the help, I'll file a nixpkgs bug for lack of a better idea of where the bug is, and the way you have your nix path set up is just as simple as the way I'm doing it, so that works for me as a workaround
schjetne has quit [Ping timeout: 245 seconds]
rcshm has quit [Read error: Connection reset by peer]
<clever>
that will make it build other things, and fail later on
<clever>
but youll get more progress done per retry
o1lo01ol1o has quit [Ping timeout: 245 seconds]
Zer000 has joined #nixos
Maxdamantus has quit [Ping timeout: 268 seconds]
fusion809 has joined #nixos
Maxdamantus has joined #nixos
<kyren>
clever: thanks a lot, your method works great and now I can upgrade and have wayland enabled alacritty
<kyren>
unrelated question though, are dead symlinks ever automatically cleaned up from /etc?
<clever>
kyren: setup-etc.pl will keep track of the symlinks it makes, and delete them when it next updates /etc/
<clever>
if they are missing from the config
<kyren>
I updated and I seem to now have a dead symlink
<clever>
the nixpkgs one?
<kyren>
yeah
<kyren>
I mention because it's possible it's related to the bug... question mark?
<clever>
add it to the ticket on nixpkgs, i'm guessing setup-etc.pl is further confused by a symlink to a symlink
<kyren>
okay, added, thanks a ton!
<kyren>
++clever
<kyren>
wait, did I do that correctly
<dramforever_>
clever++
<{^_^}>
clever's karma got increased to 102
dramforever_ is now known as dramforever
<kyren>
postincrement copies clever and increments the copy, then throws the original clever away :/
nek0 has joined #nixos
<clever>
lol
<dramforever>
that sounds functional
<clever>
ive seen a haskell guy get ++c and c++ mixed up, and translated it to haskel the wrong way, and refused to understand the difference
<Zer000>
Halp! I get "Cannot mix incompatible Qt library (version 0x50b01) with this library (version 0x50b03)" when I try to start drumkv1. It used to work!
<clever>
Zer000: that happens if you mix qt versions in nix-env vs systemPackages, uninstall everything QT in `nix-env -q`, and add anything you want to keep to systemPackages
<Zer000>
I will do that
endformationage has quit [Quit: WeeChat 2.3]
<Zer000>
worked LIKE A CHARM. Thank you clever.
<Arahael>
Hmm, ANdroid Studio doesn't work properly on nixos, afterall. Just says "Error configuring".
<clever>
Zer000: yep
<Arahael>
Presumeably, this must be at the NDK setup where it's attempting to use the NDK binaries it just downloaded.
<clever>
Arahael: probably, something like the steam-run chroot may help
<Arahael>
clever: I'm surprised that the nixos package doesn't do something like that by default. :/
<clever>
Arahael: same, but ive not been into android since i switched to nixos
<Arahael>
clever: Fair enough. :)
<Izorkin>
How to write a file in the let section with awk processing ?
<Arahael>
clever: I'm thinking that the sweet spot for a pragmatist developer is to use something like Debian or Fedora - with nix packages.
<clever>
Arahael: i would use debian under virtualbox personally
<Arahael>
(Though, then again, nixos is *nice*)
<Arahael>
clever: Also a great option! I do have a good KVM setup here.
<clever>
Arahael: of note, virtualbox doesnt use kvm, and partially conflicts with it
<clever>
Arahael: both kvm and virtualbox use the same cpu extensions, and while you can load both drivers at once, only one can be actively using those features
<Arahael>
clever: I'm not actually a fan of virtualbox, anyway. :)
<clever>
if you start a vbox vm, then qemu-kvm will hard-fail on startup
<clever>
but if you close the vbox vm, qemu-kvm instantly begins to work
<Arahael>
Been burned by it too many times - I have a fairly low tolerance for crappy software (and it *is* crappy software on windows and macos, it's Oracle, too)
<clever>
no need to disable or reset things
<clever>
i have noticed the gui has gotten worse after the latest update
<clever>
if your confortable with qemu, that also works great
<Arahael>
Yes, it's slower, and for some reason I've had windows bluescreen on it, too.
yokyolu has joined #nixos
<Arahael>
clever: I'm using the kernel kvm drivers, I believe.
<clever>
teto: `--store ssh://foo` requires that $USER (the remote user by the same name as local) be trusted on the remote box, or it cant add things to its /nix/store/
<clever>
teto: i think you want --option builders ssh://foo
slack1256 has joined #nixos
vidbina has quit [Ping timeout: 246 seconds]
rcshm has quit []
_kwstas has joined #nixos
_kwstas has quit [Remote host closed the connection]
palo1 has joined #nixos
<{^_^}>
[nixpkgs] @tadeokondrak opened pull request #56279 → btpd: init at 0.16, add module → https://git.io/fhbZn
Xal has quit [Ping timeout: 250 seconds]
palo has quit [Ping timeout: 250 seconds]
palo1 is now known as palo
Xal has joined #nixos
iyzsong has quit [Ping timeout: 246 seconds]
schjetne has joined #nixos
fusion809 has quit [Remote host closed the connection]
fusion809 has joined #nixos
schjetne has quit [Ping timeout: 244 seconds]
zupo has joined #nixos
zupo has quit [Client Quit]
emacsomancer has quit [Quit: WeeChat 2.3]
slack1256 has quit [Ping timeout: 255 seconds]
slack1256 has joined #nixos
<{^_^}>
[nixpkgs] @Izorkin opened pull request #56280 → nginx: formating the config file → https://git.io/fhbZr
<Izorkin>
Mic92: please check this PR #56280 and #56255
<sphalerite>
clever: no, --store is nice for remote building when you're *not* a trusted-user
<sphalerite>
teto: it only works with ssh-ng, not with ssh
<sphalerite>
teto: so if you just replace ssh with ssh-ng in the command it should work
<sphalerite>
clever: --store builds remotely as if it were building locally, i.e. instantiates all the drvs there and obtains all the dependencies by building or from the cache configured remotely, thus not requiring trust
<sphalerite>
clever: of course it has some disadvantages too, like not being able to use multiple builders
<slack1256>
What is ssh-ng?
<clever>
sphalerite: related, ive noticed that you can copy .drv files, even if you arent trusted, because the path is based on a hash of the contents, no trust is needed
<sphalerite>
and the long period of no user feedback while it does lots of round-trips instantiating each drv one by one (a "mass-instantiate" protocol message would be nice)
<slack1256>
I never knew of such a project
mkoenig has quit [Remote host closed the connection]
<sphalerite>
slack1256: yeah it's a bit of a confusing name, ssh-ng just refers to nix's "ng" way of communicating via SSH
<clever>
slack1256: its just the name of a protocol within nix, for accessing remote stores
<clever>
3am, and there is just so much to remember, easy to forget things sometimes
<clever>
and i dont know how i managed to overlook this on my hydra for so long
<clever>
i dont think its ever been signing things?
<{^_^}>
[nixpkgs] @dotlambda merged pull request #56268 → pythonPackages.pykerberos: take krb5 as a nativeBuildInput → https://git.io/fhbmP
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @catern to staging-next « pythonPackages.pykerberos: take krb5 as a nativeBuildInput (#56268) »: https://git.io/fhbZy
<sphalerite>
nix sign-paths --all lol
aanderse has quit [Ping timeout: 250 seconds]
<clever>
sphalerite: already started that, but i dont think hydra takes advantage of any nix2.0 features
<LnL>
clever: I pass that in the store_uri query parameters
<sphalerite>
clever: doesn't it depend on the remote builders for hydra?
xkapastel has joined #nixos
<yokyolu>
i'm trying to pass a lens package from a pinned nixpkgs
<clever>
LnL: i setup `store-uri = file:///nix/store?secret-key=/etc/nix/keys/secret-key-file` and hydra didnt add `Sig:` to its narinfo files
<LnL>
also the fact that the narinfo cache never invalidates entries is pretty annoying
<yokyolu>
but cabal doesn't see it
<sphalerite>
LnL: does it really *never* invalidate thme?
<clever>
LnL: for positive cache hits, it caches it for days, which can be a problem
<clever>
for negative hits, it caches it for a few hours i think
<clever>
narinfo-cache-negative-ttl = 3600
<clever>
narinfo-cache-positive-ttl = 2592000
<clever>
> 2592000 / 3600
<{^_^}>
720
<clever>
1h and 720 hours
<clever>
> 2592000 / 3600 / 24
<{^_^}>
30
<LnL>
clever: I mean if eg. substitution fails
<clever>
ouch, 30 days!
<clever>
LnL: yeah, thats part of where iohk ran into trouble, hydra would GC something, and then for 30 days, anything using the cache would fail hard, because the cache says it exists, but it doesnt
aanderse has joined #nixos
<clever>
LnL: we got around it by just switching to s3://, which cant GC
<LnL>
yeah, separating the cache to file/s3 is better than serving the local store directly
<clever>
also better for security
<clever>
but i dont have infinite storage for my personal hydra, so GC is a must
<teto>
sphalerite: clever: thanks for the answers, I will post your ansers on the discourse if that's ok with you
<sphalerite>
clever: I was thinking recently about how remote building performance could be much better if there were a mass path query operation (to avoid 1 RTT per path queried), and even started implementing it. But I think that might be a bit of a short-sighted approach, since it doesn't address any other operations like instantiation. I think a new daemon protocol which supports multiple operations in parallel would
<sphalerite>
be really nice, or even just opening multiple nix-daemon --stdio instances on the same ssh connection
<clever>
sphalerite: something ive thought of, that would improve the perf of hydra
<clever>
sphalerite: lets say you have 2 build slaves, A has nothing in the store, B has 80% of some closure
<clever>
then you need to build something, which slave do you pick for the job?
<clever>
sphalerite: i was thinking more to reduce how much you have to copy, to get the builder up to speed, but your issue also helps, if one of the builders happens to have a finished product
<clever>
sphalerite: pipes-async is also busted on master
Glider_IRC_ has quit [Ping timeout: 250 seconds]
<sphalerite>
but by that reasoning you'd typically end up building everything on one machine, no?
petrkr has joined #nixos
<etu>
Is there any variable (like $out) that specify the build directory? I'm building a derivation that hardcode /tmp/.whatever-something and tries to write things there during build
<clever>
sphalerite: you would also obey max-jobs, and maybe try to keep things balanced
<clever>
so it would have to use weights, and prefer something with more of the inputs, but not force it
<kyren>
I feel like I'm missing something basic, but I'm writing a script with nix-shell where several packages are made available, and I need a file out of the nix store for one of those packages
<kyren>
I could write a nix expression that references that package and does string interpolation, thus getting the path in the nix store that way
<kyren>
but if I just have a quick nix-shell script, I feel like there should be a way to find the path to the store of one of the available packages so I can reference a file inside of it
<{^_^}>
[nixpkgs] @pSub pushed 0 commits to update-parallel: https://git.io/fhbCV
Makaveli7 has joined #nixos
<{^_^}>
[nixpkgs] @xeji pushed commit from @ikervagyok to staging « xorg/xfsinfo: init at 1.0.6 »: https://git.io/fhbCS
<{^_^}>
[nixpkgs] @xeji pushed commit from @ikervagyok to staging « xorg/xstdcmap: init at 1.0.4 »: https://git.io/fhbC9
<johanot>
kyren: I think "${pkgs.foo}/path/to/file" would be the right way to reference a file in the store from a script. Perhaps I misunderstand what you are trying to do?
<kyren>
johanot: I mean a bash script using #! /usr/bin/env nix-shell
<kyren>
concretely, I need OVMF_CODE.fd out of '${pkgs.OVMF.fd}/FD/OVMF_CODE.fd'
<kyren>
I could do it with nix-build '<nixpkgs>' --no-build-output -A OVMF.fd
<kyren>
which is sort of right, but I feel like I'm doing the same thing twice
<kyren>
I could also just write a separate .nix file
<johanot>
kyren: yeah, you could do that. I think I would consider generating the script itself from a nix expression. This makes it easier to template in stuff, obviously.
emily has quit [Remote host closed the connection]
emily has joined #nixos
<betawaffle>
does anyone know how to get the nftables-compat stuff on nix? (like all the {ip,arp,eb}tables translators)
o1lo01ol1o has joined #nixos
orivej has joined #nixos
<johanot>
betawaffle: looks like they're part of the iptables package
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<betawaffle>
but that conflicts with nftables, doesn't it?
<johanot>
that depends on what rules you add with either of the cli's I guess. The compat-tools add global chains, just like iptables. Whereas nft supports better namespacing with custom tables and chain hooks.
<betawaffle>
ok, but i could probably just put in the old style rules i want, then pull them out as nftables stuff, right?
<johanot>
that's possible yes
_kwstas has quit [Quit: _kwstas]
<betawaffle>
thanks
<betawaffle>
hmmm... Problem with the specified Ethernet protocol '802_1Q', perhaps /etc/ethertypes is missing.
petrkr has quit [Read error: Connection reset by peer]
kisik21 has quit [Remote host closed the connection]
arjen-jonathan has joined #nixos
<aanderse>
in nixos mysql is pretty awesome because of services.mysql.ensureDatabases and services.mysql.ensureUsers... this can allow for some really clean module code if we wanted. i'm pretty sure postgres has the equivalent of mysqls unix socket authentication, so does anyone know why there isn't a services.postgresql.ensureDatabases and services.postgresql.ensureUsers?
<srhb>
aanderse: I can't answer your question directly (I don't know exactly what those options are) but I do know that the postgres module has been up for overhaul for a long, long time, and unfortunately hasn't gotten into master yet. People might be slightly wary of making modifications until we have that in, which should hopefully get us out of version limbo a bit more?
<srhb>
aanderse: Talking to the authors of some of those PRs might also help (check the pull requests for "postgresql")
<aanderse>
srhb: ok, thanks for the info
petrkr_ has quit [Ping timeout: 258 seconds]
<aanderse>
yeah there is quite a bit of nasty module code for postgres which could be simplified so much if postgres had the same options for having nixos setup databases... i'll ping some people :)
<srhb>
aanderse: Sounds great, thanks for your attention on the subject :)
<dminuoso>
ancarda: It's not as nice. You can provide an initialScript to create the respective database roles though.
<aanderse>
dminuoso: i've only ever used postgres as a developer, and only minimally there, so i'm not familiar with the authentication methods as much
<aanderse>
a quick search said the same end result (as mysql socket auth) could be achieved, though
kisik21 has joined #nixos
<aanderse>
the mysql module just executes the code needed to create the databases, users, and permissions every time it starts
vdemeester` has quit [Quit: ERC (IRC client for Emacs 26.1)]
<aanderse>
dminuoso: yeah, that looks good. i was very briefly reading about combining the "trust" auth-method with....
<aanderse>
something... i can't remember
<aanderse>
to make it the equivalent of mysql socket auths end result
<dminuoso>
aanderse: I dont know how mysql authentication works, so..
<aanderse>
it locks authentication to a logged in user
o1lo01ol1o has joined #nixos
<dminuoso>
aanderse: postgres calls that ident
<aanderse>
so if you create a user called aanderse in your mysql database and say they authenticate via unix sockets then they don't need a password anymore
<dminuoso>
Or peer.
<dminuoso>
aanderse: Ah yes. That's called 'peer' in postgresql hba
<DigitalKiwi>
i guess the first question is it needed a newer version of pdfminer_six, i just added the right one, but i think the existing one should be updated? but what if the things that already use it don't work with the newer version then do i update those too? and how do i test them and stuff
<betaboon>
DigitalKiwi: the need to wrap something in wine just disappeared ;)
hyperfek has joined #nixos
ixxie has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<DigitalKiwi>
oh how so?
<{^_^}>
[nixpkgs] @jluttine opened pull request #56298 → pythonPackage.XlsxWriter: 1.1.2 -> 1.1.5 and pythonPackage.can: 3.0.0 -> 3.1.0 → https://git.io/fhblp
<hyperfek>
Is there any canonical way to reexport a module's options in a modified way? Say I want to add an 'abc' option to services.xyz and export it as services.def
<[Leary]>
dminuoso: The functionality is less useful on NixOS since you can easily manage extra dependencies in configuration.nix, but xmonad does have contingencies to be recompiled via cabal or stack: when asked to recompile, if it finds an executable ~/.xmonad/build then it will run it with a target path as an argument before restarting with the binary it finds at that path.
<[Leary]>
So you make a cabal project around your xmonad.hs, and have the script build it and move/link the binary.
o1lo01ol1o has quit [Ping timeout: 255 seconds]
vidbina has quit [Read error: Connection reset by peer]
<ivegotasthma>
my node-packages.json is -> [ "@vue/clie" ] and I run `node2nix -i node-package.json`. Whenever I try to run the generated `default.nix` file I get an error: `nix-shell requires a single derivation`
<Ariakenom>
where do I change the keyboard layout?
<ivegotasthma>
which I doubt is what you want to configure
<ivegotasthma>
but you can wait around for someone to answer
petrkr has joined #nixos
<Ariakenom>
well I probably want that one fixed too. consoleKeyMap = "sv" is probably what I want
<[Leary]>
I just set i18n.consoleUseXkbConfig = true;
<Ariakenom>
Oh nice
FRidh has quit [Quit: Konversation terminated!]
<Ariakenom>
[Leary]: does that replace font, keymap, and locale?
<[Leary]>
Xkb is just about the keymap, though I'm not sure how it interacts with locale... I do set that separately. Font is orthogonal.
<infinee>
Does anyone have a shell.nix they can share with me that sets up nodejs with nodePackages?
petrkr has quit [Remote host closed the connection]
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
petrkr has joined #nixos
Ariakenom has quit [Quit: Leaving]
petrkr_ has joined #nixos
<{^_^}>
[nixpkgs] @costrouc opened pull request #56303 → Initialize python pyviz stack and intake → https://git.io/fhb4R
vidbina has quit [Read error: Connection reset by peer]
petrkr has quit [Ping timeout: 258 seconds]
petrkr_ has quit [Remote host closed the connection]
petrkr_ has joined #nixos
petrkr_ has quit [Remote host closed the connection]
hyperfek has quit [Ping timeout: 256 seconds]
petrkr_ has joined #nixos
petrkr_ has quit [Remote host closed the connection]
petrkr_ has joined #nixos
<jluttine>
How do you fetch updates to nixpkgs repo from github? i'm getting constant timeouts apparently because there's some much to fetch..
<infinisil>
jluttine: No idea what you mean, can you be a bit more specific?
<joepie91>
okay, so - cc samueldr - when you run into issues with all text just being blocks/squares and no font being rendered, for software that's running via `steam-run`, and with a bunch of Pango-related errors on the terminal... the workaround is to make a packageOverride for Steam that sets nativeOnly to true; that may break other stuff though, since it no longer uses the vendored Steam libraries
<joepie91>
guess I'll have to file a bug about this
<jluttine>
infinisil: when fetching from github.com nixpkgs, i get: "error: RPC failed; curl 18 transfer closed with outstanding read data remaining"
<samueldr>
hm, good to know I guess :) thanks joepie91
<joepie91>
samueldr: (also, keyword spam to make it googleable from the logs :P)
<jluttine>
infinisil: git fetch upstream, so i think it fetches all branches
<infinisil>
Oh well that certainly looks like a GitHub issue, nothing Nix related
<infinisil>
Or your network
<infinisil>
Or git
<infinisil>
(or the possibly hundred other parts of hardware/software involved in transmission)
Ariakenom has joined #nixos
<jluttine>
infinisil: yeah, no problem in nix. just wondering if others have similar issues as it's probably related to the huge repo size with github.
ottidmes has joined #nixos
<infinisil>
Yeah :)
<infinisil>
(I don't have any such issues right now)
vidbina has joined #nixos
alex`` has quit [Ping timeout: 255 seconds]
PirBoazo has joined #nixos
alex`` has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed 4 commits to staging-next: https://git.io/fhb4M
<{^_^}>
[nixpkgs] @xeji pushed commit from @tokudan to master « burp: 2.2.16 -> 2.2.18 (#56288) »: https://git.io/fhbBT
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
<infinee>
if I have cloned nixpkgs locally. How do I specify it in my `~with import <nixpkgs> {}~ ?
<symphorien>
replace <nixpkgs> by /path/to/clone
<samueldr>
> <nixpkgs>
<{^_^}>
/var/lib/nixbot/nixpkgs/master/repo
<samueldr>
the angled brackets syntax basically resolves to a path
<samueldr>
> <nixpkgs/nixos>
<{^_^}>
/var/lib/nixbot/nixpkgs/master/repo/nixos
petrkr__ has quit [Remote host closed the connection]
eadwu has quit [Quit: WeeChat 2.4]
<infinee>
thanks
<samueldr>
>> Paths can also be specified between angle brackets, e.g. <nixpkgs>. This means that the directories listed in the environment variable NIX_PATH will be searched for the given file or directory name.
<Izorkin>
Not work command - post-configFile = pkgs.runCommand "httpd.conf" ''awk -f ${awkFormat} ${httpdConf} | sed '/^\s*$/d' > $out''; error cannot coerce a function to a string. How to fix?
softinio has joined #nixos
petrkr__ has joined #nixos
cswl has joined #nixos
<softinio>
is there a hosted hydra service anywhere ?
<Guilherme>
I dont know how to specify stdenv.mkDerivation if is a folder or not
<dminuoso>
[Leary]: Ah I see, so I could just provide a shell script that runs `cabal v2-run myxmonad`? The reason this is useful to me is because I have some dependency not available as a nixpkg
<ryantm>
Guilherme: What happens when it doesn't work?
infinee has joined #nixos
<sphalerite>
softinio: I don't think so
<symphorien>
Guilherme: $out does not exist when you start building, so you are responsible to create it with mkdir if you want $out to be a directory, or to move a file
<symphorien>
to it if you want it to be a file
<ryantm>
Guilherme: You can also use `install` to automatically make the directory and copy the file.
kisik21 has quit [Remote host closed the connection]
kisik21 has joined #nixos
<Guilherme>
thank you, it was really the problem
<[Leary]>
dminuoso: Unless you need to hack on that dependency and have the changes seen on recompile without rebuilding your NixOS system, it's better to produce a nix derivation for that dependency and make it available via the options exposed for xmonad.
<[Leary]>
Either way, I don't think you'll be able to get v2-run to help you; xmonad needs your script to produce a binary that /it/ runs. Maybe v2-install can take a target path argument, otherwise you need to v2-build, find the binary in dist-new and move it.
<dminuoso>
[Leary]: That's actually surprisingly easy. Using v2-run you can just find the binary using 'which' :-)
<dminuoso>
Its how I have been extracting build artifacts for slim docker containers for a while now
NickSeagull has quit [Quit: WeeChat 2.2]
<[Leary]>
Ah, if you're only using it for something like which then yeah.
kisik21 has quit [Remote host closed the connection]
Guilherme has quit [Quit: Page closed]
kisik21 has joined #nixos
o1lo01ol1o has joined #nixos
<{^_^}>
[nixpkgs] @NeQuissimus pushed 4 commits to release-18.09: https://git.io/fhbBa
NickSeagull has joined #nixos
simukis has quit [Quit: simukis]
vidbina has quit [Read error: Connection reset by peer]
PirBoazo has joined #nixos
simukis has joined #nixos
<NickSeagull>
Speaking of XMonad, is there a way of linking the config through its Nix expression like one can do with i3?
<PirBoazo>
Bonjour , Hello
<NickSeagull>
I used to do a `i3Cfg = writeFile ...` and then `i3.configFile = i3Cfg`, whilst with the XMonad nix expression I cannot seem to be able to do that
<NickSeagull>
Hello PirBoazo :)
eadwu has quit [Quit: WeeChat 2.4]
eadwu has joined #nixos
<PirBoazo>
I am seeking support about an issue with kernel 4.14.102
kisik21 has quit [Quit: ERC (IRC client for Emacs 26.1)]
o1lo01ol1o has quit [Ping timeout: 255 seconds]
kisik21 has joined #nixos
<PirBoazo>
where find forms to report an issue?
<sphalerite>
PirBoazo: you can open an issue on the nixpkgs github repo (https://github.com/nixos/nixpkgs/issues) or ask your question here if you're not sure if it's a bug
<sphalerite>
PirBoazo: also, #nixos-fr speaks french :)
<PirBoazo>
I think it's bug , i don't have issue with the 4.14.98
endformationage has joined #nixos
<PirBoazo>
:-) merci sphalerite
kisik21 has quit [Remote host closed the connection]
kisik21 has joined #nixos
alex`` has quit [Read error: Connection reset by peer]
<rycee>
dminuoso: The issue with `nix-env -qaP` should be resolved now. Try updating HM and see if it works.
<dminuoso>
rycee: Absolutely, can you point me to the relevant commit (Im curious)?
kisik21 has quit [Remote host closed the connection]
<NickSeagull>
How bad is the idea of taking that file and adding to my private set of nix expressions infinisil ?
<infinisil>
"private set of nix expressions"?
<NickSeagull>
Errr, I have some Nix expressions in my dotfiles repo that I use to build specific stuff, rather than having my own fork of nixpkgs
alex`` has quit [Read error: Connection reset by peer]
shabius has quit [Quit: Leaving]
<NickSeagull>
Dont know whats the best way of calling that ^_^"
_kwstas has joined #nixos
_kwstas has quit [Remote host closed the connection]
tdbgamer has joined #nixos
<dminuoso>
rycee: Ill let you know tomorrow, since I don't want to run a GC right now (I realized that it's quite annoying if you do this right before hopping onto a train..)
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
ddellacosta has quit [Quit: WeeChat 2.2]
rcshm has joined #nixos
rcshm has quit [Client Quit]
<infinisil>
NickSeagull: Those are just Nix files I guess :P
<infinisil>
NickSeagull: I don't think the PR actually works as of now, if you really want to specify the xmonad config reliably in your configuration.nix, I recommend just making a standard haskell project out of it, buliding that, and setting services.xserver.windowManager directly
rcshm has joined #nixos
NickHu has joined #nixos
drakonis has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
oida has joined #nixos
ij has quit [Ping timeout: 255 seconds]
kenshinCH has quit [Ping timeout: 255 seconds]
<NickSeagull>
infinisil: So basically it would be like the "standard" way of having an XMonad stack project and relying on that instead of the XMonad package on nixpkgs right? :)
ddellacosta has joined #nixos
Xyliton has joined #nixos
<Xyliton>
How do I fetchurl a zip archive? right now it just tells me that it doesn't know how to unpack them
<infinisil>
NickSeagull: Yeah something like that
<infinisil>
Xyliton: fetchzip instead
<Xyliton>
ohh, thanks! infinisil
<infinisil>
:)
<NickSeagull>
Awesome infinisil thanks for the valuable help
<infinisil>
At some point I'll migrate my xmonad setup to this way too, hopefully
<NickSeagull>
^_^
eadwu has quit [Quit: WeeChat 2.4]
o1lo01ol1o has joined #nixos
<NickHu>
Hi all, would anyone be free to review one of my stalled prs? It's just a simple one for getting a printer driver in:
<dminuoso>
NickSeagull: Many discussions predate the existence of v2-*/new-* style commands.
<NickSeagull>
Yeah I have seen them, but I don't see the current advantage of using cabal over stack currently. Also I use hpack in all of my projects which stack has support for :P
<symphorien>
64 bit machines can usually run 32 bit binaries
<Xyliton>
ah, then sure. I don't care if it is pure 32 or multilib as long as I can run it :shrug:
eadwu has quit [Quit: WeeChat 2.4]
<infinisil>
Xyliton: Use pkgsi686Linux.callPackage, that should do it
<symphorien>
the solution is to use pkgsi686Linux instead of pkgs
eadwu has joined #nixos
eadwu has quit [Client Quit]
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
<Xyliton>
hm... can't figure out where to replace that. I neither use callPackage nor pkgs. I have an `with import <nixpkgs> {};` at the top of the file though
eadwu has joined #nixos
Ariakenom has quit [Read error: Connection reset by peer]
<siers>
The 19.03 release branch hasn't been created yet, has it?
Ariakenom has joined #nixos
<Xyliton>
Is it possible to supply flags to the compiler from a derivation?
<Guilherme>
let nixpkgs = import <nixpkgs> {}; in nixpkgs.callPackage ./default.nix {}
<Guilherme>
the shell.nix will be that way
rcshm has quit [Read error: Connection reset by peer]
kenshinCH has joined #nixos
agander_ has joined #nixos
rcshm has joined #nixos
<symphorien>
<Xyliton> hm... can't figure out where to replace that. I neither use callPackage nor pkgs. I have an `with import <nixpkgs> {};` at the top of the file though << replace with: with (import <nixpkgs> {}).pkgsi686Linux ;
agander has quit [Ping timeout: 255 seconds]
<symphorien>
oh it has already been said, sorry
<{^_^}>
[nixpkgs] @Ekleog merged pull request #53143 → aspell: add all missing official dictionaries → https://git.io/fhtP1
<Xyliton>
infinisil: is there something I can set to pass that flag to `configure` too? looks like that is where it actually has to go :/
_kwstas has quit [Remote host closed the connection]
agander_ is now known as agander
o1lo01ol1o has joined #nixos
b1000101 has joined #nixos
_d0t has joined #nixos
domogled has quit [Ping timeout: 245 seconds]
lekstowsky has quit [Remote host closed the connection]
hedning has quit [Remote host closed the connection]
arjen-jonathan has quit [Ping timeout: 250 seconds]
<b1000101>
Hi guys, I'd like to backport a python package to 18.09 but it depends on higher versions of click and construct packages. Now, I am more than happy to create PR for those two packages but my question is, what is the best way of doing such backport and what are the chances I break other packages by doing this?
eadwu has joined #nixos
<infinisil>
b1000101: How much higher?
<infinisil>
If it's only a minor update I doubt it would break much
<MichaelRaskin>
b1000101: also, do you know that installing pythong from unstable into a nix-shell while on stable for the system is a supported workflow?
MarcWeber has joined #nixos
rcshm has quit [Read error: Connection reset by peer]
<b1000101>
infinisil: same as in master - click 6.7->8.0 construct 2.8.x -> 2.9.x
rcshm has joined #nixos
<b1000101>
infinisil: sry, click 6.7->7.0
<infinisil>
I see a lot of packages in nixpkgs depending on click, so I'd rather not have this backported
<{^_^}>
[nixpkgs] @msteen opened pull request #56311 → nix-upfetch: init at 0.1.0 → https://git.io/fhb0R
<infinisil>
You could however add a `click_7_0` package, separate from the normal one, and use this as a dependency for your package
<b1000101>
infinisil: yeah, well, that's what I was worried about, too
<b1000101>
infinisil: aaah, interesting point, didn't know that would be tolerable :)
<infinisil>
b1000101: Or, as MichaelRaskin suggested, use some ad-hoc way to get different packages
<infinisil>
(which would be my preference)
<Xyliton>
with the default makeWrapper it my package tries to install into `/usr/local/bin`; is there any way I can change that?
<b1000101>
MichaelRaskin: I'm trying it out in nix-shell - packages from master, my version being 18.09, that works, but I'd like to make this available to other people, too
<infinisil>
Xyliton: makeWrapper doesn't have anything to do with /usr/local/bin, that's the packages makefile or so doing that, most makefiles have an option to change the destination directory (often called DESTDIR)
hedning has joined #nixos
<{^_^}>
[nixpkgs] @costrouc opened pull request #56312 → Fixes for broken python packages for python-unstable → https://git.io/fhb0u
<MichaelRaskin>
Well, you can nix-env stuff from master, too
mek42 has quit [Quit: Leaving]
<b1000101>
infinisil: sure, I could do it that way I guess - oh, by the way...would it be possible to serve such package to other users? I'm thinking about setting up a dedicated server for serving these types of packages
<gchristensen>
does dbus have some sort of in memory database which I can write read from and write to?
<b1000101>
emily: yeap, I know cachix, but I was thinking about making a local cache at my company, don't know if it makes sense
<gchristensen>
I have two programs which don't really go whining time, but I would look still like the results to be shared with each other.
arjen-jonathan has joined #nixos
<b1000101>
and I thought hydra would be a bit of an overkill since what I probably need is only sharing packages between machines as described in the nixos manual?
<aristid>
gchristensen: how about a sqlite file? :D
<MichaelRaskin>
Well, it doesn't matter where packages come from, they are built — they can be shared
<MichaelRaskin>
_Upstream_ Hydra for master can be a valuable source of binary packages
<aristid>
gchristensen: what does "whining time" mean? not an expression i'm familiar with
<gchristensen>
oops, "typo". they don't really run at the same time.
<MichaelRaskin>
Damn you Dragon?
_kwstas has joined #nixos
<gchristensen>
exactly
_kwstas has quit [Client Quit]
asymmetric has joined #nixos
<b1000101>
MichaelRaskin: thanks for the tip with the unstable channel, I didn't realize you can have both and just call the other by name when installing a package
<asymmetric>
this is kinda funny: i set the wrong initialHashedPassword for root, and now i am locked out of `su`
<asymmetric>
i have no sudo for my user
<asymmetric>
is there anything i can do?
<aristid>
gchristensen: ok, any reason why a sqlite file does not suit your needs, then? :D
<gchristensen>
sure, reboot in to the install environment and mount all your disks in to /mnt as you did when you did the installation, and then run nixos-install
<infinisil>
b1000101: Really just try it without a cache, the build might take a minute maximum (unless your python package does special compilation stuff), really not worth it
rcshm has quit [Read error: Connection reset by peer]
<gchristensen>
aristid: no, that might work fine :-)
<asymmetric>
gchristensen: you mean boot from a livecd/liveusb?
<b1000101>
infinisil: I'm on it. Thanks a million!
rcshm has joined #nixos
<aristid>
gchristensen: it doesn't involve any dbus though /o\
<gchristensen>
asymmetric: yeah, exactly, the NixOS install image.
<noonien>
i've got `imports = [ <nixpkgs/nixos/modules/installer/cd-dvd/sd-image-raspberrypi.nix> ];` this imports from the `nixos` channel from what i undedrstand, how can i import from the nixos-unstable channel(which i've setup)?
<{^_^}>
[nix] @matthewbauer closed pull request #2700 → Get relative path of NIX_DATA_DIR → https://git.io/fhb0e
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
_kwstas has joined #nixos
zupo_ has quit [Ping timeout: 250 seconds]
_kwstas has quit [Read error: Connection reset by peer]
<clever>
yeah, <nixpkgs> is mapped to the nixos channel by default
<clever>
and for all other names, it generally goes to a channel by the same name
<ryantm>
<nixpkgs> equals whatever the NIX_PATH env var has nixpkgs is set to
<noonien>
ah, that makes a lot of sense, thanks!
<ryantm>
clever: your suggestion requires a nixos-unstable channel, right?
fenlomat has joined #nixos
<clever>
ryantm: yeah
<noonien>
argh, i still get `error: linux_rpi not supported on 'x86_64-linux'`, i guess there's no way around this
vk3wtf has joined #nixos
<clever>
noonien: you need to set nixpkgs.localSystem.system i think
<ryantm>
Alternatively, you could add to NIX_PATH nixos-unstable=https://github.com/NixOS/nixpkgs-channels/archive/19eedaf867da3155eec62721e0c8a02895aed74b.tar.gz then reference it like clever suggests
<fenlomat>
hey! i'm on a lives, trying to fix an existing system, how do i do a nixes-rebuild the underlying system?
<noonien>
ryantm: i've already setup the nixos-unstable channel
<fenlomat>
i mounted the partition in /mnt
<clever>
fenlomat: nixos-enter
<fenlomat>
*i'm on a liveusb
<clever>
noonien: i was using this when testing something similar: nixpkgs.system = "armv7l-linux";
<laas>
does anyone know why nixpkgs.gcc is not the same as nixpkgs.gcc8?
ij has joined #nixos
ekleog has quit [Quit: back soon]
<clever>
6754 gcc = gcc7;
<clever>
laas: line 6754 of all-packages.nix, gcc is assigned to be gcc7
rcshm has quit [Read error: Connection reset by peer]
<dminuoso>
Perhaps someone can guide me here. So when I run X with Xfce as the desktop manager (with Xmonad), lid close events are somehow caught and turn into xscreensaver - even though logind is seemingly receiving the lid close events and its configured to suspend.
<dminuoso>
The bizarre thing is, if I shutdown the display manager, then closing the lid suspends my laptop just normally.
<dminuoso>
If logind receives the lid event even with X is running, why is it not suspending my computer?
<jonreeve[m]>
clever: Thanks. Although it says `attribute 'doJailbreak' missing`. Am I not importing that one right?
<clever>
jonreeve[m]: that is not where you find haskell.lib
<jonreeve[m]>
clever: Ah. That's just copypasta, really. What should I be using instead?
<fenlomat>
clever: my problem is that i set mutableUsers = false but set the wrong root password. can i reset it from a liveusb without doing a nixos-rebuild?
<clever>
jonreeve[m]: updated the gist again
Makaveli7 has quit [Quit: Leaving]
<clever>
fenlomat: you could just edit the configuration.nix, and then re-run nixos-install, and it will update the system for you
eadwu has quit [Quit: WeeChat 2.4]
<jonreeve[m]>
clever: It says `infinite recursion encountered`.
eadwu has joined #nixos
<clever>
jonreeve[m]: i think i got super and self backwads
<clever>
flip them around the other way
xkapastel has quit [Quit: Connection closed for inactivity]
eadwu has quit [Client Quit]
eadwu has joined #nixos
<jonreeve[m]>
clever: cool, that seems to have worked! Thanks! Is that the easiest way of jailbreaking something, just out of curiosity? Seems like there's a lot involved in circumventing a package that won't build
zupo has joined #nixos
nhbh^ has joined #nixos
<clever>
jonreeve[m]: yeah, thats fairly normal setup
<spacekitteh[m]>
So it seems, if you change the UNZIP parameter for fetchzip, it doesn't check for a new hash if it's already downloaded. So it worked on my testing image, because it already had it prefetched :(
ixxie has quit [Remote host closed the connection]
<ottidmes>
thats the danger of fixed-output derivations
<spacekitteh[m]>
Indeed :<
work_ has quit [Quit: Connection closed for inactivity]
isidore has quit [Quit: WeeChat 2.4]
random_yanek has joined #nixos
<spacekitteh[m]>
clever: any chance that could be merged within the next few hours? :3
<{^_^}>
[nixpkgs] @bendlas opened pull request #56322 → virtualbox: add warning when for ineffective nixpkgs config → https://git.io/fhbz7
<{^_^}>
[nixpkgs] @costrouc opened pull request #56323 → Move ansible to python-modules and use toPythonApplication + ansible related packages → https://git.io/fhbz5
<noonien>
does this mean an sd-image can be created from a system other than an armv6l?
<samueldr>
tbenst[m]: (cold ping) are you around? about your issue with 18.09 and switch_root
<noonien>
what's [m], btw?
<samueldr>
a common suffix for the matrix bridge
<noonien>
oh, i see
eadwu has quit [Quit: WeeChat 2.4]
arjen-jonathan has quit [Ping timeout: 252 seconds]
eadwu has joined #nixos
johanot has quit [Quit: WeeChat 2.2]
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
kvda has joined #nixos
<earldouglas>
Is there a standard way to run npm install within mkDerivation? It stopped working for me a few months back, and I haven't figured out a solution:
<noonien>
earldouglas: a fixed output derivation is a derivation that checks the hash of the result, usually you have something like "sha256 = ...", our "outputHash = ..."
<samueldr>
anyone has a more elegant way to print the configured kernel (as per configuration.nix) (and not necessarily the currently built generation)? `nix eval --raw '(import <nixpkgs/nixos> {}).options.boot.kernelPackages.value.kernel'`
<tilpner>
Huh, I ran stack2nix on prometheus-matrix, and now I have a 46kloc default.nix that depends on cuda
<clever>
tilpner: that nix file contains all of hackage, with the changes defined in your stack.yaml
<clever>
tilpner: you need to use -A to select your own package, or its going to try to build all of hackage
<tilpner>
Oh, that explains a lot
o1lo01ol1o has joined #nixos
<tilpner>
Not quite something I can put into a git repo though :/
<clever>
tilpner: thats almost always whats done with it, to save others the time spent generating
<clever>
b.x turns into an thunk containing an ExprSelect, which does a somewhat expensive map lookup
rprije has joined #nixos
<clever>
so it defers it until later
o1lo01ol1o has joined #nixos
tmaekawa has quit [Quit: tmaekawa]
sigmundv has quit [Ping timeout: 245 seconds]
rprije has quit [Client Quit]
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<siers>
How could I fill an array with trailing nulls if it's too short? I need four elements. I could do take 4 (array ++ [null null null null]), but I can't find take.