stephenjudkins has quit [Remote host closed the connection]
ryanartecona has joined #nixos
<alp>
rnhmjoj[m], yes, that's in fact more or less the only somewhat relevant link that I could find =) but I'm seeing those errors when trying to run say nixops or virt-manager (python programs). I suppose something's not right with my environment, but not sure what. my PYTHONPATH seems quite crowded in a normal shell session right from the start, but I don't know what to look for really
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] adisbladis pushed 2 new commits to master: https://git.io/vAvHP
<NixOS_GitHub>
nixpkgs/master 2e0540f Will Dietz: bfg-repo-cleaner: 1.12.15 -> 1.13.0
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master be8cc03 adisbladis: Merge pull request #34679 from dtzWill/update/bfg-1.13.0...
<tilpner>
abathur - "nixos-rebuild test: Build and activate the new configuration, but do not add it to the GRUB boot menu"
<abathur>
ah, right, that explains that part, then
<abathur>
do you know what I'm missing regarding the profiles/commands?
<tilpner>
Where/How did you add p7zip?
sigmundv_ has quit [Remote host closed the connection]
<abathur>
configuration.nix, in system packages
<tilpner>
And you then did a nixos-rebuild switch?
<abathur>
I think I tried that at one point since, though the immediate step was just nixos-rebuild test; reboot
<tilpner>
Working as intended then. nixos-rebuild test is not supposed to persist across reboots
<kini>
How does one usually deal with optional dependencies in nixpkgs? I have a package which is able to extend its functionality if various other programs are available, but not all users would necessarily want that functionality, and the check needs to be done at build time.
<abathur>
aha
<abathur>
figured I was just reasoning about the system poorly
sigmundv has joined #nixos
<abathur>
ok, so, when I rebuild, changes are active as soon as it finishes
<abathur>
and if I did a test, rebooting rolls back
<abathur>
?
<tilpner>
abathur - There is no simple "rebuild". You should always specify what mode you're using nixos-rebuild with
<abathur>
right, more accurately that changes made by nixos-rebuild switch and nixos-rebuild test would be immediately available
Drakonis has joined #nixos
<tilpner>
abathur - A common solution is adding { withFoo ? false, foo ? null } to the arguments
<tilpner>
* kini
<kini>
ah, makes sense
<tilpner>
abathur - Yes, switch and test will activate the newly-built configuration
<kini>
the choice still remains of whether to supply withFoo = false or withFoo = true in all-packages.nix though, but at least it's a bit more flexible than always requiring foo or never requiring foo
sigmundv has quit [Ping timeout: 256 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] adisbladis pushed 1 new commit to master: https://git.io/vAvQ8
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master bfa1390 Daniel Frank: burp: 2.0.54 -> 2.1.28
<tilpner>
kini - There are a few global preferences, like environment.noXlibs or config.pulseaudio. Unfortunately I'm not aware of a list of these
<abathur>
tilpner - Thanks much. I think I just ended up assuming, from working through the installer, that selecting at the bootloader was an essential step. A number of things make more sense, now. :)
ryanartecona has quit [Quit: ryanartecona]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] andrew-d opened pull request #34688: Fix build failures with DBD-SQLite (master...adunham/dbd-sqlite) https://git.io/vAvQN
NixOS_GitHub has left #nixos [#nixos]
griff_ has quit [Quit: griff_]
spietz has joined #nixos
silver_ has quit [Read error: Connection reset by peer]
freeman42x[NixOS has quit [Ping timeout: 248 seconds]
<NixOS_GitHub>
[nixpkgs] dtzWill opened pull request #34689: zsh: set configureFlags and checkFlags at nix level, also fix cross (master...fix/zsh-modern) https://git.io/vAv7u
<Lisanna>
I'd like to use nix-push in place of nix-store --export, but I'm not sure what the corresponding import function would be.
<Lisanna>
I want to use nix-push since it only updates what it needs to
<Lisanna>
whereas nix-store --export dumps the entire closure every time
abathur has quit [Ping timeout: 260 seconds]
ditadi has joined #nixos
<brennie>
So I'm trying to instal libjpeg-turbo with its headers. I did `nix-env -i libjpeg-turbo`. but `nix-env -q libjpeg-turbo --out-path` doesnt list a dev path, only bin
<brennie>
How can i get it to install headers correctly
<brennie>
(I'm very new to nix and I'm not super sure about what im doing )
<joshuaks>
Hey all! I just wrote my first deriviation(?) for software not available through the Nix package manager. Could I get some feedback on the quality of what I wrote? It all seems to work.
<joshuaks>
Also, I would like to publish this. Should I contact the author the repo, etc? I'm new to actually contributing to foss at this level.
schoppenhauer has quit [Ping timeout: 256 seconds]
schoppenhauer has joined #nixos
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
LnL has joined #nixos
ditadi has quit [Ping timeout: 255 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vAvxc
<NixOS_GitHub>
nixpkgs/master a008a9c Tuomas Tynkkynen: nixos/sd-image-armv7l-multiplatform: Port RPi config.txt changes from aarch64 image...
NixOS_GitHub has left #nixos [#nixos]
ericsagnes has quit [Ping timeout: 256 seconds]
hamishmack has quit [Quit: hamishmack]
hamishmack has joined #nixos
hamishmack has quit [Client Quit]
MP2E has quit [Remote host closed the connection]
vidbina has joined #nixos
<vidbina>
So, just a curious question... how do most of you handle updates? I started off using 17.09-small and then checkout out `nixos-version --revision` in a local version of the repo that I currently nixo-rebuild against. For updates, I pull the latest 17.09-small for example and reapply my patches in order to be able to reference any of my additions in my nixos-configuration. What is the "idiomatic" way of
<vidbina>
going about updates in NixOS? Also, does anyone here nixos-rebuild against the master branch :P?
spietz has joined #nixos
orivej has joined #nixos
<tilpner>
I avoid a local nixpkgs checkout, so updating doesn't require a rebase
<vidbina>
tilpner: how do you deal with your PR's then... it sometimes takes time before they are available upstream, right?!?
<tilpner>
I use overlays, and then cp -r /cfg/nixos/pkgs/foo ~/dev/nixpkgs/pkgs/.../foo for a PR
<tilpner>
I'll then eventually delete the package from my overlay when a PR is merged
<tilpner>
(So technically I still have multiple nixpkgs checkouts, but I don't nixos-rebuild against the ones I checkout manually)
<vidbina>
tilpner: wow... I haven't yet used OverlayFS yet so I'll look into this workflow and carefully start playing around with it until I own it. Thanks :)
<tilpner>
vidbina - Also, I don't claim this is idiomatic or anything. I don't do a lot of PRs. This is just what worked okay so far
<adisbladis>
vidbina: There are many of us who just have a local checkout of nixpkgs :)
<adisbladis>
vidbina: For work we wrap nixpkgs in a derivation and apply non-upstreamed changes as patches
<vidbina>
tilpner: rodger that, still interesting to get reminded of overlays as a feature. Perhaps I'll have another use for them but awareness already helps.
hamishmack has joined #nixos
<adisbladis>
vidbina: I mostly use overlays for packages that I dont want to upstream
<adisbladis>
Such as small personal projects that no one would end up using
<vidbina>
adisbladis: ha, so if you need to update I guess you just fetch and merge the remote branch and reapply your patches? Also, what did you mean with patches in the work-scenario -- attribute overrides or do you have patches defined in diff-ish patches and then have a way of applying them to the derivation?
badi has joined #nixos
<vidbina>
adisbladis: the phrasing wasn't too eloquent on my question there... sounded kinda stupid... basically I'm just wondering what that work-scenario rougly looks like in code. I'm aware of using overrides, and I could imagine nixpkgs and defining attributes that aren't available yet. Is that kind of what you meant?
Izorkin has joined #nixos
rauno has quit [Remote host closed the connection]
spear2 has quit [Remote host closed the connection]
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] copumpkin pushed 2 new commits to staging: https://git.io/vAvhp
<NixOS_GitHub>
nixpkgs/staging c6aa69f Ben Darnell: curl: Enable http2Support for darwin...
<NixOS_GitHub>
nixpkgs/staging 30d031b Daniel Peebles: Merge pull request #34339 from bdarnell/curl-http2-darwin...
coffeeshot has joined #nixos
abcadr[m] has joined #nixos
AllanDaemon has joined #nixos
YuFanLovezYou[m] has joined #nixos
Ralith has joined #nixos
endformationage has quit [Quit: WeeChat 1.9.1]
benasse has joined #nixos
jluttine has quit [Ping timeout: 256 seconds]
Rusty1_ has quit [Quit: Konversation terminated!]
dywedir[m] has joined #nixos
trikl[m] has joined #nixos
Rick[29SMPittsbu has joined #nixos
ma27 has joined #nixos
jluttine has joined #nixos
retrry[m] has joined #nixos
dmitry-zen[m] has joined #nixos
aaronc[m] has joined #nixos
themistocle1 has joined #nixos
leons has joined #nixos
Khorne[m] has joined #nixos
bobvanderlinden_ has joined #nixos
dpc has joined #nixos
ipj[m] has joined #nixos
aniketd[m] has joined #nixos
discontent[m] has joined #nixos
Adev[m] has joined #nixos
Barnabas[m] has joined #nixos
aspiwack[m] has joined #nixos
pstn has joined #nixos
fogbugs[m] has joined #nixos
Manish[m] has joined #nixos
mhsjlw[m] has joined #nixos
wmertens[m] has joined #nixos
efx[m] has joined #nixos
smbarte2[m] has joined #nixos
pvn[m] has joined #nixos
andyjscott[m] has joined #nixos
dkellner[m] has joined #nixos
mrdaak[m] has joined #nixos
M-dpetranek has joined #nixos
Obscurity[m] has joined #nixos
vise890[m] has joined #nixos
zoya-kosova[m] has joined #nixos
seif_madc[m] has joined #nixos
PlasmaStrike[m] has joined #nixos
paperdigits[m] has joined #nixos
chreekat[m] has joined #nixos
mudri[m] has joined #nixos
ConorCurry[m] has joined #nixos
M-berdario has joined #nixos
iohidtbhfam[m] has joined #nixos
puffnfresh has joined #nixos
bachp has joined #nixos
Wysteriary[m] has joined #nixos
grahamc has joined #nixos
sudoreboot[m] has joined #nixos
Guanin has quit [Ping timeout: 260 seconds]
asuryawanshi has joined #nixos
azertyieio has quit [Ping timeout: 256 seconds]
rauno has joined #nixos
octalsrc[m] has joined #nixos
wchresta[m] has joined #nixos
the_m4ur1[m] has joined #nixos
<adisbladis>
vidbina: Yeah I just do a "git pull --rebase origin master" :)
Guanin has joined #nixos
bendlas has joined #nixos
<adisbladis>
Hmm
ma27 has quit [Ping timeout: 255 seconds]
<adisbladis>
I need to decode base64 in a derivation
<adisbladis>
I wonder what a good way to do that is
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<vaibhavsagar>
hey, my Haskell build expression does not seem to be respecting overrides
Orbstheorem[m] has joined #nixos
<vaibhavsagar>
does anyone know why this might be?
<vaibhavsagar>
it is building a version of haskell-src-exts that I explicitly told it not to build
zhavypath[m] has joined #nixos
primeos[m] has joined #nixos
<vaibhavsagar>
and when I try to build only that it works fine but if I try to build the whole project it compiles both(?)
cornu[m] has joined #nixos
adva[m] has joined #nixos
yochai[m] has joined #nixos
noctuid has joined #nixos
chominist[m] has joined #nixos
badamson[m] has joined #nixos
Kallegro[m] has joined #nixos
hedning[m] has joined #nixos
ycy[m] has joined #nixos
mirlur[m] has joined #nixos
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
nh2[m] has joined #nixos
Wieke[m] has joined #nixos
tobim[m] has joined #nixos
AdamSlack[m] has joined #nixos
srid[m] has joined #nixos
gudea[m] has joined #nixos
danielrf has joined #nixos
johnazoidberg[m] has joined #nixos
Singularity[m] has joined #nixos
fearlessKim[m] has joined #nixos
kevincox has joined #nixos
M-fishy has joined #nixos
fogbugzz[m] has joined #nixos
Dirt[m] has joined #nixos
peterhoeg has joined #nixos
robinbb[m] has joined #nixos
y0no[m] has joined #nixos
voobscout[m] has joined #nixos
teh[m] has joined #nixos
bhipple[m] has joined #nixos
nmikhailov[m]1 has joined #nixos
moredread[m] has joined #nixos
abbafei[m] has joined #nixos
baconicsynergy[m has joined #nixos
revoltmedia[m] has joined #nixos
<dhess>
vaibhavsagar: I believe some packages like haskell-src-exts are hard-wired in the GHC config. They're in common-configuration.nix and its variants, I think.
bdimcheff has joined #nixos
corngood has joined #nixos
h30[m] has joined #nixos
NickHu has joined #nixos
ayyjayess[m] has joined #nixos
offlinehacker[m] has joined #nixos
kainospur[m] has joined #nixos
mith[m] has joined #nixos
Hook[m]1 has joined #nixos
itorres has joined #nixos
Vincenzo[m] has joined #nixos
koalaeagle[m] has joined #nixos
<dhess>
I've had trouble overriding those in the past.
<vaibhavsagar>
dhess: it's overridden in configuration-ghc-8.4.x as well though
adisbladis[m] has joined #nixos
<vaibhavsagar>
exactly the same override in fact :)
matrixkrav has joined #nixos
Drakonis[m] has joined #nixos
<dhess>
vaibhavsagar: you mean it's overridden to the version you want?
<vaibhavsagar>
yes
necronian has joined #nixos
<dhess>
Is it bootstrapping ghc-8.4 with an older ghc? maybe it's coming from there?
edef[m] has joined #nixos
imbigbrother[m] has joined #nixos
PeterRomfeld[m] has joined #nixos
rycee has joined #nixos
wzy8L-B2[m] has joined #nixos
DIzFer[m] has joined #nixos
<vaibhavsagar>
I don't think that's it, because when I try to create an environment with just haskell-src-exts it works fine
M-liberdiko has joined #nixos
jluttine[m] has joined #nixos
rnhmjoj[m] has joined #nixos
gridaphobe[m] has joined #nixos
nmikhailov[m] has joined #nixos
Deleh[m] has joined #nixos
<dhess>
vaibhavsagar: have you been able to build anything with ghc-8.4 in Nix yet? I can't get past 'happy'.
benkolera has joined #nixos
hendrik[m] has joined #nixos
rihardsk[m] has joined #nixos
Diagonal[m] has joined #nixos
apurvapavaskar[m has joined #nixos
wervenyt[m] has joined #nixos
davidar has joined #nixos
Pneumaticat has joined #nixos
<vaibhavsagar>
dhess: after pinning nixpkgs to the current master, yes
bennofs[m] has joined #nixos
jmorriss[m] has joined #nixos
<vaibhavsagar>
I do that in the file I linked above
wilornel has joined #nixos
lycium[m] has joined #nixos
EB[m] has joined #nixos
xurxo[m] has joined #nixos
florianjacob has joined #nixos
swayf[m] has joined #nixos
nickdshsadhkas[m has joined #nixos
petersjt014[m] has joined #nixos
balsoft[m] has joined #nixos
spacekitteh[m] has joined #nixos
Dezgeg[m] has joined #nixos
equalunique[m] has joined #nixos
jack[m] has joined #nixos
stuh84[m] has joined #nixos
<dhess>
hmm I just had a build against master that failed again on happy, just a few hours ago I think
fgaz[m] has joined #nixos
copumpkin has joined #nixos
admin[m]4 has joined #nixos
Phil[m]3 has joined #nixos
Elephant454[m] has joined #nixos
stites[m] has joined #nixos
<vaibhavsagar>
that's probably it then
sargon[m] has joined #nixos
olejorgenb[m] has joined #nixos
<vaibhavsagar>
could be that something else is failing
regnat[m] has joined #nixos
TimePath has joined #nixos
ArdaXi[m] has joined #nixos
sphalerite has joined #nixos
dash has joined #nixos
wizek[m] has joined #nixos
ison[m] has joined #nixos
AntonLatukha[m] has joined #nixos
freeman42x has joined #nixos
reactormonk[m] has joined #nixos
mankyKitty[m] has joined #nixos
musicmatze[m] has joined #nixos
happyente[m] has joined #nixos
joh[m] has joined #nixos
LiuWeiHua[m] has joined #nixos
davidak[m] has joined #nixos
midchildan[m] has joined #nixos
scott has joined #nixos
malteof[m] has joined #nixos
timclassic has joined #nixos
hl has joined #nixos
peel[m] has joined #nixos
AlanPearce[m] has joined #nixos
asppsa[m] has joined #nixos
spawnthink[m] has joined #nixos
r3mfKm[m] has joined #nixos
blahblah has joined #nixos
hexkey[m] has joined #nixos
codyopel has joined #nixos
indefini has joined #nixos
randomenduser[m] has joined #nixos
wak-work has joined #nixos
crystalgamma[m] has joined #nixos
dibblego[m] has joined #nixos
Li[m] has joined #nixos
WilliamHamilton has joined #nixos
ajmccluskey[m] has joined #nixos
<dhess>
yeah happy is still failing in master with ghc-8.4. There's a fix in happy master but it hasn't been released to Hackage yet, at least not as of a few days ago
steshaw[m] has joined #nixos
das-g[m] has joined #nixos
fogbugz23[m] has joined #nixos
demyan[m] has joined #nixos
tokudan[m] has joined #nixos
herzmeister[m] has joined #nixos
taohansen has joined #nixos
techtangents[m] has joined #nixos
moredhel has joined #nixos
magnap has joined #nixos
<dhess>
oh it has!
<dhess>
1.19.9
<rauno>
how to raise user open files limit in nixos configuration?
<dhess>
vaibhavsagar: hmm that's weird, it says it's using happy-1.19.9, but I don't see that in my nixpkgs checkout
<dhess>
wth is going on
xvnvx has joined #nixos
<vaibhavsagar>
do you get the same results with my repro though?
<vaibhavsagar>
very strange regardless
<vaibhavsagar>
I'm only going down this rabbit hole because it seemed like a good time to start working on GHC 8.4 compatibility for IHaskell
vidbina has quit [Remote host closed the connection]
<dhess>
yeah same for me re: 8.4
<dhess>
anyway let me try this on my end
<dhess>
I don't see why the result would be any different but why not
<dhess>
well it is getting happy-1.19.9 anyway
<dhess>
that's an improvement
<dhess>
I should re-run my build, I wonder when that dropped
<dhess>
vaibhavsagar: ok so the one with just haskell-src-exts worked
<vaibhavsagar>
yes, isn't that strange
<vaibhavsagar>
somehow hlint is depending on non-overridden haskell-src-exts
<vaibhavsagar>
instead of the one we have overridden
<dhess>
yeah but this is what I was referring to earlier. The haskell-src-exts in configuration-ghc-8.4.x is overridden, but that is only visible to ghc-8.4
<dhess>
the haskell-src-exts in configuration-common.nix is the one used everywhere else and that's 1.20.1
<dhess>
and in fact hlint uses that one specifically as well
<dhess>
sorry it's just hlint and a few other packages in configuration-common.nix that use 1.20.1
<dhess>
anyway point stands. GHC 8.4 is using a special one that is pulling from haskell-suite/haskell-src-exts on GitHub
<vaibhavsagar>
ah, I see
<dhess>
but that's just for the GHC 8.4 dependency. It is not transitive to packages you compile with ghc-8.4
<vaibhavsagar>
even if I override it in my overrides?
<vaibhavsagar>
I can try `extend` instead but that seems like different syntax for the same thing
<dhess>
if you override it in your overrides it should be visible to all the other packages
<vaibhavsagar>
yeah, that's what I though
<vaibhavsagar>
*thought
<dhess>
extend is somehow different and more composable but I don't recall the details. Gabriel Gonzalez mentioned it or got bitten by it in some issue awhile back.
<dhess>
Anyway I just use overlays now and that solves everything :D
griff_ has quit [Quit: griff_]
griff_ has joined #nixos
griff_ has quit [Client Quit]
<dhess>
vaibhavsagar: so I wonder if overrideAttrs doesn't work on a Haskell derivation like you think it should
griff_ has joined #nixos
griff_ has quit [Client Quit]
<vaibhavsagar>
dhess: I switched to overrideAttrs from overrideCabal in release-8.2.nix for e.g. cairo
<dhess>
is overrideCabal available ? That is what's used in configuration-ghc-8.4
<vaibhavsagar>
I'm trying with that now and getting the same results
griff_ has joined #nixos
griff_ has quit [Client Quit]
griff_ has joined #nixos
griff_ has quit [Client Quit]
griff_ has joined #nixos
griff_ has quit [Client Quit]
griff_ has joined #nixos
griff_ has quit [Client Quit]
griff_ has joined #nixos
griff_ has quit [Client Quit]
griff_ has joined #nixos
griff_ has quit [Client Quit]
<dhess>
vaibhavsagar: well just to work around this for now you could just cabal2nix the version from the haskell-suite GitHub and override the whole thing, rather than trying to override a few attrs
asppsa has joined #nixos
<vaibhavsagar>
good idea, let me try that
asuryawanshi has quit [Remote host closed the connection]
Itkovian has quit [Read error: Connection reset by peer]
gawa has joined #nixos
justanotheruser has quit [Ping timeout: 260 seconds]
knupfer has joined #nixos
<vaibhavsagar>
dhess: something very screwy is going on
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti closed pull request #34651: R: add libcxx to default LDFLAGS and CPPFLAGS on Darwin (master...r-darwin-libcxx) https://git.io/vNjNe
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 8ff7ac4 Peter Simons: Merge pull request #34646 from clacke/hledger-darwin...
<NixOS_GitHub>
[nixpkgs] peti pushed 2 new commits to master: https://git.io/vAfTh
i-am-the-slime has quit [Ping timeout: 256 seconds]
<srhb>
Hmm. I can't figure out why imports = [ (pkgs.path + ...) ]; leads to infinite recursion. Can anyone point me to the place where it happens?
mwahl has quit [Ping timeout: 240 seconds]
<srhb>
I'm guessing it's because pkgs itself exists as a module somehow
<tilpner>
I remember that working fine for me, it must be something else
<srhb>
Oh, really? Weird.
<jluttine>
in one nix configuration file, i import <nixpkgs/nixos/modules/profiles/all-hardware.nix>. i'm building this configuration by passing a custom nixpkgs/nixos path to nix-build. so, does these <...> refer to the paths in my installed nixos or to the nixpkgs passed as an argument to nix-build? i'd expect they refer to my current installed system because i'm passing nixos, not nixpkgs, as the argument to nix-build
<jluttine>
anyway
<dhess>
vaibhavsagar: so I wonder if the hlint that's defined in common-configuration.nix is also not overridable, and it is getting pulled in somewhere and that is causing your failure
<dhess>
jluttine: if you are doing `nix-build ... -I nixpkgs=some-other-nixpkgs` then <nixpkgs/...> referes to that one. Otherwise it looks in the usual places (NIX_PATH etc)
<tilpner>
jluttine - Something like that. It's not necessarily pointing to your "currently installed system", you could have old <nixpkgs> in your NIX_PATH
<dhess>
vaibhavsagar: well that's only a theory but glad it is leading somewhere anyway :)
<jluttine>
dhess: nix-build 'nixos.nix' -A config.system.build.isoImage -I nixos-config=iso.nix
<johnw>
i've never seen this during a nix-env -u --leq before: error: reached FD_SETSIZE limit
<jluttine>
so i suppose it refers to the system ones..
<tilpner>
jluttine - You're not passing a custom <nixpkgs> or <nixos>
<dhess>
jluttine: so that is looking in NIX_PATH and then other places. `man nix-build` for more details
<dhess>
the part towards the end about NIX_PATH and friends
<tilpner>
I can reproduce it. I changed something about my config, which might be related
<srhb>
Okay, I'm not crazy then. (Or I may be, but not for this reason :-P)
<fearlessKim[m]>
u must be a minimum masochistic to use nixos :)
<srhb>
Quite the contrary!
<srhb>
tilpner: Do you have a workaround I might adopt?
<tilpner>
srhb - I reverted to <nixpkgs/foo>
jensens has joined #nixos
<srhb>
Damn, I can't do that. Thanks anyway.
<tilpner>
srhb - I also have some imports = [ "${unstableSrc}/foo" ];
<tilpner>
srhb - Why can't you do that?
<srhb>
Because there's no NIX_PATH (or it's wrong) -- I'm relying on a pinned pkgs being passed in.
<tilpner>
Is this for some test/VM setup?
<srhb>
That, too, but also our production servers.
MP2E has joined #nixos
<srhb>
(Essentially the same thing +/- some modules)
* tilpner
checks git logs
<joko>
Hello! Could anyone tell me how to setup kodi? I see config.kodi in all-packages.nix, but I am unsure how to configure it
<srhb>
joko: That's usually your nixpkgs config
<srhb>
joko: Meaning ~/.config/nixpkgs/config.nix
<joko>
srhb: I see now
<joko>
thanks
<joko>
Having pinned pkgs is burdersome, though, you have to attend every update, right?
<joko>
burdensome *
<srhb>
Not really.
<srhb>
Quite the contrary. Things update when I want them to, so breakage happens only when I'm prepared for it.
* tilpner
didn't find anything helpful
<srhb>
tilpner: Thanks for looking anyway :)
CodeWarrior has quit [Ping timeout: 248 seconds]
<joko>
srhb: if you stay on the stable channel, then I find system.autoUpgrade quite convenient
<tilpner>
joko - For my laptop, I want to have to attend (be present for?) every update
<srhb>
joko: Our machines don't even know their own configuration, so I don't think that's possible. Nor do I want it.
<srhb>
Software version bumps are scary and break things :-)
<joko>
tilpner: point taken, it's your laptop. But what if you have to manage tenths of servers?
<adisbladis>
Pinned packages is <3
<tilpner>
Does autoUpgrade also nixos-rebuild??
<adisbladis>
joko: Then I absolutely want pinned packages.
<joko>
tilpner: yup
<srhb>
joko: It's no burden. You sit down, bump the pin, build all the machines, see that everything is working (or confirm it with tests) and deploy everything
<dhess>
srhb: you can use a fixed Nixpkgs in an import like tilpner suggested
<adisbladis>
joko: Whatever went through testing is what gets deployed.
<srhb>
dhess: Then I'd have to import it twice though.
<srhb>
dhess: And thread it through everywhere.
<adisbladis>
Sometimes I cheat and just export NIX_PATH
<joko>
srhb: this is what I am also doing on a personal channel, check that everything is built and then create a new channel, but my servers then get the updates
<joko>
on themselves
<srhb>
joko: If you're the one bumping the channel, that's essentially pinned nixpkgs anyway.
<srhb>
You're just inverting the deployment logic.
<joko>
srhb: indeed
<srhb>
Makes no difference to me. As long as nixpkgs is pinned :P
<joko>
But I am thinking of automating the bump, too
<tilpner>
joko - That might be fine if most packages are in cache. I don't my weak server suddenly recompiling llvm/rust while I'm not looking :)
<tilpner>
*don't want my
<dhess>
srhb: can you wrap nix-build or nixops or whatever it is that you're using to evaluate the expression?
<adisbladis>
joko: I just have a little script that checks which is the latest commit on the channel I want that writes a json file with nixpkgs metadata
<joko>
tilpner: it's a valid concern, had some rough times with that, too
<tilpner>
joko - I actually build everything on my laptop, and then push it to my server, so it never has to compile anything
<tilpner>
(nix-deploy is great)
<srhb>
dhess: You mean setting NIX_PATH? Possible, but nixpkgs is defined by a nixpkgs.nix which fetches a specific commit.
<adisbladis>
tilpner: That already happens with nixops
<srhb>
I think that's a little ugly, too.
<joko>
adisbladis: For me it's currently totally manual, I rebase some commits on the latest nixos-unstable-small and my hydra instance compiles the packages and the channel (no testing still)
<tilpner>
adisbladis - You're right. I used nixops for a while, but I felt there was advantage in my case
<tilpner>
*there was no
* tilpner
forgot how to type
<srhb>
dhess: Yeah, that's what we used to have.
<dhess>
in that case it's a Makefile. For nixops I wrote a little wrapper around it that does the same in shell
<adisbladis>
tilpner: Yeah it might be overkill :)
<dhess>
but in all of my projects I just import a local lib.nix that defines a fixedNixPkgs and I import it where I need it
<dhess>
usually it's only in the overlay or the module imports, so it only happens once
<srhb>
dhess: Yeah, I'm just a little weirded out by it. It seems pretty obvious to me to have a single pinned pkgs coming in from outside.
<srhb>
And I mean, it's entirely possible as long as it's <nixpkgs> rather than --arg pkgs ...
<srhb>
:P
CodeWarrior has joined #nixos
<dhess>
I don't see the difference, other than the ugly string template substitution, but maybe I'm missing something about the efficiency of this approach or something like that.
<srhb>
I mean, there's no reason to go and re-evaluate files normally, so you'd stick to pkgs. In this case it seems necessary.
asuryawanshi has quit [Ping timeout: 265 seconds]
xvnvx has quit []
asuryawanshi has joined #nixos
catch22 has joined #nixos
Ubercow has quit [Quit: Ubercow]
lsyoyom has quit [Ping timeout: 248 seconds]
i-am-the-slime has joined #nixos
i-am-the-slime has quit [Remote host closed the connection]
<NixOS_GitHub>
[nixpkgs] andir pushed 4 new commits to master: https://git.io/vAfRF
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master ea817d7 Andreas Rammhold: squid: fix CVE-2018-1000024 & CVE-2018-1000027
<NixOS_GitHub>
nixpkgs/master 291b05e Andreas Rammhold: squid4: hack to detect our libxml2
betaboon has joined #nixos
silver has joined #nixos
<rauno>
has anyone had to deal with error: * OpenSSL SSL_read: error:1408F119:SSL routines:SSL3_GET_RECORD:decryption failed or bad record mac, errno 0
<Mic92>
pie_: did you were the guy, who tried to build alacritty?
i-am-the-slime has quit [Ping timeout: 260 seconds]
Izorkin has joined #nixos
TonyTheLion has joined #nixos
<TonyTheLion>
I'm trying to build a project with my hydra server, but it fails with hydra-evaluator[19040]: Died at /nix/store/87vxcb62r4wdgzi499parzykylnrz7ck-hydra-2017-11-21/libexec/hydra/lib/Hydra/Plugin/GitInput.pm line 133.
<TonyTheLion>
has anyone seen this before?
Lisanna has quit [Remote host closed the connection]
i-am-the-slime has joined #nixos
<Mic92>
hyper_ch2: Can I force push on your pr branch? I want to make one commit out of it.
<LnL>
dtz: borg evaluates all of nixpkgs for 4 platforms, just like hydra
alex`` has joined #nixos
i-am-the-slime has quit [Ping timeout: 256 seconds]
<hyper_ch2>
Mic92: sure
<hyper_ch2>
if you know how to do that
<Lisanna>
Ugh... I'm using nix-store --export to dump the closure of a bunch of /nix/store paths to a directory, and now I need to import them. But nix-store --import refuses to import anything that doesn't have its dependencies already imported. How do I call nix-store --import on the files in the right order?
orivej has quit [Ping timeout: 240 seconds]
<dtz>
LnL: yeah, which isn't cheap! Or are you saying it can do so successfully so if a PR causes it to OOM something is wrong/ the PR might be bad?
<LnL>
it does so for every pr
hiroshi has quit [Ping timeout: 240 seconds]
<LnL>
the difference with hydra is that the nixos.org instance is configured with a special environment variable to increase the heap size of boehmgc
<nahamu>
Lisanna: how many closures did you export?
<Lisanna>
nahamu actually, just one, but it's a lot of files
<nahamu>
(I've never done this myself, but now I'm curious too)
<Lisanna>
I've read that, yes
<Lisanna>
I'm not doing it quite that way though
<gchristensen>
dtz, LnL: it is clear to me that almost every PR succeeds and this PR doesn't succeed because it is crossing the threshold. whether the difference between master and the PR is .1MB or 1MB or 10MB or 1,000MB I don't know
<LnL>
dtz: I think it's not a bad thing that borg fails if the memory usage increases in some significant way, the memory footprint is pretty large already
<Lisanna>
I can't afford to be piping the entire closure around, I'm in a very performance-constrained environment
<LnL>
gchristensen: true
<exec_>
does anyonw know about libiberty
<nahamu>
Lisanna: how are you planning on distributing the closure you generate?
<Lisanna>
so, I wrote a little program which does nix-store --export on each individual path in the closure to a separate file, and then I only want to import the ones that aren't already in the target store
<Lisanna>
nahamu shared filesystem, getting the files there isn't hard
<LnL>
dtz: if everybody disagrees with that gchristensen could use the same heap size for borg
<nahamu>
Lisanna: why not set up a binary cache?
<Lisanna>
nahamu outbound network is slow
i-am-the-slime has joined #nixos
<gchristensen>
clever: how do I tell nix to emit all those performance stats at the end of a build?
<gchristensen>
or eval or whatever
<Lisanna>
the filesystem is very fast, but networking from the environment is virtually unusable
<Lisanna>
I know nix-store --import will already not import things that already exist, but it still has to process the entire file before it can decide this, so I can't afford to even call nix-store --import on a file I already have. The logic for this is easy, the only snag I'm hitting is I don't know what order to import each file in
<gchristensen>
Lisanna: I'm fascinating to hear what you are working on =)
<LnL>
gchristensen: look at the manpage, NIX_SHOW_STATS, NIX_COUNT_CALLS, ...
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ehmry pushed 1 new commit to master: https://git.io/vAf6T
<nahamu>
(my strategy of pumping you for information to pique the curiosity of people smarter than me seems to be working)
<Lisanna>
haha
<nahamu>
I was just wondering the same
<nahamu>
Seems silly to only support the binary cache over http.
<nahamu>
oh!!
<Lisanna>
Maybe I could use nix-pull, but it's deprecated so that makes me slightly sad ):
<Lisanna>
(and hopefully it works with file://)
<nahamu>
not the greatest idea, but what if you spun up a minimal web server on the client machine serving up the NFS mount?
<Lisanna>
nahamu oh god
<LnL>
nix unstable supports all sorts of uris for substitutes
<nahamu>
2x memory copying, but the file only comes in over NFS once.
<Mic92>
pie_: did you sort it out?
<Lisanna>
nahamu disk operations are fast, so that shouldn't be an issue, it's the CPU that's slow as balls (nix-store has to process the entire file)
<nahamu>
machine A builds stuff and pushes to NFS. machine B mounts NFS, runs local web server and points nix at local web server. Machine B now has a binary cache over NFS.
<pie_>
Mic92, no I didn't
<nahamu>
oh, so 2x the CPU and memory stuff sounds like a terrible idea if it can be avoided.
<Lisanna>
LnL does nixUnstable have an equivalent to nix-pull?
<Mic92>
pie_: I did not catch what the problem was last time.
<pie_>
Mic92, something about rust dependencies but i didnt get any farther than that
<gchristensen>
dtz: this is what I get: cp ../ofborg/ofborg/src/outpaths.nix .; nix-env -f ./outpaths.nix -qaP --no-name --out-path --arg checkMeta true 2>&1 > outs
<gchristensen>
Too many heap sections: Increase MAXHINCR or MAX_HEAP_SECTS
coot has quit [Ping timeout: 256 seconds]
<dtz>
gchristensen: what nix version? store path would be ideal if not any trouble, but short of that knowing what channel built from/against would help re:pinning down the deps used
<dtz>
gchristensen: great, ty. So I'm doing something wrong then :D. I'll investigate, just wanted to bug ya to make sure I understood the borg side of things :D.
<gchristensen>
god I love nix, just nix-store -r that and hopefully it'll reproduce :P
<gchristensen>
"nix is too complicated" people have no idea
<dtz>
but yes "nix-store -r" FTW. 1 sec, appears recent-ish nixos change or something (maybe?) makes it no longer easy to run nix 1.11 on a 1.12 system QQ
<dtz>
(getting 'error: Nix database directory ‘/nix/var/nix/db’ is not writable: Permission denied' FWIW, but haven't poked at why, will just switch to nix stable for a sec)
<samueldr>
I forgot to ♥ nix yesterday, had to rebuild my workstation in a temporary VM, in ~30 minutes + time to download I had a perfect clone running within qemu managed lib libvirt. (it suffered of a non-catastrophic hardware failure)
johnsonav has quit [Remote host closed the connection]
<gchristensen>
=))
<dtz>
haha yeah that's just magical
<gchristensen>
mog: can you get me hooked up with nixops testing instructions?
<gchristensen>
dtz: it looks like your NIX_REMOTE is not correctly set there... set NIX_REMOTE=daemon then call nix 1.11 and it should work?
orivej has joined #nixos
alexteves has joined #nixos
<dtz>
oh, right, I haven't worried about that since ... using non-NixOS. nix 1.12 probably drops that var or something. good call, will try that when switch back--it's convenient to be able to nix-shell into using nix 1.11 tools (not daemon, but), would be great to not lose that.
coot has joined #nixos
<manveru>
1.12 still uses the var, afaict
<kriztw>
How do I get cargo to find e.g. the openssl or sqlite3 libraries on NixOS?
<gchristensen>
kriztw: are you using nix-shell? a shell.nix?
<kriztw>
gchristensen: right now I'm just using rust from nixpkgs-mozilla, and I'm trying to add the shared libraries to e.g. a shell.nix
Neo-- has quit [Ping timeout: 240 seconds]
<kriztw>
gchristensen: I think I tried that combination, but I'll try again, thanks
<kriztw>
gchristensen: ah, it's because rust and openssl must be included together, I had rust installed with nix-env and openssl with my shell.nix
<gchristensen>
ah
i-am-the-slime has quit [Quit: Konversation terminated!]
johnsonav has joined #nixos
rardiol1 has joined #nixos
<LnL>
dtz: oh btw, I think llvm-5 is almost ok. only a small number of failures except for qt5
tmaekawa has joined #nixos
<LnL>
when is 6 supposed to be released?
<hyper_ch2>
Mic92: so it looks good?
<dtz>
dunno, can't repro the max_sects thing. strange. will poke at later. I'll note that the nix used uses boehmgc 7.6.0 and seems to want to use a bit more memory in the eval (8219028 kbytes vs 7802832 kbytes w/1.12); ulimit'ing to 8000000 kbytes causes an error but it's a OOM not MAX_HEAP_SECTS. I am probably missing something, but blargh :P.
<dtz>
sort of error I get when ulimit'ing heap: http://ix.io/Fq4
<dtz>
I've managed to trigger the heap sects thing before, haha, why not now QQ
<dtz>
LnL: rc2 is scheduled for today, feb 21 is target release date (see column on right of http://llvm.org/)
<dtz>
lemme see if I'm sitting on some qt5 clang fixups
<LnL>
hmm, wondering if we should go crazy and switch to 6 for the release
<dtz>
:D
spietz has quit [Ping timeout: 252 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nix/master 84989d3 Eelco Dolstra: Improve filtering of ANSI escape sequences in build logs...
<NixOS_GitHub>
[nix] edolstra pushed 1 new commit to master: https://git.io/vAf9t
NixOS_GitHub has left #nixos [#nixos]
<dtz>
LnL: the linked eval on the switch-default-llvm PR shows qt4 failures too--are those resolved now?
<dtz>
fwiw when using nix 2 on latest nixos, NIX_REMOTE is not set, that was the problem (and setting NIX_REMOTE=daemon for the times I need it works nicely, good call gchristensen :))
<gchristensen>
nice
<mog>
thats how the json should look
<mog>
can you dum the network.nix too?
<mog>
er dump
<gchristensen>
sure
<gchristensen>
and if you wait around an hour or so I can send you a whole series of steps to exactly to where I am now, since that is what I was doing anyway :P
<symphorien>
is there a quide somewhere on how to install nix multi-user on non-nixos linux ?
<mog>
gchristensen, can you try the sqlite backend and see if the problem is the same?
<betaboon>
dtz: so on nix 1.16.x no chance ?
<gchristensen>
yeah in a minute, I accidentally ruined it :P
<mog>
i think it might be related to nonetype, ill double check that type of deploy at lunch. spinning up vms is just very easier for me than taking the one extra step to turn that vm to a nonetype
<dtz>
betaboon: I don't think so, no.
drakonis_ has joined #nixos
<mog>
json is so much prettier than the sqlite...
<gchristensen>
sqlite is a lot nicer of a database though :)
Drakonis has quit [Ping timeout: 256 seconds]
<mog>
mysql/postgresql will be even nicer for that though
<gchristensen>
yes
<mog>
i want to roll it out where i work so mysql was a must for here
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] edolstra pushed 1 new commit to nix-2.0: https://git.io/vAfdb
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/nix-2.0 5193807 Eelco Dolstra: VM tests: Initialize the Nix database with correct NAR hashes/sizes
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy opened pull request #34699: Add composable variant of builtins.path (master...composable-path) https://git.io/vAfFm
NixOS_GitHub has left #nixos [#nixos]
MercurialAlchemi has quit [Ping timeout: 240 seconds]
ssmike has quit [Remote host closed the connection]
ssmike has joined #nixos
<betaboon>
gchristensen: are you just actively working on the nixops/terraform integration ? oO
<jorgan>
but tar is giving a lot of errors like: "tar: nix-1.11.16-x86_64-linux/store/g1g31ah55xdia1jdqabv1imf6mcw0nb1-glibc-2.25-49/lib/libnss_compat.so: Cannot create symlink to `libnss_compat.so.2': No such file or directory"
<jorgan>
is this because I'm missing some requirements? or because I don't have root access?
<pie_>
does mkDerivation actually take a $version attribute or do people just do that
<pie_>
* version attribute
<adisbladis[m]>
pie_: It's not required.
<pie_>
but is it used for anything
<pie_>
i mean besides the user using it for stuff in their script
<pie_>
is it a nix thing or just a convention
<pie_>
im trying to decide between putting it in the mkderivation or using let syntax
<dtz>
I think using "nix build" I'm seeing it create output links that are "dead" -- like, if I interrupt/ctrl-c a build the symlink is there already? anyone else seeing anything like this?
<pie_>
i dont see a package for that strangely enough
<ij>
I couldn't install nixos with nixos-install from arch a while ago, but only now I understood why — because I didn't have /run/current-system/bin/sw in $PATH.
<makefu>
pie_: the gtk2 package may come directly from pkgs.gtk2
<makefu>
you can check out pkgs/top-level/python-packages.nix for more samples
<alp>
hello. I'm having some trouble running nixops, virt-manager and some other python programs except in pure shells (except that pure shells are not handy for running those things). https://github.com/NixOS/nixops/issues/867 I'm getting some python import error, does anyone know what this could be due to (my environment.systemPackages enforcing some PYTHONPATH that's incompatible?) and how to fix it?
ssmike has quit [Remote host closed the connection]
ssmike has joined #nixos
Sonarpulse has joined #nixos
ryanartecona has joined #nixos
asdfjkljkl has joined #nixos
Ross has joined #nixos
asuryawanshi has quit [Remote host closed the connection]
asuryawanshi has joined #nixos
coot_ has joined #nixos
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<chreekat>
what's the "right" way to have ssh-agent available in my KDE session? And why isn't it there by default? :)
xcmw has joined #nixos
<srhb>
alp: Does `which nixops` in both shells give you the same store path?
<srhb>
alp: (You'll have to add -p which to the pure one)
stephenjudkins has joined #nixos
<srhb>
Hm, nevermind, that store path looks good to me.
<srhb>
You're probably correct that it's your pythonpath
goibhniu has quit [Ping timeout: 240 seconds]
<srhb>
alp: Assuming you have a pythonpath in your regular shell, I'd go figure out where that's coming from a nix it entirely.
asdfjkljkl has quit [Quit: Page closed]
Itkovian has joined #nixos
digitus has joined #nixos
Itkovian has quit [Client Quit]
<Guest14344>
Hi all, I'm trying to make an OpenStack basic NixOS image and wondering if anyone's done this before
<srhb>
Guest14344: Did you notice the top comment? :)
<alp>
srhb, yeah I do have a pythonpath in a regular shell too, presumably inherited from my environment.systemPackages. it has paths to guake, pygtk and what not.
<srhb>
alp: As a debugging exercise, it might be useful to at least get rid of that to rule it out (clearing it in your shell will suffice I think?)
<lewo>
Guest14344: I'm using it and it works well
<srhb>
alp: Perhaps in the nix shell directly.
<srhb>
I don't remember exactly how --run works behind the scenes
<alp>
srhb, just to be clear, you're suggesting that I enter the (non-pure) shell and e.g do: export PYTHONPATH= and then try to run nixops?
<srhb>
alp: Yes, something like that :)
<alp>
if that's indeed what you suggest: I tried that and got the very same python import error unfortunately
<srhb>
Oh, shoot.
stephenj_ has joined #nixos
<lewo>
Guest14344: The command at the top the file builds a qcow2 image. This image can be uploaded by using glance. At runtime, cloud-init is started to get your ssh key for the root account.
<alp>
srhb, but the nixops wrapper script sets the python path manually too
stephenj_ has quit [Read error: Connection reset by peer]
<alp>
srhb, I've removed a lot already and the problem's still there. hmm. I hope someone will see that issue and have an idea =) I know very little about python and how it's packaged up in nixpkgs
<mog>
gchristensen, try a git pull
<srhb>
alp: Out of curiosity, can you import weakref from /nix/store/i3bx1iw2d0i3vh9sa1nf92ynlrw324w8-python-2.7.14/bin/python2.7 ?
<srhb>
I think we can rule out the PYTHONPATH at least.
<srhb>
Where's bisect-env when you need it :-)
<judson>
Does anyone reuse a project expression in packages? I'm trying to work out what's the smoothest way to be able e.g. to add git sources for deploy.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] P-E-Meunier opened pull request #34706: Xjump: init at 2.9.3 (master...xjump) https://git.io/vAJWy
NixOS_GitHub has left #nixos [#nixos]
Myrl-saki has quit [Ping timeout: 268 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vAJ8J
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 93aa83d Tuomas Tynkkynen: nixos/release-combined.nix: Make 32-bit iso_minimal a channel blocker...
<alp>
srhb, hmm, I ran the --repair-path command, it fetched python again, I started it and tried importing weakref, same error
phdoerfler has joined #nixos
<srhb>
alp: Yeah, I lost faith in that suggestion since the pure shell presumably uses the same python.
<mog>
gchristensen, for giggles could you do that again, but clearout the .nixops/ dir first?
<alp>
srhb, the pure shell does use the same python
<srhb>
alp: It must be environment leakage somehow. Not sure which, though, but anything that could affect python directly.
<gchristensen>
like my ~/.nixops?
<mog>
im trying to figure out how you keep getting the connection closed
<mog>
or move it out of the way for a second
<mog>
i just want to see a clean nixops sql file
<mog>
im having trouble reproducing issue on my machine
<alp>
srhb, I'm a bit afraid some gnome/python stuffs are messing things up
<srhb>
alp: Start unsetting envs one at a time?
swills has joined #nixos
swills has quit [Changing host]
swills has joined #nixos
<srhb>
alp: Or you know, ideally bisect, but... :-P
<radvendii>
if I want to make a default.nix for the project in the current directory, it doesn't need an 'src', since there's nothing to unpack. What's the right way to go about this, should I disable unpackPhase?
tmaekawa has quit [Quit: tmaekawa]
<coconnor>
radvendii: recall that nix only references files from the nix store.
<coconnor>
radvendii: even without an unpack phase, nix will still need to put the source into the nix store
<the-kenny>
radvendii: src = ./.; This will pull up the current directory into the nix store.
<radvendii>
ah. I tried src = ./; hahaha
<radvendii>
the-kenny: thanks!
<the-kenny>
note that ./. is a bit ugly as it will (as far as I remember) copy the `result` symlink created by nix-buil too, making it recursive :)
<coconnor>
haha aye. the ./ means "what follows is a relative path". Still needs somethingt to follow it
<coconnor>
similarly: instead of ../some-dir you use ./../some-dir
<the-kenny>
coconnor: I think ../foo works
<coconnor>
oh cool!
<coconnor>
there is a builtin to filter the copy IIRC. Nice to avoid including the build products that might be local
<the-kenny>
ah nice
<the-kenny>
It would be even more awesome if it would honor .gitignore too, but that would be too much magic
<coconnor>
builtins.filterSource does that an expression
<coconnor>
and builtins.readFile will read files...
<coconnor>
add in an expression that parses a .gitignore and, bam, done?
<gchristensen>
now all you need to do is implement the parsing of .gitignore which evidently includes 191 lines of documentation in the man page
Mateon3 has joined #nixos
BramD has joined #nixos
Mateon1 has quit [Ping timeout: 268 seconds]
Mateon3 is now known as Mateon1
<coconnor>
anybody got a parser combinator lib in nix? X-D
<gchristensen>
theres that TOML parser...
<coconnor>
maybe go the route of "load to json" then use the builtins.fromJSON
<Mic92>
hyper_ch: I currently migrate my zfs machine. Works well so far. It should be still possible to migrate an encrypted zpool, where the top-level dataset was also encrypted
kim0 has joined #nixos
digitus has joined #nixos
<Mic92>
because zfs create -o encryption=aes-256-gcm -o keyformat=passphrase -o mountpoint=none zroot/root/foo on my system works also
<Mic92>
although, zroot/root is encrypted
<Mic92>
with the old format
<pie_>
if i want to get a set that i wrote in another file in scope, how do i do that?
tertleeltret has quit [Quit: Connection closed for inactivity]
yann-kaelig has quit []
alexteves has quit [Ping timeout: 276 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] edolstra pushed 11 new commits to nix-2.0: https://git.io/vAJEb
<mojjo>
hi! does anybody know, why groups are not listed by the 'groups' command, even though they are declared in users.extraUsers,{username}.extraGroups ?
<srhb>
mojjo: Did you log out and back in?
<mojjo>
yes
<srhb>
Do the groups exist?
<mojjo>
good question... they should, no? wheel, networkmanager, audio.. however I did not create them...
<srhb>
mojjo: cat /etc/groups and see if the username is by the groups
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] edolstra pushed 1 new commit to nix-2.0: https://git.io/vAJz2
<cransom>
leotaku: there's probably a better explanation than 'its weird', but i think setting nixPath in that way would require 2 rebuilds to get an updated channel. if you set the channel once, you can `nixos-rebuild switch --upgrade` and it will update the channel before rebuilding.
<leotaku>
gchristensen: are there any other downsides to doing this? I am currently trying to rebuild nixos with the nix.nixPath set. I am currently waiting for drivers to compile. Will I have to compile every package?
<mojjo>
srhb: yes, they exist, except vboxusers
<srhb>
leotaku: For one thing I think the new channel will take effect _after_ a rebuild rather than before, which to me is confusing.
<srhb>
mojjo: I think you need to log out harder then. :)
<mojjo>
srhb: yeah.. no I re-read your msg, wait
<mojjo>
they exist, but my username is not attached to them...
<srhb>
mojjo: Did you encounter any errors while switching configuration?
<leotaku>
So is it expected that, after rebuilding with the new nixPath, I have to compile drivers? I don't know anything about drivers/compilation, but I wont have to compile all my packages from source now, right?
<mojjo>
actually, no... but I will rebuild again
<srhb>
leotaku: I tried that way around, and it's very confusing for the reasons cransom stated.
ryanartecona has joined #nixos
<srhb>
leotaku: It does not affect what's cached and what's not, however, no.
<leotaku>
Interestingly it just stopped because of an compilation error
mmlb has joined #nixos
<srhb>
mojjo: I can't see any other reason for /etc/group to not be updated.
<leotaku>
srhb, cransom: so both of you would probably not reccomend anyone to set their channels this way, correct?
<mojjo>
srhb: yes, theres an error: chown: invalid user: ‘root.vboxusers’ (it even remains after I took out the vboxusers group)
<Guest14344>
lewo: How are you using NixOS and OpenStack? I'm wondering if I should build a minimal image and use some other ops orchestration thing to put a configuration.nix in place for the different node types
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] markus1189 opened pull request #34710: docs: Explain how to login into built vm (master...doc-vm-login) https://git.io/vAJgX
NixOS_GitHub has left #nixos [#nixos]
<Guest14344>
Since uploading images seems annoying :|
<leotaku>
srhb, cransom: If this is the case I'll probably just stay with the ad-hoc channels.
<srhb>
leotaku: I would not.
<cransom>
leotaku: there are reasons for it, but i would not recommend using it to achieve your behavior
<srhb>
leotaku: Personally I prefer to point it at a local checkout. I don't like channels either, but ymmv.
aborsu has quit [Quit: aborsu]
<srhb>
mojjo: That's a quirky username.
<srhb>
leotaku: It also appears the snabblab repo no longer uses that approach either, if I'm not mistaken.
<mojjo>
srhb: the error disappears when I take out "virtualisation.virtualbox.host.enable = true;"
<srhb>
mojjo: Sounds like something is bugged in that module.
Myrl-saki has quit [Read error: Connection reset by peer]
Neo-- has joined #nixos
<mojjo>
srhb: any better news?
<srhb>
mojjo: You can get the other groups working! Yay! :-)
* srhb
always has a positive outlook
clefru has quit [Ping timeout: 240 seconds]
gamambel has left #nixos [#nixos]
aborsu has joined #nixos
radvendii has quit [Quit: WeeChat 2.0]
<srhb>
mojjo: Other than that, no. virtualbox doesn't even appear to build on unstable, at least with my current kernel.
<mojjo>
srhb: actually the groups are not working apart from that...
<srhb>
mojjo: Still nothing in /etc/group?
<mojjo>
yes.. e.g. the line: "audio:x:17:"
<mojjo>
the others are like this
<srhb>
mojjo: Can you share your user definition then?
<mojjo>
I guess it's safe to just remove the uid = 1001 part?
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mojjo>
(dont really remember why i put it there)
<srhb>
mojjo: Yes, the next uid will be allocated.
<markus1189>
is there some documentation on how the CI for nixpkgs works on pull requests? And also how to see what is buit in hydra(?) and why it failed
<srhb>
markus1189: Which CI exactly?
<srhb>
markus1189: ofborg?
<markus1189>
yes I think that's the one grahamofborg or something like that?
<mojjo>
srhb: the nixops docs only mentions virtualbox... I'd give virtd a try.. would you maybe share a minimal machine config fot virtd?
<Lisanna>
gchristensen yeah, I guess. I've got some number of layers of that, and I want ot flatten it so I can resolve all of them with --arg on the command line
<pikajude>
oh 2.0 is 1.12
<pikajude>
okay
<pikajude>
so i've already been using it
<pikajude>
neat
<Lisanna>
Unfortunately --arg only works one layer deep ):
xcmw has joined #nixos
<mojjo>
srhb: oh, its ok.. i found the part in the docs..
<clever>
Lisanna: --arg does weird things in combination with -A
<gchristensen>
Lisanna: you could maybe do something fancy with builtins.functionArgs?
<Lisanna>
gchristensen yeah I thought of that, but I'm not sure that named arguments are first-class citizens
<Lisanna>
i.e. I don't think there's such a thing as { ${arg1}, ${arg2} }: ...
<gchristensen>
Lisanna: there is { ... } @ args: ... where args contains all the passed named arguments
<Lisanna>
clever: geez, trying to grok that
<clever>
Lisanna: for each attribute in the -A, it will try to apply all --arg's to the function, if it is a function
<clever>
any unused arguments are silently ignored
<Lisanna>
gchristensen sure, but that just lets you refer to the set of arguments as a whole, not sure how it would help me specify progromatically what those named arguments are
<Mic92>
hyper_ch: looks good otherwise.
<hyper_ch>
yey
<Mic92>
I am waiting for my pool to be migrated
<hyper_ch>
migrating 100TB takes 2-3h, right? :)
<Lisanna>
clever OK, so I could get something like what I wanted if I strung each layer of arguments along with a nested attrset
mojjo has quit [Remote host closed the connection]
<clever>
hyper_ch: looks related to nixpkgs, try nix-channel --update
<hyper_ch>
clever: then I have to wait till Mic92 merges my
<hyper_ch>
PR
mojjo has joined #nixos
sud0x3 has joined #nixos
<Lisanna>
clever: actually, that might work perfectly for me. it just so happens that each layer of named arguments does correspond to a layer of a nested attrset being evaluated with -A
<mojjo>
srhb: actually... I see this in dmesg:
<mojjo>
[ 57.380288] stage-2-init: setting up /etc...
<mojjo>
[ 59.152372] NET: Registered protocol family 10
<mojjo>
[ 59.184129] ip_tables: (C) 2000-2006 Netfilter Core Team
<mojjo>
[ 59.787709] systemd[1]: Set hostname to <nixos>.
mojjo has quit [Excess Flood]
<Guest14344>
>_>
<Lisanna>
I was going to try to collect them all at the end then merge them somehow, but if I just keep them at the layer they belong it should just work I think
mojjo has joined #nixos
ssmike has quit [Ping timeout: 252 seconds]
<mojjo>
sorry for the spam !!
<Lisanna>
oh no, that won't work... making each of those functions messes up autocompletion big time, an I depend heavily on autocompletion to explore this datastructure ):
<srhb>
mojjo: However, I'm not sure where that might arise from.
<mojjo>
yeah.. that is causing the users problem
<srhb>
mojjo: Hmm. Can you paste /var/lib/nixos/uid-map ?
orivej has quit [Ping timeout: 276 seconds]
<mojjo>
that is not a text file, no?
<srhb>
.... it should be
<mojjo>
I cannot cat it..
<srhb>
mojjo: Does it not exist, or is it something else?
<clever>
mojjo: what error happens when you try to cat it?
<mojjo>
Actually I have another issue on startup... it might be related. It says that the file system is corrupted, but I can get away with the "*)" option.. I get no error, just nothing back..
<srhb>
Well, corrupted filesystem would explain why that file is bad.
<srhb>
And that's BAD.
nextos has joined #nixos
<mojjo>
I remeber that I had sth like this and I coul easily fix it with fsck or so
<srhb>
That seems hit-or-miss
<srhb>
You should backup now before doing anything else.
<nextos>
why is nixos-unstable behind nixos-17.09? is this usual? the former has linux-4.9.78 and firefox 57, whereas the latter is already at firefox 4.9.80 and firefox 58
<srhb>
nextos: Yes, it's usual.
<srhb>
nextos: nixos-unstable tracks the master branch which receives a lot more breaking changes.
<srhb>
nextos: nixos-17.09 is a separate branch that is more conservative, so its tests pass more often.
<nextos>
srhb: ok, i thought that by tracking unstable i'd get newer kernels etc more quickly
<srhb>
however, nixos-unstable also failed to update the other day when the tests did go through, so the issue is compounded a bit.
<srhb>
nextos: No, quite the opposite :)
<srhb>
nextos: security fixes and things go way faster on stable.
<nextos>
srhb: ok :)
simendsjo has quit [Ping timeout: 256 seconds]
<mojjo>
srhb: alright, I'll do so. thanx for the help. gotta leave now...
<srhb>
mojjo: Good luck!
hamishmack has joined #nixos
<gchristensen>
clever: uhh... could you put a license on the kexec work you've done?
klntsky has joined #nixos
<nextos>
srhb: but i do get new packages on unstable, whereas these will only be introduced in 18.04 right?
vaninwagen has joined #nixos
<clever>
gchristensen: thats something i always forget to do, let me see...
<gchristensen>
thanks :) I forget too
<srhb>
nextos: Right, major versions and stuff. Things are *explicitly* backported to the stable branch, not automatically.
<srhb>
nextos: (18.03 though!)
<dhess>
I always start with the LICENSE file. Easy way to get going :)
<srhb>
dhess: I guess you're one of the few people that don't suffer from crippling LICENSE indecision :P
<dhess>
not anymore! :)
coot_ has quit [Ping timeout: 265 seconds]
<hyper_ch>
oh noes... the kexec will become restriced by a licence?
<nextos>
srhb: ok! yeah i've contributed some packages, and im getting nice binary substitutions already from hydra, only on unstable of course.
<gchristensen>
hyper_ch: it is already restricted by being unlicensed and therefore unusable
mojjo has quit [Ping timeout: 256 seconds]
nextos has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
arjen-jonathan has quit [Ping timeout: 260 seconds]
JayVii has joined #nixos
<gchristensen>
clever is there any secret support in the kexec work already to take some nix config and apply it to the system after the kexec takes over? to provide exact network config, for example
<clever>
gchristensen: nothing currently to read things like /etc/network/interfaces
<clever>
gchristensen: are you on digital ocean?
Izorkin has joined #nixos
<gchristensen>
I'm playing with DO, yeah. eyeing the nixos-infect code
<clever>
gchristensen: one minute
<hyper_ch>
I noticed on kimsufi you don't have to provide network stuff just use useDHCP
<srhb>
Yeah I'll stick with the sensible expressions rather than this voodoo of arg and argstr :-P
dbe_ has quit [Ping timeout: 240 seconds]
<jadu>
Hey guys, how do I build firefox with pulseaudio support? I'm using Nix on debian, and firefox doesn't output audio at all even with pulseaudio installed
coconnor has joined #nixos
<lewo>
hyper_ch: so, I just ordered a new server, have wait a bit and let's try kexec:)
<lewo>
s/have/have to/
dbe_ has joined #nixos
<infinisil>
jadu: works for me no problem
<infinisil>
jadu: is it only the audio from ff that doesn't work?
GiGa has joined #nixos
kim0 has quit [Quit: Connection closed for inactivity]
simendsjo has joined #nixos
<GiGa>
Hello all
<GiGa>
I'm trying to mount an NFS volume but keep geeting told I need a helper
<GiGa>
sudo mount -t nfs 192.168.160.3:/volume2/Music /mnt/music
<GiGa>
I ran the nix-env command as me though, so possibly not
Lisanna has joined #nixos
Lisanna has quit [Client Quit]
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 0d65e2e sjau: zfs: add migration path for new crypto format...
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vAJDs
knupfer has joined #nixos
<srhb>
GiGa: That sounds like the culprit.
<infinisil>
jadu: hmm, no idea, what's your nix-info say
<GiGa>
srhb: looks like that was it. I've placed it in the system config now and it's mounted
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 closed pull request #34559: zfs: refactor - adding new option for support of new native encryption format (master...master) https://git.io/vNx3h
<infinisil>
jadu: sound works for me with the exact same firefox
Capprentice has quit [Ping timeout: 268 seconds]
chisui has joined #nixos
winem_ has joined #nixos
<mpickering>
I am trying to make a simple wrapper derivation that copies some additional static assets into the nix store, so I'm trying to use pkgs.runCommand, but it fails with "cannot create regular file.... Permission Denied" when I try to copy the files into $out
<infinisil>
mpickering: code?
nuncanada has quit [Remote host closed the connection]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] dtzWill opened pull request #1851: More completely fix recursive nix, unbreak tests (master...fix/unbreak) https://git.io/vAJHa
<jadu>
infinisil: I don't think ff sees pulseaudio running, most of the other people with that "failed to init cubeb" error solve the issue by running pulse. I already have pulseaudio running though (ran through `start-pulseaudio-x11` in my .xinitrc)
<infinisil>
jadu: use the dedicated nixos option: hardware.pulseaudio.enable
<infinisil>
which will with 99.9% probability solve the issue
jorgan has quit [Ping timeout: 260 seconds]
jluttine has joined #nixos
<infinisil>
mpickering: i think it's something weird with the paths of this app package, because it works with e.g. cp -r ${glibc}/lib/* $out
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] dtzWill opened pull request #1852: Add tests for "nix build --dry-run" issues, fix one. (master...fix/dry-run-partially) https://git.io/vAJQA
NixOS_GitHub has left #nixos [#nixos]
ma27 has quit [Ping timeout: 256 seconds]
<jadu>
infinisil: I'm using nix on debian, not nixos. Would I still set that in a config?
throwup has joined #nixos
<throwup>
Hi, nix-newbie here. I am currently trying to write my first custom package/overlay. However the programme I
<throwup>
want to install wants tom move stuff to /usr/local. How do I override this behavior?
jluttine has quit [Ping timeout: 255 seconds]
ryanartecona has quit [Quit: ryanartecona]
<mpickering>
infinisil: it works now, for some reason
<mpickering>
I think copying the same file twice into the nix store is not permitted?
<mpickering>
throwup: Depends on the program
<mpickering>
If there is an env var you can set to configure it then set it to something like $out/local
<mpickering>
or you might have to use sed to modify hard coded paths in the source
<throwup>
mpickering: I would do that in the patch phase, I guess?
<throwup>
Or in the configure phase?
<mpickering>
patch phase I think
<mpickering>
you probably want to set "preConfigure" or "postPatch"
<mpickering>
and not override "patchPhase" itself
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] andir pushed 2 new commits to master: https://git.io/vAJ5G
<NixOS_GitHub>
nixpkgs/master e90985e Andreas Rammhold: gitAndTools.grv: 0.1.1 -> 0.1.2
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 151f234 Andreas Rammhold: Merge pull request #34711 from andir/grv-0.1.2...
asuryawanshi has quit [Ping timeout: 276 seconds]
azertyieio has quit [Ping timeout: 240 seconds]
<infinisil>
jadu: ah yeah, then no idea, the setting is only for nixos, since it controls some part of the OS
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] andir closed pull request #34634: grv: pass in version, fix "grv -version" to at least include that (master...fix/grv-version) https://git.io/vNhAe
NixOS_GitHub has left #nixos [#nixos]
<infinisil>
jadu: I'd recommend filing an issue on github
hakujin has quit [Ping timeout: 265 seconds]
<jadu>
infinisil: alright, I'll do so. Thanks for your patience!
ssmike has joined #nixos
ssmike1 has joined #nixos
jluttine has joined #nixos
hiratara has quit [Ping timeout: 240 seconds]
ssmike has quit [Ping timeout: 264 seconds]
ssmike1 is now known as ssmike
hiratara has joined #nixos
nwspk has quit [Quit: Quit: *.banana *.split]
stephenjudkins has quit [Remote host closed the connection]
nwspk has joined #nixos
Lambda has joined #nixos
Lambda is now known as Guest50077
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
ryanartecona has joined #nixos
Ndrei2 has joined #nixos
<Ndrei2>
Hello all, not nixos related but figured it's highly likely someone here can help: I've botched my partition table leaving
ssmike1 has joined #nixos
<Ndrei2>
Leaving me with a luks encrypted partition that I don't know the start block of
<infinisil>
Ndrei2: ##linux ?
jluttine has quit [Ping timeout: 256 seconds]
<Ndrei2>
How can I find out what offset I need to recreate at?
<Ndrei2>
Okay ill try, just was expecting it to be more crowded.
<clever>
Ndrei2: one method ive had luck with before, recreate the old partitions, guessing at sizes, and confirm with blkid, make sure your partition editor doesnt try to wipe old FS's out
<clever>
Ndrei2: you can also try creating the 1st /boot one overly large, then use mount -o ro to find its real size, remake it at that, then make another after it, overly large, repeat
<clever>
Ndrei2: allowing the partition editor to use the same alignment it previously used
Guest50077 has quit [Ping timeout: 260 seconds]
bfrog has joined #nixos
winem_ has quit [Ping timeout: 255 seconds]
winem_ has joined #nixos
user0_ has joined #nixos
ndrei2__ has joined #nixos
<Dezgeg>
maybe binwalk knows how to locate luks headers?
<ndrei2__>
Dezgeg
<Dezgeg>
in case they have some recognizable magic
<ndrei2__>
Dezgeg: thanks
<Dezgeg>
(I don't actually know, but it might)
<ndrei2__>
I've tried #linux btw but no replies
hamishmack has joined #nixos
<ndrei2__>
I've found a suggestion online: grep -a -b -P --only-matching 'LUKS\xba\xbe' /dev/sdb
<ndrei2__>
the thing is, I don't know once I get the result how to translate that into an offset
<Dezgeg>
sounds like it'd work
<infinisil>
ndrei2__: it's ##linux, and you need your nick to be registered for joining ##linux
GiGa has quit [Quit: Leaving]
<Dezgeg>
divide by 512 to translate to sectors
<infinisil>
you didn't write anything in the actual linux channel
<clever>
[root@system76:~]# hexdump -C /dev/nvme0n1p2 | head
<clever>
00000000 4c 55 4b 53 ba be 00 01 61 65 73 00 00 00 00 00 |LUKS....aes.....|
<clever>
i can confirm that my luks starts with that header
<ndrei2__>
Thanks, sorry I'm a bit sloppy this took me a bit by surprise and my nick password all my setup is there so I'm a bit "naked"..
<ndrei2__>
thanks!!!
<ndrei2__>
so it should work
<ndrei2__>
but then, the question only remains
<ndrei2__>
with that result, how do I translate that into an offset for fdisk?
<ndrei2__>
well, obviously mine will be a grep result
user0_ has quit [Ping timeout: 260 seconds]
<ndrei2__>
I just don't know what it will look like, I'm on a Windows machine now...
<clever>
Ndrei2: divide by 512 i believe
<ndrei2__>
okay, sounds easy enough..
throwup has quit [Remote host closed the connection]
stephenjudkins has joined #nixos
<ndrei2__>
have an idea of how long this grep will take? on a 500 gig nvme drive
<clever>
ndrei2__: depends on what position you put the luks partition at
<clever>
and how big it is
<ndrei2__>
about 40% in
<ndrei2__>
it's been running for 15mins I think
<clever>
ndrei2__: what was using the 0 to 40% range?
<ndrei2__>
Windows
ssmike1 has joined #nixos
<clever>
ndrei2__: do you know how big that filesystem was?
<ldlework>
I'm so confused.
<ndrei2__>
well, yes, but the thing is, I actually shrunk it on purpouse, I just don't remember what size it was before
<ldlework>
~/.nix-profile/bin is on my path
<ldlework>
but zsh can't find anything in it?
<ldlework>
wtf
<ndrei2__>
Idlework: echo $PATH | tr : \n ?
Ndrei2 has quit [Ping timeout: 260 seconds]
ssmike has quit [Ping timeout: 248 seconds]
ssmike1 is now known as ssmike
<ndrei2__>
clever: so what's your estimate? I was expecting it to be done by now, nvme is supposed to be fast
<judson>
Anyone using the security.acme module with Apache?
<clever>
ndrei2__: ah, that complicates things then
<clever>
ndrei2__: one sec
<ndrei2__>
Yeah, that's why I need to brute force it
jluttine has joined #nixos
<clever>
ndrei2__: what PID is the grep process?
<ldlework>
ndrei2__ ??
<judson>
I'm trying to deploy HTML files in a derivation, and use ACME to get a cert for it.
<judson>
But letsencrypt needs to write files.
<ndrei2__>
oh I'm on the nixos recovery iso, it's in the foreground
<ldlework>
oh there's a semi-colon
civodul` has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
<ldlework>
wtf
<clever>
ndrei2__: alt+f2 to change to tty2
<judson>
I'm about to do some rewrite configs for Apache and want to make sure there's not an easier way
<ndrei2__>
Idlework: did that help?
<ndrei2__>
clever: ok
<clever>
ndrei2__: what PID is the grep process?
coot has quit [Quit: coot]
<ndrei2__>
why? 1417
<clever>
ndrei2__: ls -l /proc/1417/fd/, which symlink is pointing to the nvme drive?
<clever>
this will let you see the progress
<ndrei2__>
3
<ndrei2__>
of I just cat that ?
<clever>
watch -d cat /proc/1417/fdinfo/3
<clever>
you have to read fdinfo to get the info
<ndrei2__>
ooh amazingly cool thanks
<symphorien>
is it possible to get a debug build of libstdc++.so ?
coot has joined #nixos
ssmike1 has joined #nixos
<symphorien>
or the debug symbols
ssmike has quit [Ping timeout: 256 seconds]
ssmike1 is now known as ssmike
<clever>
ndrei2__: and alt+f1 lets you watch the grep again
<ndrei2__>
clever: thanks, I'm piping it to strings and it's all garbadge, which makes me wonder if there's something wrong as my windows partition isn't encrypted
<ldlework>
ndrei2__ yes, thank you
<clever>
ndrei2__: is there still a partition at the offset for the ntfs fs?
<ndrei2__>
oh now its fine
<ndrei2__>
it's still on windwos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vbgl opened pull request #34714: coq: default to 8.7 (master...coq-default-to-8.7) https://git.io/vAJxe
NixOS_GitHub has left #nixos [#nixos]
jluttine has quit [Ping timeout: 268 seconds]
<infinisil>
judson: you can't get a certificate for a derivation..
Ndrei2 has joined #nixos
<Ndrei2>
oh wait... I guess I'm too tired.. If I'm catting this I'm looking at the beginning of the partition, not at what grep is doing..
<infinisil>
judson: a certificate is inherently stateful
<infinisil>
if it wasn't, it would defeat its purpose
<NixOS_GitHub>
[nixpkgs] grahamc pushed 2 new commits to master: https://git.io/vAJhG
<NixOS_GitHub>
nixpkgs/master 0525e82 Markus Hauck: docs: Explain how to login into built vm
<NixOS_GitHub>
nixpkgs/master 7fb90db Graham Christensen: Merge pull request #34710 from markus1189/doc-vm-login...
<pie_>
<pie_> error: syntax error, unexpected $undefined, at /mnt/data/gscan/nixpkgs/pkgs/tools/graphics/gscan2pdf/default.nix:39:14
<pie_>
<pie_> preBuild = ``perl Makefile.PL``
<pie_>
<pie_> thats a weirdass error for missing a semicolon
<pie_>
whoops wrong channel
<dhess>
I haven't done a NixOps deployment since the LetsEncrypt stuff went down a few weeks ago. What is the story right now with acme support in nixos-unstable. Can you auto-create/renew ACME certs with just a few lines of config?
<dhess>
because that was amazingly easy before
<infinisil>
dhess: relevant link?
<dhess>
infinisil: to what, exactly?
<infinisil>
dhess: like an issue to the lets encrypt problems
<dhess>
you can apparently still renew certs "the old way" (basically by hosting a session secret on your domain) if you and your provider meet certain requirements, but you can't get new ones that way anymore
<infinisil>
ah that one, hmm
<dhess>
or possibly in the end they just disabled the TLS-SNI stuff altogether. I haven't tried since then
<dhess>
they were debating whether to grandfather in certain clients
<gchristensen>
dhess: I just got a new one that way yesterday though?
<dhess>
gchristensen: a new one? not just a renewal?
<fiatjaf>
is there a fish replacement for the ~/.nix-profile/etc/profile.d/nix.sh which I should source with bash somewhere?
<gchristensen>
I don't think we're using the SNI option
oida has joined #nixos
jluttine has joined #nixos
Ubercow has joined #nixos
<dhess>
I'll probably switch to DNS-01 anyway
ssmike has quit [Remote host closed the connection]
pareidolia has joined #nixos
smichel17 has joined #nixos
<gchristensen>
it would be good to support generating the cert on a nixops control node and distributing them from there
<Guest14344>
Hmm, the qcow2 image I created is 1GB
<Guest14344>
Seems large... compresses to 200ish MB
<gchristensen>
it is likely a sparse file
Ndrei2 has quit [Quit: Page closed]
<Guest14344>
Oh, I moved it from Linux to my Mac. That'd probably unsparsify it?
<gchristensen>
sounds likely :)
spietz has quit [Ping timeout: 265 seconds]
<clever>
some filesystems will unsparsify when copying between them
<Guest14344>
Hmm, du --apparent-size and du report the same size for the file
<Guest14344>
1 GB isn't that much in the grand scheme of things these days
<Guest14344>
(This is on the Linux box that made it)
<gchristensen>
you can get a microSD card to hold 400 of thtem
<Guest14344>
I'm more concerned about uploading it uncompressed to this openstack instance, but as long as I do it from work where I have 500 mbps symmetric I should be fine ;)
<gchristensen>
ouch, yeah
<dhess>
gchristensen: yeah that would be ideal!
<dhess>
especially for wildcard certs. Those are/will be possible with DNS-01 validation
<ndrei>
I'm back... no replies on #linux, so I'm trying a different strategy
winem_ has quit [Ping timeout: 255 seconds]
<ndrei>
for those that haven't see my earlier messages, I'm trying to recreate partition for which I don't know the start offset. I'm now thinking of looping a parted script to create partitions with 1GB increments until I find a match that cryptsetup will open
<ndrei>
can anybody help with some basic parted commands?
<ndrei>
I need to create a partition at position 4, of size 150+ <x>GB, and then another one to fill the rest of available space
<ndrei>
sorry that this isn't nixos related, just hoping I have more chances of a reply here
<clever>
ndrei: id just leave the grep running longer
taktoa has joined #nixos
<ndrei>
it's been running for .. I dunno, very long now. How long can it be? i7 6900K with nvme drive, should be fast no?
lord| has joined #nixos
<clever>
ndrei: nix-env -iA nixos.testdisk && man testdisk
<Guest14344>
Hey, got the image uploaded into OpenStack... moment of truth, time to try running it
<ndrei>
thanks, but it doesn't mention luks
jluttine has quit [Ping timeout: 248 seconds]
cnidario has quit [Remote host closed the connection]
<ndrei>
well... grep quit (or I accidentaly killed it, altough I don't think so..) with no result, now trying hexdump -C <device> | grep LUKS
<NixOS_GitHub>
nixpkgs/master 4dd3a2a Andreas Rammhold: Merge pull request #34684 from flokli/svn-all-fast-export...
<ndrei>
clever: I'll just leave it running for now because it did indeed find other instances of "LUKS" where it's clearly in a random sequence so that's probably what that was
Rusty1_ has quit [Quit: Konversation terminated!]
oida has quit [Quit: WeeChat 1.6]
<dhess>
ndrei: have you looked for forensics tools? It would surprise me if there weren't at least a few that can do this for you.