<hyphon81>
I met a strange situation. I expect to return path, but it code return nothing... "echo ${gnome-settings-daemon-ubuntu}/shara/gsettings-schemas/${gnome-settings-daemon-ubuntu.name}/glib-2.0/schemas/*.xml"
<hyphon81>
In nix code.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vmchale opened pull request #25514: tw-rs: init at 0.1.26 (master...master) https://git.io/v9ghr
NixOS_GitHub has left #nixos []
<clever>
hyphon81: the * didnt match anything
<clever>
hyphon81: and nix builds are in a special mode where *.xml returns nothing, rather then a literal "*.xml" string
<clever>
hyphon81: try doing an ls on ${gnome-settings-daemon-ubuntu}/shara/gsettings-schemas/${gnome-settings-daemon-ubuntu.name}/glib-2.0/schemas and you should confirm that
_Vi has joined #nixos
justelex_ has quit [(Ping timeout: 268 seconds)]
filterfish has quit [(Ping timeout: 240 seconds)]
<hyphon81>
clever: Yeah. This code returns the path. "echo ${gnome-settings-daemon-ubuntu}/share/gsettings-schemas/${gnome-settings-daemon-ubuntu.name}/glib-2.0/schemas/"
<clever>
hyphon81: now try running ls on that path, inside the nix-build context
<clever>
just change the echo to ls -l
<hyphon81>
clever: Hmm, there are xml files...
<clever>
strange, can you add a "set +x" before the echo, then try doing the old and new echo, and ls
<clever>
then gist the whole script and its output?
[0x4A6F] has quit [(Ping timeout: 240 seconds)]
romildo has joined #nixos
[0x4A6F] has joined #nixos
romildo has quit [(Ping timeout: 240 seconds)]
romildo has joined #nixos
acarrico has quit [(Ping timeout: 264 seconds)]
<hyphon81>
clever: Sorry, I was typing miss. I was type "shere" when "share" is correct.
<hyphon81>
nix is no problem.
Kitty- has joined #nixos
cpennington has joined #nixos
acarrico has joined #nixos
matthewbauer has quit [(Remote host closed the connection)]
ryantrinkle1 is now known as ryantrinkle
Kitty- has quit [(Ping timeout: 258 seconds)]
roconnor has quit [(Ping timeout: 246 seconds)]
jensens has quit [(Ping timeout: 255 seconds)]
<clever>
hyphon81: ah, and i see a shara in the string now
Wizek has quit [(Ping timeout: 264 seconds)]
Kitty- has joined #nixos
<hyphon81>
I'm trying to build lightdm unity greeter in my environment. Then, I have to build many ubuntu original packages...
romildo has quit [(Ping timeout: 255 seconds)]
Kitty- has quit [(Changing host)]
Kitty- has joined #nixos
romildo has joined #nixos
<Kitty->
Learning lambdas and trying to figure out why this isn't working like I'd expect. In http://pastebin.ubuntu.com/24514357/ Java says line 25 is not a statement although I'm not sure why this isn't being interpreted (sp) as a statement.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] league opened pull request #25515: zsh-git-prompt: Init at 0.5 (master...zsh-git-prompt) https://git.io/v92U5
NixOS_GitHub has left #nixos []
<Kitty->
Oh wrong channel, ops
* Kitty-
blushes
<gchristensen>
:)
<simpson>
Kitty-: No worries.
sellout- has joined #nixos
lambdamu has joined #nixos
lambdamu_ has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Profpatsch opened pull request #25517: fira-code-symbols: init at 20160811 (master...firacode-symbols) https://git.io/v92TO
NixOS_GitHub has left #nixos []
sellout- has quit [(Quit: Leaving.)]
mbrgm has quit [(Ping timeout: 246 seconds)]
eacamero_ has joined #nixos
filterfish has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
mbrgm has joined #nixos
noexcept_ has quit [(Ping timeout: 260 seconds)]
drp has joined #nixos
acarrico has quit [(Ping timeout: 240 seconds)]
<drp>
in package creation you can have src = fetchurl
<drp>
can src = anything else?
<drp>
like git
acarrico has joined #nixos
matthewbauer has joined #nixos
takle has quit [(Remote host closed the connection)]
<johnramsden>
Does anyone know, if 'environment.etc.<name>' has a directory as its target, is it possible to access the text of a file inside that directory?
<inflames>
my whole issue started here: Error loading QML file: file:///nix/store/rfrkbacw3sdx2k9qcikb7g3j6fvwdnj9-plasma-nm-5.8.6/share/plasma/plasmoids/org.kde.plasma.networkmanagement/contents/ui/main.qml:23:1: module "org.kde.plasma.networkmanagement" is not installed
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 3 new commits to master: https://git.io/v92Za
Guest92624 has quit [(Killed (verne.freenode.net (Nickname regained by services)))]
katyucha_ is now known as katyucha
katyucha_ has joined #nixos
<johnramsden>
Does anyone know, if 'environment.etc.<name>' has a directory as its target, is it possible to access the text of a file inside that directory?
<clever>
that one seems to make a list of files though, not a dir
<johnramsden>
clever: buildtime I guess. Specifically I want to change the text inside /etc/zfs/zed.d/zed.rc. I would like to access the text option using environment.etc but it seems like the target is the directory zfs/zed and not the actual file. I'm not sure if I can access that text.
<clever>
ah
<clever>
that sounds tricky, what about just editing the module that originaly made the entry?
<inflames>
clever, I started reading the urls you suggested, but so far I've had almost no time, but I'm dedicated to learning nix for my own development and work
<johnramsden>
clever: I didn't make the original entry, it's part of zfs.
<johnramsden>
clever: Do you mean edit ${packages.zfsUser}
<clever>
johnramsden: yeah, looks like you would need to override zfsUser or maybe environment.etc."zfs/zed.d".source = mkForce something
jacob has joined #nixos
systemfault has joined #nixos
<clever>
with mkForce, you would replace the entire zed.d dir, and need to make new contents
jacob is now known as Guest73740
inflames has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
<johnramsden>
clever: So if I overrode the pkg.zfsUser I would just add an extra section where I would edit the file? I'm not sure if I should replace the whole directory because zed.d needs to be there. If I were to override it, would that mean I would need to reproduce the entire package, or does it just add to it?
takle has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
MichaelRaskin has quit [(Quit: MichaelRaskin)]
matthewbauer has quit [(Ping timeout: 255 seconds)]
<clever>
johnramsden: you might be able to make a new package, that copies ${packages.zfsUser}/etc/zfs/zed.d to $out, then modifies the copy
<clever>
johnramsden: then use mkForce to put the modifiec copy into /etc
<clever>
johnramsden: that would update faster then an override on zfsUser, which would re-compile all the tools every time you change the config
takle has quit [(Ping timeout: 264 seconds)]
<clever>
johnramsden: what does that zed config file do anyways?
<avn_>
clever: btw, do you know, how to bypass kernel version asserting? (I want to build zfs master ;))
<clever>
avn_: can you paste the exact error its giving?
<clever>
johnramsden: ah, neat
simukis has quit [(Ping timeout: 240 seconds)]
zeus_ has joined #nixos
pxc has quit [(Ping timeout: 246 seconds)]
inflames has joined #nixos
ertes-w has joined #nixos
zeus_ has quit [(Remote host closed the connection)]
<johnramsden>
clever: There are other files in /etc/zfs/zed.d, if I were to modify it would it not affect those files?
<Myrl-saki>
I think I can't build nix.
<Myrl-saki>
On the master.
<clever>
johnramsden: i believe the mkForce would replace the entire zed.d, so you would need to copy all of that from the original to your custom package
<Myrl-saki>
And I don't have git either.
<Myrl-saki>
Maybe I should consider sftping to my laptop.
<clever>
Myrl-saki: why are you trying to build nix from source?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 opened pull request #25524: environment: remove lib/kde4/libexec from PATH (master...env) https://git.io/v92nb
NixOS_GitHub has left #nixos []
<Myrl-saki>
clever: It's building on its own.
<johnramsden>
clever: I see. They are symlinks though.
<Myrl-saki>
I'll use a commit that I've successfully used.
<Myrl-saki>
And see if the problem persists.
<clever>
johnramsden: one min
katyucha has quit [(Quit: Lost terminal)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/v92cB
<NixOS_GitHub>
nixpkgs/master 96c0a6d Jaakko Luttinen: pythonPackages.phpserialize: init at 1.3 (#25521)...
<johnramsden>
clever: Cool thanks. I was just wondering about the sim links that the system has put into /etc/zfs/zed.d/. Would this not remove them?
<johnramsden>
*symlinks
<clever>
that copies the symlinks over to the new config
<clever>
so they remain pointing to the same places as before
takle has quit [(Ping timeout: 240 seconds)]
<johnramsden>
clever Ahh, nice!
<clever>
if you want to edit one of the files behind a symlink, you will need to cat symlinkname > symlink.tmp ; rm symlinkname; mv symlink.tmp symlinkname
<clever>
then edit the now seperate file
<clever>
sed/cat/echo
goibhniu has joined #nixos
<johnramsden>
clever: Do I need any of the part at the bottom besides environment.etc."zfs/zed.d".source = lib.mkForce newcfg;?
<clever>
nope
freusque has joined #nixos
<clever>
that was just to convince nixos that it would be bootable, so i could build a test image
<clever>
you should already have similiar config
<johnramsden>
clever: Yep. Thanks for the help. I'll give that a shot.
<clever>
[clever@amd-nixos:~/nix-tests/zed]$ nix-build -I nixos-config=./configuration.nix '<nixpkgs/nixos>' -A system && ls -lh result/etc/zfs/zed.d/
<clever>
and with a command like this, you can easily test things without any effect on the host
Itkovian has joined #nixos
judson_work has quit [(Ping timeout: 246 seconds)]
<johnramsden>
clever: Cool. Have a good night!
johnramsden has quit [(Quit: Leaving)]
jgertm has quit [(Ping timeout: 240 seconds)]
marsel has joined #nixos
DrWaste has joined #nixos
goibhniu has quit [(Ping timeout: 240 seconds)]
<sphalerite>
clever: yes, it's a zfs system
<sphalerite>
and it says 4.6G
<sphalerite>
So I'm guessing that's what's causing it
<clever>
sphalerite: yeah, thats the zfs arc, they have their own block read cache
<sphalerite>
Ah ok. But that memory is actually available for applications if it becomes necessary?
<clever>
sphalerite: zfs will shrink the size automaticaly based on usage, ive found its a bit too aggressive at that, and it often cuts it down to 32mb of cache
<clever>
and then perf suffers from a lack of cache
<sphalerite>
Ah
<sphalerite>
Well, thanks :) good to know it's not a bug or anything
<clever>
there are config options to set both the min and max size
<clever>
it defaults to a max of 50% of your ram
Infinisil has joined #nixos
<sphalerite>
I'll just stick with the defaults for now, and tweak if I'm bored or have read performance issues. Thanks again!
jensens has joined #nixos
<clever>
sphalerite: you can also use the arcstat script to monitor the cache hit rates and cache size over time
indi_ has quit [(Remote host closed the connection)]
<clever>
sphalerite: c is the target arc size, and arcsz is the actual size
indi_ has joined #nixos
<clever>
in my case, i have 700mb in my arc, and it wants to reduce it to 32mb
takle has joined #nixos
<Infinisil>
Good morning! I got a question: How is a packages config attribute merged with configuration.nix? Because somehow an option from the former doesn't have any effect :/
<Infinisil>
I'm talking about environment.variables, which works in configuration.nix but not in the config attribute
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/v92WJ
Infinisil has quit [(Read error: Connection reset by peer)]
oida_ has joined #nixos
<clever>
ah
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gebner pushed 1 new commit to master: https://git.io/v928e
<NixOS_GitHub>
nixpkgs/master 8b6ea10 Gabriel Ebner: vscode: 1.11.2 -> 1.12.1
NixOS_GitHub has left #nixos []
<clever>
the old one appears to create a dedicated user, and runs the ipfs daemon, so you dont need +w to the dir
oida has quit [(Ping timeout: 268 seconds)]
<Infinisil_>
Where do you see +w?
<clever>
with your code, it creates a wrapper that directs ipfs to write to a single common directory
<clever>
no matter what user you run it as
<clever>
do they all have write to that dir?
amosbird has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 opened pull request #25528: nodePackages: split into v4/v6 package set (master...node-packages) https://git.io/v928w
NixOS_GitHub has left #nixos []
<Infinisil_>
Well yeah kind of, it's like a decentralized internet, it wouldn't make sense to have mutliple directories per user since the content is identified by hash (you'd just have duplicates)
<clever>
but much like the nix store, you need somebody to police access to the directory
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 closed pull request #25006: nodePackages_4_x: remove in favor of nodePackages_6_x (master...nodePackages) https://git.io/vSABs
NixOS_GitHub has left #nixos []
<clever>
or i can just ignore the rules and mess with the content of the ipfs directory, if i have my own user on your box
Infinisil_ has quit [(Read error: Connection reset by peer)]
Infinisil_ has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
<Infinisil_>
But if the directory has owner ipfs:ipfs only people in the group ipfs can use it, which has to be allowed by a root user
<Infinisil_>
Assuming the trusted users don't mess up, hmm..
<avn_>
clever: Exact error -- `Linux v4.11 is not yet supported by zfsonlinux v0.7.0-rc3` I can ofc just patch zfs/default.nix ofc (I patch it anyway to promote to current master)
<avn_>
(error is correct for rc3, but current master have 4.11 support (but idk if someone really test it ;))
<clever>
avn_: i think you need to call common with different arguments
<clever>
avn_: and common is in a let block, so the only way to do that is to modify the file
<clever>
no way to override things properly
Infinisil has joined #nixos
takle has joined #nixos
<Infinisil>
Ah hey clever I know why it's not working, I declared wrapper as a script that creates the wrapper at runtime, what I really need is that at build time
<clever>
the example in the gist you linked will make the wrapper at nixos-rebuild time
<Myrl-saki>
clever: Finally done with the re-installation.
<Myrl-saki>
clever: Can I ask how to do the copy closure?
<clever>
Myrl-saki: check the nix-copy-closure man page first
<Myrl-saki>
Thanks.
derjohn_mob has joined #nixos
eacameron has joined #nixos
thc202 has joined #nixos
<seanparsons>
Has anyone got an example of adding plugins to pidgin? I'm getting an infinite recursion error that I assume is more induced by my snot filled brain than anything else.
eacameron has quit [(Ping timeout: 245 seconds)]
derjohn_mob has quit [(Ping timeout: 268 seconds)]
goibhniu has joined #nixos
<Infinisil>
clever: How can I make makeWrapper available in my runCommand script?
<clever>
Infinisil: change the {} from my example to { buildInputs = [ pkgs.makeWrapper ]; }
<sphalerite>
How can I get the build-time closure of a package?
<sphalerite>
As in, compute and print it
<clever>
sphalerite: nix-store -qR on the .drv file that nix-instantiate spits out
<clever>
or nix-store --query --tree
<sphalerite>
awesome, thanks
<Infinisil>
clever: Thanks!
nslqqq has quit [(Ping timeout: 264 seconds)]
<Myrl-saki>
clever: How to know which library is currently in use?
<Myrl-saki>
which derivation for a library is current*
<clever>
Myrl-saki: nix-copy-closure gets all dependencies of a given storepath
<Myrl-saki>
clever: I get `imported archive of â/nix/store/98s2znxww6x7h2ch7cj1w5givahxmdna-glibc-2.23â lacks a signature at /run/current-system/sw/bin/nix-copy-closure line 104.
<clever>
Myrl-saki: are you using --to or --from?
<Myrl-saki>
Literally the same git commit.
<Myrl-saki>
clever: I'm using from.
<clever>
run it as root and it will accept unsigned things
<Myrl-saki>
Oh.
<Myrl-saki>
If I use `to`, do I use `root@...`?
<clever>
yep
<Myrl-saki>
Thanks.
<Myrl-saki>
clever: Also, I literally have 2 the same versions with a different hash, how to know which is which?
<Myrl-saki>
(To be more exact, which one should I copy?)
<clever>
Myrl-saki: there is another related option, let me find it
seanparsons has quit [(Quit: WeeChat 1.7.1)]
<clever>
Myrl-saki: i think its nix-build --option ssh-substituter-hosts user@host ....
<clever>
that will automaticaly try to copy things from user@host before building them
<Myrl-saki>
Thanks.
<Myrl-saki>
Oh right. Is there a more reliable way to remote cross compile now?
<Myrl-saki>
IIRC, if the build command was very long, it wouldn't work?
<Infinisil>
Should I always do PR's for master?
<toppler>
I think I'm having trouble setting up a clang development environment, getting error messages "Could NOT find LibClang". http://lpaste.net/355223
<toppler>
I've using pkgs.clangStdenv.mkDerivation. Is that all I need?
takle has quit [(Remote host closed the connection)]
endformationage has quit [(Quit: WeeChat 1.7)]
nslqqq has joined #nixos
reinhardt has joined #nixos
goibhniu1 has joined #nixos
goibhniu has quit [(Ping timeout: 240 seconds)]
reinzelmann has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] qknight pushed 1 new commit to release-17.03: https://git.io/v92uw
<NixOS_GitHub>
nixpkgs/release-17.03 977b320 Bas van Dijk: wordpress: replace the dbPassword option with dbPasswordFile...
<ronny>
qknight_: isnt that bascially a pr, also does that sign the guy up for something he didnt agree to?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Infinisil opened pull request #25531: ipfs service: Fix dataDir being ignored (master...fix/ipfs-dataDir) https://git.io/v922w
NixOS_GitHub has left #nixos []
<Infinisil>
Yay, second PR
Raimondii has joined #nixos
Raimondi has quit [(Ping timeout: 268 seconds)]
Raimondii is now known as Raimondi
<toppler>
Damn. Didn't realise that irony-server was already packaged.
Infinisil_ has joined #nixos
Infinisil has quit [(Ping timeout: 268 seconds)]
Infinisil_ has quit [(Client Quit)]
takle has joined #nixos
<LnL>
the wrapper might not work properly unless the stdenv also uses clang
<toppler>
LnL: Ack, sorry. I'm not sure what that means.
mudri has quit [(Ping timeout: 260 seconds)]
<toppler>
Does it mean that my system basically needs to be clang based?
<LnL>
clang.cc is the actual clang package, clang just contains the wrapper scripts
<LnL>
so if your build uses gcc to compile it won't find the clang library if you use the wrappers as build inputs
Wizek has joined #nixos
<toppler>
Thanks. It was compiling using clang, but still wasn't finding the library. I was only trying to compile irony-server, but it turns out this is already packaged, and based on the derivation, they're manually adding "-DCMAKE_PREFIX_PATH=${llvmPackages.clang.cc}" to the CMAKE flags.
<gchristensen>
the # of checkboxes you checked off
<gchristensen>
useful if they're a todo list
<hodapp>
ohhh
<hodapp>
I'm just not sure if I'm confusing people when I force-push a merged commit that addresses someone's comments
<hodapp>
as then the comments/review are a bit buried
<gchristensen>
that is pretty common behavior
<hodapp>
but I thought it was standard to rebase/amend to try to keep things to a single commit in a PR
<gchristensen>
here
<hodapp>
good to know
jophish has quit [(Ping timeout: 258 seconds)]
pie_ has quit [(Ping timeout: 264 seconds)]
mudri has quit [(Ping timeout: 255 seconds)]
derjohn_mob has joined #nixos
* gchristensen
gets too much mail :|
* hodapp
sends 1,300 pointless emails to gchristensen
* ikwildrpepper
should send an email to gchristensen thanking him for his hard work
<ben__>
git is kinda insufficient for representing changes in a changeset over time
<ben__>
youd need some sorta meta-git......
<gchristensen>
hodapp: they'd fit right in :P *eyeroll*
<gchristensen>
ikwildrpepper: that one would be a delight ;P
<gchristensen>
ben__: eh?
<hodapp>
ben__: seems to me this is the normal purpose of commits
<hodapp>
it just happens that in PRs it's often preferred that it be a cleaned-up, single commit
<ben__>
If I have a branch and rebase it a bunch of times, different versions of the branch aren't a first-class object in either git or github
<gchristensen>
hmm branches are just a pointer
<ben__>
yeah, it's a bit primitive
<hodapp>
hmmm
<gchristensen>
now if a branch was a commit object! now we're cooking!
<hodapp>
do hg, darcs, or bzr do this any differently?
<ben__>
im saying git because it's the only one i have measurably familiarity with ;)
<LnL>
yeah, mercurial embeds the branch metadata inside commits
<gchristensen>
yeah, a commit is always tied at some level to its branch
<gchristensen>
in hg
<hodapp>
interesting
<hodapp>
I haven't really used hg
<ben__>
theres phabricator where you submit changesets and if you submit a new changeset it shows differences to the prev changeset in the pull request or w/e they call it
<ben__>
but it kinda loses the commit granularity, afaik, so it's not very nice in that dimension
<hodapp>
so... a commit in mercurial could consist of, in effect, a rebase? or am I not getting that right
<ben__>
nah i think mercurials commit objects just have a "which branch was this committed to?" label
<hodapp>
but how's this change anything?
<LnL>
mercurial doesn't really have a rebase
<gchristensen>
not sure it does, hodapp
roconnor has joined #nixos
<hodapp>
huh. this makes "some sorta meta-git" seem an effective option.
derjohn_mob has quit [(Ping timeout: 240 seconds)]
obadz has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/v929H
<NixOS_GitHub>
nixpkgs/master af87b92 Frederik Rietdijk: pythonPackages.plotly: 1.9.5 -> 2.0.8
<ronny>
FRidh: any idea how hard it would be to create a "nix-python-virtualenv" command - i keep having to use python virtualenvs, and on nix they tend to break after a gc with painful repeatability
<goibhniu1>
ronny: one option is to use a persistent nix-shell environment, which you can garbage collect independently
<ronny>
dash: i need an actual mutable virtualenv since i need to install a working tox and try certain things
<ronny>
so pypi2nix is very pretty and misses my use-case
<dash>
ah, for that i have never tried to keep 'em around
jensens has quit [(Ping timeout: 240 seconds)]
<ronny>
in the teams i collaborate with, im the only nix user
<ronny>
i'd be a ass if i didnt keep things working for others without enforcing tool choices on them
<gchristensen>
indeed
<gchristensen>
gives nix a bit of a bad name too
Mateon3 has joined #nixos
<ronny>
its actually quite a pain that the nix package set ships some of those tools in a horrible broken manner
<ronny>
for example - python tox is completely mutilated
acarrico has joined #nixos
noexcept1 has joined #nixos
ilyaigpetrov has joined #nixos
mbrgm_ has joined #nixos
Mateon3 is now known as Mateon1
mbrgm_ is now known as mbrgm
leothrix has joined #nixos
<FRidh>
ronny: create a nix function that copies recursively all python packages into a single derivation. Then you can get rid of setting PYTHONPATH/PYTHONHOME. White a shell script that nix-build such a derivation, and adds $out/bin to PATH. Then there's no need for wrappers. `nix-build` creates the root for you.
katyucha_ has joined #nixos
jensens has joined #nixos
sellout- has quit [(Quit: Leaving.)]
orivej has joined #nixos
DIzFer has joined #nixos
freusque has joined #nixos
Mic92 has joined #nixos
bitonic has joined #nixos
taeradan has joined #nixos
jophish has joined #nixos
JagaJaga has joined #nixos
stepho has joined #nixos
Kitty- has joined #nixos
dridus has joined #nixos
cpennington has joined #nixos
Kitty- has quit [(Ping timeout: 258 seconds)]
DIzFer has joined #nixos
bitonic has joined #nixos
lukego has joined #nixos
jophish has quit [(Ping timeout: 240 seconds)]
<lukego>
Getting an error I don't understand. I take a derivation, update it with overrideDerivation, and end up with error 'attribute ‘drvAttrs’ missing, at .../customisation.nix'. don't get the error without the override. familiar?
derjohn_mob has joined #nixos
dridus has quit [(Quit: dridus)]
reinhardt has quit [(Remote host closed the connection)]
reinzelmann has joined #nixos
<hyper_ch>
when I try to nixos-rebuild switch --upgrade, I still get this error: error: Package ‘python2.7-os-testr-0.4.2’ in ‘/nix/store/zaphjxhxnrx2cncmbmpaacqv4w503cgf-nixos-17.09pre106916.c5badb123a/nixos/pkgs/development/interpreters/python/mk-python-derivation.nix:56’ is marked as broken, refusing to evaluate. -- but that should be solved in master since yesterday... yet I still get that error in unstable. Do I need to remove something lo
<gchristensen>
hyper_ch: a simple explanation is the fix hasn't been pushed out to unstable yet
<hyper_ch>
oh, I thought this happens automagically when it builds fine on master
<gchristensen>
it does
<hyper_ch>
I thought I might need to remove that non-working store path :)
<hyper_ch>
thx
Isorkin_ has quit [(Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)]
iyzsong has joined #nixos
dridus has joined #nixos
stepho has quit [(Quit: My MacBook Pro has gone to sleep. ZZZzzz…)]
DrWaste has quit [(Quit: Run away!)]
eacameron has joined #nixos
ryantrinkle has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/v92xL
<NixOS_GitHub>
nixpkgs/master c6385eb Matthew Daiter: openmvg: init at v1.1
<NixOS_GitHub>
nixpkgs/master 587ed89 Michael Raskin: Merge pull request #25480 from mdaiter/openmvg...
<cmacrae>
Hey guys, writing an expression which exposes a systemd user service. I need to set the PATH variable for the service so it can find some binaries when it's running. What's the best value I can use for `environment.PATH`? Are there any variables I can use to populate this with the usual dirs?
<gchristensen>
I think you can just do like myservce = { /* your usual systemd service options*/ path = [ extra, paths, go, here]}?
tnias has quit [(Quit: leaving)]
cpennington has quit [(Ping timeout: 255 seconds)]
<cmacrae>
Sure, you can indeed do that. What I'm asking is: is there an existing variable I can use in my expression that would evaluate to the common PATH string?
takle has quit [(Remote host closed the connection)]
<cmacrae>
For instance, I might write it like so: environment.PATH = "${systemPath}"
Isorkin has joined #nixos
<cmacrae>
If 'systemPath' existed as a var in NixOS that evaluated to something like '/home/cmacrae/bin:/run/wrappers/bin:/run/wrappers/bin:/home/cmacrae/.nix-profile/bin:/home/cmac'....
MercurialAlchemi has quit [(Ping timeout: 240 seconds)]
<mudri>
Where are the Haskell packages built with profiling enabled?
<cmacrae>
Basically I'm just asking if there's a shorthand variable in Nix that would evaluate to a common PATH string :p
<mudri>
Using haskellPackages.ghcWithPackages, ghc gives me a “Perhaps you haven't installed the profiling libraries for package ...” error.
erasmas has joined #nixos
darlan has quit [(Remote host closed the connection)]
<niksnut>
cmacrae: generally systemd services should not have directories like /run/current-system/sw/bin in their PATH since that causes undeclared dependencies
<gchristensen>
cmacrae: what is your goal that these paths accomplish?
<niksnut>
however, systemd.services.*.path by default contains a few packages (coreutils, findutils, grep, sed, systemd)
<cmacrae>
Hmm, alright, thanks niksnut
<cmacrae>
gchristensen: I'm launching a binary which uses an exec. It's a FUSE component that's trying to execute fusermount, but it can't find it in the PATH - so I need a way to tell it where it is
<gchristensen>
cmacrae: hmmm sounds like you need to add whatever provides fusermount to the path
<gchristensen>
that is a setuid wrapper though... not sure how to handle this, than adding /run/wrappers/bin to the path.
<cmacrae>
Yep, exactly
<cmacrae>
Just figured I'd ask if there was an already existing solution to this sorta thing, rather than hardcoding a PATH value
cpennington has joined #nixos
stepho has joined #nixos
georges-duperon has joined #nixos
dfranke has joined #nixos
stepho has quit [(Ping timeout: 240 seconds)]
<cmacrae>
Okay, cool - so I'm trying to use 'systemd.user.services.<name>.path' to add the fuse package's bin & sbin paths to my service PATH
<cmacrae>
I'm using 'path = [ pkgs.fuse ];'
<cmacrae>
But it doesn't seem to be adding the paths
orivej has quit [(Ping timeout: 255 seconds)]
orivej has joined #nixos
<gchristensen>
does anyone have upstart versions of the systemd services for nix-daemon? :)
mizu_no_oto has joined #nixos
<gchristensen>
maybe I should spelunk in the nix source for when nixos used upstart ... (did nixos use upstart ever?)
acertain has quit [(Remote host closed the connection)]
ertes has quit [(Ping timeout: 240 seconds)]
ryanartecona has joined #nixos
ertes-w has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
katyucha has quit [(Quit: leaving)]
lukego has joined #nixos
<lukego>
Just a passing thought: Can you put a breakpoint into a shell script running in a derivation? It would be awfully nice to drop to an interactive shell at exactly the point where things are going wrong in my build
<lukego>
(with nix-shell I would need to step through all the phases manually to reach the point of interest, right?)
<LnL>
lukego: nix-shell -A hello and run genericBuild
<LnL>
that will go through the entire builder
<lukego>
Thanks for the tip, LnL! do you also have a tip for if the place that I want to stop is in a dependency rather than the derivation I name with -A ?
<cmacrae>
Okay, looks like I just wasn't running 'systemctl --user daemon-reload' so my systemd manifest wasn't updating :) My mistake for assuming this would happen naturally. Now to work out some other kinks!
<LnL>
not sure what you mean, do you want to debug the build of a dependency?
<lukego>
yep
<lukego>
really I have a fairly complex network of derivations, some of which are really parts of others, and are not that easy to address. it's not nixpkgs but a nix-based application.
<LnL>
then just do the same but with the attribute of the dependency instead
<lukego>
I don't think it has an attribute really.
<lukego>
or maybe it does... but seems like it wouldn't necessarily, since derivations can be hooked up using closures rather than attribute sets
<LnL>
you can also use nix-build -k if you want to keep the build directory around
<lukego>
hm, yeah, that is probably a better option actually
freusque has quit [(Quit: WeeChat 1.7)]
<lukego>
I'd forgotten about that, thanks for the reminder :)
derjohn_mob has quit [(Ping timeout: 268 seconds)]
pie_ has joined #nixos
<sphalerite>
I've got a server here with two smaller and two larger hard drives. I want to set it up with LVM and some form of redundancy because I don't trust the hard drives that much. Any suggestions for a partitioning setup?
<gchristensen>
what are the sizes & specs?
pie_ has quit [(Ping timeout: 240 seconds)]
noam has quit [(Read error: Connection reset by peer)]
noam has joined #nixos
metaphysician has joined #nixos
goibhniu has quit [(Ping timeout: 240 seconds)]
olejorgenb has joined #nixos
Tucky has quit [(Remote host closed the connection)]
cpennington has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 2 new commits to master: https://git.io/v9aYf
<NixOS_GitHub>
nix/master 465cb68 Eelco Dolstra: Figure out the user's home directory if $HOME is not set
<NixOS_GitHub>
nix/master bb50c89 Eelco Dolstra: Make the location of the build directory in the sandbox configurable...
NixOS_GitHub has left #nixos []
<sphalerite>
gchristensen: 18GB/10k RPM and 70GB/???. The server is a PowerEdge 4600. Basically ancient tech that was donated to the computing society at my uni :p
Kitty- has joined #nixos
<gchristensen>
hmm I'd probably RAID-1 the 18gb and RAID-1 the 70GB
cpennington has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gnidorah opened pull request #25536: winusb: init at 2017-01-30 (master...winusb) https://git.io/v9aOZ
NixOS_GitHub has left #nixos []
Kitty- has quit [(Ping timeout: 240 seconds)]
mbrgm has quit [(Ping timeout: 260 seconds)]
justelex_ has joined #nixos
marsel has quit [(Ping timeout: 240 seconds)]
MichaelRaskin has joined #nixos
gloaming has joined #nixos
bin7me has quit [(Remote host closed the connection)]
simukis has joined #nixos
acertain has joined #nixos
jgertm has quit [(Ping timeout: 260 seconds)]
lukego has quit [(Ping timeout: 260 seconds)]
civodul has quit [(Quit: ERC (IRC client for Emacs 25.2.1))]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin pushed 1 new commit to openssl-1.1: https://git.io/v9aZu
<NixOS_GitHub>
nixpkgs/openssl-1.1 3d4ba40 Robin Gloster: Merge remote-tracking branch 'upstream/master' into openssl-1.1
NixOS_GitHub has left #nixos []
digitus has joined #nixos
ryanartecona has quit [(Quit: ryanartecona)]
pie_ has joined #nixos
darlan has joined #nixos
orivej has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] judah opened pull request #25537: haskell: work around linker limits on Mac OS X Sierra. (master...sierra-haskell) https://git.io/v9aCf
NixOS_GitHub has left #nixos []
<codedmart>
Is there a exit hook from nix-shell? So I can run a command when I exit nix-shell?
<gchristensen>
you can use bash's `trap`
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin pushed 1 new commit to openssl-1.1: https://git.io/v9aC6
<NixOS_GitHub>
nixpkgs/openssl-1.1 dd2f2bc Robin Gloster: aircrack-ng: add unreleased patch for openssl 1.1
NixOS_GitHub has left #nixos []
darlan has quit [(Remote host closed the connection)]
reinzelmann has quit [(Quit: Leaving)]
cmacrae has quit [(Ping timeout: 264 seconds)]
<codedmart>
gchristensen: That is great thanks! Have never used trap but works great.
<gchristensen>
well do I have a treat for you
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin pushed 1 new commit to openssl-1.1: https://git.io/v9aWO
<NixOS_GitHub>
nixpkgs/openssl-1.1 c936bc8 Robin Gloster: httrack: 3.48.22 -> 3.49.1...
<spluko>
I can't figure out how to use ghc-mod in nixos. From what I've found it seems nix-shell -p "haskellPackages.ghcWithPackages (pkgs : with pkgs; [ghc ghc-mod shake])" --pure --command "ghc-mod list | grep Shake" should list the shake packages, but it doesn't find them. What am I not understanding?
orivej has joined #nixos
<nixy>
I hope the wiki comes down soon. I am tired of being led astray
<gchristensen>
as soon as niksnut merges those PRs
ryantrinkle has joined #nixos
DIzFer has quit [(Remote host closed the connection)]
tokudan has joined #nixos
sellout- has joined #nixos
freusque has joined #nixos
DIzFer has joined #nixos
spluko has quit [(Quit: Page closed)]
jsgrant_om has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
ryantrinkle has quit [(Ping timeout: 258 seconds)]
<hodapp>
hrmph, why does nix-collect-garbage seem to lose environments in shell.nix that I'm still using?
<hodapp>
sometimes, literally in the middle of using
matthewbauer has joined #nixos
mizu_no_oto has joined #nixos
pie_ has quit [(Quit: Leaving)]
taktoa has joined #nixos
jmeredith has quit [(Quit: Connection closed for inactivity)]
Kitty- has joined #nixos
freusque has quit [(Ping timeout: 246 seconds)]
ryanartecona has quit [(Quit: ryanartecona)]
simp__ has joined #nixos
mkoenig_ has joined #nixos
mojjo has joined #nixos
mkoenig has quit [(Ping timeout: 255 seconds)]
freusque has joined #nixos
ryantrinkle has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/v9aX7
<NixOS_GitHub>
nixpkgs/master 959695d Frederik Rietdijk: pythonPackages.grequests: 0.2.0 -> 0.3.0
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh opened pull request #25540: pythonPackages.requests: point to requests2 (master...requests) https://git.io/v9aXp
NixOS_GitHub has left #nixos []
wak-work has quit [(Remote host closed the connection)]
<nixy>
hodapp: I don't think there is a garbage collection root for any nix-shell environments even while in use
georges-duperon has quit [(Ping timeout: 240 seconds)]
wak-work has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 2 new commits to master: https://git.io/v9aMv
<NixOS_GitHub>
nixpkgs/master f63eb58 Frederik Rietdijk: pythonPackages.requests: point to requests2...
<NixOS_GitHub>
nixpkgs/master ef1149e Frederik Rietdijk: pythonPackages.umemcache: mark as broken
NixOS_GitHub has left #nixos []
freusque has quit [(Ping timeout: 260 seconds)]
<asymmetric>
hey, what's the `platforms` for a python application that runs on linux/osx/win ?
jmeredith has joined #nixos
<nixy>
I think that would be platforms.all
<asymmetric>
thx, makes sense
derjohn_mob has joined #nixos
<asymmetric>
and what is better: fetchurl or fetchPyPi?
<LnL>
platforms.all also includes bsd, but generally if something works in cygwin it works everywhere
<LnL>
I think fetchPyPi is preferred, you just have to give the pname/version for that
nh2 has quit [(Ping timeout: 258 seconds)]
nh2 has joined #nixos
acowley2 has quit [(Ping timeout: 245 seconds)]
matthewbauer has quit [(Remote host closed the connection)]
matthewbauer has joined #nixos
<ronny>
hmm
<asymmetric>
thx
<alibabzo>
Anyone here managed to build intero with stack recently?
<alibabzo>
It's having a problem finding the curses dependency but I've tried putting curses into global config, local config, global stack config, local stack config and nothing is working
<simpson>
FRidh: Are there any wishlist refactorings that you want but don't have time to do? I'm willing to chip in a couple hours of refactoring Python packages.
<FRidh>
simpson: I don't know what you would like to work on, but I've been trying to write a Python-specific override function but couldn't get it working. https://github.com/NixOS/nixpkgs/pull/24155
mbrgm has joined #nixos
<simpson>
FRidh: I just want to ease the maintenance burden, maybe pay off some longstanding tech debt.
<asymmetric>
how do i remove something from /nix/store that was put there by nix-build?
<asymmetric>
i want to force a re-build, even though the definition hasn't changed
<gchristensen>
asymmetric: you can pass `--check` to `nix-build` to rebuild itn
<MichaelRaskin>
nix-store --delete
<LnL>
you can just use nix-build --check then
<simpson>
asymmetric: Curious about the larger context. What happened?
<asymmetric>
--check is not mentioned in `man nix-build`.. weird
<gchristensen>
sounds like a bug in the making :)
<asymmetric>
simpson: nothing important, i just deleted `result/bin` :)
<gchristensen>
oh just nix-build again and it should restore your ./result link?
<FRidh>
simpson: another import improvement would be an update script that updates the version and hash of packages in pkgs/python-modules/<package>/default.nix
<gchristensen>
nix-store --delete as MichaelRaskin said
<asymmetric>
error: cannot delete path ‘/nix/store/a1qbh2bq1f3nw9xnn1s9bz5xjkp7754c-electrum-ltc-2.6.4.2’ since it is still alive
takle has joined #nixos
<simpson>
FRidh: Exciting. I presume that you'd like that incrementally, so I won't stay up all night making a one-off 30000-line patch.
<MichaelRaskin>
delete result symlink
<gchristensen>
delete your result symlink
<MichaelRaskin>
This extra «your» cost you tempo loss
<gchristensen>
:)
<FRidh>
simpson: Unless you find a reliable way to just do it automatically I think the best approach here is indeed incrementally, e.g. together with a package update
<asymmetric>
still doesn't work.. is this what i should run? `nix-store --delete /nix/store/a1qbh2bq1f3nw9xnn1s9bz5xjkp7754c-electrum-ltc-2.6.4.2`
<asymmetric>
by the way /nix/store/a1qbh2bq1f3nw9xnn1s9bz5xjkp7754c-electrum-ltc-2.6.4.2 is an empty directory
takle has quit [(Ping timeout: 240 seconds)]
* gchristensen
isn't sure how to fix this
<LnL>
something else is keeping it alive
<simpson>
Probably the result symlink?
<simpson>
What's the query, $ nix-store --query --referrers /nix/store/a1qbh2bq1f3nw9xnn1s9bz5xjkp7754c-electrum-ltc-2.6.4.2
<LnL>
I think there's a symlink in the gcroots that points to the result symlink
<LnL>
that's how it's a root
<LnL>
if you run nix-store --gc those stale roots will get garbage collected
takle has joined #nixos
<LnL>
asymmetric: what nix-build command dis you run? I'ts probably easier to figure out what drv you should build with --check
<asymmetric>
LnL: i thought what was keeping it alive was that the application was actually running
<asymmetric>
d'oh
<asymmetric>
but even after quitting it, it still cannot remove with --delete
<LnL>
ah yes
takle has quit [(Ping timeout: 240 seconds)]
<asymmetric>
ok, it might have worked
<asymmetric>
i just needed to run all the steps in order
phreedom has quit [(Ping timeout: 240 seconds)]
<asymmetric>
quit app; delete symlink; run --delete; build again
<LnL>
a running process also keeps a path alive, deleting it could cause all sorts of strange issues
<asymmetric>
thanks for the help guys
Jackneilll has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] asymmetric opened pull request #25544: Electrum-LTC: init at 2.6.4.2 (master...electrum-ltc) https://git.io/v9apJ
NixOS_GitHub has left #nixos []
matthewbauer has quit [(Ping timeout: 240 seconds)]
<MichaelRaskin>
gchristensen: I have reasons to believe that if _these_ people would know something about systemd, their hate would not be any less
<jmg_>
MichaelRaskin: thanks, I'll try that. Could you give me a hint why it is not just part of the nixpkgs collection?
<jesper>
hi. anyone got pointers as to how i could load a nixos module from unstable? i found how to for packages but it's not straightforward how to for modules and i don't find anything on the web.
<MichaelRaskin>
Well, it is just too much NixOS-specific
<MichaelRaskin>
So it lives inside nixos/ subtree in a directory with a few scripts relevant only to NixOS
<gchristensen>
MichaelRaskin: yeah but knowing it lets you really hate it
<MichaelRaskin>
It definitely lets me
<jmg_>
I am trying to create an installation image for Scaleway (a cloud provider), and all usual methods of installation do not work.
<MichaelRaskin>
But these people would have a head start on me w.r.t. of hat intensity, I guess
<MichaelRaskin>
Some of the scaleway offers are interesting, yes
<khoi>
hi! Why does imapfilter prompt me to accept certificates? It doesn't do that on other distributions. Thanks :)
<jmg_>
It seems to work. Just for curiosity, who would I add that to the current profile? Your command just installed it into the nix-store. I know I can use it from there
<MichaelRaskin>
Isn't this a command that you need to install once?
<MichaelRaskin>
This greps inside Nix, so this all can be put into a Nix file and used however you want
<MichaelRaskin>
nix-env -i -f
<MichaelRaskin>
Or whatever
<MichaelRaskin>
I am not sure I am willing to explain what is going on without your reading the Nix language manual first (not NixOS manual), and I don't see anything that reading the language manual should not make clear.
tokudan has quit [(Quit: Leaving)]
Filystyn has joined #nixos
takle has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
arnm has joined #nixos
bennofs1 has joined #nixos
<arnm>
LnL: ever tried to get mysql service up and running?
calvertvl has joined #nixos
<LnL>
not with nix I think
pxc has joined #nixos
<LnL>
no wait, I have a nixos server that runs a mysql db
<calvertvl>
looks like qtbase-5.8.0 in current master is failing on darwin (travis build failure)
takle has joined #nixos
<jmg_>
MichaelRaskin: thanks for your help
<arnm>
LnL: mysqld is running properly? where could I see an example of that configuration?
takle has quit [(Ping timeout: 240 seconds)]
<gchristensen>
arnm: I have several dozen mysql deployments, what problem are you finding?
acertain has quit [(Remote host closed the connection)]
<LnL>
anyway, darwin is a different story I have a project that implements some basic modules for osx (including launchd)
<LnL>
but that only has a few actual services so you'd have to rewrite the nixos module
<gchristensen>
back in a while
hiratara has quit [(Ping timeout: 255 seconds)]
hiratara has joined #nixos
<ericnoan>
anyone know of a packaged node app i can copy?
<LnL>
alternatively you could probably just write a regular service plist that uses the mysql version of you profile or something
<mbrgm>
!seen fpletz
<ericnoan>
nodejs
takle has joined #nixos
acarrico has quit [(Ping timeout: 240 seconds)]
JagaJaga has quit [(Ping timeout: 255 seconds)]
takle has quit [(Ping timeout: 260 seconds)]
mkoenig_ has quit [(Remote host closed the connection)]
<sphalerite>
Is it just me or would it be very useful to actually have some sort of statistics reporting? For example, I suspect from the fact that I discovered the issue with the firefox module when I wanted to start using it that nobody is actively using it at the moment.
<sphalerite>
Similarly for amd-hybrid-graphics — nobody else was even able to test my changes
<sphalerite>
Or perhaps we just need more tests :)
<sphalerite>
But actually... Shouldn't the build failure have appeared on hydra as well?
MichaelRaskin has quit [(Remote host closed the connection)]
svij1 has joined #nixos
simp__ has quit [(Quit: Connection closed for inactivity)]
<svij1>
hi, i'm trying to override buildInput for php memcached package but i get an error that memcached doesn't have overrideAttrs attribute. documentation suggests that it should be there -- https://nixos.org/nixpkgs/manual/#sec-pkg-overrideAttrs
<svij1>
does anyone know why it is not there?
MP2E has joined #nixos
<sphalerite>
svij1: what exactly are you writing, and where?
<gchristensen>
sphalerite: the major obstacle is we would need to have that system be built, with an awareness that there are many users with a deep desire for privacy
<sphalerite>
gchristensen: yeah, I have the feeling that there's quite a bit of stuff in nixpkgs that nobody uses and as such it's prone to bitrot
<gchristensen>
certainly
<gchristensen>
that said, the fact that hyrda builds our entire tree all the time points it out way faster than, say, debian can
<gchristensen>
as debian depends (or did) on contributors compiling and uploading, not a central builder
<ekleog>
copumpkin: do I correctly remember? nixos-prepare-root will be out just after nix 1.12, right?
<ekleog>
(and hey!)
hiratara has joined #nixos
<sphalerite>
Yeah, I think a potential good option for that would be forced choice — require a setting for some "data collection" in the system configuration
<sphalerite>
I don't know
<sphalerite>
svij1: that file doesn't have valid syntax.
<aneeshusa>
svij1: You need parens around the (oldAttrs: {}) function
<aneeshusa>
svij1: Also, what version of nixpkgs do you have?
simp__ has joined #nixos
<svij1>
re: syntax, it is an error in copy and paste
<svij1>
re: nixpkgs version, 16.09
exarkun has left #nixos []
<sphalerite>
That's no longer supported, you should upgrade to 17.03
<sphalerite>
overrideAttrs was introduced in 17.03
<nixy>
gchristensen: Bit of a late response on this one, but I have been messing with nix on and off for a year. I feel like I get it, but every once in a while someone says something that blows my mind
<svij1>
sphalerite: hrm, it is present in other packages...ruby, sbt etc.
<gchristensen>
sphalerite: that is another problem for sure
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy pushed 2 new commits to release-17.03: https://git.io/v9Vl5
<NixOS_GitHub>
nixpkgs/release-17.03 92ed6cb Judah Jacobson: haskell: work around linker limits on Mac OS X Sierra....
<NixOS_GitHub>
nixpkgs/release-17.03 f287842 Judah Jacobson: Fix use of `isDarwin` conditionals....
NixOS_GitHub has left #nixos []
<gchristensen>
nixy: I try and comment on ops-y things with a nixos perspective, and would love it if other people did too. I don't want to be the lone voice haha
<svij1>
sphalerite: meh, weird. i wonder what changed in 17.03, but thanks! let me dig further.
<sphalerite>
svij1: It might be 9d55479e07421d1c02c4bbab9508bf5d0913c21b that's key for your case
takle has quit [(Ping timeout: 260 seconds)]
<nixy>
gchristensen: I have seen that going on for a while. I imagine being the sole representative of NixOS can be tiring ;) I'll try to represent the nix perspective lol
<sphalerite>
svij1: but you should upgrade to 17.03 regardless, 16.09 isn't getting security updates anymore
<gchristensen>
or any updates :)
<sphalerite>
well security updates are the last to go in most OSes and the most critical
filterfish has quit [(Remote host closed the connection)]
filterfish has joined #nixos
<sphalerite>
Anyway...
<svij1>
sphalerite: memcached doesn't even has overrideDerivation method. re: security, yeah i'm planning to do it soon. just don't want to rush and break my dev env.
<sphalerite>
If it breaks anything you can roll back :)
<svij1>
the docs say "Most standard derivation-producing functions, such as stdenv.mkDerivation, are defined using this function, which means most packages in the nixpkgs expression, pkgs, have this function."
<NixOS_GitHub>
[nixpkgs] shlevy closed pull request #25537: haskell: work around linker limits on Mac OS X Sierra. (master...sierra-haskell) https://git.io/v9aCf