<joepie91>
clever: do you know if there exists some sort of debugger for Nix?
<clever>
joepie91: and in here, its using a map function over the list of all systemd.services, to make unit files
<clever>
closest thing is nix-repl
<joepie91>
clever: I mean with breakpoints and all
<clever>
nothing like that
<joepie91>
alright
<clever>
joepie91: but there is builtins.trace
<clever>
nix-repl> builtins.trace "print me" "return me"
<clever>
trace: print me
<clever>
"return me"
<joepie91>
clever: yeah, but I want to inspect the flow of existing code :P
<joepie91>
without dumping traces everywhere
<joepie91>
hence the debugger question
<clever>
you can sorta do that by inserting an assertion into things like systemPackages
<clever>
and then the backtrace shows the path from top-level to the failure
<joepie91>
clever: is there an easy way to print all the values in scope at that point?
<clever>
dont think so
<joepie91>
hrm.
newhoggy has quit [(Read error: Connection reset by peer)]
newhoggy_ has joined #nixos
<joepie91>
clever: so, another question. I'm tired of the inscrutable output from a nixos-rebuild - it just dumps all the output from everything into one terminal, it's difficult to figure out what output comes from where, failures are difficult to trace back reliably, and it's not always clear what dependencies are built because of what other things. do you know of any tools or methods to make the build output more usable, or even provide some sort of custom
<joepie91>
hook for recording/piping the build output elsewhere?
<clever>
i build with -Q to silence it all
<joepie91>
I don't mind taping together my own solution, but I need a starting point :P
<joepie91>
right, but I *do* want the output, just not in the way it provides it to me now
<clever>
and it already saves the output to /nix/var/log/
<clever>
you can run "nix-store -l /nix/store/foo" to retrieve the log for any storepath
<joepie91>
clever: also, how do I reconcile those logs with an ongoing installation process?
schoppenhauer has joined #nixos
<joepie91>
like, I'd like to be able to script my way around the current rebuild process in a way that I can generate more useful output as the rebuild is progressing
<clever>
it doesnt save the log for failed builds
<clever>
i usualy build with -j8 -Q for speed, then re-build with -j1 to get the failure
<joepie91>
that... seems like precisely the opposite of what you'd want :P
<clever>
and also, the storepath you tried to get the log for never even started
<clever>
builder for ‘/nix/store/p2s2900lznd33vnz63gz9slfrb8nkk4h-factorio.tar.gz.drv’ failed with exit code 1
<joepie91>
clever: right, but are you aware of any way to provide hooks or whatever to build things on top of the existing rebuild process
<clever>
the tarball is the one that failed first
<clever>
so the path you queried refused to even start
<joepie91>
right.
<joepie91>
yeah, there *is* a log for the tarball apparently
<clever>
dont know of any hooks to do that right now
<joepie91>
clever: any way to replicate the rebuild process, within a reasonable amount of time/effort, that allows me to control the process more? I'm not dead-set on wrapping nixos-rebuild, and if I can invoke things more directly, all the better
<clever>
nixos-rebuild basicaly just boils down to "nix-build '<nixpkgs/nixos>' -A system"
<joepie91>
clever: is there a way to control the output by eg. specifying a custom build target? (in the sense that there are system and VM etc. targets right now)
newhoggy has joined #nixos
<clever>
not really
<joepie91>
hrm,.
<joepie91>
I was hoping to avoid rewriting half of Nix :P
<joepie91>
anyhow, I'll sleep over it and look into things more tomorrow...
<joepie91>
clever: thanks for the help
<clever>
yep
takle has joined #nixos
<Welkin>
nix is working :D
<Welkin>
it was just a problem with some specific packages for haskell that need to be updated
newhoggy has quit [(Ping timeout: 240 seconds)]
newhoggy has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
osa1 has joined #nixos
osa1 has quit [(Changing host)]
osa1 has joined #nixos
osa1_ has quit [(Ping timeout: 260 seconds)]
newhoggy has quit [(Ping timeout: 252 seconds)]
matthewbauer has joined #nixos
newhoggy has joined #nixos
matthewbauer has quit [(Remote host closed the connection)]
matthewbauer has joined #nixos
newhoggy has quit [(Ping timeout: 260 seconds)]
ryantrinkle has quit [(Ping timeout: 260 seconds)]
newhoggy has joined #nixos
newhoggy has quit [(Ping timeout: 260 seconds)]
mizu_no_oto has quit [(Quit: ["Textual IRC Client: www.textualapp.com"])]
newhoggy has joined #nixos
wkennington has quit [(Quit: Leaving)]
newhoggy_ has joined #nixos
newhoggy has quit [(Read error: Connection reset by peer)]
<unlmtd[m]>
It seems the nix store, unsafe (world readable) issue is showing everyone, is that a clear line must be drawn between two separate stores. I believe they could both be nix stores
<unlmtd[m]>
So in essence I would have a ${user} private store and a system global store which strictly contains the Operating System TM and its Exclusive Libraries TM
newhoggy_ has quit [(Ping timeout: 240 seconds)]
newhoggy has joined #nixos
<simpson>
unlmtd[m]: Well, you're attempting to shoehorn POLA into standard filesystem permissions, which doesn't work.
<simpson>
What you really want is something more like each process getting its own VFS.
<simpson>
(What you really want is something more like KeyKOS or Genode!)
<unlmtd[m]>
didnt plan9 have something like namespace?
<simpson>
Not namespacing, but actual private process space, with true encapsulation and no ambient authority. (This implies no root filesystem, in a sense!
newhoggy has quit [(Ping timeout: 268 seconds)]
dridus has quit [(Quit: znc? wassat?)]
newhoggy has joined #nixos
newhoggy has quit [(Ping timeout: 255 seconds)]
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/v9WT7
<NixOS_GitHub>
nixpkgs/master 03f939e Peter Hoeg: kirigami2: 1.90.0 -> 2.1.0...
newhoggy has joined #nixos
luigy has left #nixos []
kampfschlaefer has joined #nixos
newhoggy has quit [(Ping timeout: 252 seconds)]
justelex has joined #nixos
newhoggy has joined #nixos
eacameron has quit [(Remote host closed the connection)]
jacob has joined #nixos
jacob is now known as Guest48735
newhoggy has quit [(Ping timeout: 240 seconds)]
vandenoever has quit [(Ping timeout: 258 seconds)]
newhoggy has joined #nixos
takle has joined #nixos
wkennington has joined #nixos
newhoggy has quit [(Ping timeout: 240 seconds)]
takle has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
nicknovi1 has quit [(Ping timeout: 260 seconds)]
chessai has quit [(Quit: Connection closed for inactivity)]
<NixOS_GitHub>
[nixpkgs] domenkozar force-pushed nixos-datadog-process-config from 106f262 to 5830e71: https://git.io/v9nHp
<NixOS_GitHub>
nixpkgs/nixos-datadog-process-config 5830e71 Domen Kožar: nixos: datadog: correctly pass conf.d to the daemon...
eacamero_ has joined #nixos
eacameron has quit [(Ping timeout: 260 seconds)]
newhoggy has joined #nixos
Welkin has quit [(Quit: leaving)]
eacameron has joined #nixos
eacamero_ has quit [(Ping timeout: 258 seconds)]
newhoggy has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
<avn>
Folks, anyone know -- how to deal with `error: Package ‘virtualbox-modules-5.1.18-4.10.13’ in ‘/home/avn/nixos/nixpkgs/pkgs/os-specific/linux/virtualbox/default.nix:21’ has an invalid meta attrset: - key 'priority' has a value of an invalid type; expected integer, refusing to evaluate.`?
<avn>
looks like it introduced by recent changes
<joachifm>
there were some metadata checks added yesterday or so
<joachifm>
I suppose the way to deal with the error is to fix what its complaining about :)
newhoggy has joined #nixos
eacamero_ has joined #nixos
jasom has quit [(Ping timeout: 260 seconds)]
eacameron has quit [(Ping timeout: 245 seconds)]
derjohn_mob has quit [(Remote host closed the connection)]
takle has quit [(Ping timeout: 260 seconds)]
eacameron has joined #nixos
newhoggy has quit [(Ping timeout: 246 seconds)]
derjohn_mob has joined #nixos
eacamero_ has quit [(Ping timeout: 268 seconds)]
johnsonav has quit [(Ping timeout: 245 seconds)]
newhoggy has joined #nixos
eacameron has quit [(Ping timeout: 246 seconds)]
eacameron has joined #nixos
jasom has joined #nixos
newhoggy has quit [(Ping timeout: 260 seconds)]
eacameron has quit [(Ping timeout: 252 seconds)]
newhoggy has joined #nixos
filterfish has quit [(Ping timeout: 258 seconds)]
newhoggy has quit [(Ping timeout: 240 seconds)]
endformationage has quit [(Quit: WeeChat 1.7)]
eacameron has joined #nixos
<avn>
joachifm: specified file have `meta = main_meta // { description = .... };` (and main_meta also haven't meta)
takle has joined #nixos
newhoggy has joined #nixos
<avn>
so I just haven't idea, from where meta.priority come
eacameron has quit [(Ping timeout: 260 seconds)]
newhoggy has quit [(Ping timeout: 252 seconds)]
takle has quit [(Ping timeout: 260 seconds)]
<joachifm>
avn: hrm, there's no meta.priority in the applications/virtualization/virtualbox either, from what I can see
<joachifm>
joy
<avn>
joachifm: It why I ask, not silently make PR ;)
newhoggy has joined #nixos
<avn>
btw, if I put { config }: argument to package -- it will be nixpkgs.config from configuration.nix?
<joachifm>
avn: I expect it would be, for the package set instantiated as part of configuration.nix eval
<joachifm>
but not for package sets outside of that context
newhoggy has joined #nixos
<avn>
I want to put list of host/realhost/port/key attrsets somewhere in configuration.nix and refer them from custom version of fetchgitPrivate
osa1_ has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
osa1 has quit [(Ping timeout: 260 seconds)]
<joachifm>
avn: maybe specify a custom fetchgitPrivate wrapper via packageOverrides, or is that what you had in mind?
newhoggy has quit [(Ping timeout: 240 seconds)]
newhoggy has joined #nixos
justelex has quit [(Ping timeout: 268 seconds)]
<avn>
joachifm: I have few fetchgit statements, which refer "gitlab.private" host, I want to add indirection layer to refer different host/port and also specify "deployment key"
osa1 has joined #nixos
osa1 has quit [(Changing host)]
osa1 has joined #nixos
<avn>
joachifm: so overriden fetchgitPrivate can substitute only all hosts at once, proper implementation can allow put { host= ...; hostname= ...; port= ... } (and optionally key) in host configs
newhoggy has quit [(Ping timeout: 240 seconds)]
osa1_ has quit [(Ping timeout: 260 seconds)]
matthewbauer has quit [(Ping timeout: 260 seconds)]
<joachifm>
avn: so I take it you want to set nixpkgs.config.hostConfig (or whatever the parameter would be called) in configuration.nix, on a per machine basis
newhoggy has joined #nixos
ambro718 has joined #nixos
<avn>
joachifm: is my current proposition ;) Except I was unsure -- where it should be in configuration nix
<Filystyn>
ericoan th eproblem is when i start alsa mixer i see i have the intel card but the default set is some pulseaudio thing
<avn>
Filystyn: qasmixer work well, event if PA enabled
<Filystyn>
?
<avn>
qastools package
<Filystyn>
how is this related?
<avn>
well, possible I catch your problem wrong
<Filystyn>
i just want to have by default intel card
<ericnoan>
Filystyn: pulseaudio is the audio subsystem, thats what you use
<mojjo>
joachifm: ok, I see. For instance in modules/system/activation/activation-script.nix, apply is not a function but another set {script = "some bash.."} . Where's the magic coming from?
<Filystyn>
Card: HDA Intel PCH but by default i get
<Filystyn>
pulseaudio
<ericnoan>
try installing the qastools package and running "qasmixer"
<Filystyn>
ok
Filystyn has quit [(Read error: Connection reset by peer)]
Filystyn has joined #nixos
Filystyn has quit [(Changing host)]
Filystyn has joined #nixos
jabesed has joined #nixos
<Filystyn>
ok got it installed running
<Filystyn>
what I am supposed to do now?
dejanr has joined #nixos
<ericnoan>
is anything muted?
<ericnoan>
is it a laptop?
<Filystyn>
laptop
<ericnoan>
Filystyn: check on your laptop if you see a mute button
<Filystyn>
i found the card;-)i have sound
<ericnoan>
Filystyn: it might be a FN button (one of the F1-F12 buttons)
<Filystyn>
stop**
<Filystyn>
i mean
<joachifm>
mojjo: the only apply I see in activation-script.nix is a function
<Filystyn>
i have sound normaly the problem is
<Filystyn>
there can not be played two sounds
<Filystyn>
so i wanted the card to be enabled
<ericnoan>
what fixed it?
DIzFer has joined #nixos
<Filystyn>
It was not fixed i wrote chaoticaly
<ericnoan>
only one sound? hmm
<Filystyn>
The problem is if i want to listen to music and start some other thing that gives sound
<Filystyn>
the other thing wont get sound
<Filystyn>
wont give*
<Filystyn>
I can have only one program giving away sound at times
<joachifm>
mojjo: basically, it receives the final attrset and generates the complete activation script
<mojjo>
joachifm: oh, you're right. It's a function taking a set and returning a set. I was wondering about the return value.. But considering, what you just wrote: I guess it will be merged to ther other parts of the activation script...
<joachifm>
mojjo: my understanding is that all merging will have taken place at the point apply is evaluated; the script attribute is consumed by top-level.nix and boot/stage2.nix
cfricke has joined #nixos
ambro718 has quit [(Ping timeout: 260 seconds)]
<Filystyn>
hmm seems i can have multiply sounds
<Filystyn>
but i get some strange error on sound init
<Filystyn>
in one game
<Filystyn>
probably game fault seems the sounds actualy works fine
<Filystyn>
anyway thx gyuys
<mojjo>
joachifm: alright, I'm getting an idea of what it does. was a bit misleading what I wrote before. The reason I'm digging here is that I figured out that the example value for systemActivation script does not compile. I'm getting "error: Cannot merge definitions of `system.activationScripts.stdio' given"
<mojjo>
joachifm: I think the example value is wrong, it does not need the text/deps extra layer. it works without
<Filystyn>
btw do I have alternative to pulseaudio?
<joachifm>
mojjo: hm, briefly scanning it, seems the apply procedure handles both cases (string or { text = ...; deps = [ ... ]; }). one possibility here is that stdio is already specified as string, but if you additionally specify it as a { text = ...; } then it won't be able to merge the two
<joachifm>
mojjo: you'll note that stdio is a standard activationScript fragment
<joachifm>
mojjo: that's what I'd hypothesise, but then my head tends to spin whenever I look at the innards of the module system :)
marsel has joined #nixos
maikklein[m] has left #nixos ["User left"]
thineye has quit [(Remote host closed the connection)]
eacameron has joined #nixos
<mojjo>
joachifm: yeah, looking in there always makes your head spin.... Yeah, I also read the condition you point to. But what's the sense for it when the mergin already happened before the apply function. I think it's kinda a bug. Since the default options are already in the string fashion, merging with the other style cannot happen.
<joachifm>
mojjo: that's not what I meant; the stdio fragment is already defined as a string; when you provide an additional definition using the { text = ... } syntax, the module system will attempt to merge the two values; this occurs before apply enters into the picture
<joachifm>
hence the error
<sphalerite>
Filystyn: pulseaudio usually helps with getting sound playing from multiple applications at once. It's somehow possible without pulseaudio as wlel, but I don't know anything about it. If an application is misbehaving and you're using pulseaudio, it often helps to wrap it in padsp (e.g. I launch minecraft using "padsp minecraft")
romildo has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
<mojjo>
joachifm: I agree 100%.
<joachifm>
mojjo: it's a little unfortunate that the example value would produce this result, though
<joachifm>
mojjo: maybe propose a new example that additionally showcases what the deps thing is for; I don't know that its too clear from the example
<srhb>
Is Thomas Tynkkynen here under some nick?
thc202 has joined #nixos
<joachifm>
srhb: I believe that is Dezgeg
newhoggy has quit [(Remote host closed the connection)]
ibor has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
newhoggy has joined #nixos
<srhb>
joachifm: Thank you :)
<srhb>
Oh, Tuomas. Sorry I butchered your name.
jabesed has quit [(Ping timeout: 260 seconds)]
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
newhoggy has quit [(Ping timeout: 255 seconds)]
<Heffalump>
I have a recent checkout of nixpkgs in /home/ganesh/nixpkgs/general, and I'm trying to setup a build environment for cabal-install with "nix-shell -I nixpkgs=home/ganesh/nixpkgs/general '<nixpkgs>' -A haskellPackages.cabal-install --pure". The resulting shell doesn't even have ghc in the path. What am I misunderstanding about the right way to do this?
<Heffalump>
sorry, that's -I nixpkgs=/home/ganesh/nixpkgs/general - the path is correct, I just mis-re-typed it
<sphalerite>
Heffalump: you'll want 'ghc.withPackages (ps: with ps; [cabal-install])'
<obadz>
gchristensen: do they have a separate /dev? could we use them to spit put ext4 images?
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/v9WG6
<NixOS_GitHub>
nixpkgs/master eb4792a Vladimír Čunát: nixos manual: add a note about "nofail" FS option...
<obadz>
in other words, can one be root inside the container without being root on the host?
eacameron has joined #nixos
ajp_ has joined #nixos
<gchristensen>
no
<gchristensen>
at least, not as far as I know :)
ajp has quit [(Ping timeout: 258 seconds)]
eacameron has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
<thineye>
For certainty: Runtime Dependencies are also set in buildInputs and they remain after build?
<gchristensen>
no, if the output of a build refers to a store path, it is considered a run time dependency. run time dependencies may be explicitly added via propagatedBuildInputs
<thineye>
oh, ok, thanks :)
newhoggy has joined #nixos
<NixOS_GitHub>
[nixpkgs] lheckemann opened pull request #25344: dolphin-emu: include more dependencies (master...dolphin-deps) https://git.io/v9WZL
eacameron has quit [(Ping timeout: 252 seconds)]
ajp_ has quit [(Ping timeout: 245 seconds)]
mt has quit [(Remote host closed the connection)]
eacameron has joined #nixos
ajp has joined #nixos
<joachifm>
obadz: user namespaces are supposed to let you do that
<joachifm>
obadz: I think recent systemd have --private-users or something for nspawn, if that's what you're referring to by "containers"
hoverbear has quit [(Remote host closed the connection)]
BramD_ has quit [(Remote host closed the connection)]
dgpratt has quit [(Remote host closed the connection)]
gleber_ has quit [(Remote host closed the connection)]
carter has quit [(Remote host closed the connection)]
rodarmor has quit [(Remote host closed the connection)]
feepo has quit [(Remote host closed the connection)]
<joachifm>
used to be that root in container was root on host but perhaps that's improved lately
<obadz>
could we run some of the nixos tests in containers rather than vms then?
<gchristensen>
absolutely
<joachifm>
looking at the systemd-nspawn manpage now it seems to have gained lots of capabilities since I last looked :)
eacameron has quit [(Ping timeout: 260 seconds)]
<Unode>
hi all, is there any way to install a dependency of a package that is not available as an attribute name without building the entire package?
<gchristensen>
which one?
<obadz>
Unode: got a specific usecase?
<Unode>
In the past I've installed android-platform-tools-r24.0.2 which is pulled in with the android-sdk
<Unode>
I only want the 'adb' binary
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 3 new commits to master: https://git.io/v9WZu
<NixOS_GitHub>
nixpkgs/master 2bd39ab Daiderd Jordan: csdp: fix darwin build...
<NixOS_GitHub>
nixpkgs/master f240276 Daiderd Jordan: blas: fix darwin library id
<joachifm>
gchristensen: nor any other artefact that would imply rights to source code; you can only get it by paying them, which is fine per GPL, I think
<copumpkin>
obadz: nope containers are fine
<gchristensen>
ahh makes sense
<obadz>
copumpkin: i haven't kept up with all your work on image generation but would it be worth considering containers?
justanotheruser has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
<joachifm>
gchristensen: but if you buy an appliance that runs a grsec enabled kernel, I believe you'd be within your right to ask the vendor for the complete source, including grsec
nh21 has joined #nixos
<joachifm>
gchristensen: which you would then be free to redistribute &c
<gchristensen>
makes sense
nh2 has quit [(Ping timeout: 245 seconds)]
<gchristensen>
how much does grsec cost? :P
<joachifm>
I've read about 200 USD per month but now it's "tailored to your needs" or whatever
<joachifm>
it's ridiculous ...
<joachifm>
the 200USD price, that is, though I don't have a source for that one, just something I read :)
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/v9Wn0
<NixOS_GitHub>
nixpkgs/master 0c4de3c Tim Steinbach: linux: 4.4.64 -> 4.4.65
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.03: https://git.io/v9WnE
<NixOS_GitHub>
nixpkgs/release-17.03 41c2084 Tim Steinbach: linux: 4.4.64 -> 4.4.65...
<gchristensen>
disappointing.
<NixOS_GitHub>
[nixpkgs] FRidh closed pull request #25159: wxWidgets: move wxGTK-* to one wxWidgets folder (master...wxwidgets-refactor2) https://git.io/v9kWV
<copumpkin>
obadz they don't help unfortunately. With a separate file system namespace I can in theory have separate mounts but in practice those can only be bind mounts and similar things. If you try to mount e.g., ext4 inside a mount namespace you'll still fail unless you're superuser
noam_ has joined #nixos
noam_ has quit [(Remote host closed the connection)]
<copumpkin>
I think upstream is concerned about the filesystem code being exploited to escalate privileges
samotarnik has joined #nixos
<gchristensen>
seems reasonable
newhoggy has quit [(Remote host closed the connection)]
<samotarnik>
hey everyone, nixos newbie here. just updated my configuration with instructions here https://github.com/NixOS/nixpkgs/issues/24172 to make it work with gnome 3.22 but now all my profile settings inside gnome are gone. it is as if gnome-tweak-tool does not work, for instance i try to change the background image and it does not change. can someone help me with the correct configuration please? my config is at https://github.com/samo
<jophish>
What's the neatest way of testing a program which manipulates the nix store? The two ways I can think of are: to test it with a storepath of /tmp/xxxx/nix/store, or to test it in some container/vm with store path /nix/store
<jophish>
What can anyone suggest?
<makefu>
jophish: i think the best way would be to test it in a throw-away container
<jophish>
makefu: yeah, that's the way I'm leaning
<jophish>
I wonder what the nicest way of setting this up using nix will be
<jophish>
I know that there is a whole bunch of machinery for this, but I've never used it before
cfricke has quit [(Quit: WeeChat 1.7.1)]
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to master: https://git.io/v9Wct
<NixOS_GitHub>
nixpkgs/master a167826 Joachim Fasting: nixos/hardened profile: disable user namespaces at runtime
<NixOS_GitHub>
nixpkgs/master 56e1133 Joachim Fasting: nixos/lock-kernel-modules: fix typo in unitConfig...
eacameron has quit [(Ping timeout: 240 seconds)]
pie_ has joined #nixos
thaega has quit [(Remote host closed the connection)]
eacameron has joined #nixos
<copumpkin>
obadz: yeah, otherwise it would be perfect :(
vaibhavsagar has quit [(Ping timeout: 245 seconds)]
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/v9WW3
<NixOS_GitHub>
nixpkgs/master 5717417 Vladimír Čunát: lib.makeScope: sync comment after rename in #25285
mbdb92 has joined #nixos
kuznero has joined #nixos
<kuznero>
Hi All
<obadz>
kuznero: hello
<obadz>
copumpkin: makes sense :-/
<kuznero>
When building a new package I keep getting following error: "cp: missing destination file operand after '/nix/store/z1gphh5f416k6v4bvlxm7a76z4jlkr2x-FSharp.AutoComplete-0.32.0/lib/dotnet/FSharp.AutoComplete'..." How can I troubleshoot this type of error? The build phase is completed successfully...
aanderse has quit [()]
bennofs has joined #nixos
aanderse has joined #nixos
nithor has quit [(Quit: WeeChat 1.0.1)]
mbdb92 has quit [(Client Quit)]
mbdb92 has joined #nixos
<obadz>
kuznero: have you applied changes to the nix pkg tree?
mbdb92 has quit [(Client Quit)]
<kuznero>
obadz: yes
<obadz>
are you calling cp in those changes?
<kuznero>
nope, that is not part of my package
<obadz>
are you touching anything dotnet/fsharp related?
<kuznero>
well, it might be part of the content I am pulling from GitHub though
<kuznero>
Yes
<joko>
ok, I'm trying to rebuild Hydra and I get "Can't locate Nix/Store.pm in @INC", any idea what's missing?
<obadz>
kuznero: hard for me to give a generic debugging recommendation. want to push your changes to a branch on github and I can make a recommendation?
<kuznero>
obadz: no, `cp ...` is not part of source from GitHub, just checked
<kuznero>
obadz: great! just a moment...
eacameron has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
justelex has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
nh2 has quit [(Ping timeout: 260 seconds)]
nithor has joined #nixos
MrBiTs has joined #nixos
nithor_ has joined #nixos
newhoggy has joined #nixos
nithor_ has quit [(Client Quit)]
<MrBiTs>
Hi, everybody. Just installed NixOS Gorilla with success on an Acer Aspire M. Full hardware support.
justanotheruser has quit [(Read error: Connection reset by peer)]
the-kenny has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
bennofs has quit [(Ping timeout: 252 seconds)]
Itkovian_ has joined #nixos
derjohn_mob has quit [(Ping timeout: 245 seconds)]
<MrBiTs>
Sometimes I install packages thru nix-env -iA nixos.package, sometimes putting them into configuration.nix. Anyway I do, I manually put package reference onto configuration.nix. Is there a way to generate a configuration.nix file with current NixOS configurations?
Itkovian has quit [(Ping timeout: 260 seconds)]
vaibhavsagar has quit [(Ping timeout: 260 seconds)]
nh2 has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
<MrBiTs>
Cool, gchristensen. I'm trying to figure out how to write this kind of tool
sellout- has quit [(Quit: Leaving.)]
<kuznero>
obadz: replied... "Absolutely! This is how NuGet packages are getting restored into current solution in order to then build it. I am just linking to existing content for Nix packages. And it actually works."
kragniz has joined #nixos
MrBiTs has quit [(Quit: Leaving)]
justanotheruser has quit [(Remote host closed the connection)]
vaibhavsagar has quit [(Ping timeout: 260 seconds)]
justanotheruser has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
<ambro718>
Is firefox borken in nixos 17.03?
<MichaelRaskin>
obadz: gchristensen: user namespaces means that one can be root inside the container but not on the host, but this is not as secure as desired
<gchristensen>
ambro718: what are you experiencing? my firefox on 17.03 seems to be ok
vaibhavsagar has joined #nixos
<MichaelRaskin>
copumpkin: is the user namespace ext4 mount limitation a current limitation, or is it just because you cannot access the devices anyway?
<ambro718>
gchristensen: I deleted the .mozilla and now it seems fine, the problem was "tab crashed" immediately at startup. I must have turned on that new multiprocess thingy, that is probably broken.
<gchristensen>
ack
<obadz>
syn ack
<ambro718>
ack+data
<ambro718>
that's what I was complaning about yesterday in ##networking, it's pretty inefficient to just send an ack then some data, nobody seems to figure out the ack could also have data and save a packet :)
ertes has quit [(Ping timeout: 260 seconds)]
<obadz>
ambro718: it's not like applications have any control over this :()
vaibhavsagar has quit [(Ping timeout: 240 seconds)]
<ambro718>
yes, applications just do connect()+send(). Though the OS could have a period where it will wait for a send() before sending the ack, hoping to combine the data with the ack.
* gchristensen
has no data to send
<obadz>
kuznero: (building)
<kuznero>
obadz: thanks for spending time!
<obadz>
kuznero: glad someone is taking over the F# stuff :)
<obadz>
it was getting a little stale
eacameron has joined #nixos
newhoggy has joined #nixos
<kuznero>
obadz: that is the only way to keep using it at work ;)
<obadz>
nix is?
<kuznero>
I am using NixOS on VirtualBox on Windows. Others coding in Visual Studio ;) still
jsgrant- has quit [(Ping timeout: 240 seconds)]
<kuznero>
yes
<obadz>
can't say that I'd rather write F# in vim+nixos+vbox than in vstudio directly :)
<gchristensen>
vscode + nixos + vbox? :)
<kuznero>
:) I am trying to push for more functional languages, not only F#. And NixOS/Nix is another perfect example to help me with that. Hence, the effort :)
<obadz>
kuznero: your output is in src/FsAutoComplete/bin/Release/* not in FSharp.AutoComplete/bin/Release/*
<obadz>
kuznero: nix-shell -A dotnetPackages.FSharpAutoComplete_0_32_0
<obadz>
kuznero: then set -x ; genericBuild
<steveeJ>
it seems that (builtins.getAttr x pkgs) is not called but instead returns a function, but I wouldn't know why or how to chagne it since it works standalone
<kuznero>
I see. I always felt blind without it... Was always just using `nix-build -A ...` which does not show a lot of details...
<MichaelRaskin>
\\
<gchristensen>
steveeJ: well sometimes it'll be a function
<kuznero>
obadz: we are running a ton of F# services under mono without a single issue
<kuznero>
obadz: but I am happy to see that dotnet coreclr is there. Will need to try at some point!
<obadz>
kuznero: as you can see a simple Http.RequestString() just hangs :-/
<obadz>
well not anymore with my workaround, but it was a symptom of a greater issue imho
<kuznero>
obadz: There were a lot of issue wrt proxies...
<kuznero>
obadz: by the way what `set -x` is for?
<obadz>
re the dotnetPackages, I see that you're using fetchNuget a lot but would suggest you try to transition to builds from source whenever possible as that's the nix way and enables patches, etc.
<steveeJ>
gchristensen: cool, what's isAttrs?
<obadz>
kuznero: tells bash to print everything it does
<kuznero>
obadz: thanks! I will try to do more builds then going forward. Still getting used to things in Nix land
<obadz>
kuznero: also I see you're splitting packages by versions like FSharpAutoComplete_0_32_0 ⇒ unless you have a very specific reason to do that, best to just have the latest stable version
<steveeJ>
gchristensen: I wish there was a proper tool that could search for definitions
<kuznero>
obadz: that is problemmatic, since other packages require specific versions
<gchristensen>
steveeJ: I usually search "fooFn =" here: http://search.nix.gsc.io/ and if that is missing, I search the nix manual for it
<steveeJ>
in nix, "name =" is not a good search term because of sets have such-named arguments
<obadz>
kuznero: ok so that would be a specific reason, but often times you can get away with 'latest'
<gchristensen>
I've almost never run in to an ambiguity problem with "x ="
<obadz>
kuznero: when you can't it's fine to pin versions but then maybe factor things out so there's no copy/pasting when possible
<kuznero>
obadz: agree, that is why (after some suggestions) I upgraded mono46(.0) to mono46(.2)
<obadz>
yes saw that :)
nckx has quit [(Quit: restarting my GuixSD server)]
<steveeJ>
gchristensen: ok, I'll stop complaining now ;-) thanks
nckx has joined #nixos
<obadz>
kuznero: may I ask where you work (that uses Nix & F#) ?
m0rphism2 has joined #nixos
<kuznero>
obadz: that is not official. Only F#/mono/linux runtime is official and only for my team as I was trying to get it approved for about 2 year by now. It is DanskeBank in Denmark
<gchristensen>
steveeJ: to be clear, I don't love our docs situation.
newhoggy has joined #nixos
vaibhavsagar has joined #nixos
eacameron has quit [(Ping timeout: 258 seconds)]
nckx has quit [(Client Quit)]
m0rphism1 has quit [(Ping timeout: 268 seconds)]
nckx has joined #nixos
<kuznero>
obadz: I am trying to locate that error after `set -x; genericBuild` but it is huge... Can I redirect output from genericBuild to a file somehow?
<obadz>
maybe try genericBuild 2>&1 | vim -
osa1_ has joined #nixos
<obadz>
or genericBuild 2>&1 | tee -a some_file.log
nckx has quit [(Client Quit)]
nckx has joined #nixos
vaibhavsagar has quit [(Ping timeout: 260 seconds)]
newhoggy has quit [(Ping timeout: 260 seconds)]
<kuznero>
obadz: thanks
osa1 has quit [(Ping timeout: 240 seconds)]
<NixOS_GitHub>
[nixpkgs] bjornfor pushed 1 new commit to master: https://git.io/v9W8F
justanotheruser has quit [(Read error: Connection reset by peer)]
justanotheruser has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
justanotheruser has quit [(Read error: Connection reset by peer)]
<kuznero>
obadz: I will definitely have a look, but frankly without yet reading the material I can tell that the biggest problem is the mindset of folks working for very long time in a big organization. Though I must say that it is changing slowly but surely lately.
justanotheruser has joined #nixos
frigate_freedom has joined #nixos
<obadz>
yes I think this is addressed by the "baby steps" approach
takle has quit [(Remote host closed the connection)]
<obadz>
though in banks info sec can often be the bottleneck to adopting new technologies…
vaibhavsagar has quit [(Ping timeout: 240 seconds)]
<kuznero>
It is monumental to get this monster moving ;)
dfranke has quit [(Remote host closed the connection)]
<gchristensen>
kuznero: does your company use npm?
<kuznero>
gchristensen: some parts of it does as far as I know
justanotheruser has quit [(Remote host closed the connection)]
<gchristensen>
garbas' approach of treating Nix as a build tool made it very easy to introduce to my work
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to release-17.03: https://git.io/v9W4Y
<NixOS_GitHub>
nixpkgs/release-17.03 1536b8d Shea Levy: Backport halvm from master
vaibhavsagar has joined #nixos
justanotheruser has joined #nixos
takle has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
takle has quit [(Remote host closed the connection)]
<kuznero>
gchristensen: we might have a chance with that. Though we have a very polished and streamlined setup wrt CI/CD for now. Surely nothing comparable to Nix capabilities and guarantees, but still management will not grant us that amount of time for now.
justanotheruser has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
takle has joined #nixos
takle has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
takle has joined #nixos
<gchristensen>
aye. I solved my own problems with nix, and eventually built something nobody thought possible. now we use nix.
justanotheruser has quit [(Remote host closed the connection)]
sellout- has joined #nixos
Itkovian_ has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
takle has quit [(Remote host closed the connection)]
<kuznero>
gchristensen: do you have a showcase that you can share? Maybe slides or some sort of github repo that is public?
justanotheruser has joined #nixos
<gchristensen>
not yet, still going through legal and security review.
<gchristensen>
I can describe though:
justanotheruser has quit [(Remote host closed the connection)]
<kuznero>
gchristensen: that would be great...
takle has joined #nixos
justanotheruser has joined #nixos
gchristensen has left #nixos ["WeeChat 0.4.2"]
gchristensen has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
vaibhavsagar has quit [(Ping timeout: 240 seconds)]
justanotheruser has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
<gchristensen>
my work has a tool for managing our mysql infrastructure, where we have billions of rows and hundreds of database machines. it used to be very difficult and expensive to test this tool. in about a week I built a test harness that spawns several mysql servers in VMs, registers them with an internal asset tracker, runs the tool, and asserts final state
alx741 has joined #nixos
<gchristensen>
tests run in about 2 minutes, a similar puppet-based workflow would take hours. using actual hardware takes days.
<kuznero>
gchristensen: that should inspire a lot of folks! You should consider show casing it on one of the conferences around Nix...
takle has quit [(Ping timeout: 240 seconds)]
frigate_freedom has left #nixos []
<gchristensen>
not many conferences on nix around me
<kuznero>
obadz: I was considering exposing a few meta packages for Vim-folks. For example today in order to enable integration with F# 4.0 I need to install:
<kuznero>
dotnetPackages.FSharpAutoComplete
<kuznero>
dotnetPackages.FSharpCompilerCodeDom
<kuznero>
dotnetPackages.FSharpCompilerService
<kuznero>
dotnetPackages.FSharpData
<kuznero>
dotnetPackages.FSharpFormatting
<kuznero>
obadz: same applies to another set of packages for F# 4.1 (actually not to Vim, but in general - compiler services).
<gchristensen>
that said, I'm hoping fpletz will let me speak at nixcon 2017
<mitchty>
what is the best way to have nix setup PKG_CONFIG_PATH for multiple things? I notice graphviz sets lib/pkgconfig to its lib/pkgconfig directory
<kuznero>
obadz: So, what if to wrap those into `dotnetPackages.FSharpCompilerServices40` and `dotnetPackages.FSharpCompilerServices41`?
<obadz>
kuznero: why would you need FSharp.Data for vim ?
<kuznero>
obadz: It is needed indirectly by one of the other packages
<kuznero>
if I remember correcly
<kuznero>
s/correcly/correctly/
<obadz>
kuznero: shouldn't that get picked up transparently when you install that other package then?
<kuznero>
obadz: that is correct, will clean it up. Thanks! But what do you think about aggregating meta packages?
derjohn_mob has joined #nixos
earldouglas has quit [(Remote host closed the connection)]
<kuznero>
obadz: the thing is that it is not really for vim folks. It is a generic compiler services.
<kuznero>
So, it can be used by Emacs, or another plugin in VS Code.
<obadz>
but why doesn't it depend on all the right things then?
<obadz>
dotnetPackages.FSharpCompilerService should have all the appropriate dependencies to be usable no?
Infinisil has quit [(Ping timeout: 246 seconds)]
<mitchty>
so how do package config files get installed normally to ~/.nix-profile/lib/pkgconfig?
jsgrant-_ has joined #nixos
<mitchty>
or am I supposed to set PKG_CONFIG_PATH manually
takle has quit [(Ping timeout: 260 seconds)]
<kuznero>
I think it might actually be the problem of vim-fsharp plugin. That has `make fsautocomplete` line at the time it is installed. It might be that it fails to properly bind to the right version perhaps?
<Fare>
obadz, hi
Itkovian has joined #nixos
<Fare>
I can't get an IDE fsharp working in nixos, neither monodevelop nor vscode
<kuznero>
Fare: I think I have Vim setup working nicely with F# 4.0 and mono46 if that might be of any interest.
roconnor has joined #nixos
<obadz>
Fare: halo
<obadz>
kind of gave up on F# dev myself
<NixOS_GitHub>
[nixpkgs] kuznero opened pull request #25354: Enable editor integration with F# 4.1 (master...dotnetPackages) https://git.io/v9WBA
jsgrant-_ has quit [(Ping timeout: 240 seconds)]
<obadz>
Fare: monodevelop seems to be built and runs, i think one has to install the appropriate F# plugin
<obadz>
kuznero: why create a separate package for F# 4.1 rather than bump the F# 4 package?
<mitchty>
gah, turns out i had a stash that updated the offending nix pkg to install the pc files correctly which it doesn't do right now for... dumb reasons
<hodapp>
hm, I need to look at what F# support is like in NixOS
<kuznero>
obadz: at the time I was doing it, it seemed safer. Now I think it might make sense to make it an upgrade instead. One small thing though is that fsharp is currently built on `mono40` whhereas fsharp41 requires mono42 at least - preferrably mono46.
<hodapp>
in case I get drunk and F# looks like something I should learn
<obadz>
kuznero: but we've just merged the Mono 4.6 version bump right?
<obadz>
kuznero: I'd suggest avoiding splitting packages when possible as it increases the support burden very rapidly. (Again, unless there's a good reason for it like python2 vs python3)
takle has joined #nixos
<hodapp>
haaaah, python, good reason
* hodapp
runs away
<kuznero>
correct. I can try doing it as upgrade and let's see if it is breaking something else :) It was breaking before. The location of `Microsoft.FSharp.Targets` file change from 40 to 41. That file is normally referenced from solutions and that exact thing will break (should break).
oida has joined #nixos
<obadz>
hodapp: well a distro that only supported one of { python2, python3 } wouldn't be very popular…
<hodapp>
has there been any push to use .NET Core with F# on Linux?
<kuznero>
obadz: I can see that F# should be one single package. For now they haven't introduced non-backwards compatible features (at least 4.1 can compile all 4.0 solutions, not the other way around).
<hodapp>
obadz: that was more me rolling my eyes at Python 2.x still being in so much use than having any problem with now it's done in Nix.
<kuznero>
obadz: I will submit another PR to fix it then
<simpson>
hodapp: Some of us will never leave Python 2. Is it really that big of a deal to you?
jsgrant-_ has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
<obadz>
kuznero: great, and same thing for the dotnet packages, we should avoid version splits unless absolutely necessary and then write in the comment why the version split was required
<hodapp>
simpson: Apparently it's a far bigger deal to you.
<kuznero>
obadz: I don't think there is the case for single package in case of dotnetPackages for now. But I will keep an eye on it going forward.
<simpson>
hodapp: Well, yes, in that I'm a pig here and you sound like a chicken.
<obadz>
kuznero: you should remove the version numbers from the name and the the clashes will be obvious :)
<kuznero>
obadz: I know what you mean, I am just saying that all versions needed
<obadz>
ah
<hodapp>
simpson: No, in that sense, I'm also a pig.
<simpson>
hodapp: Oh, okay. Then surely you understand: Some of us will never leave Python 2.
takle has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
Infinisil has joined #nixos
<hodapp>
I'm stuck there from libraries that are.
justanotheruser has joined #nixos
<jluttine>
how do i enable openvpn support? i try to connect with network manager, but it doesn't work. in journalctl i see: "VPN plugin: failed: connect-failed (1)"
<obadz>
kuznero: hmm so who requires F#.Data 2.2.3 that cannot accomodate 2.3.3 ?
<obadz>
kuznero: also if we're splitting, the default should be the latest version with only packages that can't handle the new version pointed to a fully spelled out older version
newhoggy has quit [(Ping timeout: 240 seconds)]
<srhb>
jluttine: Looks like there's a networkmanager_openvpn package. Did you get that installed?
romildo has quit [(Quit: Leaving)]
takle has quit [(Ping timeout: 240 seconds)]
justanotheruser has quit [(Read error: Connection reset by peer)]
justanotheruser has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
<jluttine>
srhb: i think that's installed by default with network manager
<kuznero>
obadz: vim-fsharp will require new version. I like the idea for having latest version and explicitly pointing to old ones. Will try changing towards it.
justanotheruser has joined #nixos
justanotheruser has quit [(Read error: Connection reset by peer)]
<jluttine>
(not sure how to check if some package is installed in the currently active system)
justanotheruser has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
<srhb>
jluttine: I don't think it is unless that's the version of networkmanager you explicitly installed.
<srhb>
jluttine: I may be mistaken though
justanotheruser has quit [(Remote host closed the connection)]
<srhb>
jluttine: Try anyway :)
justanotheruser has joined #nixos
<jluttine>
srhb: ok, i'll try soon when i boot back to nixos. now i already booted to my parabola so i could get my vpn-required stuff done immediately :)
<jluttine>
srhb: pptp worked without any additional packages. i just had to open some holes to my firewall
<srhb>
jluttine: I switched away from using the networkmanager openvpn stuff,but I'm _pretty_ sure that's what I had to do. Oh hang on, my work config should still have it...
justanotheruser has quit [(Remote host closed the connection)]
<srhb>
jluttine: In systemPackage on my work config that does openvpn with networkmanager, I have networkmanagerapplet, networkmanager_openvpn, networkmanager_pptp, networkmanager_vpnc
justanotheruser has joined #nixos
<srhb>
jluttine: And networkmanager.enable = true and my user is in the networkmanager group.
<srhb>
jluttine: But that looks like all :)
takle has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
justanotheruser has quit [(Remote host closed the connection)]
justanotheruser has joined #nixos
takle has quit [(Ping timeout: 258 seconds)]
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/v9W0a
<NixOS_GitHub>
nixpkgs/master 2c54fa0 Alexander V. Nikolaev: nixos: allow supply customized locale package...
<NixOS_GitHub>
nixpkgs/master a5d3642 Michael Raskin: Merge pull request #22489 from avnik/nixos-locales...
ryantrinkle has joined #nixos
jsgrant-_ has quit [(Remote host closed the connection)]
oida has quit [(Ping timeout: 240 seconds)]
nicknovi1 has joined #nixos
takle has joined #nixos
thineye has joined #nixos
<thineye>
Is it right, that if I am developing on a channel, I should do a git rebase ..., like said in https://nixos.org/nixos/manual/#ch-development for a pull request? Because I read on a tutorial that I should never use rebase on public branches where other people are working on. Or is this rule not valid for a PR?
<Fare>
obadz: the plugin provides syntax highlighting correctly, but when I try to run, I get an error: System.IO.FileNotFoundException: Could not load file or assembly 'FSharp.Core, Version=4.4.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a' or one of its dependencies.
<obadz>
seen that one before
<obadz>
don't really recall the solution, I think it might have involved symlinking FSharp.Core.dll in the plugin directory?
<NixOS_GitHub>
nixpkgs/master 0168e48 Aneesh Agrawal: passlib: switch from insecure pybcrypt to bcrypt...
<NixOS_GitHub>
nixpkgs/master 189479a Aneesh Agrawal: radicale: Add NixOS test with Python 2...
<kuznero>
If I need to test if my change broke other packages what should I do (other than pushing and creating PR and letting TravisCI show me if it will happen to be working at that time)?
takle has quit [(Ping timeout: 260 seconds)]
<Fare>
obadz, thanks. Can you find -name '*.dll' ~/.cache/Mono*
<Fare>
also, monodevelop is schizophrenic about whether to heed Xdefaults' X
<Fare>
Xft.dpi: 280 or not
<kuznero>
MichaelRaskin: I am not sure how does it work... I need to test a specific package against the rest...
<MichaelRaskin>
kuznero: what do you mean?
<MichaelRaskin>
nox-review tries to build everything changed by the PR
<kuznero>
MichaelRaskin: I see. And if I am running nixpkgs from master on GitHub, it might be failing due to some other reason, right?
<MichaelRaskin>
Well, if you change a dependency of a failing package, yes
takle has joined #nixos
<MichaelRaskin>
I think -k should let you get a list of failing packages, but not sure
<kuznero>
MichaelRaskin: not sure I understood... I made an upgrade, but have no overview of what packages depend on it, thus do not know if I broke anything else. When I will run this nox-review it will detect such outer dependencies and then build them? Or will it just build everything?
<MichaelRaskin>
It will only build transitive reverse dependencies
<kuznero>
MichaelRaskin: ok, thanks
<kuznero>
MichaelRaskin: btw, how much TravisCI reliable? I noticed that on the same packages it can sometimes error and sometimes not...
takle has quit [(Ping timeout: 246 seconds)]
<MichaelRaskin>
Well, assume there is a 50% chance of an irrelevant failure… But build logs can be useful even in this case
schoppenhauer has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
schoppenhauer has joined #nixos
fresheyeball has quit [(Quit: WeeChat 1.7.1)]
eacameron has quit [(Ping timeout: 246 seconds)]
peacememories has joined #nixos
takle has joined #nixos
eacameron has joined #nixos
takle has quit [(Ping timeout: 245 seconds)]
eacameron has quit [(Ping timeout: 260 seconds)]
marsel has joined #nixos
takle has joined #nixos
<jophish>
fetchTorrent might not be so terrible on a lan or other private networ
<jophish>
k
eacameron has joined #nixos
<jophish>
It would be so nice to be able to hook up all the machines in the office with one big nix store
takle has quit [(Ping timeout: 245 seconds)]
<jophish>
which would be mirrored nice and quickly hard link information
<jophish>
perhaps ipfs is what we want
leat has quit [(Read error: Connection reset by peer)]
eacameron has quit [(Read error: Connection reset by peer)]
<MichaelRaskin>
The horrible part is that I have started with a cutoff around the beginning of the year and haven't even reached April
<MichaelRaskin>
And I skip the stuff where I have negative amount of clue, too!
Infinisil has joined #nixos
<gchristensen>
I know the feeling >.<
<gchristensen>
I pushed too hard to merge a bunch of PRs a bit ago and burned out, am taking a bit of time off.
bennofs has quit [(Ping timeout: 260 seconds)]
<MichaelRaskin>
With PRs I have a periodic competition between having burnt out (which doesn't go away with time, no) and horror when looking at the counter
<clever>
i have a PR that has been sitting unmerged for about a week now
<MichaelRaskin>
I was under impression that you have asked for access already.
<clever>
taktoa has been saying that i need to just get commit access
Infinisil has quit [(Ping timeout: 240 seconds)]
<LnL>
clever: what!
k2s has joined #nixos
<MichaelRaskin>
clever: is ioport Linux-specific?
<clever>
MichaelRaskin: probably, checking the source to confirm
k2s has quit [(Client Quit)]
<MichaelRaskin>
Wouldn't be too suprised if it was basically «inb», but I would hope that requires kernel code, not just root
<clever>
You must compile with -O or -O2 or similar. The functions are defined as inline macros, and will not be substituted in without optimization enabled, causing unresolved references at link time.
<clever>
iopl - change I/O privilege level
<clever>
iopl() is Linux-specific and should not be used in programs that are intended to be portable.
Infinisil has joined #nixos
<clever>
i believe its using a linux syscall to allow userland direct ioport access
<MichaelRaskin>
Is this conversation we just had supposed to freak out half the newbies at the channel, I wonder
<clever>
somebody in #ghc recently said "after seeing a bunch of questions asked here by Nix users, I've sworn off the idea of using Nix for anything ever"
<MichaelRaskin>
What were the questions that scared this person so much?
<clever>
dont know, only joined the channel recently and it sounds like a long-term trend
<clever>
description = "direct access to I/O ports from the command line"; look good for ioport?
seanparsons has joined #nixos
newhoggy has joined #nixos
<MichaelRaskin>
Should I nitpick about capitalisation or ask what meaning is added by the word «Direct»?
<clever>
nix-repl> ioport.meta.deserror: getting status of ‘/home/clever/apps/nixpkgs/pkgs/pkgs/os-specific/linux/ioport’: No such file or directory
<clever>
seems i can still crash nix-repl, heh
<clever>
nix assertion throw during tab completion
<clever>
$ sudo ./ioport/bin/outb 0x80 0x01
<MichaelRaskin>
Did it actually crash?
<clever>
MichaelRaskin: this command causes the LED display on my motherboard to display 01
<clever>
nix-repl returned status 1, and didnt enable local-echo
<clever>
so i'm now typing blind in that shell
<MichaelRaskin>
As if you don't know how to run reset
<clever>
i do, but that shows how badly it crashed
<MichaelRaskin>
Did you just admit the possibility that I wouldn't understand simply «didn't reenable local-echo»?
<MichaelRaskin>
But yes, such an immediate crash does sound like a bug.
<clever>
just stating how badly nix-repl crashed, it failed to restore terminal settings on exit
<sziszi>
Guys, I don't really know what to do. I'm using nixOS almost 3 months now. I'm having issues with third party binaries(some closed source ones) because they expect the FHS. I managet to patch some, and I used nspawn to make another few working but this Nix lang doesnt quite click for me. A part of me want to leave this system behind, but I don't want to leave this deterministic system which nix provides.
<sziszi>
I'm also fascinated by Disnix/hydra/nixops. I'm sure others experienced difficulties, but how did you came them over?
<clever>
i didnt notice it had left echo off until i tried to get the exit status
newhoggy has quit [(Ping timeout: 240 seconds)]
<simpson>
sziszi: Have you considered putting your third-party stuff in FHS wrapper environments? nixpkgs does this for e.g. Steam.
<MichaelRaskin>
Well, it is hard to notice echo being off if you don't type
<clever>
yeah
<simpson>
sziszi: For me, the main thing was realizing that you can be *incremental* in your changes. You can make your changes as small as one or two lines, save, roll out with a canary/tests/etc., and if it doesn't work, just roll back.
<DIzFer>
sziszi: steam-run is mighty useful for that
simony has joined #nixos
<MichaelRaskin>
But I would expect that using Nix will continue to require using the Nix language, too, for the forseeable future
<Judson>
`
<sziszi>
simpson: that rollback stuff is what I aint gonna leave that easy. That stuff is dope.
<MichaelRaskin>
As for FHS environment construction — I would say we probably have more than one, but the Steam-targeted one is the most polished
<sziszi>
simpson, DIzFer I'm looking into more in steam then.
<simpson>
sziszi: Look at it this way. I know Puppet. I've done Puppet at all three of my most recent jobs. My best friend wrote a Puppet book and I have a signed copy.
<simpson>
It took me maybe a week of using Nix for me to decide that I would never use Puppet for personal stuff again.
<gchristensen>
^ same but chef
<DIzFer>
yeah, steam-run is a package that puts the same FHS environment used for steam at your service
<Judson>
+1 me too also
<LnL>
^ same but ansible :)
<DIzFer>
`steam-run proprietary-bin` should work for most stuff
<clever>
MichaelRaskin: and PR is open
<NixOS_GitHub>
[nixpkgs] cleverca22 opened pull request #25370: ioport: init at 1.2 (master...ioport) https://git.io/v9W1j
<MichaelRaskin>
I just used a complicated tangle of funionfs mounts to install each package on A/BLFS into its own personal slice, and then came across Nix…
<gchristensen>
MichaelRaskin: you are fascinating
<clever>
i used to run LFS on my router
<sziszi>
clever: holy hell!
<MichaelRaskin>
Don't have any advice about the language, though: after J, Scheme and long-form Bash, learning Nix language by reading the manual did sound like a plan. It worked out.
<clever>
sziszi: and it wasnt a normal router, it was a pentium1 with 128mb of ram and 2 ethernet cards
<clever>
sziszi: maxing out the DSL modem brought the cpu usage up to about 60%
<sziszi>
later it will be my workstation AND my router
zarel has quit [(Quit: Leaving)]
<clever>
i had to retire the LFS router when i upgraded to fiber, 370mbit download speed wouldnt suit a pc with only 100mbit NIC's lol
<MichaelRaskin>
Ah right, I can beat the encouragements by comparing with Chef etc. At some point I got really annoyed by some properties of NixOS (not NixPkgs). Writing my own set of Nix expressions to generate all the bootscripts is still a better idea than using a package manager other than Nix on my main personal laptop.
<clever>
but the isp provided router was just too garbage to keep using
<sziszi>
okay guys, you're persuasive :) and I got some motivation to look into some problems
hiratara has quit [(Ping timeout: 252 seconds)]
<gchristensen>
sziszi: also ask around for help :)
<MichaelRaskin>
On the one hand, you should remember about survivor bias. On the other hand, if you infer that we have solutions — maybe sometimes complicated — for a lot of things, you will be right.
<sziszi>
gchristensen: I'm not that kind of person. I'm sick of people who beg for help constantly, and I dont want to become one. At least till I learn how to contribute.
hiratara has joined #nixos
<gchristensen>
there are places between the two extremes, and we _want_ to help. don't set yourself up for failure
<MichaelRaskin>
sziszi: take a random conversation from the logs that you can understand and that is not a retelling of documentation; propose a writeup as a PR for Nix or NixPkgs manual; receive a lot of nitpicking feedback and try to accomodate it
neeasade has joined #nixos
<MichaelRaskin>
Given that you are a person who does understand what is there not to understand, that would often be a very valuable piece of documentation
neeasade has quit [(Remote host closed the connection)]
justan0theruser has joined #nixos
kragniz has quit [(Remote host closed the connection)]
justanotheruser has quit [(Ping timeout: 268 seconds)]
<sziszi>
gchristensen: That is not the full story, I'm also try to solve my problems with search engines, or figure it out by myself. When I fail I still learn something, and I try again till I do it properly.
newhoggy has joined #nixos
simukis has quit [(Quit: simukis)]
simukis has joined #nixos
marsel has quit [(Ping timeout: 240 seconds)]
simukis has quit [(Client Quit)]
Guest48735 has quit [(Quit: Lost terminal)]
<sziszi>
MichaelRaskin: Good idea, I'll keep an eye on it.
simukis has joined #nixos
<MichaelRaskin>
If you do have a commitment to doing things properly, Nix language manual (distinct form NixOS manual) is something you will end up reading anyway
simukis has quit [(Client Quit)]
<MichaelRaskin>
Reading it earlier than absolutely inavoidable could be an upfront cost, but could also save some effort overall
simukis has joined #nixos
<MichaelRaskin>
I don't know if you want to find and read Eelco Dolstra's thesis about Nix: it explains the model of the world that Nix uses, it may help you get a coherent idea of Nix
thc202 has quit [(Ping timeout: 268 seconds)]
gengwird has joined #nixos
<Guest54573>
jews did 911
<Guest54573>
jews did 911
<Guest54573>
jews did 911
<Guest54573>
jews did 911
<taktoa>
MichaelRaskin: if sziszi is an academic person used to reading papers then I'd recommend it
Guest15141 has left #nixos [requested by gchristensen (Guest15141)]
<MichaelRaskin>
taktoa sziszi: well, for a person with academic background there is no doubt, it maybe be useful to try reading parts of the thesis even without such background, but it could be hard
<Supersonic112_>
suck my dick michaelraskin
<sphalerite>
I need to extract some version info from its canonical source, which is a CMake script. What's the best approach to take for that?
Supersonic112 has left #nixos [requested by gchristensen (Supersonic112)]
<sphalerite>
Low flying trolls today...
<joepie91`>
lol fuck you faggot
simukis_ has joined #nixos
<taktoa>
gchristensen: keep the +o for a little bit lol
DIzFer has quit [(Quit: DIzFer)]
simukis has quit [(Ping timeout: 268 seconds)]
<joepie91`>
NIGGERS
<joepie91`>
NIGGERS
<joepie91`>
NIGGERS
<joepie91`>
NIGGERS
<joepie91`>
NIGGERS
<NixOS_GitHub>
[nixpkgs] garbas pushed 1 new commit to master: https://git.io/v9Wy3
<Infinisil>
Thanks for the info about nickserv everybody! Learning more everyday :D
<sziszi>
MichaelRaskin: our company using a Windows "server"
<sphalerite>
just going to repeat my question now that the chaos is over: I need to extract some version info from its canonical source, which is a CMake script. What's the best approach to take for that?
<gchristensen>
now now, no distro/OS hate
hiratara has joined #nixos
newhoggy has quit [(Remote host closed the connection)]
<MichaelRaskin>
gchristensen: no hate only contempt
<Judson>
All the same, there's folks who will take it personally. "Oh, I use Ubuntu; I guess I must be dumb."
<Infinisil>
I am using windows for gaming, but for everything else it's distgusting (sorry not sorry)
<Infinisil>
sphalerite: I don't think anybody knows an answer right now, I for one don't even know what you mean exactly
<Infinisil>
Question: When I make changes to <nixpkgs/nixos>, how do I test it? I know I can nix-build -A <package> for a package, but that doesn't work for that
<clever>
-rwxr-xr-x 1 root root 1.1M Oct 26 2014 /usr/bin/irssi
<steveeJ>
this one's interesting: /nix/store/v5fgaa2nsvpmp5z76a13sr283z3hzlwq-y0r98fq15wk65a6698s2zqj9qha3prg9-zqjrx28j59h2hnlw7xyhrmz44lz7zja6-1qi370kx2606q7pky5swlvqq3pswf544-ap969333517ycs957vsmlzhr2ghrpl8g-43l7x72rcgp5nhj3qivn771r1gpwfzj6-m52s6zbzcyq48agq6ac74y1250hc91i6-2f2ajcrsw43qgpx7zq7ml0bihz6gjq96-nixos.lock
<steveeJ>
how is this filename generated?
<clever>
gchristensen: and criminaly out-dated gentoo systems are in close 2nd! :P
<steveeJ>
nix-instantiate says it's too long :-)
<gchristensen>
clever: :P
<sphalerite>
MichaelRaskin: what about if I want to put it in the derivation name?
ma9e has quit [(Remote host closed the connection)]
ryanartecona has joined #nixos
<MichaelRaskin>
sphalerite: but won't the cmake file exist only after fetchurl has been launched and not during evaluation?
<sphalerite>
MichaelRaskin: this is for a default.nix to ship with a project
<sphalerite>
so it'll already have the source available
Twix has joined #nixos
Twix has left #nixos ["Verlassend"]
tv has joined #nixos
newhoggy has joined #nixos
<MichaelRaskin>
I guess you still don't want to launch anything in the evaluation phase…
<MichaelRaskin>
Could you do an equivalent of readFile in cmake?
Infinisil has quit [(Quit: Good night!)]
<sphalerite>
well how am I meant to avoid that when I want to compute the derivation's name based on that file's contents?
<sphalerite>
I figure something along the lines of readFile and some regex magic would be what I want
<sphalerite>
but it has to be at evaluation time
cheshircat has joined #nixos
<MichaelRaskin>
splitString could work, but it would be royally annoying
mekeor` has joined #nixos
nicknovi1 has joined #nixos
mekeor has quit [(Ping timeout: 252 seconds)]
wkennington has quit [(Quit: Leaving)]
<steveeJ>
I really have 0 clue what is wrong with that lockfile name. What user error could possibly lead to this?
rdesfo[m] has left #nixos ["User left"]
Jookia has quit [(Ping timeout: 248 seconds)]
pie_ has quit [(Ping timeout: 240 seconds)]
Jookia has joined #nixos
<gchristensen>
I hate replacing ubuntu servers ᕙ(⇀‸↼‶)ᕗ
estewei has joined #nixos
justan0theruser has joined #nixos
<steveeJ>
is there something builtin I can use to write a textfile in the nix-store? I'm using writeTextFile from nixpkgs now and seem to have trouble with the lock name
<Ralith>
gchristensen: are you aware of having emitted some suspicious DCC messages about an hour ago?
<gchristensen>
yes :(
<Ralith>
what's the story
<gchristensen>
someone caused my old weechat to segfault by sending malformed DCC messages, they then took over my nick (I've since enabled nick protection) and started spamming people.
__20h__ has joined #nixos
<Ralith>
interesting
justanotheruser has quit [(Ping timeout: 240 seconds)]
<Ralith>
I assume freenode staff have been notified
<gchristensen>
they stepped in when my client segfaulted
thineye has joined #nixos
<Yaniel>
they sent a global NOTICE two days ago or so
<clever>
oh, didnt notice that
<sphalerite>
hah
noam has joined #nixos
nahamu has joined #nixos
gengwird has quit [(K-Lined)]
ericsagnes has joined #nixos
takle has quit [(Read error: Connection timed out)]
abrar has joined #nixos
takle has joined #nixos
cpennington has joined #nixos
thineye has quit [(Quit: Konversation terminated!)]