<clever>
Lisanna: under most unix based OS's, the file will keep serving after being deleted
<clever>
and the kernel will finish the delete when the handle is closed
<Lisanna>
sure, I'm not worried about that side of it. I'm guessing that query-then-download from a binary cache is not an atomic operation
<clever>
Lisanna: yeah, there is a chance it may fail there
<Lisanna>
so if it queries the binary cache webserver, sees that it exists, binary cache has the entry deleted, and then it tries to fetch a nonexistent entry
<Lisanna>
ugh
<clever>
Lisanna: with `nix-store -r /nix/store/foo` you can forcibly download something from a cache
<clever>
with nix-store --delete, you can then delete it, but it remembers that the binary cache had it
<srid>
Lisanna: I get `is not of type `package`
<clever>
remove it from the cache manually, and try to -r again
<clever>
srid: you may need (pkgs.callPackage ./foo.nix {})
<srid>
Aha, it is a function
<clever>
and callPackage deals with giving it the right args
<clever>
you can also just do foo = pkgs.callPackage ./foo.nix {}; within an overrides or overlay, and then pkgs.foo later on
<Lisanna>
clever so if I want high availability I'll have to do all of that lift manually?
asuryawanshi has joined #nixos
<clever>
Lisanna: you may want to analyze the access.logs and see when something was recently used
<srid>
is there a way to determine the sha256 of a fetchFromGithub thing automatically? especially if i plan to add multiple repos
<Lisanna>
srid nix-prefetch-git
<Lisanna>
clever okay, deletion aside, what about adding to a binary cache, with nix copy --to file://? Is that safe? (i.e., could a client come along and think something is available when it's not since it's not done being added?)
<clever>
Lisanna: i'm assuming nix will only create the narinfo after the nar is valid
<Lisanna>
okay
asuryawanshi has quit [Ping timeout: 265 seconds]
<clever>
it would also be trivial to write to a tmp file with the wrong name, then use rename() to atomicly put it into place
<clever>
havent confirmed if it actually does such things though
asuryawanshi has joined #nixos
<srid>
Lisanna: nice - that returns a JSON. do i manually convert it to nix exprs, or is there a function that takes that json as is?
<clever>
srid: pkgs.fetchFromGitHub (builtins.parseJSON (builtins.readFile ./foo.json)) i think
coconnor has joined #nixos
<srid>
fetchFromGit?
<clever>
Hub
asuryawanshi has quit [Ping timeout: 265 seconds]
<srid>
that takes 'owner' instead of 'url'
<clever>
ah, i dont think there is a took that outputs the right json for that
kelleyNif has joined #nixos
<clever>
but you can still use nix, (builtins.parseJSON (builtins.readFile ./foo.json)).sha256
<clever>
that reads the sha256 field, and ignores the rest
<srid>
I guess I could write a code generator :P
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Myrl-sak1 has quit [Ping timeout: 240 seconds]
asuryawanshi has joined #nixos
<clever>
tnks: [root@system76:~]# sqlite3 -header -column /nix/var/nix/db/db.sqlite 'select path,registrationTime,narSize,sigs from ValidPaths where sigs != "" limit 10'
asuryawanshi has quit [Ping timeout: 276 seconds]
<tnks>
clever: yeah, I've been doing variants of that on my side.
<tnks>
but now, I need to write helper functions to check signatures.
<tnks>
clever: I'll take shortcuts to that if you have them.
<tnks>
getting all the bits in line for a proper signature check can sometimes be finicky.
pkill9 has joined #nixos
<ottidmes>
tnks: clever: Just tried again to make it work with 2.0, double checked everything I could think of, but I am still getting: warning: substituter 'ssh://root@server' does not have a valid signature for path '/nix/store/...-samba-4.7.5'
thc202 has quit [Ping timeout: 256 seconds]
<ottidmes>
I think it is obvious that it is not working, but I have no clue what to do in order to fix it, in the Nix manual under "require-sigs" it says: "must have a valid signature, that is, be signed using one of the keys listed in trusted-public-keys or secret-key-files", so my open question at the moment is, how/where do I specify that the private key I list in secret-key-files is used to sign the nix store
<tnks>
clever: I worry about this idea you have that trusted-users automatically ignore signatures.
Arcaelyx has quit [Ping timeout: 265 seconds]
<ottidmes>
Got it, not sure how I missed it, but: nix sign-paths --all --key-file <path/to/secret-key-file>
<clever>
tnks: the rules may have changed in 2.0
<tnks>
clever: is there a good person to ask? or code to look up?
fragamus has quit [Ping timeout: 240 seconds]
<clever>
tnks: not sure yet
<tnks>
clever: in the meantime, I suppose I could leave the trusted-users bit and remove the secret-key-files and trusted-public-keys stanzas, right?
<tnks>
that should prove out this theory, I'm hoping.
<tnks>
s/stanzas/fields/
kelleyNif has quit [Quit: Leaving.]
<ottidmes>
But I still got plenty of questions, like, from what I read it all points to it that I have to do this for all new entries in the /nix/store, so do I need to call it every time I add stuff to the nix store I want to share? Does it resign all existing things?
<clever>
tnks: i would keep the secret-key-files at least, since that signs everything automatically
<clever>
tnks: so signatures are just always present in db.sqlite
kelleyNif has joined #nixos
<clever>
ottidmes: what i said above
<ottidmes>
clever: Yeah, was about to respond, I am quite sure it does not do what you say it does
<clever>
ottidmes: i manually deleted a path, set that option, then nix-build'ed it, and it was signed
<ottidmes>
clever: I had secret-key-files with my private key, but I got "valid signature" errors
<ottidmes>
clever: Ahh!
<ottidmes>
clever: It probably signs automatically for all new builds
<ottidmes>
clever: I was testing with an existing one
asuryawanshi has joined #nixos
<ottidmes>
I had made the assumption that it would sign on request
<tnks>
clever: yeah, but it could do a double calculation and also trust things signed with the secret-key-files, right?
<clever>
tnks: that could probably be done with some changes to the src
asuryawanshi has quit [Ping timeout: 256 seconds]
<tnks>
clever: meaning it's not implemented that way right now?
<tnks>
clever: I don't know what I read that convinced me it might be that way.
acarrico has quit [Ping timeout: 245 seconds]
<tnks>
clever: actually, I found it... the text in the man page for nix.conf for "require-sigs"
<tnks>
"...must have a valid signature, that is, be signed using one of the keys listed in trusted-public-keys or secret-key-files."
<tnks>
clever: what do you think? Am I reading too deeply into the docs?
<clever>
tnks: back it up with src
<tnks>
I know the source code far less.
roblabla has quit [Remote host closed the connection]
<shapr>
shlevy: heard of an effort to get GHC on risc-v?
asuryawanshi has joined #nixos
roblabla has joined #nixos
<Lisanna>
is there a way to get the outpath that a nix expression would produce if built with nix-build without actually doing the build?
<shlevy>
shapr: Yep, actually met up with bgamari- yesterday for some cross-compilation related hacking :) though not specifically this
<shapr>
oh wow! awesome!
<shlevy>
IIRC the blocker at the latest attempt was no llvm port
<shapr>
is there a trac page or github repo recording this effort?
<tnks>
now, I just need to find code that reveals how far trusted user privileges can go.
pie__ has joined #nixos
asuryawanshi has joined #nixos
roblabla has quit [Remote host closed the connection]
pie___ has quit [Ping timeout: 260 seconds]
asuryawanshi has quit [Ping timeout: 260 seconds]
roblabla has joined #nixos
acarrico has joined #nixos
asuryawanshi has joined #nixos
dkao has joined #nixos
asuryawanshi has quit [Ping timeout: 256 seconds]
asuryawanshi has joined #nixos
yorick has quit [Ping timeout: 256 seconds]
asuryawanshi has quit [Ping timeout: 256 seconds]
ryanartecona has joined #nixos
<Lisanna>
is there a way to list all of the storepaths in a binary cache
<Lisanna>
i.e., some nix tool that effectively iterates over all the narinfo files and prints the StorePath entry in each
<Lisanna>
closest thing I've found so far is the concatenation of nix-store --gc --print-live and nix-store --gc --print-dead
<Lisanna>
(Assuming nix 2 here)
asuryawanshi has joined #nixos
<Lisanna>
although, --gc doesn't work on binary caches ):
hellrazor has joined #nixos
Synthetica has quit [Quit: Connection closed for inactivity]
acarrico has quit [Ping timeout: 268 seconds]
coot_ has quit [Quit: coot_]
asuryawanshi has quit [Ping timeout: 260 seconds]
hellrazo1 has quit [Ping timeout: 256 seconds]
jtojnar has quit [Quit: jtojnar]
jtojnar has joined #nixos
ryanartecona has quit [Quit: ryanartecona]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to master: https://git.io/vAHFe
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master eaed0ec Nikolay Amiantov: xgboost: 0.60 -> 0.7...
pkill9 has quit [Ping timeout: 256 seconds]
asuryawanshi has joined #nixos
asuryawanshi has quit [Ping timeout: 240 seconds]
asuryawanshi has joined #nixos
<Ralith>
how the hell do I get nix to give me libraries with working debug info? :|
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] romildo opened pull request #36320: [WIP] [need help] gsettings-qt: init at 0.1.20170824 (master...new.gsettings-qt) https://git.io/vAHFO
NixOS_GitHub has left #nixos [#nixos]
<Ralith>
I tried dontStrip = true, I tried separateDebugInfo = true, the package is certainly being built with -O0 -g, but nothing works
<Ralith>
there are no source locations
asuryawanshi has quit [Ping timeout: 268 seconds]
winem_ has quit [Ping timeout: 252 seconds]
<Lisanna>
Ralith what do you mean "working debug info"?
<Lisanna>
I usually use dontStrip = true
<Lisanna>
if you're trying to get the source code to show up in gdb, that's a whole other story
<Ralith>
I mean I can drop into gdb and step through the library in question line by line
<Ralith>
I can deal with it not showing up but there's no line numbers at all
<Ralith>
it's not even failing to resolve references to source, it's just not looking
<Lisanna>
Ralith that's because nix builds do the build in a temporary directory, which is destroyed after the build completes
<Ralith>
normally gdb generates an error in that circumstance and prompts you to provide an alternative source root
<Lisanna>
nix might be doing something special there... not sure. But either way, you won't be able ot get the source code unless you do something special
<Lisanna>
this is what I do, so that when I attach gdb to the process I get the full source code
<Ralith>
and you would still get lin enumbers regardless
<Ralith>
gdb is not getting any line numbers at all
asuryawanshi has joined #nixos
acarrico has joined #nixos
endformationage has joined #nixos
<Ralith>
hm, looks like the derivation in question was defaulting separateDebugInfo to true for some reason, that's probably why there was no info showing up at all
<Ralith>
unclear how that option's supposed to be used
<Lisanna>
dunno, I never use it
asuryawanshi has quit [Ping timeout: 256 seconds]
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
<Lisanna>
'cause I'm lazy and like things to "just work"
Supersonic112 is now known as Supersonic
<Ralith>
+1 to that
<Ralith>
Lisanna: that made the build fail with something about an erroneous dependency on perl o_O
<Lisanna>
wat
<Ralith>
looks like it's treating the build scripts in `src` as executables and catching stuff from `nativeBuildInputs` there, maybe?
<Ralith>
wonder if there's a way to blacklist that dir from fixup
asuryawanshi has joined #nixos
jtojnar has quit [Remote host closed the connection]
plakband has quit [Remote host closed the connection]
d4g_ has joined #nixos
asuryawanshi has quit [Ping timeout: 240 seconds]
<Ralith>
okay, the build fails even if I explicitly add perl to `buildInputs` :|
ottidmes has quit [Ping timeout: 240 seconds]
<Ralith>
anyone have any idea what exactly nix considers an "erroneous dependency"?
jmeredith has quit [Quit: Connection closed for inactivity]
d4g has quit [Ping timeout: 256 seconds]
acarrico has quit [Ping timeout: 260 seconds]
pie__ has quit [Ping timeout: 276 seconds]
asuryawanshi has joined #nixos
<Ralith>
oh wait that's literally hardcoded into the derivation, lol
owickstrom has quit [Remote host closed the connection]
owickstrom has joined #nixos
asuryawanshi has quit [Ping timeout: 265 seconds]
asuryawanshi has joined #nixos
<Ralith>
Lisanna: what's the `postInstall = ''make clean'';` for?
<Lisanna>
Ralith so that you don't have a bunch of intermediate build products in your /nix/store/asfklsjdfh-mypackage/src
<Lisanna>
it should actually probably be make distclean
<Ralith>
Lisanna: seems like the copy happens during `setSourceRoot`, not after the build, though?
<Lisanna>
that thing moves the build location from the /tmp build location to $out/src
jtojnar has joined #nixos
<Lisanna>
so the actual build happens in $out/src
<Lisanna>
so that the debug info points to the correct location for the source files
<Ralith>
ah, right
<Ralith>
um
<Ralith>
where does it do that?
<Ralith>
oh, implicit cd to $sourceRoot
markus1189 has joined #nixos
asuryawanshi has quit [Ping timeout: 268 seconds]
leat has joined #nixos
<Lisanna>
yeah
markus1199 has quit [Ping timeout: 256 seconds]
Tobba has joined #nixos
asuryawanshi has joined #nixos
<Li[m]>
its 2018 and most people use non-rebroducible build systems! ah!
<Li[m]>
most people dont even have a good bicycle!
<Li[m]>
isHavvy: try boskey, chinese brand really nice for the price. I highly recommend them
<Li[m]>
steel is read
<Li[m]>
real
<isHavvy>
If I wanted a bicycle, I would already have one.
<Li[m]>
at least you use a reproducible build system
<Li[m]>
:P
<isHavvy>
Yes, where I can reproducibly fail to compile the Rust compiler over and over again.
<Li[m]>
isHavvy: are you trying to compile nightly ?
<Li[m]>
i wanted to bring back the nightly expression
<Li[m]>
id like to package some stuff that needs nightly. with mozilla overlay and carnix iv gotten pretty far, but Iv hit some roadblocks
mbrgm has quit [Ping timeout: 248 seconds]
<isHavvy>
Li[m]: Yeah, I want to compile nightly so I can work on it.
<catern>
I have a project with a default.nix which I use to build it; is there a straightforward way to switch between a nixpkgs-pinned version of a default.nix and a version which uses <nixpkgs> from NIX_PATH?
mbrgm has joined #nixos
<catern>
like, an idiomatic typical way to do that?
<isHavvy>
Li[m]: I'd rather not have the nightly expression in nixpkgs.
<Ralith>
Lisanna: okay, after like ten iterations of me discovering wacky shit this specific derivation does, that hack solved my problem; thanks!
<Li[m]>
isHavvy: well thats fine, as long as the expression is up to date people can use it
<isHavvy>
Li[m]: That requires updating it _every day_.
<Li[m]>
isHavvy: I didnt know that
<Li[m]>
you cant compile nighly on a weekly basis?
<isHavvy>
You can, but then you're potentially using a week-old nightly.
<Li[m]>
thats usually what I do
<Li[m]>
since internet is expensive here
<isHavvy>
In general, you either want to keep with the head nightly or pin to a specific one.
<Li[m]>
yea I pin
<Li[m]>
with the mozilla overlay
ryanartecona has joined #nixos
<Li[m]>
probably the wrong way :P
<fearlessKim[m]>
while rebuilding home manager with nixos-unstable : nix-repl.cc:13:10: fatal error: common-opts.hh: No such file or directory #include "common-opts.hh"
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] shlevy opened pull request #1946: ssh-ng: Only forward settings that make sense to forward. (master...forwardable-settings) https://git.io/vAHNN
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar opened pull request #36322: libcloudproviders: init at 0.2.5 (master...cloudproviders) https://git.io/vAHNp
NixOS_GitHub has left #nixos [#nixos]
b has quit [Ping timeout: 260 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] andrew-d opened pull request #36323: pillow: Fix build on non-NixOS systems (master...adunham/fix-pillow) https://git.io/vAHAv
NixOS_GitHub has left #nixos [#nixos]
asuryawanshi has joined #nixos
<tnks>
clever: I think I found the difference between trusted semantics of Nix 1.0 and 2.0.
<tnks>
although, looking at the code, it seems to be doing the same thing to me.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar pushed 1 new commit to master: https://git.io/vAHxo
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 0f2dc6c Jan Tojnar: chrome-gnome-shell: 9 → 10
asuryawanshi has joined #nixos
MP2E has quit [Remote host closed the connection]
asuryawanshi has quit [Ping timeout: 252 seconds]
acertain has quit [Ping timeout: 276 seconds]
<disasm>
I'm playing with the lib function commitIdFromGitRepo. What's the reasoning to check in .git/HEAD first instead of looking in .git/packed-refs? Is there a case where a git repo wouldn't have a packed-refs file or the packed-refs file would be out of date compared to .git/HEAD?
acertain has joined #nixos
<disasm>
never mind, I looked at git documentation. looks like you have to run `git pack-refs` on a repo for packed-refs to exist
asuryawanshi has joined #nixos
ryanartecona has joined #nixos
schoppenhauer has quit [Ping timeout: 240 seconds]
<tnks>
liminal18: make sure to do it in the right profile.
<tnks>
I believe from your Mac comment, you're not on NixOS. I'm not sure the proper way to upgrade to Nix 2.0 there.
<tnks>
And all the installations of Nix on other OS's get snowflaked really easily.
asuryawanshi has joined #nixos
asuryawanshi has quit [Ping timeout: 265 seconds]
<liminal18>
tho cinnamon you gave me installed a nix cli
<liminal18>
command
<liminal18>
but not the 2.0 one
<liminal18>
anyways I'm out of time for this issue today
asuryawanshi has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar pushed 13 new commits to master: https://git.io/vAHhp
<NixOS_GitHub>
nixpkgs/master 3648db8 Jan Tojnar: gnome3.gexiv2: 0.10.7 → 0.10.8
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master a3a8209 Jan Tojnar: gnome3.libgepub: 0.5.2 → 0.5.3
<NixOS_GitHub>
nixpkgs/master dc40604 Jan Tojnar: gnome3.libgee: 0.20.0 → 0.20.1
liminal18 has quit [Quit: Page closed]
spear2 has quit [Remote host closed the connection]
<fearlessKim[m]>
anyone having experience with latex ? I have `texlive.combine ( cleveref )` installed through home manager, yet running latexmk doesnt find cleveref.sty even though `/nix/store/y6bn4fly4w71w7nhhnsmsb6qgfmrm7hl-texlive-combined-2017/bin/kpsewhich cleveref.sty` finds it /nix/store/y6bn4fly4w71w7nhhnsmsb6qgfmrm7hl-texlive-combined-2017/share/texmf/tex/latex/cleveref/cleveref.sty
<tnks>
I've not paid attention to Disnix. I just looked at the manual... Is this an alternative to NixOps?
<fearlessKim[m]>
tnks: it's supposed to be different, more focused on services instead
<jtojnar>
fearlessKim[m]: I think using nix expressions is the default, if some package works globally, it usually had to be modified with regards to plug-in paths
<tnks>
fearlessKim[m]: yeah. I wonder why I never hear anything about it. Everyone seems to be using Nixops instead.
<tnks>
fearlessKim[m]: okay, this seems to be the biggest difference to me: "Moreover, the target machines to which Disnix deploys are neither required to run NixOS nor Linux."
<fearlessKim[m]>
jtojnar: that's kind of crazy to run nix to compile my pdf, especially while writing it. I want to enjoy vim + vimtex. What surprises me is that kpsewhich finds the file so it seems properly configured
MP2E has joined #nixos
<jtojnar>
fearlessKim[m]: maybe you can run vim inside nix-shell
rauno has quit [Ping timeout: 268 seconds]
nocoolnametom has quit [Quit: ZNC 1.6.5 - http://znc.in]
nocoolnametom has joined #nixos
<jtojnar>
vcunat might be able to tell you more
gspia has quit [Read error: Connection reset by peer]
gspia has joined #nixos
<Lisanna>
can nix-build for a system that doesn't match the current system still use binary caches to substitute the result before turning to remote builders?
<elvishjerricco>
Lisanna: I believe so
<Lisanna>
seems like my nix-build is going to a remote builder before checking the binary cache
kannan has joined #nixos
kannan has quit [Client Quit]
cakoose has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg opened pull request #36324: okular: add support for CHM, ZIP and markdown files (master...f/okular) https://git.io/vAHji
NixOS_GitHub has left #nixos [#nixos]
cakoose has quit [Client Quit]
<elvishjerricco>
How can I migrate a Hydra cache from an old location to a new one?
cakoose has joined #nixos
<Lisanna>
elvishjerricco use nix copy?
<cakoose>
I noticed that the Nix installer just installs in multi-user mode on MacOS. Is there a way to install in single-user mode?
<elvishjerricco>
Well if you used a custom path instead of `/nix/store` for the Hydra cache, then the cache is not a normal nix store that you can copy out of, IIRC
<cakoose>
(I don’t know much about the difference, but it seems like single-user mode has fewer moving parts, which seems preferable if I don’t need multi-user capabilities.)
<elvishjerricco>
cakoose: I think there are some problems with single user mode on macos IIRC. If multi user mode is working for, it's probably best to just leave it
<Lisanna>
elvishjerricco not sure that that would matter to something like nix copy... otherwise, recompile nix with the path prefix set to whatever you changed it to as a configure option
<Lisanna>
then that should work
<elvishjerricco>
Lisanna: A hydra cache isn't like a normal store. It's in a binary cache format with compressed, archived store paths
Rusty1_ has quit [Quit: Konversation terminated!]
<Lisanna>
elvishjerricco nix 2.0 can handle binary caches and store paths the same
<Lisanna>
I can't think of any reason why recompiling nix with your custom /nix/store prefix and then using nix copy wouldn't work.
<Lisanna>
ugh, yep, nix-build is building it on a remote rather than pulling it from the binary cache ):
<fearlessKim[m]>
jtojnar: thanks for the help I'll open an issue as the integration with vimtex requires a few steps (generating a custom .latexmkrc to set paths to biber ) and I have the feeling there is just a little line missing somewhere to make it all work
<elvishjerricco>
Lisanna: Ah, using `nix copy --from file:///old/path --to file:///new/path` worked. I was missing the `file://`
<Lisanna>
yeah
asuryawanshi has joined #nixos
<elvishjerricco>
Now I just gotta get a full listing of the old store
<elvishjerricco>
Is that possible?
<tnks>
okay, I just grepped a bit through the Nix code base, and I don't think the priority of a substituter is overrideable client-side.
<tnks>
which seems more inflexible than I'd expect. I'd have to proxy a server and change it's /nix-cache-info.
qmm has quit [Ping timeout: 256 seconds]
asuryawanshi has quit [Ping timeout: 245 seconds]
asuryawanshi has joined #nixos
qz has joined #nixos
robstr has joined #nixos
asuryawanshi has quit [Ping timeout: 245 seconds]
qmm has joined #nixos
blankhart has quit [Quit: WeeChat 1.9.1]
asuryawanshi has joined #nixos
justbeingglad has joined #nixos
asuryawanshi has quit [Ping timeout: 252 seconds]
pie__ has joined #nixos
justbeingglad has left #nixos [#nixos]
muzzy_ has joined #nixos
muzzy has quit [Ping timeout: 252 seconds]
asuryawanshi has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar opened pull request #36326: meson: set multiple output dirs in setup hook (staging...meson-fuck-prefix) https://git.io/vAQeP
NixOS_GitHub has left #nixos [#nixos]
cakoose has quit [Quit: cakoose]
dkao has quit [Ping timeout: 260 seconds]
asuryawanshi has quit [Ping timeout: 276 seconds]
asuryawanshi has joined #nixos
asuryawanshi has quit [Ping timeout: 240 seconds]
reinzelmann has joined #nixos
ma27 has quit [Ping timeout: 256 seconds]
asuryawanshi has joined #nixos
asuryawanshi has quit [Ping timeout: 256 seconds]
<Lisanna>
crap, I think I hosed my /nix/var/nix/db/db.sqlite
rauno has joined #nixos
<Lisanna>
ran out of disk space during a write or something ):
<edef>
it's transactional, it should deal with that fine
<edef>
it should deal with pulling the plug during a tx fine
aarvar has joined #nixos
<Lisanna>
....okay...
<Lisanna>
so what happened?
<edef>
what's it saying
<etu>
edef: Full filesystems may wipe a file empty on write
<Lisanna>
"error: setting synchronous mode: unable to open database file"
<edef>
hm
<Lisanna>
well, it's not empty
<edef>
i assume you've fixed the full disk issue by now?
<Lisanna>
no... that's a complicated problem that takes a long time to fix
guibou has quit [Ping timeout: 256 seconds]
asuryawanshi has joined #nixos
oahong has quit [Ping timeout: 268 seconds]
<Lisanna>
but, you're right, I should try to fix that before I worry about the database
aarvar has quit [Client Quit]
MercurialAlchemi has joined #nixos
<edef>
it'll generally fail to open for writes until you've got a few megs free at least
oahong has joined #nixos
aarvar has joined #nixos
<Lisanna>
the things taking up most of the disk space are only reclaimable with nix-collect-garbage... and I can't run nix-collect-garbage until this issue is fixed...
<Lisanna>
okay, deleting some misc. files fixed it
<srhb>
Lisanna: You can be nasty to the store and delete them manually, then repair afterwards.
<srhb>
Or that :-)
<Lisanna>
I really need to get to the bottom of what's creating terabytes of symlinks in /nix...
<srhb>
terabytes of _symlinks_? O_o
<Lisanna>
yes
<srhb>
That sounds very weird indeed.
<Lisanna>
nix-collect-garbage deletes them, but it takes *forever*
<Lisanna>
and it doesn't even report them as reclaimed space :)
<edef>
they're technically only taking up inodes, not real disk blocks
<Lisanna>
since they're just symlinks
<edef>
like, real filesystem data blocks
<Lisanna>
edef symlinks take up disk space, don't they
<Lisanna>
not hard links
<edef>
they do, yeah
asuryawanshi has quit [Ping timeout: 260 seconds]
<edef>
but i think most filesystems jsut store the target in the inode
<Lisanna>
I'm assuming they're symlinks... they might be hard links, I'm not sure actually
robstr has quit [Remote host closed the connection]
<Lisanna>
they should be symlinks, since nix-collect-garbage reports no space freed
<edef>
nix will hardlink things itself and recognise hardlinked things
robstr has joined #nixos
<Lisanna>
regardless, I've got about 2 terabytes worth of those things, and the only thing that's changed is I call nix-prefetch-url, nix-hash, and nix-prefetch-git in loops
gay has joined #nixos
<Lisanna>
but the results of those don't change too often...
<gay>
hi
<edef>
hm
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 38373bb Peter Hoeg: falkon: 2.1.99.1 -> 3.0.0
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vAQvj
NixOS_GitHub has left #nixos [#nixos]
<Lisanna>
so, maybe there's a bug in one of them where it doesn't clean up everything perfectly or something... no direct proof yet, I'll have to do some experimenting
robstr has quit [Ping timeout: 240 seconds]
asuryawanshi has joined #nixos
* edef
ponders running an experiment of her own
<edef>
you're just fetching the same, rarely changing repos frequently, right?
<Lisanna>
yeah
<edef>
and you're certain the space usage is from /nix/store, since GCing the store fixes it
<srhb>
I'd also like to know how you determined that there was 2 TB of links.
<gay>
So i am a big fan of functional program design. I like the choices haskell has made. I am new to NixOs. But i just don't get how Nix is a functional package manager. What i am talking about is, when i write down in my configuration that i want to install package 'x', how does the system know how to install it and what settings to use? And won't changing those settings alter the output? For example if i want to
<gay>
enable bluetooth on my device i may write down hardware.bluetooth.enable = true, but what will happen if i completely omit the line, is bluetooth enabled or not?
<Lisanna>
yes, gc takes forever, and I can watch the strace output showing it unlinking stuff, and I also watch df and show the space freeing up slowly as the gc runs
<edef>
gay: it's not enabled by default, as you can find by reading the relevant NixOS module, or checking the relevant options property
<edef>
gay: running `nixos-option hardware.bluetooth.enable` will tell you this, for example
<Lisanna>
I've had to fix this once before, the gc took several days at least
<gay>
see then there is something else other than 'configuration.nix' which was incharge of it, since i never said it in my configuration to turn it to false or true ever
<srhb>
gay: Notice for instance that it adds several packages to systemPackages
<edef>
this sounds like a fun pathological case
<edef>
gay: well, yes, and you can control what's in $NIX_PATH yourself
<srhb>
gay: The modules may do a _lot_ of things once you flip a switch
<srhb>
gay: Your configuration.nix is just part of the program. Not all of it.
<edef>
gay: you're essentially complaining that there is a standard library
<srhb>
gay: But you could, for instance, disable _all_ the module imports and import them by hand.
<gay>
i am not complaining, i just don't understand it very well and want some clarity
<srhb>
Sure :)
<srhb>
gay: Basically most modules do a Lot Of Things if they are enabled. They are self-contained programs that control part of the system.
<edef>
Lisanna: i assume the "several days" is without strace
<Lisanna>
edef yes
<edef>
(just checking my assumptions here)
<Lisanna>
once it gets to the point where it's deleting the bulk of it, I'll be sure to write down the exact things it's unlinking
muzzy_ has quit [Ping timeout: 265 seconds]
<gay>
so this 'standard library' + my configuration is the input for Nix and the output is the working system? with no other hidden inputs?
<edef>
yep
<gay>
cool
<srhb>
gay: RIght! The version os nixpkgs, which contains the module system, and your configuration.nix
asuryawanshi has joined #nixos
<srhb>
gay: You can have a look at nixpkgs/nixos/modules :)
<edef>
you could consider the other hidden input being fetches from the internet that may fail or succeed
<edef>
but in the same sense, pure FP langs have outcomes like "out of memory" that are also execution failures
<Lisanna>
but those all have hashes attached to them, to make sure they don't change
<edef>
Lisanna: are any of the repos you're using public / how large are the checkouts / how large are the full repos?
<gay>
thanks
<Lisanna>
edef I mean, I prefetch a lot of stuff, some public, some private
<NixOS_GitHub>
[nixpkgs] jtojnar opened pull request #36327: freedroidrpg: init at 0.16.1 (master...freedroidrpg) https://git.io/vAQfl
NixOS_GitHub has left #nixos [#nixos]
<Lisanna>
the experiment I plan to do is, in several days once the GC is done, with my prefetchers turned off, do some manual prefetching and then check to see if the GC cleans anything up afterwards
<etu>
Hmm, 2.0.2 update of vagrant broke some ruby thing. Haven't had the time to look into it yet.
asuryawanshi has quit [Ping timeout: 265 seconds]
<etu>
Temporarly changed the version in my nixpkgs checkout... :)
<Lisanna>
and measure if the amount of cleaning is proportional to the number of times it was prefetched
aarvar has quit [Remote host closed the connection]
asuryawanshi has joined #nixos
<gay>
i do have a complaint, nix manuals are hard to read.
<edef>
Lisanna: sounds like a plan
magnetophon has joined #nixos
<edef>
as i remember it, nix-prefetch-{url,git} are the same scripts that run in fetchurl/fetchgit, just outputting to a tmpdir and performing nix-store --add rather than just outputting to $out
<edef>
Lisanna: once upon a day, it was a bash script that called curl
<jtojnar>
cool
<Lisanna>
I believe that
<magnetophon>
edef: thanks!
<MP2E>
basically to work around it, you gotta rollback to an earlier nixpkgs version, then set 'nix.package = pkgs.nixUnstable;' in your configuration.nix for that version, then remove the line and rebuild off of the latest unstable
<MP2E>
and it will then work
<MP2E>
I am currently doing that with both my nixos machines heh
asuryawanshi has joined #nixos
<edef>
ah, nix-prefetch-zip has been replaced by nix-prefetch-url --unpack
<edef>
i was not aware of that
<edef>
and/or nix-prfetch-url -A foo.src, TIL
<edef>
*prefetch
<edef>
actually, hm, i might be wrong, nix-prefetch-url might've never been a script
aarvar has joined #nixos
<magnetophon>
MP2E: thanks
<edef>
ah! right. it was a perl script before it was a C++ program
<jtojnar>
edef: is not there something like python.withPackages for perl?
<edef>
Lisanna: oh, nuts
<edef>
Lisanna: oh hm, apparently grahamc is gchristensen-on-matrix
<edef>
¯\_(ツ)_/¯
<edef>
i should probably go to sleep anyhow
justbeingglad has joined #nixos
<Lisanna>
edef ohh, okay
periklis has joined #nixos
civodul has joined #nixos
zzamboni has joined #nixos
MP2E has quit [Remote host closed the connection]
MP2E has joined #nixos
robstr has joined #nixos
noam has quit [Read error: Connection reset by peer]
noam has joined #nixos
<mfiano>
Hello
jensens has joined #nixos
<mfiano>
Can someone tell me how I can merge a line onto a string for an option that was previously defined in a file that was brought in from an import? It's a double single quote string of lines, and I'd like to add a line to it from another file.
<ij>
Can I get nix 2.0 print the path of the build?
<mfiano>
jtojnar: Not really. I'm in the livecd on another computer trying to write a config. Basically I have file-a.nix that declares services.xserver.displayManager.sessionCommands = ''line 1\nline 2\n'', and then I have a file-b.nix that imports this file, and it would like to append to that option with a ''line 3\n''
<jtojnar>
hmm, not sure if it is possible
robstr has quit [Remote host closed the connection]
robstr has joined #nixos
<jtojnar>
maybe it is possible to change the merge rule so it behaves like lists
<jtojnar>
mfiano: there is mkMerge in lib/modules.nix, not sure if it works for strings
<mfiano>
I was hoping there would be some trick with '+' concatenation
chocopuff has joined #nixos
nick_l has joined #nixos
<jtojnar>
looks like mkMerge is only for modules
robstr has quit [Ping timeout: 256 seconds]
ertes-w has joined #nixos
periklis has quit [Ping timeout: 256 seconds]
Myrl-saki has joined #nixos
jensens has quit [Ping timeout: 245 seconds]
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
justbeingglad has left #nixos [#nixos]
thc202 has joined #nixos
asuryawanshi has quit [Ping timeout: 240 seconds]
MP2E has quit [Remote host closed the connection]
Itkovian has joined #nixos
aarvar has quit [Ping timeout: 245 seconds]
atu1 has quit [Ping timeout: 240 seconds]
logotags has quit [Read error: Connection reset by peer]
logotags has joined #nixos
<nioncode>
how can I append to the cmakeFlags variable that is defined in a package's mkDerivation from within an overlay? I either get 'function called with unexpected argument 'mkDerivation'/'cmakeFlags'' or 'cmakeFlags' already defined
Izorkin has joined #nixos
Izorkin_ has quit [Read error: Connection reset by peer]
MP2E has joined #nixos
brodul_ has joined #nixos
brodul has quit [Ping timeout: 255 seconds]
lord| has quit [Ping timeout: 260 seconds]
statusfailed has quit [Ping timeout: 240 seconds]
dev1 has quit [Ping timeout: 240 seconds]
statusfailed has joined #nixos
lord| has joined #nixos
dev1 has joined #nixos
orivej has joined #nixos
jtojnar has quit [Quit: jtojnar]
jtojnar has joined #nixos
<joko>
nioncode: use package.overrideAttrs and inside there sth like cmakeFlags = oldAttrs.cmakeFlags ++ [ "newflag" ];
<nioncode>
joko: sweet, works like a charm
<nioncode>
how can I now combine my override block (where I change the stdenv) with my overrideAttrs block?
<joko>
nioncode: wrap one with parantheses, e.g. (package.override { ... }).overrideAttrs( ... );
apurvapavaskar[m has quit [Ping timeout: 256 seconds]
<joko>
It may seem ugly, though :D
justanotheruser has quit [Ping timeout: 256 seconds]
<nioncode>
better than having a tmp variable like I have now :)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ekleog opened pull request #36329: clamsmtp: fix build after linuxHeaders update to 4.15 (master...fix-clamsmtp) https://git.io/vAQmk
NixOS_GitHub has left #nixos [#nixos]
aarvar has joined #nixos
robstr has joined #nixos
FRidh has quit [Remote host closed the connection]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/staging fe5f012 sjau: sqlite: Raising max variable number and max expr depth values...
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to staging: https://git.io/vAQmW
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vAQYE
NixOS_GitHub has left #nixos [#nixos]
<nixer101>
does anyone have any fullscreen opengl games, preferably quake3/idtech based? Like openarena or urbanterror. They seem to detect resolution incorrectly, causing monitor to turn off on stable. The session then seems to keep and also save that, so after relogin the monitor turns off permanently (unless kscreen config file is purged).
<nixer101>
I have mesa enabled, using opensource dri driver and dxtn also enabled. glxinfo looks good to me.
robstr has quit [Remote host closed the connection]
robstr has joined #nixos
<genesis>
nixer101 : i packaged assaultcube, but it's cube engine based ...
<nixer101>
genesis: thanks! is it in stable or unstable? I am eager to try it, if it also causes issues. freeorion does not cause any, even when in true fullscreen.
<genesis>
i think in unstable
<nixer101>
it should help narrowing down the issue, if assaultcube works good for you. Really appreciate the response!
<nixer101>
genesis: thanks!
<genesis>
BUT i donno well the nixos process .
<genesis>
the fact is i'm not on nixos and openarena is such opengl that failed on my gl trick
<nixer101>
genesis: interesting! Could you dump your logs to pastebin any chance? I can't even get to the logs. It just shuts down my display. Even tty is blacked out.
<genesis>
SDL_GL_LoadLibrary ...
<genesis>
i'm not on nixos, so it's different pb we have
chocopuff has quit [Ping timeout: 268 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz closed pull request #30416: nixos/luksroot.nix: fallback to interactive password entry when no keyfile found (master...luksnokey) https://git.io/vd13j
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 81fed4d adisbladis: Merge pull request #36316 from DougBurke/fix/cfitsio-3430-update...
<NixOS_GitHub>
nixpkgs/master 21a1b76 Douglas Burke: cfitsio: 3.41 -> 3.43
<NixOS_GitHub>
[nixpkgs] adisbladis pushed 2 new commits to master: https://git.io/vAQ3B
rindvieh has quit [Read error: Connection reset by peer]
rindvieh_ has joined #nixos
<hyper_ch>
Mic92: so, I changed base to staging and now fborg has errors
<Mic92>
hyper_ch: I already pushed it to staging
<Mic92>
you can close or keep the pr open
<hyper_ch>
I close it
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] sjau closed pull request #36293: sqlite: Raising max variable number and max expr depth values (staging...master) https://git.io/vAHGi
NixOS_GitHub has left #nixos [#nixos]
rindvieh_ has quit [Read error: Connection reset by peer]
rindvieh has joined #nixos
Mateon3 has joined #nixos
Mateon1 has quit [Ping timeout: 252 seconds]
Mateon3 is now known as Mateon1
ixxie has joined #nixos
<mfiano>
How can I copy the channel nix expressions from one pc to another? I'm trying to build duplicate of another machine running unstable, but the latest channel update doesn't compile, so i want to copy the existing channel revision to build
<hyper_ch>
mfiano: you could test which commit was used on your "old" machine, download nixpgks onto new machine, make checkout at same commit and then nixos-rebuild switch -I nixpkgs=/path/to/local/dir
<mfiano>
sounds good. how can i check the commit of old machine?
<nixer101>
hyper_ch: is there a way to see which unstable tree revision compiles?
<hyper_ch>
727... is the commit... but not sure if it's channel commit or nixpkgs commit as well
<hyper_ch>
nixer101: what do you mean?
<mfiano>
hyper_ch: thanks. looks like same commit as you
<nixer101>
hyper_ch: the specific of nixos, is that whole nixexpression either builds or fails. If its possible to see which tree commit passes compilation, that would allow more or less troublefree running of unstable configuration.
<hyper_ch>
nixer101: no idea
<nixer101>
hyper_ch: I would be really interested if there exists an automatic hint, as compared to trying the luck in building the tree manually each time.
<nixer101>
I have seen users, that tried to migrate to unstable in this chat, but it all came to luck of the current moment.
<nixer101>
hyper_ch: the users wanting to migrate or who use unstable, would then be able to use the last revision of tree that has successfully been evaluated.
zzamboni has joined #nixos
timon37 has joined #nixos
<hyper_ch>
no idea
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fakhreddinebb90 opened pull request #36333: dd-agent: exclude tracefs from datadog disk checks (master...master) https://git.io/vAQCY
NixOS_GitHub has left #nixos [#nixos]
<nixer101>
hyper_ch: fair enough..
chisui has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] chrisburr opened pull request #36334: libpqxx: Add dependency on doxygen and xmlto (master...patch-2) https://git.io/vAQW8
NixOS_GitHub has left #nixos [#nixos]
humanoyd_ has joined #nixos
<timokau>
Why does hydra sent me a build failure notification 3 days after the build, is that normal?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fakhreddinebb90 closed pull request #36333: dd-agent: exclude tracefs from datadog disk checks (master...master) https://git.io/vAQCY
NixOS_GitHub has left #nixos [#nixos]
humanoyd has quit [Ping timeout: 260 seconds]
<chisui>
I'm running nixos 17.09 on the unstable channel and am expericing the closure-info bug. The Github issue isn't really clear on how to proceed. I can't go back to stable and I don't want to have my system unpatched for too long.
<mfiano>
chisui: Go back to another unstable commit
<chisui>
mfiano: than my system is locked on that commit until the bug is fixed
<mfiano>
The risk of running unstable has a hint in the name. You have to be prepared to fix things yourself or go back to a working state
troydm has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
<chisui>
Yeah, I know and I am ok with that.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vAQlt
<NixOS_GitHub>
nixpkgs/master a6d9548 Jörg Thalheim: pythonPackages.python-ptrace: init at 0.9.3
NixOS_GitHub has left #nixos [#nixos]
<clefru>
I would like to open a discussion around the closure-info bug and that I suggest we adopt a rollback-first-ask-questions-later policy.
<clefru>
Having something that breaks the tree for that long for a large portion of users isn't really hitting my expectations. Even for HEAD I would rush to a fix or if none is easily available for a rollback.
<nixer101>
clefru: that would be nice.
<clefru>
(I used to work for a 40k+ engineers company with global integration tests. If you break the company globally, you unlocked an "achievement" and if the build cops didn't rollback your change without asking any question you could be sure to get a phone call on Sunday night at midnight on your holidays.)
raynold has quit [Quit: Connection closed for inactivity]
<niksnut>
clefru: that's fine for people who get paid to work on something, not volunteers doing this in their spare time
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to release-17.09: https://git.io/vAQ8n
<clefru>
niksnut: I don't see how payment and quality are related.
jensens has quit [Ping timeout: 268 seconds]
<clefru>
niksnut: also my point was to rollback breaking changes first, then ask questions later. rollbacks are cheap, so I don't find the argument compelling that this is any amount of work
robstr has joined #nixos
robstr has quit [Remote host closed the connection]
archaeron has joined #nixos
robstr has joined #nixos
<archaeron>
hi. has anyone here used nix for ocaml development?
<symphorien>
There is opam2bix for that
<symphorien>
*opam2nix
<symphorien>
Not perfect though
<nixer101>
niksnut: every distribution I know has received benefits and grow in userbase after setting the quality standards. This lead to lesser breakages and lesser stress for everyone. I think the higher standard on global level is actually requirement to become widespread.
<archaeron>
symphorien: do you know if that can also provide things like merlin?
Guest42148 has quit [Ping timeout: 240 seconds]
Guest42148 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 5b66c2b Moritz Ulrich: pythonPackages.mt-940: init at 4.10.0.
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] the-kenny pushed 1 new commit to master: https://git.io/vAQ4E
<symphorien>
Merlin is in nixpkgs and works
<symphorien>
Sorry afk
robstr has quit [Ping timeout: 276 seconds]
<nixer101>
is there a tool for fast comparison same package version in different standard channels? thanks.
<archaeron>
symphorien: thanks!
<goibhniu1>
nixer101: you can add multiple channels and then query them locally
goibhniu1 is now known as goibhniu
chpatrick has joined #nixos
<chpatrick>
hey, do you know if there's a way to run stuff on a nixops machine before the nix paths get copied to it?
<nixer101>
goibhniu: thats a good idea, thanks!
<chpatrick>
I'm using the libvirtd target and I'd like to mount the host's /nix/store like it does in stage 1
<goibhniu>
yw!
logotags has quit [Ping timeout: 256 seconds]
<archaeron>
symphorien: sorry to bother you again, but I get "permission denied" when trying out the examples of opam2nix. Do you know anything about that?
humanoyd has joined #nixos
logotags has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 closed pull request #36334: libpqxx: Add dependency on doxygen and xmlto (master...patch-2) https://git.io/vAQW8
NixOS_GitHub has left #nixos [#nixos]
humanoyd_ has quit [Ping timeout: 268 seconds]
robstr has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fakhreddinebb90 opened pull request #36335: dd-agent: Excluding tracefs from disk checks (master...master) https://git.io/vAQBP
NixOS_GitHub has left #nixos [#nixos]
robstr has quit [Read error: Connection reset by peer]
<clefru>
chpatrick: try to look at stage-1.nix, and the attribute "neededForBoot = true" on the "filesystems" attributes. see modules/virtualisation/qemu-vm.nix for an example
<clefru>
chpatrick: pardon, I might not have understood your question correctly
<chpatrick>
clefru: my problem is that when I use nixops each machine's closure is copied over quite slowly
<chpatrick>
it would be better if /nix/store was mounted with libvirt
<chpatrick>
which seems to happen anyway, but only when the machine is being set up
sigmundv has joined #nixos
<clefru>
chpatrick: I am not sure what the implications are for what you are suggesting. presumably nix/store would need to be writable by two machines (your host nad your guest). I am not sure if /nix/store is designed to be written to by two independent uncoordinated nix-daemons for instance.
<chpatrick>
clefru: I think a read-only mount is ok, since nixops machines aren't supposed to add to the nix store usually
clefru has quit [Remote host closed the connection]
nixer101 has quit [Quit: Konversation terminated!]
<ZoomZoomZoom>
nixer101, Ok, so If I understand it correctly, only the systemd unit will fail? But what If I'd like the boot process to fail?
<TweyII>
What's in scope when the expression given to commands' --arg arguments is evaluated?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg opened pull request #36338: nixos home-assistant: a couple of fixes (master...f/hass) https://git.io/vAQz7
NixOS_GitHub has left #nixos [#nixos]
atu1 has joined #nixos
humanoyd_ has joined #nixos
deanman has quit [Remote host closed the connection]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] adevress opened pull request #36339: Fix a dirty hack in the python wheel extractor that did not work if a… (master...master) https://git.io/vAQgr
NixOS_GitHub has left #nixos [#nixos]
humanoyd has quit [Ping timeout: 256 seconds]
xcmw has joined #nixos
<gchristensen>
edef: a great idea!
<gchristensen>
edef: can you open an issue on nixos/ofborg?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] shlevy opened pull request #1948: ssh-ng: Don't forward options to the daemon. (master...no-forward) https://git.io/vAQ2v
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] shlevy closed pull request #1946: ssh-ng: Only forward settings that make sense to forward. (master...forwardable-settings) https://git.io/vAHNN
NixOS_GitHub has left #nixos [#nixos]
ma27 has joined #nixos
peacememories has joined #nixos
clefru has joined #nixos
Thra11 has joined #nixos
coot has joined #nixos
roblabla has quit [Remote host closed the connection]
rindvieh has quit [Remote host closed the connection]
<symphorien>
but in my exprerience mixing both does not usually work well
<archaeron>
symphorien: I'm just trying to get a very basic thing working with base and core
<archaeron>
and then utop and meerlin that workn with it
<xnaveira[m]>
in my list of packages in configuration.nix i do (termite.override { configFile = "mynewfile" }) but i keep getting the default config filr
<xnaveira[m]>
what am i doing wrong?
zzamboni has quit [Quit: Leaving.]
zzamboni has joined #nixos
<archaeron>
symphorien: it's probably a mistake to learn ocaml and nix at the same time :)
<symphorien>
archaeron: if all your dependencies are in nixpkgs try with pure nixpkgs first
<avn_>
TweyII: we renamed mesa to libGL/libGLU/libGL_combined everywhere, except mesa itself
<avn_>
so it first little step to have "generic" reference to use in builds, and mesa is only default implementation now
<TweyII>
Ooh
<woffs>
btw, is libreoffice broken on i686? Last time I tried it didn't compile
rindvieh has quit [Read error: Connection reset by peer]
rindvieh has joined #nixos
<gchristensen>
fun bug: security.acme.certs."events.nix.gsc.io" = { ... The option `security.acme.certs.events.nix.gsc.io.webroot' is used but not defined.
<NixOS_GitHub>
[nixpkgs] bignaux opened pull request #36343: soulseekqt: init at 2016-1-17 (last official stable) (master...soulseekqt) https://git.io/vAQ17
NixOS_GitHub has left #nixos [#nixos]
fuyuuri has quit [Ping timeout: 265 seconds]
peacememories has joined #nixos
Rusty1_ has joined #nixos
szicari has joined #nixos
Thra11 has quit [Ping timeout: 240 seconds]
Thra11 has joined #nixos
mounty has quit [Ping timeout: 268 seconds]
mounty has joined #nixos
magnetophon has quit [Ping timeout: 260 seconds]
<avn_>
TweyII: it open us ability to play with libglvnd for both build-time reference, and run-time driver dispatching
<gchristensen>
is there a fast way to determine if a string contains a specific character in nix 1.11?
pie__ has quit [Ping timeout: 265 seconds]
<ottidmes>
gchristensen: builtins.match maybe?
<symphorien>
gchristensen: I have used replace char "" string == string
<ottidmes>
symphorien: lol, I was about to suggest that one as an alternative
reinzelmann has quit [Quit: Leaving]
<ottidmes>
gchristensen: str: builtins.replaceStrings [ chr ] [ "" ] str == str, but I do not know if that will be faster than builtins.match
<gchristensen>
hmm does 1.11 have .match? cool
<ottidmes>
gchristensen: Not sure... I just updated my systems to 2.0, so I would have to downgrade or see if I can get a older version of the manual
<gchristensen>
no worries
alex`` has quit [Ping timeout: 256 seconds]
<ottidmes>
gchristensen: It seems to, at least, nix-repl gives me 1.11.16 for its version, so I assume it is still using 1.11.16 and it does work: nix-repl> builtins.match ".*a.*" "abc"
ZoomZoomZoom has quit [Read error: Connection reset by peer]
ZoomZoomZoom has joined #nixos
elvishjerricco_ has joined #nixos
aanderse has joined #nixos
brodul_ has joined #nixos
flyx|znc has joined #nixos
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] fpletz pushed 6 new commits to master: https://git.io/vAQH4
<NixOS_GitHub>
nixpkgs/master 48bd537 Franz Pletz: batman-adv: 2017.4 -> 2018.0
<NixOS_GitHub>
nixpkgs/master e406460 Franz Pletz: mbedtls: 2.6.1 -> 2.7.1
<NixOS_GitHub>
nixpkgs/master 94db63f Franz Pletz: nginxMainline: 1.13.8 -> 1.13.9
Olgierd_ has joined #nixos
aanderse_ has quit [Remote host closed the connection]
ocharles has quit [Ping timeout: 276 seconds]
elvishjerricco has quit [Ping timeout: 276 seconds]
bollu has quit [Read error: Connection reset by peer]
<xnaveira[m]>
i tried to add the following in my configuration.nix environment.systemPackages = with pkgs; [ ( termite.override { configFile = "blablabla" }; )] but it is not working
<xnaveira[m]>
help please :)
shoogz has quit [Read error: Connection reset by peer]
<gchristensen>
you probably don't need that ; after the }, but you do need it before the }
andymandias has joined #nixos
<xnaveira[m]>
yes sorry gchristensen , i transcribed wrong, but when i run the switch command the syntax is fine
<xnaveira[m]>
i couldn't copy and paste my multiline code in irc
<gchristensen>
ok, perhaps then you can tell us what you mean when you say it doesn't work
<Lisanna_>
that's odd... sometimes I'm getting runaway build-remote.pl processes when using nixops, and I have to manually kill them before I can do additional nixops deploys ):
<xnaveira[m]>
but the point is that the syntax is fine, is just that i don't think i'm understanding the scopes right
<xnaveira[m]>
sorry, i mean that the changes i make to the config file don't show after the switch
<xnaveira[m]>
everything seems to run fine but the changes don't get applied, the package keep having the default config
rauno has joined #nixos
erasmas has joined #nixos
<aminechikhaoui>
maybe check if the package produced during nixos-rebuild is the same you're using
<xnaveira[m]>
is that for me aminechikhaoui ? If so, how do I check that?
MercurialAlchemi has quit [Ping timeout: 268 seconds]
johnsonav has joined #nixos
<aminechikhaoui>
yeah for you :) when you do nixos-rebuild it will likely output the store path of the package you did the override for, maybe check if the output of ll `which <pkg-bin>` is the same store path
ma27 has joined #nixos
<aminechikhaoui>
xnaveira[m]: also is the binary not running as expected both as root and as your user ?
halfbit has joined #nixos
<xnaveira[m]>
aminechikhaoui: indeed it seems like the store created by build and the one pointed by the result of doing ls -l $(which termite) are different, how do I fix this?
cross has quit [Remote host closed the connection]
<aminechikhaoui>
a workaround could be to do nix-env -i <store_path_of_needed_binary>
<aminechikhaoui>
but there is probably a better way to fix this kind of issues
<xnaveira[m]>
well it is my main terminal so i was kind of hoping to be able to handle it in my configuration.nix
iyzsong has quit [Ping timeout: 248 seconds]
hotfuzz has joined #nixos
liminal18 has joined #nixos
<liminal18>
hey how do I install pip via nix if my current python is python36?
<liminal18>
nix-env -i python36Packages.pip error: selector 'python36Packages.pip' matches no derivations
<TweyII>
liminal18: Looks like you want -iA
<TweyII>
liminal18: -i searches for packages by name; -A says to use the actual Nix attribute path
cement has joined #nixos
archaeron__ has quit [Ping timeout: 240 seconds]
<TweyII>
nix-env -iA nixpkgs.python36Packages.pip
<nick_l>
How can I make something like this work? #! /usr/bin/env nix-shell #! nix-shell -I nixpkgs=./nixpkgs --pure -i sh -E 'with import <nixpkgs> {}; (((terraform_0_10.withPlugins (p: [ p.aws ])).withPlugins (p: [ p.azurerm p.google ])).overrideAttrs (orig: { name = orig.name + \"foo\"; })).withPlugins (p: [ p.kubernetes ])'
<TweyII>
liminal18: Or nix-env -i python3.6-pip
<liminal18>
TweyII that worked thanks!
<gchristensen>
nick_l: my goodness, can I recommend a shell.nix?
<nick_l>
gchristensen: doesn't that require people to know that nix-shell exists?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] matthewbauer opened pull request #36345: emacs: give melpa and melpa-stable own scope (master...emacs-melpa-fix) https://git.io/vAQFv
NixOS_GitHub has left #nixos [#nixos]
<liminal18>
lol
<liminal18>
now I need python3 setuptools
<gchristensen>
nick_l: your system requires people to know that your script exists
ZoomZoomZoom has quit [Quit: Leaving]
<ottidmes>
Hmm, thats a new one: error: serialised integer 7161674624452356180 is too large for type 'j', when trying: nixos-rebuild switch
<ottidmes>
Anyone any idea what might cause that particular error?
<liminal18>
ok got it thanks
<nick_l>
gchristensen: ./my_domain_specific_script VS "You need to run cd directory && nix-shell".
<nick_l>
gchristensen: I don't know everything, which is why I asked, but do people need a manual step or not? If the solution involves seeing 'nix-shell' in their screen and then typing something, it changes their workflow and doesn't count as a solution.
<nick_l>
gchristensen: now, that's just a work of art!
<ottidmes>
My error is caused by setting networking.networkmanager.enable = true; it seems. I tried enabling it earlier today, then I got no such error, weird
<gchristensen>
=)
<fyuuri>
How can I use plain xinit as "displayManager" in NixOS?
<TweyII>
liminal18: It sounds like you're building something. I'd suggest making a nix-shell (and maybe a derivation) for that thing instead of trying to install its dependencies into your user environment
<rauno>
any idea why the networking.localCommands fails to add a networking route ?
<liminal18>
how do I uninstall python36 from nix
<ottidmes>
fyuuri: As also mentioned by a comment on that reddit post I linked, you might also take a look at: https://nixos.org/nixos/options.html#services.xserver.displaymanager.session
<fyuuri>
thank you :)
<fyuuri>
that helps!
<ottidmes>
rauno: Any errors? Seems unrelated to NixOS though, I would just double check the command in your shell to be sure they are working properly for your network setup
<Lisanna_>
liminal18 nix-env -e python36
<liminal18>
<Lisanna_> Tried that
<Lisanna_>
result?
<rauno>
ahh, found the reason, it's only triggered after network.service reload
aarvar has joined #nixos
<liminal18>
it's still there
<rauno>
thought it would be execute at the change of the value also
<Lisanna_>
liminal18 are you on NixOS?
<liminal18>
using Mac osx
<Lisanna_>
rauno that sounds like a bug, if you change the value, it should reload anything that depends on it...
<liminal18>
which python /Users/andrewjones/.nix-profile/bin/python
<Lisanna_>
are you sure that's python36?
<liminal18>
python --version Python 3.6.2
<ottidmes>
liminal18: nix-env -q, and see what is installed
<liminal18>
turns out the python package I need is available as nix package in python27
<liminal18>
@ottidmes thanks
<liminal18>
python3-3.6.2-env
<liminal18>
that might be it
<liminal18>
I have the env installed
<liminal18>
anyways figured out a way around it
<Lisanna_>
yeah, you might have "python3" installed also, which is also a python36
yegortim1 has quit [Write error: Connection reset by peer]
<liminal18>
Lisanna_ that is probably it
knupfer has quit [Ping timeout: 245 seconds]
yegortim1 has joined #nixos
globin_ is now known as globin
atu1 has joined #nixos
rindvieh has quit [Remote host closed the connection]
<sonarpulse>
basically now that cross is good, it would be cool to cross compile to illumos
<edef>
so like, this is kind of a tall order
<sonarpulse>
hehe yeah
<edef>
illumos kind of expects to be its own world
<sonarpulse>
years of accumulated layer violations
<sonarpulse>
build system wise
<sonarpulse>
from living in isolation
<edef>
well, there's a pretty narrowly defined public interface
Thra11 has quit [Ping timeout: 260 seconds]
<sonarpulse>
libc not stable kernel abi?
<edef>
yeah
<sonarpulse>
Darwin is the same thing
<sonarpulse>
so there's some nixpkgs precedent
<edef>
also things along the lines of "there is a cc, it does the things, takes these -l's, etc"
<sonarpulse>
they don't require Sun Studio anymore at least
<edef>
but no real public stable notion of what that cc involves
<edef>
well, have we cut the linter out yet? i don't think so
<sonarpulse>
hmm?
<edef>
it doesn't require Studio for compilation, this has been true for years though
<fearlessKim[m]>
I want to test the instalPhase of gpgme in a nix -shell so I `export out=$PWD/toto` (which usually works) but the configurePhase is still using --prefix=/nix/store/... is it because of autoreconfHook ?
<sonarpulse>
oh linTer not linKer
alex`` has joined #nixos
<edef>
Sonarpulse: yeah
vcunat has joined #nixos
hotfuzz_ has joined #nixos
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Thra11 has joined #nixos
knupfer has joined #nixos
<ottidmes>
anyone any idea what can solve this error, I am at a loss: error: serialised integer 7161674624452356180 is too large for type 'j'
hotfuzz has quit [Ping timeout: 260 seconds]
m0rphism has quit [Quit: WeeChat 2.0]
<gchristensen>
ouch, how'd you get that?
<ottidmes>
gchristensen: Just by building my system
<ottidmes>
gchristensen: Weird thing is, on my desktop I get it, and on my server (exact same integer), while on my desktop I can remove the error if I disable networking.networkmanager.enable = true; which my server does not have (yet still getting the error there)
<ottidmes>
gchristensen: And on another server with almost identical config as the other server, I do not get the error
<ottidmes>
The other server (without error) has an AMD CPU, while my desktop and the server (with error) have an Intel CPU, could be related, could not be...
Thra11 has quit [Ping timeout: 256 seconds]
m0rphism has joined #nixos
<ottidmes>
gchristensen: I thought maybe my networking.hostId (because on the server it traces to that), but on my desktop it does not show a trace, and I after replacing the hostId's I still get the error
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 5 new commits to master: https://git.io/vA7TU
<NixOS_GitHub>
nixpkgs/master d21e682 Franz Pletz: virtmanager: 1.5.0 -> 1.5.1
<NixOS_GitHub>
nixpkgs/master cda1e6c Franz Pletz: urbanterror: 4.3.2 -> 4.3.3 for CVE-2017-6903
<NixOS_GitHub>
nixpkgs/master 1a0f84d Franz Pletz: wireshark: 2.4.4 -> 2.4.5
NixOS_GitHub has left #nixos [#nixos]
typeggze1o has quit [Quit: leaving]
Thra11 has joined #nixos
<ottidmes>
gchristensen: Thanks for the pointer, I guessed as much, but now I at least know for sure that I assign an integer somewhere that overflows
m0rphism has quit [Client Quit]
<ottidmes>
gchristensen: Got it, seems to be a bug in Nix 2.0
<ottidmes>
gchristensen: When I add substituters = ssh://nix-serve@server it crashes on that error
<gchristensen>
neat .....
phdoerfler has joined #nixos
<LnL>
ottidmes: I ran into a very similar error once when messing with the pre-build-hook
Thra11 has quit [Ping timeout: 240 seconds]
<LnL>
just assumed it was my bug, but maybe not?
cmacrae has joined #nixos
ryanartecona has quit [Quit: ryanartecona]
yann-kaelig has joined #nixos
shoogz has joined #nixos
shoogz has quit [Max SendQ exceeded]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra closed pull request #1948: ssh-ng: Don't forward options to the daemon. (master...no-forward) https://git.io/vAQ2v
NixOS_GitHub has left #nixos [#nixos]
__Sander__ has quit [Quit: Konversation terminated!]
shoogz has joined #nixos
shoogz has quit [Max SendQ exceeded]
<ottidmes>
LnL: It could very well be the same bug, I just replaced my nix-serve user with root again, and everything works again without the error
rindvieh has quit [Remote host closed the connection]
shoogz has joined #nixos
shoogz has quit [Max SendQ exceeded]
shoogz has joined #nixos
shoogz has quit [Max SendQ exceeded]
shoogz has joined #nixos
shoogz has quit [Max SendQ exceeded]
<LnL>
I was using it in combination with distributed builds which substitutes over ssh
shoogz has joined #nixos
shoogz has quit [Max SendQ exceeded]
<LnL>
so that's definitively possible :)
<ottidmes>
LnL: That is seems very likely to be the same bug
pxc2 has joined #nixos
pxc2 has quit [Client Quit]
shoogz has joined #nixos
shoogz has quit [Max SendQ exceeded]
<ottidmes>
LnL: I had my nix-serve user restricted with: ForceCommand nix-store --serve
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<nick_l>
gchristensen: looks like it doesn't work as intended, although I could add an additional indirection to make it work. https://github.com/NixOS/nix/issues/1344
<ottidmes>
LnL: Maybe it has to do with nix-serve not being allowed to call nix-store --serve, still that should not crash like it does now
<gchristensen>
nick_l: I think you missed that I was using -f ./shell.nix
tertle||eltret has quit [Quit: Connection closed for inactivity]
<LnL>
ottidmes: yes, assuming you're using 2.0 on both sindes but I wouldn't recommend mixing versions
<ottidmes>
LnL: yeah, I am using the same version on all relevant machines
<nick_l>
Is there some default service to make sure that the disk doesn't run about of space just because there is 100GB of NixOS software on it with 10 versions of Libreoffice?
<nick_l>
That resulted in not being able to login on one machine (and even possibly losing a wifi connection).
Quickshot has joined #nixos
<nick_l>
It seems weird to require a non-empty harddisk for wifi to work, though.
<nick_l>
It seems weird to require available space on a harddisk for wifi to work.
<nick_l>
That would be a good test to include in the release tests, if it is at all possible.
<Quickshot>
Though I'm probably not remembered, some weeks ago I was here over a multi-monitor issue, and I'm back with the exact same issue still. My specific problem isn't that I can't set it up, or control them in a very minimal easy way. Instead it's if the primary monitor is turned off that these settings are lost and need to be reapplied.
<Baughn>
Quickshot: Please pastebin the exact configuration you're trying.
<Quickshot>
Sufficient configuration in xorg.conf can I believe prevent this, as it works on a previous system like that
<nick_l>
Baughn: what's the name of that option?
<Quickshot>
Baughn, You mean the original Xorg.conf?
<Baughn>
Quickshot: That too, but I meant your configuration.nix
<Baughn>
nick_l: nix.gc.automatic
pxc has joined #nixos
<nick_l>
Baughn: cool, also found it on the wiki.
<Quickshot>
It's a bit of a mess of different attempts mixed through each other and then all '#' out
<Quickshot>
But I'll paste it I guess
<nick_l>
Baughn: can't really think of any reason to not enable this by default.
<Baughn>
nick_l: Note that since ~every filesystem will near-permanently degrade in performance if you fill them above 85-90%, there's no point in testing what happens if it hits 100%
<Baughn>
nick_l: Well, the problem is it may delete inputs needed to rebuild your system.
<Baughn>
(Which then get redownloaded, but still.)
<ottidmes>
LnL: Nope, changing it to: ForceCommand nix-daemon --stdio, still gives me the error :(
<nick_l>
Baughn: but it won't delete what you are currently using, right?
<Baughn>
nick_l: Also it'll break the cache for nix-shell -p, etc. Most people here prefer to run it manually... but it should be mentioned, I agree.
<nick_l>
Baughn: including input dependencies, that is.
<Baughn>
No, not that.
nocoolnametom has quit [Quit: ZNC 1.6.5 - http://znc.in]
<Baughn>
If you just want to fix your current system, run `sudo nix-collect-garbage -d`.
<Baughn>
That'll delete everything *but* your running system.
sanscoeur has joined #nixos
sanscoeur has quit [Read error: Connection reset by peer]
nocoolnametom has joined #nixos
<nick_l>
Baughn: I guess there should be high-level configuration like "this is a desktop on a fast network".
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Quickshot>
I could probably manage to fully replicate that one if need be in Nixos I suppose, though I'm pretty sure not all those options should be necessary
<Quickshot>
But the other one requires being able to set multiple monitor sections and I have found no ways to do so in Nixos
<Baughn>
That one only has a single monitor section, you'll notice.
<Baughn>
You can add the 'metamodes' field to the screen section easily enough.
<Baughn>
Otherwise, I think xrandrHeads?
<LnL>
ottidmes: hmm, one more thing to check isthe ssh user a trusted user?
<Lisanna_>
Does Nix uses the first matching entry in NIX_PATH or the last matching entry?
<Lisanna_>
e.g., if NIX_PATH="nixpkgs=./dir1:nixpkgs=./dir2", which will it use?
<ottidmes>
LnL: Without any restrictions and with it added to nix.trustedUsers, I am still getting the error
<ottidmes>
LnL: I checked the journal again and saw this: Attempted login by nix-serve on UNKNOWN (in red)
<Baughn>
Lisanna_: Interesting question. Let me know if you figure it out?
<Baughn>
From first principles, I would guess the last one.
<coconnor>
If it works like PATH then the first one
<Quickshot>
That's my current generated nixos xorg conf
<Baughn>
Right. I have no clue what it's doing.
<coconnor>
Quickshot: I'd try xrandrHeads for output config. That's the one that worked for me to change the primary head at least
acertain has joined #nixos
<Quickshot>
coconnor, I've used xrandrheads, yes
<Quickshot>
Changing the primary isn't the issue
<coconnor>
Quickshot: understood. That's only the closest option I've used to what you are trying haha
<Quickshot>
It's that if I turn the primary off it will lose config and default to mirror view
<Quickshot>
I see
<Quickshot>
I get the feeling most people don't really try dual monitor setups
<ottidmes>
Lisanna_: Baughn: Here is the proof: NIX_PATH=foo=/bin/sh:foo=/usr/bin/env nix-instantiate --eval -E '<foo>'
<coconnor>
I have two monitors and an nvidia if you'd like me to try
<coconnor>
Just got up so not caught up with the chat. Not sure of hte issue
kitemikaze has joined #nixos
<Quickshot>
Ok, my issue is basically I can setup a dual monitor setup fine. I even have a simple command to force it to it if need be. But if I turn off my primary monitor it will lose config and jump to mirror mode instead.
<Quickshot>
This is kind of annoying, as then you have to reset the config, via for instance the simple command
<Quickshot>
It's not some kind of major issue, but it's one of those niggling things
atu1 has joined #nixos
<ottidmes>
LnL: Where would I find this nix-daemon log? I don't see anything in the journal about it, it only contains connection and disconnections stuff for the nix-serve user
<coconnor>
Quickshot: "turn off my primary" specifically how? In xorg config or xrandr commands?
<LnL>
sudo journalctl -u nix-daemon
<LnL>
ottidmes: ^
<Quickshot>
coconnor, Via the power button
<coconnor>
Quickshot: OH
<Quickshot>
Right, something one might do in the real world if one leave the computer for a long time
<coconnor>
Quickshot: In which case the secondary monitor would then be mirroring the (now off) primary
<Quickshot>
Right
<coconnor>
that is pretty annoying
arximboldi has quit [Quit: arximboldi]
<Quickshot>
Well technically it's worse
<ottidmes>
LnL: Ah thanks, it only mentions root connecting, not nix-serve, so it does not get to nix-daemon before disconnecting: nix-daemon[13089]: accepted connection from pid 17981, user root (trusted)
dnovosel has joined #nixos
<Quickshot>
Because it in fact remembers the setup of your windows in dual monitor mode
<Quickshot>
And just sets your coordinates to 0,0
<coconnor>
why "windows entirely outside of display" is still an issue on all desktops in 2018 is a mystery...
ixxie has quit [Quit: Lost terminal]
<coconnor>
what desktop are you using? EG: KDE?
<Quickshot>
XFCE
<Quickshot>
It has a menu to setup dual monitor if one wants, but as one can guess, doesn't persist
<ottidmes>
Quickshot: Is is limited to NixOS, or did you also encounter it with other distros? Could it be a GPU driver issue?
<nick_l>
Quickshot: if you work with Plasma5 it sort of works to the point that I can connect to multiple different external monitors and use it as a laptop.
<Quickshot>
You can find this in other distro. But in other distro I just write the entire xorg.conf myself. With for instance two monitor sections
<Quickshot>
Or alternately I let say the nvidia configuration tool write one out
<Quickshot>
Either way has worked for me
<LnL>
ottidmes: ok, then it's probably a configuration issue
<nick_l>
Quickshot: you can write your own xorg.conf if you want.
* LnL
searches for his config
<Quickshot>
nick_l, There's a way to write ones own config entirely?
<nick_l>
Quickshot: you just need to change the module that currently writes the xorg.conf file.
<nick_l>
Quickshot: yes
<Baughn>
No need to replace the module. ^^;
<Quickshot>
And how does one do this? I only found options for changing subsections of xorg.conf
<nick_l>
Quickshot: You could also write a patch allowing this override everything behavior you seem to want.
<Baughn>
Quickshot: services.xserver.config
<Quickshot>
Sadly I'm not so proficient in programming to be able to write a patch
<Quickshot>
So that's not a real option
<Baughn>
But you shouldn't do that, because you won't get the right /nix/store paths in it.
<Quickshot>
Which is a problem then
<Baughn>
Pretty much have to use xrandrHeads, I think.
<Quickshot>
Which didn't work
sigmundv has quit [Ping timeout: 252 seconds]
<nick_l>
Baughn: he could just comment the required store paths in the configuration, right?
<Quickshot>
Unless I made a mistake in my xrandr command that you can see?
<coconnor>
Quickshot: have you tried seeing what happens with only the xterm desktop?
<Quickshot>
coconnor, You mean TWM?
<Baughn>
nick_l: COmment? Do you mean copy?
<nick_l>
Baughn: so, that would result in the store paths actually being available.
<coconnor>
Both KDE and XFCE have daemons that respond to xrandr events and manipulate xrandr IIRC
MarcWeber has quit [Ping timeout: 248 seconds]
<Quickshot>
coconnor, Do I really have to... ^^;
<nick_l>
Baughn: but I must admit that I don't understand exactly why replacing xorg.conf wouldn't work.
<ottidmes>
LnL: Ah, I think I found it
<coconnor>
only if you want! otherwise: patience
MarcWeber has joined #nixos
<Baughn>
nick_l: The x server won't know where to find the fonts unless it's told to.
<Quickshot>
coconnor, Though on a previous system I used to use windowmaker and I think it had the same issue... thogh admittedly one needs to test to be entirely sure
<nick_l>
Baughn: but that would simply be the responsibility of whoever writes the xorg.conf file then.
<Quickshot>
Hand entering the store paths for the fonts probably has as problem that if the fonts get changed, they won't work any more, right?
<nick_l>
Quickshot: this was a trick question, since there is no xorg.conf file!
<Baughn>
Quickshot: Quite.
<nick_l>
Quickshot: on my system I just have 10-evdev.conf
<Quickshot>
So basically we're at a point where perhaps nixos ability to modify the xorg.conf is insufficient for certain multi monitor commands?
<ottidmes>
LnL: Well, I got closer at least, I now at least see: nix-daemon[13089]: accepted connection from pid 20026, user nix-serve (trusted)
<nick_l>
Quickshot: in the xorg.conf.d directory.
<Quickshot>
That only contains quirks, doesn't it?
<coconnor>
Quickshot: comparing the nvidia generated and your current config I do see the nvidia one disables Xinerama
<Quickshot>
Oh hey, that dir does still exist
<Quickshot>
Well there's probably multiple ways to the solution
<nick_l>
Quickshot: anyway, I think your mindset should be that NixOS can do everything.
<coconnor>
Quickshot: this will impact multimonitor mode. Have you tried serverLayoutSection = ''Option "Xinerama" "0"''; in nixos?
<nick_l>
Quickshot: it's just not as direct, because it solves a more complex problem.
<Quickshot>
But I probably need to be able to define two monitor sections for the xinerama one I think
<ottidmes>
LnL: Just had to apply the same fix on the second substitutor as well and it solved the bug
<Quickshot>
coconnor, I don't think so. Let me check
<nick_l>
Quickshot: the problem being defining an implicit function how to build a system as opposed to just writing a file to an existing system.
<srhb>
Are you sure config even contains store paths normally?
<srhb>
Mine does not.
<LnL>
ottidmes: yeah, that looks better what was the problem?
<Quickshot>
coconnor, Looks like I didn't, might be worth a shot
<nick_l>
Quickshot: I recommend you write a simple test module that writes the file /helloworld to the root folder in your very own module.
<ottidmes>
LnL: Thanks for the help! I will report this and another bug I encountered with substitutors
<Quickshot>
nick_l, As I noted, my programming ability isn't that good. So that's an impossible request practically speaking
<coconnor>
Quickshot: Ideally you should be able to do services.xserver.config = ./your-custom-config.cfg;
<nick_l>
Quickshot: hmm, I guess that's good news. I didn't know NixOS was getting that popular.
<coconnor>
buuut. that config option is not set up to be overriden. Which is a bit annoying.
<nick_l>
Quickshot: you can open an issue on GitHub if you want.
<nick_l>
Quickshot: that only requires being precise in English.
<nick_l>
Quickshot: I personally think programming of this kind is easier than configuring Xorg ;)
<srhb>
coconnor: overridden how?
<Quickshot>
nick_l, Perhaps, but I figured out xorg many years ago to an extent already. So I guess I'm already used to it.
<coconnor>
srhb: Some options have a default value unless configured otherwise.
humanoyd has joined #nixos
<Quickshot>
The strange ways one develops I suppose.
<ottidmes>
LnL: I needed to set a correct shell for nix-serve, otherwise it would use nologin, causing it to immediately exit. Yet getting such an error message from Nix in response to this is not OK, so I will report an issue for it
<srhb>
coconnor: Yes?
rindvieh has quit []
<coconnor>
srhb: I would expect this option to behave the same. Default to the generated config if there is no module setting a specific config.
<srhb>
Ah
<nick_l>
Quickshot: what kind of magic setup do you want anyway?
<coconnor>
srhb: IIRC that is the "mkDefault" function.
<nick_l>
Quickshot: are you willing to dump XFCE?
<Quickshot>
nick_l, Just a dual monitor setup that persists. Rather then getting cycled to mirror mode if you switch the monitors off with the power button
<coconnor>
srhb: which is not applied to the value set to services.xserver.config by the xserver module.
Myrl-saki has quit [Ping timeout: 256 seconds]
<Quickshot>
It's nothing what one would really call exotic
<coconnor>
not sure if the "lines" type has any specific handling of that tho
<srhb>
coconnor: They are orthogonal. mkDefault just creates a very low priority override.
<nick_l>
Quickshot: ok, so if I now turn off my main screen and then turn it on again, I should see the same as I do now?
<nick_l>
Quickshot: I am fairly sure that works.
<Quickshot>
Basically I want to see the same then, yes
<coconnor>
srhb: excellent. Then having mkDefault applied to services.xserver.config would enable a user to avoid the generated config entirely.
<Quickshot>
And it doesn't
<nick_l>
Quickshot: it does for me.
<nick_l>
Quickshot: using Plasma5.
<Quickshot>
With dual monitors you get cycled to mirror mode instead. And coordinates for both monitors gets set to 0,0
<Quickshot>
nick_l, I'm using XFCE though
Myrl-saki has joined #nixos
<nick_l>
Quickshot: I have used XFCE too for sometime, but at some point you just need to consider what you want to spend time on.
<srhb>
coconnor: I think it's confusing that it's lines to begin with. Having no merge in the outer config is easier to understand, but I'm sure a lot of modules add to it.
<nick_l>
Quickshot: XFCE should fix their software to work at least as well as Plasma5.
<Quickshot>
nick_l, It's more a feature of simple desktops
<coconnor>
srhb: thus is mkDefault was used Quickshot would be able to go services.xserver.config = builtins.readFile ./totally-custom-xorg-config.cfg;
<Quickshot>
They just rely more on X11 to do the heavy lifting
<Quickshot>
Blaming everything on software, rather then the ability to set it correctly for people who want such simple desktops is unproductive
<nick_l>
Quickshot: then open a ticket.
<coconnor>
srhb: I don't see other modules that add to services.xserver.config directly... Still searching. Other modules add to the generated config options tho.
<nick_l>
Quickshot: er issue. :)
TweyII has quit [Ping timeout: 256 seconds]
<Quickshot>
nick_l, I'd rather first explore possible solutions. Which is why I asked for possible solutions.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat tagged 18.03-beta at c6fea88: https://git.io/vA7nR
NixOS_GitHub has left #nixos [#nixos]
<Quickshot>
Escalating with out doing so is kind of incomplete research anyway
<vcunat>
^^ cutting 18.03 from master right now
<fpletz>
it's happening \o/
<srhb>
vcunat: omg
<srhb>
coconnor: Doesn't filterOverrides get applied anyway?
<srhb>
coconnor: In which case mkForce would suffice.
pkill9 has quit [Ping timeout: 240 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat created release-18.03 from master (+0 new commits): https://git.io/vA7nr
NixOS_GitHub has left #nixos [#nixos]
<gchristensen>
woooooohooo!!!!
<genesis>
:)
<srhb>
coconnor: Or any mkOverride with a priority higher of whatever the default settings are. 100?
xcmw has quit [Ping timeout: 245 seconds]
<nick_l>
Man, where is the party?
<coconnor>
srhb: ahhh yea.. that does sound right
<vcunat>
gchristensen: are you among github admins? We need to protect the new branch.
<LnL>
\o/
<srhb>
coconnor: So your trick would work with no module hacking required :)
<gchristensen>
vcunat: somehow I've evaded that duty. better ping domenkozar, ikwildrpepper, or nick_l
<gchristensen>
vcunat: somehow I've evaded that duty. better ping domenkozar, ikwildrpepper, or niksnut
<coconnor>
srhb: excellent!
<vcunat>
We should have a list that maps such permissions to people.
jtojnar has quit [Quit: jtojnar]
<coconnor>
Quickshot: there looks to be a way for you to use your "known working" config exactly.
<globin>
shlevy has that permission too afaik
<coconnor>
Quickshot: tho, I think the output link names would still need to be changed.
<coconnor>
Quickshot: try services.xserver.config = builtins.readFile ./your-working-config.cfg; Where "your-working-config.cfg" is the xorg config you want. In the same directory as the nix module.
<Judson>
Is codetriage an officially sanctioned NixOS thing?
xcmw has joined #nixos
<vcunat>
gchristensen: right, thanks. This permission does have a list :-)
ThatDocsLady has quit [Remote host closed the connection]
<gchristensen>
Judson: I don't think so, what would it mean to be sanctioned or not?
ThatDocsLady has joined #nixos
<srhb>
I think we had the badge on the front of the nixpkgs repo for a long while.
<srhb>
Oh, we still do
<srhb>
So "yes" I guess :-P
<ottidmes>
LnL: Now that I fixed my issue, could you perhaps point me to the documentation about nix-daemon --stdio, since I would like to know what the impact is of using that one over nix-store --serve
<LnL>
euh.... :p
kitemikaze has quit [Ping timeout: 240 seconds]
<srhb>
coconnor, Quickshot: you still need mkForce. lib.mkForce (builtins.readFile ...
<srhb>
careful though :)
<srhb>
It's not the final config value for determining the contents of the file, despite the name
<Quickshot>
I see... so a bit iffy if you go that way. Aside of probably also not maintaining font paths?
<srhb>
Actually the font parts are exactly untouchable, so they will survive. As the only thing.
xcmw has quit [Client Quit]
<Quickshot>
So those are always dynamically updated?
<coconnor>
Quickshot: fontpaths and modules will be included in the final config. Likely as expected
<srhb>
Quickshot: Logically it goes something like this: xserver.conf = config + fonts
<coconnor>
(+ module paths)
<shlevy>
globin: Nope
<srhb>
coconnor: Oh, yes. Thank you. :)
<coconnor>
just being pedantic ;)
<srhb>
Nah, I missed it entirely. :)
<shlevy>
Oh, maybe I can do protected branches actually... lemme check
<shlevy>
Nope
<coconnor>
from an end user perspective.. it really does seem to do "the right thing"
<shlevy>
niksnut: ^^ Can I get that access?
<srhb>
coconnor: Maybe someone put some thought into this! Heresy...
<coconnor>
hahaha
<Synthetica>
How do you set a cursor theme globally?
<gchristensen>
!!! #nixos is > 800users!
<srhb>
Quickshot: Oh, and you might not have lib in scope. You'll need to add it to the top level function of configuration.nix :) { config, pkgs, lib, ... }: ...
<LnL>
what!
<LnL>
gchristensen: when did that happen? :D
<srhb>
gchristensen: 2018, year of the declarative distro!
<gchristensen>
9 users ago! ;)
kitemikaze has joined #nixos
<Quickshot>
srhb, Ahh, this is all starting to sound a bit more involved.
<Quickshot>
I'm still trying some of the earlier things suggested though
kitemikaze has quit [Remote host closed the connection]
<srhb>
coconnor: What was the name of that ancient tool to configure XFree86 which would warn you not to much it up or be prepared to lose your monitor?
<srhb>
Those were the days...
<srhb>
s/much/muck
kitemikaze has joined #nixos
ThatDocsLady has quit [Ping timeout: 256 seconds]
<clever>
srhb: i remember that one, what was it...
<coconnor>
srhb: oof. I remember reading, with excitement, about an upcoming "-configure" option to, wait for it, auto configure X. freakin magic
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 1 new commit to release-17.09: https://git.io/vA7W0
<fpletz>
!m vcunat
<[0__0]>
You're doing good work, vcunat!
<srhb>
X -probeonly perhaps
<srhb>
Terrifying stuff.
<coconnor>
anyways, nixos is better than that by far. ;)
<srhb>
:-)
<coconnor>
let the domain experts create the modules. let the users have a uniform config language. much rejoicing
<srhb>
grmbl. Smart test returns no errors, and yet I still see random permafreezes when zfs scrubs.
<srhb>
How do you debug something when the logs aren't written to disk because... The disk is probably done working for this power cycle.
<srhb>
Derp.
<coconnor>
in theory: echo kernel messages over network. Never learned how tho
<srhb>
Good idea.
<coconnor>
some mobos have a remote access to serial port as well. Also, never learned how to use that
<coconnor>
now that I think about it... I should unplug that module
<srhb>
netconsole... woo
<Quickshot>
Well I've had enough fiddling with my monitors for a bit now, have some other things like food to do still. I'll try some more options suggested next time. Thanks for the help everyone.
<vcunat>
Ah, I thought you linked to something completely different :-)
<gchristensen>
ah :)
<vcunat>
This is a problem of `inherit name;`, I suspect
<gchristensen>
quite likely
<gchristensen>
boy I quickly fell off the deep end trying to use pandoc to convert a mardown file to a pdf file. pretty sure I'm downloading the entire nixos cache as part of it.
<vcunat>
Hmm, I wonder: Jackal, Jackrabbit, Jaguar, Jay, Jellyfish :-)
<gchristensen>
jellyfish could be fun
<vcunat>
yes, I thought so, so far
<vcunat>
+JackRussel
<vcunat>
I'm not sure about "too exotic" ones like
<vcunat>
Now we got rid of that nix-2.0 upgrade problem, I hope.
<avn_>
vcunat: yep, eelco' commit fix everything. I already built from master today earlier
<vcunat>
Yes, I meant we just now past that commit on nixos-unstable, I think.
<vcunat>
*got past
sanscoeu_ has quit [Remote host closed the connection]
MichaelRaskin has joined #nixos
sanscoeur has joined #nixos
dan_b has quit [Ping timeout: 260 seconds]
mg- has quit [Read error: Connection reset by peer]
hamishmack has joined #nixos
michas_ has joined #nixos
nkaretnikov has left #nixos [#nixos]
rauno has joined #nixos
asuryawanshi has quit [Ping timeout: 240 seconds]
hiratara has quit [Ping timeout: 240 seconds]
fuyuuri has joined #nixos
<elvishjerricco>
I can't `nix-store --delete --store s3://my-bucket /path`?
michas_ has quit [Ping timeout: 265 seconds]
hiratara has joined #nixos
MichaelRaskin has quit [Ping timeout: 248 seconds]
fuyuuri has quit [Ping timeout: 240 seconds]
<elvishjerricco>
is it possible to remove a signature from a store path?
katona has joined #nixos
mkoenig has quit [Ping timeout: 256 seconds]
mkoenig has joined #nixos
dnovosel has quit [Ping timeout: 240 seconds]
<ottidmes>
Why is there an option called extra-substitutors, how does it differ from substitutors? Is the difference that untrusted users can use those listed in substitutors, but cannot list those in extra-substitutors? However, isn't that what trusted-substitutors is for? Also, this description of trusted-substitutors seems open to interpretation to me: "Unprivileged users are only allowed to pass a subset of the URLs
<ottidmes>
listed in substituters and trusted-substituters."
<ottidmes>
Does it mean "the subset" instead of "a subset", if it is "a subset", which subset? And then, why not specify extra-substitutors as well
mekeor has joined #nixos
knupfer has joined #nixos
<mekeor>
any ideas why my sound only works with flash-videos and VLC?
<symphorien>
maybe enable pulseaudio
<symphorien>
firefox only supports it for example, but it is not NixOS default
<mekeor>
oh
<mekeor>
thanks
vcunat has quit [Quit: Leaving.]
bpa has quit [Remote host closed the connection]
<mekeor>
thank you, symphorien :)
knupfer has quit [Remote host closed the connection]
Neo-- has joined #nixos
Judson has quit [Quit: Leaving.]
justanotheruser has quit [Ping timeout: 256 seconds]
pkill9 has joined #nixos
jmob has joined #nixos
self has joined #nixos
self is now known as Guest77574
pxc has quit [Ping timeout: 268 seconds]
<jmob>
Is there a way to run an "installPhase" script for a nix-shell?
simukis has quit [Ping timeout: 265 seconds]
<symphorien>
there is shellHook
<symphorien>
it is executed when you enter the shell
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] matthewbauer opened pull request #36357: demo: make virtualbox more usable (master...vbox-usable) https://git.io/vA7DR
NixOS_GitHub has left #nixos [#nixos]
<symphorien>
so shellHook = '' eval "$installPhase" '' should be possible
<tnks>
so I switched my nix-serve implementation from bzip2 to pxz, and my "nix build /nix/store/..." works fine when I have 127.0.1.0 as my cache server, but not 127.0.0.1.
<jmob>
symphorien: I'm using that, but it doesn't let me modify the environement
<symphorien>
what do you mean
<tnks>
It's complaining of things like this: file 'nar/phcnxfcmnzdy52k0696d5zpjnrlpz6k4.nar.bz2' missing from binary cache
<symphorien>
?
<tnks>
which suggests to me there's some dirt in a cache somewhere.
<symphorien>
can you be more specific/paste code ?
<jmob>
$ nix-shell
<jmob>
mkdir: cannot create directory ‘/nix/store/v9986vp8n528md4nj8692qg6axlzrg2m-piksi-buildroot-env’: Read-only file system
<jmob>
ln: failed to create symbolic link '/nix/store/v9986vp8n528md4nj8692qg6axlzrg2m-piksi-buildroot-env/lib/libtinfo.so.5': No such file or directory
<jmob>
Let me gist the default.nix I have
justanotheruser has joined #nixos
mahalel_ has quit [Remote host closed the connection]
<symphorien>
ahhhh
Guest77574 has quit [Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org]
<symphorien>
you cannot run the installPhase because it will try to write to the store which is read only
<symphorien>
yes but why do you insist in doing it in a nix-shell ?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] nicknovitski opened pull request #36358: cloud-sql-proxy: init at 1.11 (master...init-cloud-sql-proxy) https://git.io/vA7yC
NixOS_GitHub has left #nixos [#nixos]
<jmob>
We're trying to run buildroot on NixOS, would there be a better way to do that without nix-shell?
<symphorien>
what is buildroot ?
ryanartecona has quit [Quit: ryanartecona]
<jmob>
symphorien: it's a package for building embedded firmware
<jmob>
https://buildroot.org/ -- includes toolchains and packages for building firmware for small embedded devices
<symphorien>
you should use LD_LIBRARY_PATH I am afraid
<symphorien>
you cannot write to $out from a nix-shell
<katona>
jmob: try to run with steam-run package
<symphorien>
the link where you took the ln -s from is only valid in a derivation meant to be build with nix-built
<avn_>
steam-run should be good for it (or some variations like fhsUserEnv)
<jmob>
Using steam run seems kind of heavy handed? I'm already using fhsUserEnv
<tnks>
With Nix 2.0, I don't thinks binary-cache-v3.sqlite is used any more
<tnks>
am I right about that?
<jmob>
But it looks like I can take pieces from steam-run that I need...
<elvishjerricco>
tnks: Yea I just discovered that
<tnks>
elvishjerricco: I moved it, and it never recreated!
<elvishjerricco>
It's in `$HOME/.cache/nix/` now
<MP2E>
hello, does anyone here have an example of how I could use a remote NixOS machine as a binary cache with Nix 2.0 and ssh-ng? Before, I used sshServe on my remote NixOS machine, and simply used 'nix-env -iA package -option ssh-substituter-hosts nix-ssh@example.ddns.net' and it worked as long as I had the SSH key added to my ssh-agent. Now if I do something similar it doesn't seem to work
<jmob>
avn_: symphorien: katona: thanks!
Tobba_ has joined #nixos
<tnks>
elvishjerricco: nice! I was confused about where it was.
<elvishjerricco>
tnks: There's one in `/root/.cache/nix/` too. Not really sure what the difference is.
<elvishjerricco>
I guess it's just per-user, and that's the one for the root user
<tnks>
elvishjerricco: I think it's just a matter of whether you call the tool with root or your user.
<tnks>
yeah.
<elvishjerricco>
So that's probably the one used for, e.g., `sudo nixos-rebuild`
<tnks>
yeah, I'm running Nix in a Docker container and on top of Debian.
<symphorien>
jmob: fwiw nixpkgs has recently better cross compilation. depending on your use case, it might require less efforts to use.
<symphorien>
+gained
<jmob>
symphorien: I have been looking into that, switching to nixpkgs might be an option, but definitely a longer term goal
Tobba has quit [Ping timeout: 256 seconds]
<MP2E>
the main issue is that I don't know how to figure out how to get the binary cache working over SSH because I just get non descriptive error messages
<MP2E>
nix-env is giving me "error: unexpected end of file"
<avn_>
MP2E: nix-serve should works
<avn_>
which simulate hydra' cache over http port (*:5000 by default)
<ottidmes>
MP2E: ssh-substitutor-hosts has been removed in Nix 2.0 you will have to use substitutors ssh(-ng)://example.com
chisui has quit [Ping timeout: 260 seconds]
<catern>
is there anything like https://github.com/sovereign/sovereign for Nix? a single easy one-stop-shop for setting up a personal NixOS server with a bunch of useful services?
<MP2E>
unfortunately, I don't think I can use HTTP because of my ISP. What I've tried so far : I added 'nix.sshServe.protocol = "ssh-ng";' to my configuration.nix. I also have a valid public SSH key listed for my client machine. Then I go to the client, use ssh-add to add the key, and try 'nix-env -iA package -option substituters ssh-ng://nix-ssh@example.ddns.net'
<MP2E>
and it just gives me the same error...
<elvishjerricco>
catern: NixOS has modules for a ton of useful things. It's not like that, where a bunch of cool stuff comes on by default. But it's pretty easy to just write a config with all the parts you want
<ottidmes>
MP2E: I recommend first trying to make it work with ssh
civodul has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
<MP2E>
alright, will do
<ottidmes>
MP2E: Have you succesfully logged in yourself via SSH to that host?
<MP2E>
Yep, actually my irc client is running on the remote machine, and I'm typing through SSH
<MP2E>
works fine
<ottidmes>
MP2E: Have you setup a secret and public binary cache key at secret-key-files and trusted-public-keys (ow and don't forgot to add the public binary cache key of cache.nixos.org, it is there by default, but you overwrite by setting it)
<MP2E>
ah, I have not actually
<elvishjerricco>
There must be a bug in the Nix signature stuff...
<ottidmes>
MP2E: Remember you can generate them with: nix-store --generate-binary-cache-key key-name secret-key-file public-key-file
<elvishjerricco>
I've signed a path in a local `file://` cache, and it verifies fine
<elvishjerricco>
But I put that path on an `s3://` cache, and it no longer verifies
<MP2E>
I will try generating some public cache keys next and see if that helps
<ottidmes>
MP2E: If you have generated a pair with the command I showed, and you have set the options I mentioned in nix.extraOptions, you will have to update the existing nix store to sign what is already there (new additions will be signed automatically). You can do this with: nix sign-paths --all --key-file <secret-key-file>
<MP2E>
thanks! I found that a bit confusing, appreciate the explanation
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] alunduil opened pull request #36359: bump behave to version 1.2.6 (master...bump-behave-version) https://git.io/vA7Hh
NixOS_GitHub has left #nixos [#nixos]
<ottidmes>
MP2E: Took me a while to figure it all out
jrolfs has joined #nixos
<musicmatze[m]>
Running on nix 2.0 now! :-)
mekeor has quit [Remote host closed the connection]
<catern>
ugh
<elvishjerricco>
How do you import a nar file into a `--store file:///path`?
<catern>
how much memory should I expect Nix to use?
<catern>
because, it's using way too much :)
<catern>
can I reduce this somehow
<catern>
I only have 182M of free memory
<ottidmes>
catern: out of?
<catern>
512M
<catern>
I'm running some other services
noam has quit [Ping timeout: 268 seconds]
<elvishjerricco>
catern: I would not expect to run any sort of modern day computer on 512M :P
gammarray has joined #nixos
<catern>
it's a server? :)
<ottidmes>
catern: Would nixops help?
<catern>
ottidmes: why would nixops help?
<elvishjerricco>
catern: yea I would not recommend doing builds on servers directly
<catern>
elvishjerricco: if you don't expect to be able to run on 512M, you're the problem :)
<elvishjerricco>
Do the build elsewhere and push it up
<catern>
I'm not talking about builds
<catern>
catern@caerus:~$ nix-env -i hello
<catern>
Killed
<elvishjerricco>
Well even nix evaluation can be avoided this way if necessary
<ottidmes>
catern: Does the heavy lifting on your dev machine and only deploys, which should be relatively dumb (at least, that is how I understand it works)
<catern>
meh, I suppose I could do that, but I kind of consider that a hack
<catern>
Nix should certainly be able to run in 100MB
<catern>
pretty weak that it can't :)
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 closed pull request #36323: pillow: Fix build on non-NixOS systems (master...adunham/fix-pillow) https://git.io/vAHAv
NixOS_GitHub has left #nixos [#nixos]
<elvishjerricco>
catern: That could be partly nixpkgs. nixpkgs is not very well performance optimzed
<clever>
catern: nix-env -iA nixos.hello
<clever>
!-A
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vA77I
<ottidmes>
catern: I had it run out of GC space, it was evaluating nixpkgs and nixpkgs unstable, at least twice (so 4x nixpkgs) and then a complicated config
<catern>
whaaaaaaaaaat!!!!
<clever>
catern: -i has to eval all of nixpkgs, -iA hello only has to eval hello
<catern>
digital ocean increased their entry-level VPS memory size :)
<catern>
I should reboot my VPS and get my free 512MB :)
<catern>
clever: ah! thanks, yeah! I should have thought of that
<Acou_Bass>
eey folks, I've just rebooted to latest nixos-unstable and plasma is now booting to a black screen and not moving beyond this =/ anyone else having this problem? seems sddm got fixed with its black screen now plasma itself got the same problem :P
<clever>
Acou_Bass: if you do ctrl+alt+f1 can you get a text console?
<Acou_Bass>
yeah
<elvishjerricco>
catern: Holy crap. `nix-env -i hello` takes 741M max. That IS excessive
<clever>
Acou_Bass: then check if X is running
<clever>
elvishjerricco: it has to eval every single attribute in pkgs. and check the .name of each
<catern>
elvishjerricco: well, it's like clever said, that evaluates all of nixpkgs
<catern>
yeah
<elvishjerricco>
`-iA` only uses 36M though
<catern>
I don't really blame nix for that being huge
<catern>
I guess I blame it for being the default, maybe :)
<Acou_Bass>
hmm it seems to be, I've got an xcursor ( set to the theme I've got set in plasma)
<MP2E>
I think that's fair, I don't think I ever use -i without -iA now
<MP2E>
or... uhh... ever
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] volth closed pull request #36147: [WIP] perl-packages.nix updated to actual CPAN versions [staging] (staging...perl-packages-update-3) https://git.io/vA1rS
NixOS_GitHub has left #nixos [#nixos]
<clever>
Acou_Bass: ah, journalctl -u display-manager, any errors near the end?
<ottidmes>
yay, updated my nixpkgs-unstable, finally out of the 2.0pre version: nix (Nix) 2.0
<elvishjerricco>
Anyway, how do I import a nar archive to a non standard store like `--store file:///path`?
<mpickering>
Is there a way to set an override when using -iA? In particular to set "withGrass" to true in the qgis derivation
<clever>
elvishjerricco: do you want it to land in /nix/store or just unpack it anywhere and peek inside?
<clever>
mpickering: nix-env -i -E 'with import <nixpkgs> {}; foo.override { bar = true; }' i believe
<Acou_Bass>
clever: doesn't look like it, session starter and sddm helper exited successfully
<clever>
Acou_Bass: not sure then, i also had problems on the last update, but X refused to even start, amdgpu cant find the card, and simple rollbacks dont help, have to boot the older kernel
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dtzWill opened pull request #36360: nix: fix conditional so brotli is used w/2.0 builds, needed for logs! (master...fix/nix-use-brotli-for-sanity) https://git.io/vA75f
NixOS_GitHub has left #nixos [#nixos]
<elvishjerricco>
clever: Well I WAS trying to debug a complicated signature issue. But the problem seems to have disappeared... So nevermind :P
erasmas has quit [Quit: leaving]
<sonarpulse>
nix build --dry-run with 2.0 fails due to lack of store access
<sonarpulse>
what should I do?
<clever>
elvishjerricco: nix-store --dump <path> prints a nar to stdout, and nix-store --restore <path> reads a nar from stdin
<clever>
Sonarpulse: is the daemon also 2.0?
<sonarpulse>
clever: i assume so?
<sonarpulse>
maybe i need to manually restart
<dtz>
Sonarpulse: run it without --dry-run and then ctrl-c and run again with --dry-run
<clever>
Sonarpulse: how did you install 2.0?
<sonarpulse>
clever: nixos-rebuild
<dtz>
Sonarpulse: damn annoying but it's a known issue
<dtz>
1 sec
<clever>
Sonarpulse: then it should have restarted itself
<Acou_Bass>
strange, I'll roll back for now see if it helps
lopsided98 has quit [Remote host closed the connection]
<clever>
mpickering: lib.optional returns a list in either case
lopsided98 has joined #nixos
<sonarpulse>
dtz: is there like a --vverbose to make `nix build` spit out more things?
<sonarpulse>
i miss the noise
<LnL>
-vvvvvv
<LnL>
:p
<abathur>
if I'm using nix-shell to work on building a MUD server (ldmud), which has optional build/runtime dependencies like mysql/postgres: am I on-my-own when it comes to starting the database service after loading the shell (manually or in shellHook)?
<Acou_Bass>
clever: FWIW i seem to have some sort of btrfs corruption in my home folder that may be causing the problem
<Acou_Bass>
excellent eh
<clever>
Acou_Bass: ive got zfs with daily snapshots and mirroring, so i would expect things to be recoverable
<clever>
gammarray: on my laptop (not a macbook), nothing ive done has been able to disable tap to click, and every now and then i find a new un-documented feature
<clever>
tapping with 2 fingers is right click
<abathur>
heh
<clever>
you cant tripple click, because the touchpad firmware thinks your trying to so a single click&drag
<clever>
single click has a decent amount of latency, so it always messes up if i try to control+tap
<gchristensen>
whoa, mine is too ...
<clever>
i let go of control too fast,because the dumb firmware is checking for a double-tap
<clever>
and its hard to simply scroll down, because 2 finger drag is a 2-axis mouse "wheel"
<gammarray>
@clever are you using synaptics?
<clever>
gammarray: i tried both synaptics and libinput, neither could disable tap to click
<clever>
the laptop also has a totally unmarked hotkey, fn+1 just maxes out the cpu fan