ttuegel has quit [(Remote host closed the connection)]
<NixOS_GitHub>
[nixpkgs] benley pushed 2 new commits to master: https://git.io/vykk3
<NixOS_GitHub>
nixpkgs/master 59bb3d1 Itai Zukerman: bazel: remove dependence on buildFHSUserEnv
<NixOS_GitHub>
nixpkgs/master 0c3afbd Benjamin Staffin: Merge pull request #23074 from izuk/bazel...
takle has joined #nixos
<unlmtd[m]>
so im in the nix-shell now, cargo build works. If I do cargo install, it puts it in my home folder. Can I do a nix-style install from the nix-shell with this rust application?
edvorg has quit [(Ping timeout: 260 seconds)]
takle has quit [(Ping timeout: 240 seconds)]
<c74d>
unlmtd[m]: AFAIK you'll need to write a Nix expression for it, and it shouldn't matter whether you're in a nix-shell or not
<unlmtd[m]>
I thought I could do stuff like ```mkdir -p $out/bin``` in nix-shell, but it's read-only. So i guess it *has* to be a full nix expression that does the work. ```nix-env -f . -i cobalt-rs``` complains about the source not set. So fetchgit is the only way?
<c74d>
for a GitLab project I'd use fetchFromGitLab
<unlmtd[m]>
then Ill just write an expression in the nixpkgs
<c74d>
which should take the same arguments as fetchFromGitHub
<unlmtd[m]>
ahh no thats my fork, its on github, but thanks
<unlmtd[m]>
ya ill use the original on github
mguentner2 has joined #nixos
zraexy has quit [(Quit: Leaving.)]
mguentner has quit [(Ping timeout: 255 seconds)]
proteusguy has joined #nixos
<unlmtd[m]>
" error: attempt to call something which is not a function but a set, at /home/unlmtd/src/nix/nixp
<unlmtd[m]>
kgs/lib/customisation.nix:56:12"
<c74d>
unlmtd[m]: what's your Nix expression?
<unlmtd[m]>
clone of the one you sent me
<unlmtd[m]>
not much in there
<unlmtd[m]>
gitub sucks
<c74d>
unlmtd[m]: you haven't modified the expression? how are you building it?
<unlmtd[m]>
one second, im trying to push it
takle has joined #nixos
takle has quit [(Ping timeout: 268 seconds)]
takle has joined #nixos
smw_ has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
Wizek has quit [(Ping timeout: 268 seconds)]
alx741 has quit [(Quit: alx741)]
dramforever has quit [(Ping timeout: 260 seconds)]
cybrian has quit [(Quit: cya)]
cybrian has joined #nixos
<tnks>
how simple is it to nix-build a Linux binary on a Mac?
<tnks>
I figured I'd ask here before poking around.
<tnks>
(also, I don't have direct access to a Mac right now)
sigmundv has quit [(Quit: Leaving)]
<copumpkin>
hard
<copumpkin>
we've put no effort into cross-compilation from mac; it might work but I doubt it
tempeh has quit [(Quit: Page closed)]
<copumpkin>
also, not sure if it matters, but the hash will be different from the same thing built on linux natively
<smw_>
I should probably ask when more people are around, but has anyone gotten nixos running on a raspberry pi 3? If not, I don't want to go down a rabithole.
<smw_>
yeah, stuff for old RPi is nearly useless for v3.
<copumpkin>
oh boo
<smw_>
My searching has turned up people saying they can't get 3 to work... but I don't know if that has since been solved.
<smw_>
I just ordered an RPi after seeing how well my Dad's RPi works. Before I go installing debian (or worse, raspian), I was hoping I could take a crack at nixos again :-).
<tnks>
copumpkin: okay, good to know... hard to motivate going down that path when building within a Docker container works just fine.
<copumpkin>
tnks: yeah, that's what I'd recommend
<copumpkin>
smw_: most of this channel is in Europe so you might have more luck asking around when most of them are awkae
eacameron has quit [(Remote host closed the connection)]
dramforever has quit [(Quit: Leaving)]
filterfish has quit [(Ping timeout: 255 seconds)]
<endformationage>
OK. Anybody trying to get vscode to inherit a nix-shell env must also export VSCODE_CLI=1
iyzsong has joined #nixos
takle has joined #nixos
filterfish has joined #nixos
dustmote has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
dustmote has quit [(Client Quit)]
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 268 seconds)]
dustmote has joined #nixos
dustmote has quit [(Quit: bye)]
dustmote has joined #nixos
systemfault has joined #nixos
endformationage has quit [(Quit: WeeChat 1.6)]
zml has quit [(Quit: WeeChat 1.5)]
dustmote has quit [(Client Quit)]
takle has joined #nixos
dustmote has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
tempeh has joined #nixos
<tempeh>
Hey - how do I use nix-shell for qualified packages like gnome.scrollkeeper? "nix-shell -p gnome.scrollkeeper" gives "error: undefined variable 'gnome'"
edvorg has joined #nixos
dustmote has quit [(Quit: bye)]
takle has joined #nixos
dustmote has joined #nixos
dustmote has quit [(Client Quit)]
takle has quit [(Ping timeout: 240 seconds)]
dustmote has joined #nixos
takle has joined #nixos
dustmote has quit [(Client Quit)]
takle has quit [(Ping timeout: 260 seconds)]
eacameron has joined #nixos
tempeh has quit [(Ping timeout: 260 seconds)]
takle has joined #nixos
eacameron has quit [(Ping timeout: 260 seconds)]
takle has quit [(Ping timeout: 260 seconds)]
jluttine has joined #nixos
eacameron has joined #nixos
<NixOS_GitHub>
[nixpkgs] gebner pushed 1 new commit to master: https://git.io/vykmH
<NixOS_GitHub>
nixpkgs/master cc64780 Gabriel Ebner: lean: 3.0.0 -> 3.1.0
eacameron has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 268 seconds)]
takle has joined #nixos
filterfish has quit [(Ping timeout: 240 seconds)]
mkoenig has quit [(Remote host closed the connection)]
hotfuzz has quit [(Read error: Connection reset by peer)]
hotfuzz has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
<yurrriq>
What's the best way to derive an alternate package from an existing one, changing only a few properties and maybe some dependencies? Any links or hand-holding would be greatly appreciated.
<pikajude>
yurrriq: you could just use overrideDerivation
<yurrriq>
Real examples: I want to add an alternate LilyPond package that builds the dev branch. And I want to add an alternate MuseScore for Darwin that uses the published .dmg
<yurrriq>
The former just changes a few props and the later needs to add the undmg dependency
<yurrriq>
pikajude: Are there any good examples in nixpkgs?
<pikajude>
probably about 500
<yurrriq>
Ha, I can grep then
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vykOk
<NixOS_GitHub>
nixpkgs/staging 2f726fe Vladimír Čunát: findutils: fixup sandboxed build after #23152
takle has joined #nixos
<yurrriq>
Is adding a dependency (maybe the wrong parlance) any trickier?
<pikajude>
no
<pikajude>
you just add it to buildInputs
<pikajude>
or nativeBuildInputs actually when using overrideDerivation because the evaluation order is WRONG
<yurrriq>
Aha!
<yurrriq>
What's the preferred method for adding a separate platform-specific package?
<yurrriq>
There's a MuseScore one for Linux and I want to add one specific to Darwin.
takle has quit [(Ping timeout: 268 seconds)]
<NixOS_GitHub>
[nixpkgs] edwtjo pushed 2 new commits to master: https://git.io/vykOt
<NixOS_GitHub>
nixpkgs/master 1aaf2be Christoph Hrdinka: retroarch: libretro-mednafen-* -> libretro-beetle-*...
<NixOS_GitHub>
nixpkgs/master a78b801 Edward Tjörnhammar: Merge pull request #23176 from hrdinka/retroarch/mednafen2beetle...
<yurrriq>
IIRC there's a similar situation with the jdk so I can look there for guidance.
<yurrriq>
Thanks for the tips
<NixOS_GitHub>
[nixpkgs] ericsagnes opened pull request #23202: styx and themes: 0.5.0 -> 0.6.0 (master...pkg-update/styx) https://git.io/vykOn
fhoffmeyer has quit [(Ping timeout: 268 seconds)]
<NixOS_GitHub>
[nixpkgs] vcunat pushed 2 new commits to staging: https://git.io/vykO4
<NixOS_GitHub>
nixpkgs/staging 6988d2d Tim Steinbach: curl: 7.53.0 -> 7.53.1
<sphalerite>
I tried mounting a new tmpfs, rsyncing the entire nix store in, and then bind-mounting it on /nix/store, but that broke everything for some reason
<sphalerite>
as in no external commands were working anymore
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vyksG
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vykcn
<NixOS_GitHub>
nixpkgs/master e3f18ec Eric Sagnes: parsimonious: 0.6.2 -> 0.7.0
takle has quit [(Ping timeout: 240 seconds)]
Jackneill_ has joined #nixos
<sphalerite>
viric: gosmore might also work
<sphalerite>
oh, offroad is based on OsmAnd? Nice
ilyaigpetrov has joined #nixos
takle has joined #nixos
indi_ has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
iyzsong has quit [(Quit: bye)]
<NixOS_GitHub>
[nixpkgs] ericsagnes opened pull request #23208: linkchecker: fix darwin build (master...pkg-fix/linkchecker) https://git.io/vykC4
hiberno has quit [(Quit: WeeChat 1.6)]
<pierron>
LnL: Now just add a changelog entry and it would be all good for merging. I would prefer if you can make the changelog entry before merging knowing that the fork is soonish
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vykCa
<NixOS_GitHub>
nixpkgs/master b4d8edd Eric Sagnes: linkchecker: fix darwin build
<gchristensen>
lol, powertop is reporting my CPU has 7019268433962.48GHz
<Mic92>
gchristensen: your pc can travel through time because it is faster then light.
<gchristensen>
I rate this computer a Good Buy ther, definitely more than I bargained for
roundhouse has joined #nixos
takle has joined #nixos
<roundhouse>
hi guys, I'm trying to write a haskell program that uses the sdl2-mixer library. Unfortunately I get "ALSA: Couldn't open audio device: Connection refused" when trying to start the program. I tried both enabling and disabling "sound.enable", i.e. alsa. Any ideas?
<gchristensen>
I wonder if I can implement an mcelog service between now and branch-off
<gchristensen>
my new laptop throws mcelog events...
bennofs1 has joined #nixos
<FRidh>
tried latest kernel/
<FRidh>
?
<pierron>
LnL: I referreing to nixos/doc/manual/release-notes/rl-1703.xml
<gchristensen>
I think I'm running 4.10, yeah, I can check for sure after this powertop calibration run
<LnL>
pierron: yeah I was looking at the wrong manual before :p
<pierron>
LnL: that's our only changelog so far.
<pierron>
LnL: or I am not aware of any other
takle has quit [(Ping timeout: 260 seconds)]
<LnL>
I added an entry there
<ben__>
What's the process for reporting "i said nix-env -i ... and got a compilation error" kinda issues?
<ben__>
do i just dump all the output into a github issue?
<gchristensen>
ah, 4.9 FRidh. think 4.10 might help? I suppose I could switch to unstable a day before branch-of...
takle has joined #nixos
fre has joined #nixos
alx741 has joined #nixos
takle has quit [(Ping timeout: 268 seconds)]
ndowens08 has joined #nixos
<pierron>
LnL: ok, minor doc fixes, if you want you can leave the follow-up remarks for future patches.
<pierron>
LnL: Sorry for asking you to iterate that many times on these modifications, but I am trying to be as careful as I am when I review Firefox changes.
<NixOS_GitHub>
[nixpkgs] grahamc pushed 1 new commit to master: https://git.io/vyk0K
<NixOS_GitHub>
nixpkgs/master 8b40d2e Graham Christensen: mcelog: 144 -> 148
<hodapp>
https://github.com/NixOS/nixpkgs/pull/22219 I wrote this and it's merged, but can anyone think of ways that I might get this to allow a user to add packages to RStudio without requiring an entire rebuild of RStudio (which is lengthy enough to be annoying)?
<ben__>
I can't seem to install obconf, any advice? http://ix.io/nN5 - nixpkgs 17.03pre101839.53a2baa which i believe is current nixos-unstable
<unlmtd[m]1>
dont know much abouut the workflow of keeping track of issues like these, should I open an issue in my fork, or open a pull request and comment about the errors?
<ben__>
hodapp: does it need to rebuild all of rstudio, not just the wrapper?
<unlmtd[m]1>
something about .cargo-index-lock being, well, locked out of reach I guess
<LnL>
pierron: no problem, I appreciate the feedback. But perhaps I should improve the documentation in a separate pr so this can be merged before the release branch is created?
<gchristensen>
FRidh: I'm building an mcelog service, where would you put that in the nixos modules?
<peterhoeg>
gchristensen: you didn't have any problems getting 4.10 up and running?
<gchristensen>
peterhoeg: I haven't tried it yet
<peterhoeg>
Ah
<hodapp>
ben__: the wrapper is changed to modify some environment variable, but what my also inserts some R code into a startup file. This is all past all of the actual compilation for RStudio as it's in installPhase/postInstall.
<hodapp>
what my patch*
<gchristensen>
peterhoeg: I'm still running the 350 rounds of powertop :|
<peterhoeg>
Now THAT sounds like fun.
<hodapp>
ugh. grammar. "...my patch also inserts..."
<ben__>
hodapp: can you let RStudio install as one derivation, and then have another derivation that takes the output dir of RStudio as its source, copies/symlinks it over to its own output dir and then adds that R code?
<ben__>
so not touching RStudio's own phases at all
JagaJaga has quit [(Ping timeout: 260 seconds)]
<hodapp>
ben__: that might be an option if I can figure out how. All it has to do is invoke RStudio in a wrapper with --set R_PROFILE_USER pointed to a script that it generates
<peterhoeg>
ben__: do you know of a good, clean implementation of such a wrapper currently in nixpkgs? I have another package I need to give that treatment.
<ben__>
peterhoeg: I don't, sorry
<hodapp>
dang
<hodapp>
I feel sort of like R itself works this way but I'm not sure
<hodapp>
see https://nixos.org/nixpkgs/manual/#r-packages, for instance. As I understand it, one can install normal R, or one can install this rWrapper which is intended to be overwritten
<hodapp>
and this is why my patch allowed one to override RStudio's R package for an R wrapper created as described in the manual (with other R packages bundled with it), but the rest of the patch was because of the idiocy RStudio did which completely negated that wrapper's effect
<peterhoeg>
gchristensen: please let me know if you're getting the same message when trying out 4.10:
<peterhoeg>
inconsistent kernel versions: 4.10.0
<peterhoeg>
4.9.12
<copumpkin>
niksnut_: I see why the service isn't starting up on the EC2 box, but am unsure what to do to fix it
<pierron>
LnL: I don't think the documentation is good enough in yet, and I am ok to delay these changes because of the documentation issue
<pierron>
LnL: I honestly don't like merging changes with a promess of upcoming documentation, especially since we constantly seems to get complain about the lack of documentation.
takle has joined #nixos
<gchristensen>
peterhoeg: what do you mean?
<pierron>
Quick vote everybody, should we hold features until they are properly documented?
nckx has quit [(Ping timeout: 255 seconds)]
<pierron>
pierron: hold on doc +1
<gchristensen>
I agree with that, nothing like the drive of the merge to get the documentation done
smw_ has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
<ben__>
hm, firefox does the thing where they have an "unwrapped" derivation that other derivations build wrappers around
<gchristensen>
peterhoeg: code, it seems it needs to access /proc/cpuinfo, its config file in its own store dir, /sys/firmware/dmi/entries/0-0/raw, /dev/mem, /sys/firmware/efi/systab, and /dev/mcelog
<gchristensen>
peterhoeg: should I just add that whole list and see where it gets me?
Chattered has quit [(Remote host closed the connection)]
<peterhoeg>
I wil realistically not have time to work on this until later in the week, so if you want it in before the release deadline I'm not your guy. My approach would be simply enable as much as possible and then start removing bits if the daemon doesn't run. Very scientific....
<gchristensen>
I guess I'm also not sure why start with this one, unless the answer is "gotta start somewhere"
takle has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 3 new commits to staging: https://git.io/vykzE
<NixOS_GitHub>
nixpkgs/staging 4bc1d02 Frederik Rietdijk: Python 2.7: increase priority...
<NixOS_GitHub>
nixpkgs/staging 1b66b6a Frederik Rietdijk: pythonPackages.pyparsing: 2.1.8 -> 2.1.10
<NixOS_GitHub>
nixpkgs/staging 57afc0f Frederik Rietdijk: pythonPackages.packaging: 16.7 -> 16.8
smw_ has quit [(Quit: Computer has gone to sleep)]
takle has quit [(Ping timeout: 268 seconds)]
smw_ has joined #nixos
chattered has joined #nixos
RchrdB has quit [(Ping timeout: 240 seconds)]
<NixOS_GitHub>
[nixpkgs] fkz opened pull request #23216: phpfpm: eliminate build at evaluation time (master...patch-1) https://git.io/vykzN
smw_ has quit [(Ping timeout: 268 seconds)]
<gchristensen>
peterhoeg: where did you get the kernel error?
<gchristensen>
peterhoeg: fwiw I'm using the default service definition which comes with mcelog, same as many distros. while I see the value in adding the restrictions, it is not an easy thing to do for a few reasons: I don't want to just take a list of "best practices" and apply them without understanding what they mean (not to mention the service doesn't work with them.), the service can probably operate with more in some
<gchristensen>
ways and less in others. I'm also frustrated by the seemingly arbitrary "well let's start with this PR" application of the list, without you being able to help with it.
takle has quit [(Ping timeout: 240 seconds)]
MinceR has quit [(Ping timeout: 240 seconds)]
<peterhoeg>
gchristensen: apologies if I seemed dismissive - that was not the intention. Applying that laundry list of items is something that takes time if upstream hasn't done any of the work.
MinceR has joined #nixos
<peterhoeg>
And really not something one can just do in 5 minutes - that's why I said I only have time later in the week.
<gchristensen>
I would like to merge this (to have it in 17.03) and then would love to work with you to lock it down
<gchristensen>
perhaps even work on a grander locking down of services
<peterhoeg>
And it would be a very cool project to take on
<hodapp>
peterhoeg: if you ever figure out a good example of a wrapper (like you were asking before), let me know
<peterhoeg>
If you haven't tried with PrivateTmp, PrivateNetwork and PrivateUsers, it might be worth a shot just with those. They should not break anything and it's better than nothing.
<gchristensen>
what was your problem with the kernel, though? I'm not seeing any kernel version problems.
<peterhoeg>
hodapp: sure
<peterhoeg>
regd the kernel, I haven't really looked into it - not an awful lot of time this weekend (or this coming week). ```nixos-rebuild boot``` just errors out with the message I mentioned.
<gchristensen>
I don't seem to have that problem
<gchristensen>
sorry
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to release-16.09: https://git.io/vyk2u
<NixOS_GitHub>
nixpkgs/release-16.09 bda6b18 Frederik Rietdijk: diffoscope: 63 -> 77...
nixy has joined #nixos
<peterhoeg>
I have a hunch that maybe virtualbox doesn't work properly with 4.10, but I haven't tried. It sounds like a 3rd party kernel module that is blowing up.
<peterhoeg>
hodapp: Just FYI, I noticed that kodi has a very straight forward wrapper, but I have not looked into it in detail.
<hodapp>
peterhoeg: thanks, I'll take a look
<hodapp>
I really should probably look at how R/rWrapper does it
simendsjo has quit [(Quit: ERC (IRC client for Emacs 24.5.2))]
<obadz>
but I'm certain it won't affect me negatively
<obadz>
since that was reverted from 16.09
<obadz>
and I was on 16.09 up until last week
<obadz>
copumpkin: stalling on technical grounds? or niksnut_ not paying enough attention?
<copumpkin>
nah, technical grounds, although sort of for an iffy reason
nckx has quit [(Ping timeout: 240 seconds)]
<obadz>
I've seen the comments but haven't had a chance to dive in yet
<gchristensen>
obadz: I also don't know :) I just don't want people's WMs to be broken.
<obadz>
hmmm
<obadz>
let me have a quick look at what other DMs are doing
<obadz>
we should sort this out before branch out tomorrow
<gchristensen>
sounds great, I'm around to help most of the day )
<gchristensen>
:)
<copumpkin>
I'm unsure how I want to handle it, basically. Currently, nixos-install (which I need to refactor to avoid duplicating logic with the image build) _builds_ the target system configuration on against the target store, which has the "side effect" of putting build-time dependencies into it. The installer tests try to do an offline nixos-rebuild, which fails on mine but not on theirs, because I build my closure outside of the target store
<copumpkin>
(sort of like the --closure argument you added, which would have similar issues)
<copumpkin>
obadz: my current inclination is to just amend the installer tests to explicitly ask for the build-time closure on the image, since getting that programmatically appears to be impossible right now
<copumpkin>
you can get a vast overapproximation
<ndowens08>
I noticed you guys were talking a bit about services running as root, at least seen peterhoeg mention it, though I try to create a seperate user for services I create when I can; but for this one I am currently writing, there is no option in the server to run as a certain user, so is there a way to make it run as it's own user vs root even though there is no built in option?
<copumpkin>
but that puts crap that has no business being in the image into it
nckx has joined #nixos
<copumpkin>
ndowens08: yeah, systemd can run it as anything you ask it to
bennofs1 has joined #nixos
<copumpkin>
see man systemd.service
<ndowens08>
copumpkin: ah, will look at that then :)
<obadz>
gchristensen: thruth be told, even if we fix this in master, it's not enough time to test it until stable. let's just revert and kick can down the road
takle has joined #nixos
<NixOS_GitHub>
[nixpkgs] obadz pushed 1 new commit to master: https://git.io/vyka9
<obadz>
gchristensen: have you noticed that in the ligthdm screen, you can select a window manager in the top part, and it'll remember which one was selected on a per-user basis??
<peterhoeg>
ndowens08: if you don't need persistent data, using DynamicUsers=true would be very good. Then we don't have to pollute the main ids.nix file as well.
<obadz>
gchristensen: that's what cf94cdb59b did
<ndowens08>
peterhoeg: hmm, if that is ok to do, though this is a webserver, so idk if that could be used?
<ndowens08>
since i guess it would be persistent data right
takle has quit [(Ping timeout: 260 seconds)]
<obadz>
copumpkin: I fail to understand why the build-time deps are relevant at all? just because the tests happen to depend on it and they can't download more stuff?
<peterhoeg>
ndowens08: well it depends. If it doesn't need to persist any data, you might be fine. Of course whatever the application is serving would then have to have world-read permissions.
Fare has joined #nixos
<copumpkin>
obadz: well, sort of
<copumpkin>
I could understand actually wanting build-time dependencies (I'd make it a flag of the image builder code)
takle has joined #nixos
<NixOS_GitHub>
[nixpkgs] johbo opened pull request #23218: mysql57: Add developer_cmds as build input on darwin (master...darwin-mysql) https://git.io/vykVk
<copumpkin>
if I want to produce an image that's intended to be reconfigured in a fully offline manner
<FRidh>
Could someone trigger a staging build?
<ndowens08>
peterhoeg: hmm probably the only thing to change much of course is the dir that has the web data to serve
<copumpkin>
FRidh: done
<copumpkin>
nixos:staging?
<copumpkin>
I'm assuming
<peterhoeg>
ndowens08: you can always try it out ;-) I'm signing off from here. Getting rather late.
<copumpkin>
obadz: yeah that's what I'm going to do to unstall it, but it's an interesting broader question, since it does change the composition of the image somewhat
<Dezgeg>
I think breaking the nixos manual build (due to invalid XML) is quite common one to check for
JagaJaga has joined #nixos
<gchristensen>
looks good from here
dustmote has joined #nixos
<gchristensen>
thanks, Dezgeg
<LnL>
gchristensen: I think we should start requesting people to add meta.maintainers
<gchristensen>
you can do that? I didn't know that
<LnL>
I think only a small number of modules have it at the moment
<LnL>
yeah, even less then I expected
<gchristensen>
ok, I added myself as a maintainer :) merging. thank you!
dustmote has quit [(Client Quit)]
<NixOS_GitHub>
[nixpkgs] grahamc pushed 2 new commits to master: https://git.io/vykVp
<NixOS_GitHub>
nixpkgs/master 1430506 Graham Christensen: mcelog: init Machine Check Exception Logging Daemon service
<NixOS_GitHub>
nixpkgs/master 4f3d06d Graham Christensen: Merge pull request #23214 from grahamc/mcelog-service...
<ndowens08>
How do you counter the issue when it says value is a string, while expecting a set; trying to set option for configFile location; I have default = "${pkgs.hiawatha}/etc/hiawatha"; as option and in the config area under systemd, the command looks like ExecStart = "${pkgs.hiawatha}/bin/hiawatha -c cfg.configFile";
takle has joined #nixos
<sphalerite>
clever: why do you dd zeroes onto the disk before generating the new partition table?
<NixOS_GitHub>
[nixpkgs] husnoo opened pull request #23220: copy gatttool for bluez (master...master) https://git.io/vykwL
<bennofs1>
Is the output name always appended as -<name of output> to the store path of a multiple output derivation?
davidak has joined #nixos
najib has quit [(Ping timeout: 240 seconds)]
takle has quit [(Remote host closed the connection)]
takle_ has joined #nixos
<gchristensen>
bennofs1: except for $out I think, yes
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vykrv
<NixOS_GitHub>
nixpkgs/master 1d9fa60 Michael Raskin: wavefunctioncollapse: 0.0pre20160930 -> 0.0pre20170130
<NixOS_GitHub>
nixpkgs/master a872e22 Michael Raskin: speed_dreams: 2.0.0-a3 -> 2.2.1
<BarrOff>
i allready looked it up in the freebsd handbook how to use "extern char **environ"
<BarrOff>
however it doesnt work
nautillus has joined #nixos
<nautillus>
Hi, do you have any problems with Wine on Unstable/master? I am not sure if that is related to X bump to 1.19, but it is very similar to https://github.com/NixOS/nixpkgs/issues/22760 and wine gives 'libGL error: unable to load driver' too.
<NixOS_GitHub>
[nixpkgs] fpletz pushed 4 new commits to master: https://git.io/vykr2
<NixOS_GitHub>
nixpkgs/master 5e937b1 Franz Pletz: gstreamer: 1.10.3 -> 1.10.4 for multiple CVEs...
<NixOS_GitHub>
nixpkgs/master 8a75569 Franz Pletz: linux: 4.10 -> 4.10.1
<NixOS_GitHub>
nixpkgs/master 62857b1 Franz Pletz: linux: 4.4.51 -> 4.4.52
<joko>
Hello, is anyone using boot.initrd.network.ssh? I have trouble getting an IP address during boot
<bennofs1>
copumpkin: shouldn't the service in the EC2 case get activated then as soon as multi-user.target is reached?
<sphalerite>
Has anyone here used steamCMD on nixos?
<bennofs1>
copumpkin: or is this because the "transaction" starting multi-user.target is still in process, and that transaction isn't changed on daemon-reload?
<NixOS_GitHub>
[nixpkgs] bennofs pushed 1 new commit to master: https://git.io/vykot
<copumpkin>
bennofs1: I don't really know why it isn't starting. I'd have expected it to get retroactively added to the in-progress multi-user.target startup, but it doesn't
<NixOS_GitHub>
[nixpkgs] joachifm pushed 1 new commit to master: https://git.io/vykoq
<NixOS_GitHub>
nixpkgs/master f3a6991 Joachim Fasting: grsecurity: 4.9.12-201702231830 -> 4.9.13-201702261126
<copumpkin>
if I explicitly tell it to start multi-user.target from switch-to-configuration.pl, it works again
<copumpkin>
so perhaps it needs a nudge?
<bennofs1>
copumpkin: well I assume it is because systemd already created the transaction, and after that, doesn't look at the unit files etc again even if you've got a daemon-reload
<copumpkin>
sounds plausible... I don't really know how it works though
xwvvvvwx has joined #nixos
<NixOS_GitHub>
[nixpkgs] bennofs pushed 1 new commit to master: https://git.io/vyko8
<Dezgeg>
I still need to send some patches to u-boot... but otherwise it works great, including X11 and the mesa vc4 driver
xwvvvvwx has quit [(Client Quit)]
wangoe has quit [(Ping timeout: 255 seconds)]
<smw_>
viric: thanks
<smw_>
Dezgeg: is there a guide somewhere on putting nixos on a rpi3? If not, I plan to go the bootstrap method. I hear that only takes a few days to compile :-)
mudri has joined #nixos
<viric>
Dezgeg: great
<viric>
Dezgeg: can you post an openssl speed run somewhere?
<NixOS_GitHub>
[nixpkgs] vbgl pushed 19 new commits to master: https://git.io/vykKd
<NixOS_GitHub>
nixpkgs/master 0e06c90 Vincent Laporte: ocamlPackages.sexplib-113.33.00+4.03: fix version number
<NixOS_GitHub>
nixpkgs/master 3a9ae7e Vincent Laporte: ocamlPackages.ppx_compare: init at 113.33.00+4.03
<NixOS_GitHub>
nixpkgs/master b4abe04 Vincent Laporte: ocamlPackages.ppx_inline_test: init at 113.33.00+4.03
mudri has quit [(Client Quit)]
<nautillus>
Any feedback or there is no Wine users on master/Unstable?
<simpson>
nautillus: Do you have 32-bit DRI enabled?
mudri has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 3 new commits to release-16.09: https://git.io/vyk6Y
<NixOS_GitHub>
nixpkgs/release-16.09 d2ceedd Franz Pletz: linux: 4.4.51 -> 4.4.52...
<NixOS_GitHub>
nixpkgs/release-16.09 fcc6e3a Franz Pletz: linux: 4.9.12 -> 4.9.13...
<NixOS_GitHub>
nixpkgs/release-16.09 2296799 Franz Pletz: gstreamer: 1.10.3 -> 1.10.4 for multiple CVEs...
ertesx has joined #nixos
x1n4u has quit [(Remote host closed the connection)]
thaega has quit [(Read error: Connection reset by peer)]
ertes has quit [(Ping timeout: 240 seconds)]
ertesx is now known as ertes
derjohn_mob has joined #nixos
x1n4u has joined #nixos
thaega has joined #nixos
<nautillus>
simpson: I have hardware.opengl.driSupport = true; but not hardware.opengl.driSupport32Bit. But it worked for me before recent upgrade so dunno what is the correct one.
<simpson>
nautillus: You should set both to `true` and rebuild.
<nautillus>
Maybe defaults changed.
mudri has quit [(Ping timeout: 240 seconds)]
<simpson>
nautillus: What's happening is that Wine needs 32-bit GL libraries in order to emulate 32-bit hardware rendering.
<nautillus>
simpson: I will try, thank you for suggestion, but I am surprised that now with hardware.opengl.driSupport32Bit = false I can play e.g. World Of Warships via Wine.
fre has quit [(Ping timeout: 240 seconds)]
<simpson>
nautillus: You only need it for DRI stuff, emulating GL or D3D. Maybe the game you're playing with is a DD game? I no longer know anything interesting about Wine internals.
<simpson>
But I would expect that you need to have `dirSupport32Bit = true;` in order to have accelerated games in Wine.
derjohn_mob has quit [(Ping timeout: 240 seconds)]
<nautillus>
I wonder why the same config but on the more recent Nixpkgs result in libGL error: unable to load driver error.
<nautillus>
The big thing is that X org jump to 1.19 ABI.
<obadz>
gchristensen: spaceFM->ifuse->libplist don't build because of security advisory
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vykiO
<NixOS_GitHub>
nixpkgs/master 8fcd1a0 Franz Pletz: faac: fix build with glibc-2.25...
<copumpkin>
fpletz1: you seem to have grown a new character!
<fpletz1>
the one and only!
fpletz1 is now known as fpletz
<copumpkin>
:)
fpletz has quit [(Changing host)]
fpletz has joined #nixos
<fpletz>
Mic92: yeah, ericsagnes should have access if the wants :)
<copumpkin>
fpletz: so staring at switch-to-configuration.pl, I'm thinking more and more that we should remove unitsToStart altogether and just start default.target
<copumpkin>
and otherwise leave the restart/reload logic in place
<fpletz>
copumpkin: yeah, I always wondered why we are doing it this way, probably leftovers from the upstart era
<copumpkin>
the main thing we lose is the "starting the following units:" message
<fpletz>
right :?
<copumpkin>
no idea, 'twas before my time
<copumpkin>
but yeah
<copumpkin>
we'd still get the "started the following NEW units" message, of course
<fpletz>
and before mine… I've been using nixos for only 2.5 years or so
<copumpkin>
err, "the following new units were started:" that one
<fpletz>
hm, but then again, I had some other dependent services restart (when using partof) that also didn't show up there
<fpletz>
I guess we could argue that these messages are not realiable anways
<copumpkin>
they filter out stuff
<ben__>
Is there a flowchart that tells me whether I want override, overrideDerivation or overrideAttrs
<copumpkin>
ben__: not that I know of, but I can tell you the difference between the first two :)
<ben__>
this was really confusing
<ben__>
I wanted to add something to buildInputs
<ben__>
but I guess some of these swap around buildInputs and nativeBuildInputs
<ben__>
override is like if i added callpackage args, overrideDerivation is like if I added stuff to the derivation { ... } attrs and overrideAttrs is like if I added stuff to the stdenv.mkDerivation invocation?
<copumpkin>
fpletz: there's a bit of special logic in there to catch units that were started manually and need updating
jgertm has quit [(Ping timeout: 240 seconds)]
<copumpkin>
so I guess it's not all upstart holdover
<sphalerite>
manveru: a mighty oracle told me you were trying to get steamcmd working on nixos 2 and a half years ago. Did you get anywhere with it? :p
<NixOS_GitHub>
[nixpkgs] obadz pushed 1 new commit to master: https://git.io/vykPW
<NixOS_GitHub>
nixpkgs/master daaf817 obadz: SpaceFM: (optionally) remove dependency on ifuse due to broken dep...
<copumpkin>
ben__: I've never seen overrideAttrs, but overrideDerivation lets you change the attribute set being passed to mkDerivation. override lets you change the parameters passed into the function passed to callPackage
<copumpkin>
obadz: I think I have a better solution than hardcoding for the test
<obadz>
copumpkin: talking about libplist or nixos-install?
<copumpkin>
actually no, never mind
fingerzam has joined #nixos
<copumpkin>
the latter :)
derjohn_mob has joined #nixos
<copumpkin>
shlevy: I had a thought about the build-time closure
mudri has joined #nixos
Fare has quit [(Ping timeout: 240 seconds)]
derjohn_mob has quit [(Ping timeout: 240 seconds)]
<copumpkin>
fpletz: the default.target thing seems to fix my issue. Do you know if we have any VM tests that specifically test edge cases of that logic?
<fpletz>
copumpkin: there are some tests that invoke nixos-rebuild switch but those probably aren't testing the edge cases affected by this change
<copumpkin>
gah, how do I gain confidence in it? :) I guess I'll submit a PR and highlight our resident systemd gurus?
<copumpkin>
if only we had Hydra testing PRs :D
mudri has quit [(Ping timeout: 260 seconds)]
<cstrahan>
copumpkin: your nixos-install work looks promising :)
<copumpkin>
yay
<copumpkin>
thanks
<copumpkin>
I think I can fix the install test failure now
<copumpkin>
I'm going to do that once I sort out this systemd crap
<cstrahan>
copumpkin: any developments in the "runtime closure of the build time deps" issue?
<copumpkin>
basically means I think I can retain the current behavior for nixos-install, but won't able to do that for images I build
<copumpkin>
which is fine, I can wait until I get the nix feature I need
<copumpkin>
ideally I'd like some sort of "rebuildable offline" setting for images/systems we install
<cstrahan>
Oh, so addressing something similar to the issues I've had where installing NixOS from CD requires an internet connection, despite installing pretty much exactly what was on the live CD, I suppose?
fre has quit [(Ping timeout: 240 seconds)]
<bennofs1>
sorry for the snide remark, but can we make a commitment about contributions / issues at least getting *some* comment, even if all it says is "i will look at this later when I have time"? Opening an issue and not getting *any* reaction in months is just a little frustating to me, I mean, it's not the end of the world, but it would be *much* nicer I there was some sort of acknowledgment.
smw_ has quit [(Quit: Computer has gone to sleep)]
<copumpkin>
bennofs1: the issue is who would commit to that :/
smw_ has joined #nixos
<copumpkin>
but I agree
<copumpkin>
have something specific in mind here?
<cstrahan>
copumpkin: I wonder if there's a legitimate reason not to have some sort of primop that exposes these closures. Has that been discussed at all? Wouldn't mind implementing it if I suspected it would be merged.
<copumpkin>
cstrahan: yeah, shlevy mentions it in that #1245 ticket
<NixOS_GitHub>
[nixpkgs] copumpkin opened pull request #23224: switch-to-configuration: don't try to figure out active targets ourselves (master...switch-to-configuration-fixes) https://git.io/vyk9k
<fpletz>
Mic92: fsck, yeah :/
<fpletz>
I should've thought of that :)
ryanartecona has joined #nixos
<fpletz>
they're now in resolv/resolv-internal.h, hopefully that header is available
<jyp[m]>
Does this ring a bell: "ld: framework not found Cocoa", when buiding a Hakyll site?
<Mic92>
fpletz: if you fix it, ideally in patch form, so it can be submitted to <lennart@poettering.net>
<Mic92>
ah wait
<Mic92>
he does not maintain the python package
<NixOS_GitHub>
[nixpkgs] joachifm opened pull request #23225: webkitgtk24x: mark as broken (master...webkitgtk24x-broken) https://git.io/vyk9R
ttuegel has joined #nixos
<fpletz>
turns out the fix was easier than expected
justbeingglad has joined #nixos
<fpletz>
I should stop nixing for today, been at it for hours now and keep making silly errors :)
<fpletz>
Mic92: only ns_t_zxfr isn't available anymore
<NixOS_GitHub>
[nixpkgs] vbgl pushed 1 new commit to master: https://git.io/vyk9H
ryanartecona has quit [(Ping timeout: 255 seconds)]
hiratara has quit [(Ping timeout: 255 seconds)]
<zraexy>
Is there any specific reason as to why some packages are defined directly 'top-level/python-packages.nix' instead of having folders in 'development/python-modules'?
bennofs1 has quit [(Quit: WeeChat 1.7)]
bennofs has joined #nixos
hiratara has joined #nixos
Kingsquee has joined #nixos
ndowens08 has quit [(Ping timeout: 240 seconds)]
<FRidh>
zraexy: python-modules is for python libraries, or for tools that need to be available for each python interpreter version. Packages that are called in all-packages.nix are applications.
<FRidh>
oh, nevermind, I misread
<FRidh>
zraexy: they're supposed to be in python-modules/<name>default.nix
<FRidh>
and not python-modules/<name>.nix
FRidh has quit [(Remote host closed the connection)]
goibhniu has quit [(Ping timeout: 255 seconds)]
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vykFn
<NixOS_GitHub>
nixpkgs/master 78d9414 Tuomas Tynkkynen: findutils: Fix cross compilation...
<yorick_>
I would like to be able to just save my nix-shell -p invocations to shell.nix for later automagic use
yorick_ is now known as yorick
<yorick>
it could make a .nix file out of it
<bennofs>
yorick: I think just putting that into a .nix file with `with import <nixpkgs> {}; stdenv.mkDerivation { name = "env"; buildCommand = ":"; buildInputs = [ ...nix shell -p flag arguments.. ]; }` ?
wizeman has quit [(Quit: Connection closed for inactivity)]
<yorick>
bennofs: yeah, I wrote it down, but I keep forgetting the exact incantation
<bennofs>
yorick: make a script for it :)
hiratara has quit [(Remote host closed the connection)]
hiratara has joined #nixos
takle has joined #nixos
<yorick>
is sage broken for anyone else?
<18VAAN137>
[nixpkgs] globin closed pull request #23216: phpfpm: eliminate build at evaluation time (master...patch-1) https://git.io/vykzN
<NixOS_GitHub>
[nixpkgs] globin pushed 1 new commit to master: https://git.io/vykbB