<judson>
Or I could just say "someone else can run the tests for now" :)
<samueldr>
that's more convenient :)
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
goibhniu has quit [Ping timeout: 246 seconds]
shabius has joined #nixos
georges-duperon has quit [Ping timeout: 264 seconds]
<endformationage>
--arg / --argstr options seem not to work with e.g. `nix-shell <nixpkgs> -A myTestEnv --argstr stringArg "overridden string"` where myTestEnv is a trivial env overlayed with callPackage.
Rusty1 has joined #nixos
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fx3Ey
slack1256 has joined #nixos
semilattice has joined #nixos
orivej has joined #nixos
<semilattice>
Any procedure for when a package release removes compatibility with 32-bit linux
<semilattice>
Should it be kept as the older version for 32-bit?
thc202 has quit [Ping timeout: 268 seconds]
jasongrossman has joined #nixos
semilattice has quit [Remote host closed the connection]
alunduil has quit [Quit: leaving]
<{^_^}>
[nixpkgs] @yrashk opened pull request #47949 → webkitgtk222x: init at 2.22.2 → https://git.io/fx3ua
<endformationage>
Ah, I see --arg / --argstr options only operate top level.
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
jasongrossman has joined #nixos
<Purple-mx>
has anyone ever tried to compose multiple trees (nixpkgs & overlays) into a single one? I guess nix could do it, or something like dhall, although I still havent played with dhall so my opinion on that isnt well researched
alex`` has quit [Quit: WeeChat 2.2]
<Purple-mx>
adding lots of overlays has its complexity and performance issues. it seems more logical to me to compose all the overrides into a single monolithic tree for evaluation
<Purple-mx>
that way nixpkgs could also be a lot more modular
<Purple-mx>
i would even compile it to bytecode if nix defined one
jperras has joined #nixos
<Purple-mx>
well, I either will one day, unless a better lang comes along, or unless I pass away first
<Purple-mx>
there could be a lot of optimization done there, like removing duplicate expressions (from things like carnix, node2nix, etc)
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
Supersonic112 is now known as Supersonic
lassulus_ has joined #nixos
<avn>
sevanspowell: better to keep last 32bit as own expression. (if it binary only, I am curious, if source build restricted to 64)
<samueldr>
the announcement for discourse is ready
<gchristensen>
nice
<samueldr>
well, I can probably tag it
<samueldr>
(the commit)
<samueldr>
not sure why one would wait for a channel update for the tag
<gchristensen>
well .... I'd wait just to make sure the tag is actually a good release
<samueldr>
the instructions don't say *what* should be tagged!
<gchristensen>
don't bother staying up too late: eelco will be awake when we wake back up to merge homepage PRs, so staying up late doesn't get everything finished
<samueldr>
I'm not doing anything different I would otherwise
sigmundv has quit [Ping timeout: 252 seconds]
<gchristensen>
+1 on nights like this I have a hard time going to bed ;) (in other words, that instruction is as much for me...)
<samueldr>
not even doing release stuff!
<samueldr>
I was looking at xpra
<gchristensen>
oh wow that looks so cool
<samueldr>
though it'll need some love in nixpkgs... currently broken :/
<samueldr>
half-broken
<samueldr>
shadowing a complete screen works
<samueldr>
and I would bet connecting to an non-nixos xpra works
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/fx3V6
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #46068 → Prometheus Alertmanager: Allow the definition of extra options on commandline → https://git.io/fABxF
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/fx3Vi
<{^_^}>
[nixpkgs] @samueldr opened pull request #47956 → [backport] Miniupnpd and bittorrent improvements (#46443) → https://git.io/fx3wQ
jb55 has quit [Quit: WeeChat 2.1]
Rusty1 has quit [Quit: Konversation terminated!]
endformationage has quit [Quit: WeeChat 1.9.1]
<sphalerite>
I'm trying to delete a drv from the store. nix-store -q --referrers outputs nothing for it, nix-store -q --roots prints a couple of generations of my user profile. But AFAIU I should still be able to delete it using nix-store --delete --no-keep-derivations..?
<sphalerite>
(which I can't, it still complains that it's still alive)
<{^_^}>
[nixpkgs] @emmanuelrosa closed pull request #47916 → CONTRIBUTING.md: -bin as suffix for packages not build from source → https://git.io/fxOE7
<{^_^}>
[nixpkgs] @andir pushed 2 commits to release-18.03: https://git.io/fx3ip
alex`` has quit [Quit: WeeChat 2.2]
simukis has joined #nixos
slack1256 has quit [Ping timeout: 252 seconds]
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
<{^_^}>
[nixpkgs] @andir opened pull request #47957 → [18.09] firefox - fix i686 & mark old esr as unsupported → https://git.io/fx3PI
<CMCDragonkai>
I have a Haskell package that also depends on Clang. I'm using cabal2nix. Is there a way to specify this in the cabal file so that cabal2nix picks it up, or a way to override the cabal2nix derivation to get the clang compiler into the build environment?
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<cyris212>
Is there a sane way to add jdk's to an IDE without it breaking on every update?
<CMCDragonkai>
I'm also confused why that when I use `.env` of the haskell package, and I add `buildInputs` or `nativeBuildInputs` via `overrideAttrs`, they are not preserved in `.env`
<CMCDragonkai>
In the `default.nix`. If I add in `nativeBuildInputs = attrs.nativeBuildInputs ++ [ clang ];` and then use `shell.nix`, there's no clang.
<srhb>
CMCDragonkai: Ah. I believe env is an entirely separate derivation. It doesn't recursively update.
<CMCDragonkai>
Why not? Shouldn't env update based on the derivation I just overrode?
rfold has joined #nixos
<srhb>
I think you'll need to dig into generic-builder.nix to get the Real Answer.
<srhb>
After all, when you overrideAttrs, you're altering the result, not the call to mkDerivation..
<srhb>
Hence my guess.
<rfold>
Hi, I was wondering if it is safe to evaluate untrusted Nix expressions. I don't think so.
<srhb>
rfold: evaluation can be made mostly safe with the right settings.
xok has quit [Ping timeout: 272 seconds]
<srhb>
rfold: Building and running cannot.
<jasongrossman>
rfold: Not completely safe, but any bad effects will be a little bit more delayed than if you run e.g. an untrusted bash script.
<jasongrossman>
srhb: Overlapped.
<srhb>
:-)
<andi->
cyris212: what I do for those kinds of things (e.g. some libraries that some steam game needs) is have an activation script symlink them to a well-known location on my system and point the IDE to that location.
<rfold>
srhb: Thanks. Are the details documented somewhere?
<srhb>
rfold: Of what exactly?
<Myrl-saki>
Is there any case where Nix's infinite recursion checker fails?
<srhb>
rfold: The nix manual has all the flags related to restricting what is allowed during evaluation.
<rfold>
srhb: Thank you!
<srhb>
rfold: Regarding building and running something: In short, it doesn't matter whether you or the malicious person builds the malware that is run on your computer. :-)
<Myrl-saki>
Oh, there is lol
<Myrl-saki>
nix-repl> let f = x: f (x + 1); in f 0
<andi->
Myrl-saki: I am pretty sure there is. I had one of those (or what seemed to me since cpu was on 100% and RAM fully utilized after a few minutes..) after modifying a carnix generated expression myself..
<andi->
It would eventually run OOM but not detect it..
<Myrl-saki>
andi-: Shh, let me delude myself that nix will always work.
<andi->
maybe I just need another 100GB of RAM ;-)
<Myrl-saki>
> let xs = [1] ++ xs; in xs
<{^_^}>
infinite recursion encountered, at (string):196:17
<Myrl-saki>
This one's interesting though.
<Myrl-saki>
> let xs = [1] ++ xs; in builtins.head xs
<{^_^}>
infinite recursion encountered, at (string):196:17
<Myrl-saki>
Not as interesting now.
<rfold>
It seems like runInLinuxVM works in sandbox mode but the program running in the VM can still see all of the Nix store.
<Myrl-saki>
Oh cool. This one's pretty damn awesome.
xok has joined #nixos
<Myrl-saki>
> let list = { head = 1; tail = list; }; in list
<{^_^}>
{ head = 1; tail = { head = 1; tail = <CYCLE>; }; }
<Myrl-saki>
That <CYCLE>
<sphalerite>
srhb: well, environment variables don't get updated without the process they belong to playing a role...
<CMCDragonkai>
srhb: If I use `build-tools` or `system-build-tools`, I can see `clang` in the `buildInputs` now for both default.nix and shell.nix. That's cool, but what about the distinction between `nativeBuildInputs` and `buildInputs`?
<Myrl-saki>
CMCDragonkai: Are you asking what they're used for, or
<CMCDragonkai>
I mean in Nix I can specify whether a dependency is `nativeBuildInputs` or `buildInputs`, but it seems if I use the package.yaml `build-tools` or `system-build-tools`, I cannot specify such distinction?
<Myrl-saki>
CMCDragonkai: Ah, don't use the tool. :P
<Myrl-saki>
Err
<Myrl-saki>
I* don't use the tool
<{^_^}>
[nixpkgs] @andir opened pull request #47959 → [18.03] oraclejdk8psu: mark as insecure → https://git.io/fx3Pi
<CMCDragonkai>
What do you use?
<Myrl-saki>
CMCDragonkai: I bootstrapped it with cabal2nix, but nowadays I just modify the nix file.
<CMCDragonkai>
Do you have an example?
jasongrossman has quit [Remote host closed the connection]
<lingeeal>
my nixos has a weird behaviour, I reboot the xfce and the following boot does not start my x11. Then if I log in and do 'shutdown -h now'. I wait 3 min for "A stop job is running for X11 Server", it completes, then it hangs on "Starting mdadm-showdown.service.." I turn off the laptop with the power button. I turn it on and my X11 starts the session well, I start my xfce. However, if I then restart, it fails again to
<lingeeal>
start X11. Why is this happening? How can I fix this?
<srhb>
d1rewolf: Looks like a lot of paths in that package don't get fixed correctly. Definitely needs work.
<hodapp>
hmm... adding libxml2 support to gdal... make it optional and not there by default, optional and there by default, or just always present? I feel like optional & not there makes for an easier commit but have no idea what's normal here
mixfix41 has quit [Ping timeout: 246 seconds]
<srhb>
hodapp: It's mostly a closure size vs. "how often do people need that" thing, so very subjective.
<hodapp>
I wouldn't guess libxml2 that adds substantially to anything
<srhb>
optional & not there is almost always completely uncontroversial.
<hodapp>
but on the other hand it was a bit annoying to be using gdal and discovering that some really trivial build-time stuff hadn't been enabled
<srhb>
Check the cost on the closure. :)
<hodapp>
how does one do this?
<srhb>
nix-store -qR path | du -sch
<srhb>
Also worth looking through the list from -qR just to see if there's any dependencies in there that are probably unwanted.
<srhb>
er.. du -sch $(nix-store -qR path) actually :P
<srhb>
Looks like it already has libxml2 in its closure.
<d1rewolf>
srhb: so you're thinking the derivation is broken?
<srhb>
d1rewolf: Not broken per se, but certainly not very good either
<d1rewolf>
K, thx
<srhb>
d1rewolf: My guess is that *some things will work* but only if it's installed in your PATH
<srhb>
hodapp: In which case it's probably quite silly that it's not enabled :)
<d1rewolf>
Yeah, I'm finding a lot of these problems lately. I'm thinking for my work desktop I may have to switch back to Ubuntu sadly for now
<srhb>
d1rewolf: It doesn't look very maintained. The version is from november 2017.
<d1rewolf>
But i assume if i install nix i can use most of my existing nix config?
<hodapp>
srhb: the entire field of GIS is, arguably, defined by a lot of really silly software...
<srhb>
d1rewolf: Depends what you mean by "nix config"
<srhb>
d1rewolf: If you mean NixOS config, then basically none of it.
<srhb>
hodapp: Hehe. :P I just meant, if it already has the dependency in its closure, it's silly to not enable the features.
<d1rewolf>
Not the hardware config, but systemPackages, etc
<srhb>
d1rewolf: Nope. You can probably get some of that with home-manager though.
<d1rewolf>
srhb: oh...ok
<d1rewolf>
Thx
<srhb>
hodapp: For instance, right now it depends on libspatialite, which depends on libxml2. So since libxml2 is needed, using it directly it won't affect closure size.
<hodapp>
srhb: ohhhh I see what you mean
<hodapp>
I hadn't realized libxml2 was already present
<etu>
Hmm, how do I set options like this one: "config.kodi.enableSVTPlay"?
__monty__ has quit [Quit: leaving]
<etu>
When I add it to my config I get this error: "error: Module `/etc/nixos/configuration.nix' has an unsupported attribute `boot'. This is caused by assignments to the top-level attributes `config' or `options'."
<sphalerite>
samueldr: maybe we could take the nix-info bit out of the topic line? Would leave more space for the extra channels and we can just jam the nix-info stuff into a factoid to invoke when appropriate
<sphalerite>
,info = Please share the output of nix-shell -p nix-info --run nix-info to help us help you.
<{^_^}>
info defined
<gchristensen>
or drop #nixos from the topic
<sphalerite>
oh yeah that would also make sense :D
<samueldr>
hmm, that #nixos is probably me typing the channel name two times for setting the topic!
<samueldr>
(I'm not opped anymore btw)
<obadz>
hi all, has anyone connected a printer to nixos recently? something maybe wifi based via google cloud print or the like
<samueldr>
recently no, but did do it with a wireless brother printer, but not via google cloud print
<samueldr>
though, chrome itself was able to print using google cloud print on it
<samueldr>
I ported the brother printer driver and was then able to print wirelessly with it; though I'm no expert on linux printing, maybe it would have worked without its specific driver
<obadz>
seems like a lot of work? feels like google cloud print is nicely generic..
<samueldr>
obadz: as I'm no expert, I didn't know if google cloud print can work with cups, though if it would, that probably would have been easier than with the driver :/
<{^_^}>
[nixpkgs] @aneeshusa opened pull request #47975 → containers: Make systemd journals available from the host → https://git.io/fx3xd
phreedom has quit [Ping timeout: 256 seconds]
grumble has joined #nixos
<Mateon1>
Hi, I may be misunderstanding something, but why doesn't this work?: nix-shell -p 'python3.withPackages (p: [p.tensorflow.override { sse42Support = true; }])' - on import Python throws a ModuleNotFoundError
<Mateon1>
Without the .override, python sees tensorflow just fine
<Mateon1>
I'm probably misunderstanding something, tensorflow has some optionals for CPU feature flags that I'd like to set. Is .override not the correct way to do this?
iqubic has joined #nixos
<vcunat>
Mateon1: you are missing parentheses
<vcunat>
within lists, space separates arguments
<vcunat>
s/arguments/list elements/
<vcunat>
call of the override function has a lower "operator priority"
<Mateon1>
Alright, hm. Even with the correction: python3.withPackages (p: [p.tensorflow.override({ sse42Support = true; })]) - I get the same import error
<Mateon1>
I just tried that and it's now actually copying tons of paths
<Mateon1>
Thank you, that's somewhat not obvious
<vcunat>
It's strange that it wasn't throwing any errors.
<Mateon1>
Also explains why I couldn't override the python with a python.withPackages in sage earlier
aanderse has quit [Quit: aanderse]
<vcunat>
Perhaps it would have been clearer to require comma as list separator, e.g. like Haskell, but such incompatible changes are rather hard *now*.
<Mateon1>
Ok, so if I understand correctly, the packages list was [<function override>, <map>]. Really weird that it didn't report any errors, but explains why tensorflow was absent
<vcunat>
packages are specific maps actually, but it's possible to recognize them easily
<clever>
Mateon1: haskell packages runs a filter function over the entire list, to restrict it to just haskell libraries
<clever>
python might do the same thing
<Mateon1>
Huh, I thought packages were derivations
<clever>
and that can silently ignore all functions
<clever>
Mateon1: all derivations are technically sets
<clever>
> "${ {outPath = "foo";} }"
<{^_^}>
"foo"
<Mateon1>
Ok, for some reason I thought derivations were a built-in type in nix
<clever>
if you try to treat a set as a string, it will return the .outPath attribute
<Mateon1>
Hm, another small question, is it possible to make nix renice itself while building, by default?
<Mateon1>
Or should I just alias all nix-* commands I use to use nice?
<clever>
on nixos, it will run the builds under the nix-daemon, so even if you renice nix-build, it wont have any effect
<clever>
so you would instead just renice the current nix-daemon process
<{^_^}>
[nixpkgs] @P-E-Meunier opened pull request #47976 → Rust build-support: enabling cross-compilation of Rust code → https://git.io/fx3hL
nDuff has joined #nixos
<nDuff>
Howdy -- I'm trying to test a change to my local nixpkgs tree (intended to fix gnupg20's ability to be used by gnupg1compat), but when I run nix-build -I ~/VC/nixpkgs -E 'with import <nixpkgs> { }; pkgs.gnupg1compat.override { gnupg = gnupg20; }', I get an error implying that it's using my system's nixpkgs-18.09pre rather than ~/VC/nixpkgs (the locally-patched tree with the proposed change). How should I be doing that instead?
FRidh has quit [Quit: Konversation terminated!]
<clever>
nDuff: -I ~/VC/nixpkgs and <nixpkgs> makes it look for ~/VC/nixpkgs/nixpkgs, which doesnt exist
<shreyansh_k>
Hi, quick question: When linux 4.19 comes out, would the default kernel on 18.09 updated to 4.19 or will it stay on 4.14?
<clever>
think of it like #include <foo.h> and -I/usr/include in c/c++
<clever>
nDuff: nix has a -I nixpkgs=/path to help with things
iqubic has quit [Remote host closed the connection]
<clever>
nDuff: you can also just ignore the search path and directly `import ~/VC/nixpkgs`
<sphalerite>
shreyansh_k: it will stay on 4.14 I believe, but linux_latest will be updated to point to 4.19
<sphalerite>
I think
<samueldr>
not sure
* nDuff
plays around with the syntax a bit, and eventually gets something kludgy but working: nix-build -E 'let pkgs = import '"$PWD"' { }; in pkgs.gnupg1compat.override { gnupg = pkgs.gnupg20.override { guiSupport = false; }; }'
<samueldr>
default kernels are always an LTS kernel
<samueldr>
I would guess that it's going to stay 4.14 since 18.09 will have released with 4.14 as the default
<samueldr>
but 4.19 should be made available
<samueldr>
and kept when 4.19+1 releases, while non-lts releases aren't kept around once there's no support
<symphorien>
nDuff: import ./.
<nDuff>
much cleaner, thank you
<shreyansh_k>
Oh! okay. Thank you.
Ariakenom has joined #nixos
xok has quit [Quit: Leaving.]
<{^_^}>
[nixpkgs] @charles-dyfis-net opened pull request #47977 → gnupg20: set version variable to fix gnupg1compat → https://git.io/fx3ju
nDuff has quit [Quit: zzz]
mayhewluke has quit [Ping timeout: 252 seconds]
mayhewluke has joined #nixos
<vcunat>
+1 about the kernel
<sphalerite>
the one case where it does get upgraded is if the current one goes EOL during the release's maintenance period, right?
<sphalerite>
wait no that never happens AFAIK, I'm confusing things
Izorkin has joined #nixos
jedahan has joined #nixos
<samueldr>
kernels do get removed when they get EOL'd, but as you said, never happened with LTS kernels
<samueldr>
so e.g. 4.18 will get removed at some point from stable and unstable releases, but 4.14 will not and neither would 4.19
DigitalKiwi has joined #nixos
DigitalKiwi has left #nixos [#nixos]
<vcunat>
Yes, that can't ever happen, as we use latest longterm and their support cycle is way longer than ours.
<jtojnar>
could anyone using plasma please check if enabling virtual pulseaudio device for simultaneoous output in paprefs work for them? https://github.com/NixOS/nixpkgs/issues/47947
<{^_^}>
#47947 (by Chiiruno, 18 hours ago, open): Paprefs does not work
<srhb>
Assuming we never end up with, like. a .0 release just before our release freezes :-)
<srhb>
Still, the windows are way bigger.
<vcunat>
The windows are known in advance.
<srhb>
Right. I just meant I could conceive a situation where we ended up with latest-1 instead of latest.
<srhb>
Still plenty of leeway in the LTS windows for that to not be a problem.
<vcunat>
Yes, that will happen very soon.
<samueldr>
latest lts - 1 is going to happen soon with 18.09 :)
<srhb>
Well, 18.09 is released.
<srhb>
So we did start with latest latest ;)
<samueldr>
we always will :D
<srhb>
Really? even if, say, next LTS comes out days before our feature freeze?
<srhb>
That's what I was referring to above.
<vcunat>
Ah, freeze vs. release time.
<srhb>
Yeah..
<samueldr>
chances are it would get in, unless they somehow made so much change that the update can't happen quickly in nixpkgs
<vcunat>
that's too short a period when compared to longterm kernel kernel support
* srhb
nods
<samueldr>
the LTS release would have to happen the day the freeze happens :)
<srhb>
Or fail to inspire confidence during the release stabilization.
<srhb>
But yeah.
<samueldr>
and even then, probably can sneak it past
<srhb>
Doubt it'll happen :)
<vcunat>
I hope we don't have any other important package whose default will get unsupported during 18.09.
<vcunat>
We can probably deal with it somehow when a significant vulnerability comes.
<{^_^}>
[nixpkgs] @lopsided98 opened pull request #47980 → lib/systems: use correct config for armv7l-hf-multiplatform → https://git.io/fxsJr
magnetophon has quit [Quit: ERC Version 5.3 (IRC client for Emacs)]
zduch4c has joined #nixos
<zduch4c>
hello!
<zduch4c>
can I upgrade to 18.09 simply by changing the version in configuration.nix?
<samueldr>
no!
<samueldr>
:)
<samueldr>
,stateVersion
<{^_^}>
Setting stateVersion to the latest release doesn't update anything and can only break your setup at best. If you want to regardless, Ctrl-F for "stateVersion" in https://nixos.org/nixos/manual/release-notes.html to see things that need to be manually upgraded with the new value
jedahan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<zduch4c>
is it me or is the newest iso larger than the one before?
<zduch4c>
wasn't 18.03 ~300MB or something?
Anton-Latukha has joined #nixos
<lassulus>
hmm, nix repl has a disappearing prompt for me, tab completion also seems broken, anyone else with those errors? using urxvt here
slack1256 has joined #nixos
zarel has quit [Ping timeout: 272 seconds]
<manveru>
lassulus: that was fixed quite a while ago, what's your nix version?
<lassulus>
2.1pre6338_45bcf541
<lassulus>
I'm currently upgrading to 18.09, maybe it's fixed afterwards
<manveru>
hmm, no
<manveru>
i'm on the same version, and it's happening here too :(
<manveru>
it was working for a while...
<manveru>
upgrading to 18.09 atm, will see
<Anton-Latukha>
Does anyone remembers more clearly then me why LICENSE files distribution was removed. Maybe even words to direct me to find the evidence.
<manveru>
which license?
magnetophon has joined #nixos
sigmundv__ has quit [Remote host closed the connection]
magnetophon has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @8573 opened pull request #47984 → nixos/bash: Use `escapeShellArg` for shell aliases → https://git.io/fxsTW
<magnetophon>
I'm trying to update libclthreads, but am getting: "ldconfig: Can't open cache file /nix/store/fg4yq8i8wd08xg3fy58l6q73cjy8hjr2-glibc-2.27/etc/ld.so.cache
<magnetophon>
: No such file or directory"
<magnetophon>
There's plenty of people on the web having this issue, but I can\t find the solution.
<{^_^}>
[nixpkgs] @timokau merged pull request #47984 → nixos/bash: Use `escapeShellArg` for shell aliases → https://git.io/fxsTW
<{^_^}>
[nixpkgs] @jqueiroz opened pull request #47985 → pythonPackages.sslib: init at 0.2.0 → https://git.io/fxsIR
fragamus has joined #nixos
marusich has quit [Ping timeout: 240 seconds]
wpcarro has joined #nixos
<wpcarro>
I wrote a shell script to integrate my window movement and closing between i3 and Emacs using a shell script. I'd like a way to package the shell script in a way that allows users to whitelist it in the `systemPackages` as something like `i3-and-emacs`. Once they have this, I'm hoping to have an executable named `i3-and-emacs` available on that user's `$PATH`. I imagine this involved making a derivation.
<wpcarro>
Am I on the right track?
phreedom_ has quit [Ping timeout: 256 seconds]
phreedom has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kyren has quit [Remote host closed the connection]
fragamus has joined #nixos
[Leary] has joined #nixos
Lears has quit [Read error: No route to host]
<srhb>
wpcarro: Yup.
<srhb>
wpcarro: But that doesn't say very much. You might be specifically interested in eg. writeScriptBin.
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jasongrossman has quit [Ping timeout: 268 seconds]
nekroze has joined #nixos
<nekroze>
I am trying to update from 18.03 to 18.09 via `sudo nixos-rebuild boot` but it fails towards the end with "perl: warning: Setting locale failed." I have found some stuff with that error on google but nothing pans out, am I missing something from the update process?
<rawtaz>
i presume you updated the channels after adding them to change the version number?