<clever>
Myrl-saki: i havent found the code yet, but i believe it goes thru a buildEnv with ignoreCollisions = true;, so the new i915.ko will just override the old
<eacamero_>
Is there a common way to chmod a derivation to a certain user? I'm cloning a PHP app and need to move things around and then chmod it to a certain user. But everything is in nix/store
<eacamero_>
chown is what I mean
<clever>
eacamero_: when the build is done, nix-daemon will chown everything to 0:0
<clever>
so no chown stuff you do will stick
<eacamero_>
clever: I kinda figured.
eacamero_ is now known as eacameron
<eacameron>
clever: sooo...Everything has to run as root?
Fare has quit [(Ping timeout: 245 seconds)]
<Myrl-saki>
clever: Okay, thanks.
<clever>
eacameron: and /nix/store is mounted read-only, so you cant write to it even with root
<clever>
eacameron: once something has finished building, it can never be modified
<eacameron>
clever: Ah right...which means I'm even more screwed.
<clever>
eacameron: you need to patch it to write to somewhere else, or put a symlink into the store
<clever>
so foo/upload is a symlink to /var/lib/service-name
<clever>
so when it tries to write to the upload directory, it winds up in /var/lib/service-name
zraexy has joined #nixos
<eacameron>
clever: Ah that's helpful. But I'm not even there yet. I want the PHP files themselves to be owned by the same user running phpfpm
<clever>
eacameron: this generates a config.php, and inserts its path into a copy of index.php, along with the path to the core source, and makes some symlinks for static files
Fare has joined #nixos
<eacameron>
clever: Awesome. I think I get the idea. But this is pointing to aedb folder somewhere...which is presumably writable?
<clever>
nope, aedb is just a clone of the source from github
<clever>
which has the core of the code that makes everything work
<eacameron>
clever: but $out won't be writable either.
<eacameron>
So nothing is writeable?
<NixOS_GitHub>
[nixpkgs] grahamc opened pull request #20876: ecryptfs test: use TTY output to stabilize test (master...ecryptfs) https://git.io/v1Wvi
<clever>
yeah, none of this is writeable, but config.php could contain paths to the writeable folders
<clever>
and this public_html derivation is faster to rebuild when you change that, then the aedb derivation
<gchristensen>
clever: ^ vcs1 is paying off
<clever>
gchristensen: vcs1 all the tests!!
<gchristensen>
:)
<eacameron>
clever: Ah ok. In my case I'd symlink to /var/craftcms/site or something.
herzmeister has quit [(Quit: Leaving)]
herzmeister has joined #nixos
<gchristensen>
clever: your kexec script and the way nixos tests work are two reasons why I <3 <3 <3 nixos, it is really very elegant
<eacameron>
Now I'm curious: how do nixos tests work?
<gchristensen>
they start a VM using Qemu
<gchristensen>
and share the appropriate /nix/store paths
markus1199 has joined #nixos
<gchristensen>
check out nixpkgs/nixos/tests/
<eacameron>
Oh that's neat.
<eacameron>
Actually that's sick.
<gchristensen>
yup
<gchristensen>
they are effectively any other derivation
<eacameron>
So the tests are run when building the derivation to various os components?
<clever>
the tests are a special derivation, that depends on the package
<gchristensen>
no, they are run by hydra
<clever>
and will "compile" the test results by running scripts against qemu
<clever>
and if all of the inputs are unchanged, it will reuse the old test results
markus1189 has quit [(Ping timeout: 246 seconds)]
Shou has joined #nixos
systemfault has joined #nixos
<eacameron>
clever: I see so your os doesn't itself depend on the tests...but hydra "does"
<clever>
eacameron: yeah, the release's of a channel depend on every nixos test passing
<clever>
so nixos-unstable wont update until all tests pass
<eacameron>
That reminds me: Is there a turnkey solution for a hydra server? say with nixops?
<eacameron>
gchristensen: Awesome. One of these days I'd like to have a nixops config that you can just...run and viola have a hydra server on your network.
<gchristensen>
same
<eacameron>
gchristensen: Maybe this is it.
rardiol has quit [(Remote host closed the connection)]
<gchristensen>
offlinehacker: ping
proteusguy has quit [(Ping timeout: 244 seconds)]
agjacome has quit [(Quit: leaving)]
sigmundv_ has quit [(Ping timeout: 245 seconds)]
redmq has quit [(Ping timeout: 265 seconds)]
wak-work has quit [(Ping timeout: 260 seconds)]
<eacameron>
gchristensen: clever: So I can mkdir -p /var/.../ folders and then chown them to someone inside of a derivation?
<NixOS_GitHub>
[nixpkgs] romildo opened pull request #20877: adapta-backgrounds: init at 0.4.0.6 (master...new.adapta-backgrounds) https://git.io/v1WJh
page_ has joined #nixos
page has quit [(Ping timeout: 252 seconds)]
<clever>
eacameron: nope, the derivation only has access to $out when building
<clever>
eacameron: but you can then make a systemd unit that creates the folders before the php services start
Fare has quit [(Ping timeout: 268 seconds)]
<eacameron>
clever: oh my...
<eacameron>
clever: how does symlinking work then...if the folders don't exist yet?
<clever>
symlinks can point to things that dont exist
newhoggy2 has quit [(Remote host closed the connection)]
<clever>
a symlink is basicaly just a string, with a special type on it
<clever>
and when you try to open it with a normal syscall (cd, open, stat) it will treat that string as a path relative to the symlink
<clever>
so the directory has to exist only when the php code is being ran
newhoggy2 has joined #nixos
wak-work has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
eacamero_ has joined #nixos
Wizek has quit [(Ping timeout: 250 seconds)]
Fare has joined #nixos
tv has joined #nixos
tv has quit [(Changing host)]
tv has joined #nixos
eacameron has joined #nixos
eacamero_ has quit [(Ping timeout: 250 seconds)]
<eacameron>
clever: hmm...I'm not sure I like it, but I'll do it.
zraexy has quit [(Ping timeout: 256 seconds)]
angerman has joined #nixos
saintromuald has joined #nixos
Shou has quit [(Ping timeout: 246 seconds)]
Shou has joined #nixos
Seichi has joined #nixos
ThatDocsLady has quit [(Ping timeout: 260 seconds)]
georges-duperon has quit [(Ping timeout: 260 seconds)]
tv has quit [(Quit: WeeChat 1.5)]
tv has joined #nixos
angerman has quit [(Ping timeout: 244 seconds)]
tv has quit [(Quit: WeeChat 1.5)]
newhoggy2 has quit [(Remote host closed the connection)]
Fare has quit [(Ping timeout: 248 seconds)]
Seichi has quit [(Quit: Konversation terminated!)]
tv has joined #nixos
tv has quit [(Changing host)]
tv has joined #nixos
glines has quit [(Ping timeout: 258 seconds)]
johnsonav has joined #nixos
proteusguy has joined #nixos
<NixOS_GitHub>
[nixpkgs] aszlig pushed 2 new commits to master: https://git.io/v1WtI
<NixOS_GitHub>
nixpkgs/master 6343fef aszlig: python/pillow: Only disable failing tests...
<NixOS_GitHub>
nixpkgs/master a76f524 aszlig: vimiv: init at 0.7.2...
paperd has joined #nixos
mbrgm has quit [(Ping timeout: 260 seconds)]
tv has quit [(Quit: WeeChat 1.5)]
mbrgm has joined #nixos
tv has joined #nixos
tv has quit [(Changing host)]
tv has joined #nixos
tv has quit [(Quit: WeeChat 1.5)]
herzmeister has quit [(Quit: Leaving)]
herzmeister has joined #nixos
lverns has joined #nixos
tv has joined #nixos
hexagoxel has quit [(Ping timeout: 258 seconds)]
NEo4v8_ has joined #nixos
hexagoxel has joined #nixos
eacameron has quit [(Remote host closed the connection)]
paperd has quit [(Quit: WeeChat 1.0.1)]
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 268 seconds)]
eacameron has joined #nixos
jgrasser has joined #nixos
Wizek has joined #nixos
eacameron has quit [(Remote host closed the connection)]
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 245 seconds)]
PragCyphr has joined #nixos
PragCyph1 has quit [(Ping timeout: 256 seconds)]
mguentner has quit [(Quit: WeeChat 1.6)]
magnetophon has quit [(Ping timeout: 260 seconds)]
glines has joined #nixos
mizu_no_oto has joined #nixos
mguentner has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
newhoggy2 has joined #nixos
mguentner2 has joined #nixos
glines has quit [(Ping timeout: 260 seconds)]
mguentner has quit [(Ping timeout: 245 seconds)]
newhoggy2 has quit [(Ping timeout: 248 seconds)]
ericsagnes has quit [(Ping timeout: 245 seconds)]
xadi has quit [(Ping timeout: 258 seconds)]
Shou has quit [(Ping timeout: 250 seconds)]
<NixOS_GitHub>
[nixpkgs] aszlig pushed 1 new commit to master: https://git.io/v1WOU
<NixOS_GitHub>
nixpkgs/master 1f9cbcd aszlig: beets: Set Python version via top-level attribute...
xadi has joined #nixos
det has quit [(Remote host closed the connection)]
<Rotaerk>
why is "-e" uninstall? does it stand for "evict"?
<c74d>
maybe "erase"
<c74d>
or "eliminate"
eacameron has joined #nixos
AllanEspinosa has quit [(Ping timeout: 260 seconds)]
justanotheruser has joined #nixos
justan0theruser has quit [(Ping timeout: 256 seconds)]
<cfricke>
Ahoy! I am using a simple `with import <nixpkgs> {}; (python35.withPackages (ps: [ps.ipython ...])).env` to instantiate a python dev env. I would like to override the prompt inside the script, but fail to do so.
latk has joined #nixos
mathias_ has left #nixos ["WeeChat 1.5"]
NEo4v8_ has quit [(Ping timeout: 240 seconds)]
<cfricke>
As in, can I place a shellHook outside a derivation?
ebzzry has quit [(Ping timeout: 245 seconds)]
ebzzry has joined #nixos
Wizek has quit [(Ping timeout: 260 seconds)]
Itkovian has joined #nixos
newhoggy2 has joined #nixos
edvorg has quit [(Remote host closed the connection)]
jensens has joined #nixos
newhoggy2 has quit [(Ping timeout: 260 seconds)]
<cfricke>
Using `buildEnv.override` with the postBuild hook with an 'export PS1' command, still nothing. Hm.
RchrdB has joined #nixos
Hunner has quit [(Ping timeout: 240 seconds)]
ixxie has quit [(Ping timeout: 260 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
RchrdB has quit [(Ping timeout: 250 seconds)]
alx741_ has joined #nixos
Itkovian has joined #nixos
markus1199 has quit [(Ping timeout: 250 seconds)]
alx741 has quit [(Ping timeout: 258 seconds)]
markus1189 has joined #nixos
lverns has joined #nixos
<BlessJah>
1
iyzsong has quit [(Quit: night.)]
Shou has joined #nixos
<NixOS_GitHub>
[nixpkgs] grahamc closed pull request #20876: ecryptfs test: use TTY output to stabilize test (master...ecryptfs) https://git.io/v1Wvi
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<NixOS_GitHub>
[nixpkgs] grahamc pushed 1 new commit to release-16.09: https://git.io/v1WgR
<NixOS_GitHub>
nixpkgs/release-16.09 b5b1959 Graham Christensen: ecryptfs test: use TTY output to stabilize test...
Itkovian has joined #nixos
<gchristensen>
domenkozar: many tests on release failed due to timing out when accessing the VM (error: timed out waiting for the VM to connect) can you restart them?
newhoggy2 has joined #nixos
<obadz>
!m gchristensen !!!!
<[0__0]>
You're doing good work, gchristensen !!!!!
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/v1Wg6
<NixOS_GitHub>
nixpkgs/master 4162623 Frederik Rietdijk: pythonPackages.cython: 0.24.1 -> 0.25.1
<gchristensen>
:]
<obadz>
very much appreciating the ecryptfs test patch
<obadz>
do you use ecryptfs btw?
systemfault has joined #nixos
systemfault has quit [(Remote host closed the connection)]
<gchristensen>
I don't, clever gave me the idea to use `/dev/vcs1` to read the terminal though, which lets us have way more reliable tests :)
systemfault has joined #nixos
edvorg has joined #nixos
<obadz>
!m clever !!!!
<[0__0]>
You're doing good work, clever !!!!!
<obadz>
well thanks for cleaning up my mess, I was never happy with the sleeps in my test
ebzzry has quit [(Ping timeout: 248 seconds)]
xadi has quit [(Quit: Leaving.)]
xadi has joined #nixos
xadi has quit [(Client Quit)]
latk has quit [(Ping timeout: 250 seconds)]
ixxie has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<NixOS_GitHub>
[nixpkgs] tomberek opened pull request #20884: PostgREST: 0.3.2 -> HEAD (master...patch-4) https://git.io/v1Was
xwvvvvwx has quit [(Quit: Page closed)]
<FRidh>
is there a function to split a store path string?
danharaj has quit [(Quit: Leaving.)]
alx741 has joined #nixos
Itkovian has joined #nixos
seppellll has quit [(Ping timeout: 244 seconds)]
mizu_no_oto has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 5 new commits to release-16.09: https://git.io/v1WVr
<NixOS_GitHub>
nixpkgs/release-16.09 ce812ca Joachim Fasting: Revert "grsecurity: work around for #20490"...
<NixOS_GitHub>
nixpkgs/release-16.09 88c8ffd Joachim Fasting: linux_grsec_nixos: use the "modinst arg list too long" patch...
<NixOS_GitHub>
nixpkgs/release-16.09 368ebd6 Joachim Fasting: grsecurity: 4.8.10-201611210813 -> 201611232213...
newhoggy2 has quit [(Ping timeout: 260 seconds)]
phanimah1sh has quit [(Quit: leaving)]
RchrdB has joined #nixos
kampfschlaefer has quit [(Quit: No Ping reply in 180 seconds.)]
kampfschlaefer has joined #nixos
guanjian[m] has joined #nixos
newhoggy2 has joined #nixos
jgertm__ has joined #nixos
pstn has quit [(Ping timeout: 250 seconds)]
newhoggy2 has quit [(Ping timeout: 250 seconds)]
systemfault has quit [(Quit: Bye!)]
ryantm____ has quit [(Quit: Connection closed for inactivity)]
Itkovian_ has joined #nixos
Itkovian has quit [(Ping timeout: 256 seconds)]
<tnks>
so right now I'm on a fresh-enough version of nixpkgs, and everything is cached on Hydra.
<tnks>
but as my project goes into maintenance, I think the old builds will fall off the cached.
<tnks>
so I think people would recommend I run a company-internal Hydra?
<FRidh>
tnks: you can track a channel (e.g. nixos-16.09 or nixpkgs-unstable) and use the cache, or build indeed yourself
<tnks>
that seems kind of needless, because the Nix stores of my build agents is really all I think I need to save away.
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 5 new commits to master: https://git.io/v1WoV
<NixOS_GitHub>
nixpkgs/master 851b594 Thomas Tuegel: qt5: factor out common setup hook for qtbase
<NixOS_GitHub>
nixpkgs/master 53a694e Thomas Tuegel: qt5: factor out common qtSubmodule setup hook
<NixOS_GitHub>
nixpkgs/master 2c4995c Thomas Tuegel: qt5: factor out common qmake setup hook
<tnks>
for our JVM builds, we've purchased Artifactory, so we're using that as our proxy cache, so we're not too reliant on the internet for dependencies.
<tnks>
and we can let our local build caches get deleted if convenient/necessary.
<tnks>
I could set up a NixOS server and just push closures to it after building?
ambro718 has quit [(Quit: Konversation terminated!)]
<tnks>
right now, I'm building in Docker with Jenkins, because that's allowed me to slip in Nix without anyone in Ops questioning.
<tnks>
so if I tell them to run Hydra, they may start asking annoying questions.
xwvvvvwx has joined #nixos
<xwvvvvwx>
Hey :)
<xwvvvvwx>
So I just installed nixos today
<xwvvvvwx>
and I **love** it
<Ralith>
tnks: if you expect to often be spinning up new dev envs without updating to recent packages, hosting an internal binary cache may be helpful, yeah
<xwvvvvwx>
ok so if I want the latest versions then I should switch to unstable?
<tnks>
FRidh: I just read this: "Actually, there is currently *no* garbage collection on cache.nixos.org at all"
<FRidh>
tnks: or the nix.sshServe.enable option
<tnks>
is that still true?
<gchristensen>
xwvvvvwx: you should use fetchTarball
<FRidh>
tnks: I don't know. Items are now uploaded to s3.
<xwvvvvwx>
ok
<xwvvvvwx>
thanks for all the help! :)
<tnks>
FRidh: okay, thanks for the pointers, I'll probably think about this later.
<tnks>
if the retention policy is to retain everything forever, then I might not have to worry about this too much.
<tnks>
I just thought I saw some stuff building from scratch.
<tnks>
but I didn't experient too much.
xwvvvvwx has quit [(Ping timeout: 260 seconds)]
eju has quit [(Quit: WeeChat 1.6)]
pstn has joined #nixos
edvorg has quit [(Remote host closed the connection)]
pstn has quit [(Ping timeout: 268 seconds)]
<mguentner>
Is there an easy way to inspect a closure using something like a (graphical) dependency graph?
gilligan_ has quit [(Disconnected by services)]
gilligan_ has joined #nixos
mizu_no_oto has quit [(Quit: ["Textual IRC Client: www.textualapp.com"])]
<LnL>
tnks: you can also use nix-push to generate files that can be served by any http server as a cache
<LnL>
mguentner: nix-store -q --graph
xwvvvvwx has joined #nixos
<LnL>
that gives you graphviz to render
Itkovian_ has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<LnL>
it gives output that graphviz can render
<mguentner>
LnL: very nice. thx
Wizek has joined #nixos
<mguentner>
LnL: dotty is freaking out over the dependency graph of qtbase :>
<mguentner>
maybe I should too
<LnL>
there's also --tree if you want a textual representation
Fare has joined #nixos
<Rotaerk>
hmm, I like /homeless-shelter... cute
<NixOS_GitHub>
[nixpkgs] grahamc pushed 1 new commit to master: https://git.io/v1Wio
<NixOS_GitHub>
nixpkgs/master 3e61793 Graham Christensen: Merge branch 'staging'
<Fare>
hi. Somehow, upgrading nixos, I find that the kernel can't find the block device anymore. The configuration worked before I updated the channels.
<NixOS_GitHub>
[nixpkgs] grahamc merged master into staging: https://git.io/v1WiP
<mguentner>
LnL: thanks!
<NixOS_GitHub>
[nixpkgs] grahamc pushed 2 new commits to staging: https://git.io/v1Wi1
<NixOS_GitHub>
nixpkgs/staging cdafc8e Graham Christensen: Merge pull request #20882 from lsix/update_libunistring...
jgertm__ has quit [(Quit: WeeChat 1.6)]
slack1256 has joined #nixos
NeQuissimus has quit [(Ping timeout: 260 seconds)]
NeQuissimus has joined #nixos
JonReed has joined #nixos
<JonReed>
How to add a python package? Here is what I don't get. In nixpkgs repo I search some packages like requests and results come up. I search some packages like "pyquery" and nothing comes up, despite that `pyquery` package exists. I would assume that all packages from pypy are added automatically, but this does not seem to be the case as I don't see x
<JonReed>
lsxwriter there.
<JonReed>
Anybody has any links to examples of how python packages are added?
<JonReed>
which exists in pypy already
gilligan_ has quit [(Quit: Leaving)]
pstn has joined #nixos
<JonReed>
Ah, nevermind, I had temporary amnesia. I think I've added packages in the past, now I remember. It's still puzzling why some packages are not seen on nixpkgs.
<JonReed>
Is there some mechanism where some of them are added automatically and some not?
<gchristensen>
we don't auto-package anything in pypi.
<LnL>
yeah, I'm pretty sure we don't have that yet
<JonReed>
It does not come up when I search nixpkgs, but I can install it in nixos
<LnL>
it's hard to automate for python packages because they don't have static metadata
<JonReed>
What do you mean by static metadata?
<JonReed>
I assume haskell packages are added automatically?
<JonReed>
What type of metadata is present in hackage than is not present in pypy
<JonReed>
As for pyquery, it seems that github has a problem showing all search results. The string "pyquery" comes up in python-packages.nix, but github is not showing it.
<LnL>
dependencies, you have to evaluate setup.py to get the dependencies of a package
<LnL>
and setup.py is just python code so that can have dependencies of it's own
<JonReed>
Yeah, that sucks
xwvvvvwx has quit [(Ping timeout: 260 seconds)]
<LnL>
a lot of pacakges on hackage even use the extra-libraries field correctly
<LnL>
so we can even detect system dependencies correctly in most cases
xadi has joined #nixos
<JonReed>
I can imagine that through some magic you can deduce dependencies even when they are not specified fully. If a package has tests, then you can fail, parse error message, see which module is missing, try to find the package that contains that module, etc. But this sounds like a chore.
<JonReed>
I'm sure tests are even not needed for this, as it will fail just if you try to import any module from the package
<JonReed>
But then dependencies will have to be all known, because otherwise we would know which module name belongs to which package
<JonReed>
Or we would have to parse whole pypy to index modules
<JonReed>
and link them to packages
<JonReed>
This does not sound like a fun adventure though
<NixOS_GitHub>
[nixpkgs] aristidb pushed 1 new commit to master: https://git.io/v1WXx
<aristid>
everybody needs to make a special snowflake nowadays
<Ralith>
at least cargo plays pretty well with nixpkgs
<Ralith>
unlike, say, npm
<aristid>
Ralith: so maybe i'm missing something because i'm not a rust expert
<aristid>
but... i'm trying to package rq
<aristid>
and it fails because rq specifies some of its dependencies with git references instead of registry
<aristid>
and then it fails to do network access during build
<aristid>
because it's nix
<aristid>
Ralith: is there a way around that?
systemfault has joined #nixos
<Ralith>
aristid: git references in a release are bad practice on rq's part; you could easily hack around it by replacing them in Cargo.toml with a path dependency and then getting the paths with fetchgit or similar
jensens has joined #nixos
<aristid>
Ralith: is there an example of something in nixpkgs doing that?
<aristid>
then i could copy that :D
<Ralith>
not that I'm aware of
<aristid>
nor can i :(
lverns has quit [(Ping timeout: 260 seconds)]
arjen-jo1 has joined #nixos
<Ralith>
do you know how to use fetchgit or similar to download the git deps?
<Ralith>
because all you need to do is do that, then use sed or something to replace each dep in Cargo.toml with something like { path = "${thing}" }
<Ralith>
where thing is the matching dep
<Ralith>
of course, the proper fix would be to nag upstream into using stable releases of their dependencies
<aristid>
Ralith: i know how to use fetchgit, but i don't know how to correctly change Cargo.toml
<Ralith>
but this workaround isn't complicated
<aristid>
how can i add a sed phase into buildRustPackage?
<aristid>
hmm, i could try cargoUpdateHook
<Ralith>
aristid: is there no patch phase?
<aristid>
not sure
<Ralith>
I'd try that and see if it works
<Ralith>
if there's a hook explicitly for before cargo is ran that works too though
<aristid>
there is postUnpack
<aristid>
Ralith: do i need to adjust both Cargo.lock and Cargo.toml?
<aristid>
what is the difference, even?
<Ralith>
postUnpack should do fine, I think
<Ralith>
if you supply the dependencies correctly I do not think you need to touch Cargo.lock
<Ralith>
Cargo.lock is where cargo stores the output of dependency solving, so it contains the exact version numbers and checksums of every single rust package in the transitive closure
<Ralith>
if you're replacing git dependencies with path dependencies that point to the exact same git checkout, it should be the same
<Ralith>
that might be a little cleaner/less fragile
<Ralith>
effect should be equivalent though so w/e
seanz has joined #nixos
<aristid>
Ralith: postUnpackPhase is too late, as it fails in fetchDeps
<aristid>
i think i _do_ need to use cargoUpdateHook
<savanni>
So... just did a clean installation of nixos on a bare disk!
<savanni>
Took a while. Had to backtrack some. But I have a list of commands that I ran that include setting up an encrypted volume.
<aristid>
savanni: yay :)
<savanni>
Now it's just a matter of getting up enough of the system that I can attach my external drive (whole-disk encryption) and copy the bits and pieces of config into place and then restore the backup and then merge my two systems!
<savanni>
Soon to be a lot of "hurry up and wait", really.
rly has quit [(Ping timeout: 240 seconds)]
seanz has quit [(Quit: Leaving.)]
amarsman has joined #nixos
msd has joined #nixos
arjen-jo1 has quit [(Ping timeout: 246 seconds)]
<aristid>
Ralith: hmm, this is starting to exceed the patience i have for a program i just wanted to try out of curiosity :(
<aristid>
Ralith: maybe it's better to replace in-place than to use '[replace]'
saintromuald has quit [(Ping timeout: 244 seconds)]
<Ralith>
aristid: yeah, if cargo checks the original source first then that seems like the best option
<LnL>
msd: looks good, I'll merge it in when my build finishes
<msd>
Thanks LnL!
<aristid>
if somebody ever makes a tarball with base32 hash 0000........., i'm so screwed :D
<Fare>
What could cause nixos's kernel to now fail to find the root device?
<Fare>
not sure whether it's before or after initrd. Maybe before. I don't see much difference in principle in the grub entries before and after it stopped working.
<LnL>
msd: where did you get the current version from? I wonder why they recommend an older one
<msd>
LnL: got the current version from their GitHub repo. They--for some reason--recommend R02basho8 on their Riak/Riak CS/Stanchion pages
seanz has joined #nixos
seanz has quit [(Client Quit)]
<LnL>
hmm, I'm wondering if the documentation is just out of date
seanz has joined #nixos
<msd>
LnL: maybe; but on their GitHub page, they specifically recommend (here's where it gets stranger) R02basho10
<msd>
LnL: I'd just suggest to stick with the version documentation states for now
<msd>
LnL: and this way, if we really *do* need to upgrade it, we can just change the number from 8 to 10 and the sha256 version
<aristid>
Ralith: still not working, but i feel like i'm close :D *removes git references from .lock file*
<Ralith>
aristid: may be easier to generate a fresh lockfile with cargo than to correct it by hand
<Ralith>
not sure
eacamero_ has joined #nixos
<aristid>
Ralith: well, i think i just need to remove the reference, because it's now a "local" path
<aristid>
i mean remove the part in ()
<Ralith>
maybe!
eacameron has quit [(Read error: Connection reset by peer)]
<aristid>
Ralith: oh, actually nix already patchifies the lock file
agjacome has quit [(Read error: Connection reset by peer)]
agjacome has joined #nixos
<aristid>
Ralith: ok, i did phrase that needlessly inflammatorily
<aristid>
i assume they do it to avoid the equivalent of DLL hell where nix is not available
<aristid>
BUT we do have nix!
<Ralith>
most people don't
hiratara has quit [(Ping timeout: 258 seconds)]
<aristid>
chances are, cargo supports doing it in a nixier way
<aristid>
but we haven't implemented it yet
<aristid>
Ralith: even if system-installed dependencies are not the default, there should (maybe is?) a way to force them
heyho has joined #nixos
<Ralith>
it's certainly possible in principle to compile libraries of things, but you'll always end up having bits of most libraries compiled into whatever binary anyway, because that's how generics get compiled
hiratara has joined #nixos
<aristid>
sure, but that's the same in c++ and i'm happy that we have compiled libraries there :)
<aristid>
(i actually didn't know how rust generics work until just now :D)
<gchristensen>
it is important to rust's adoption that it cary with it all its deps, I think
<Ralith>
I imagine it would be possible to rewrite the nix infrastructure to work in terms of libs, but it might require ungodly amounts of autogenerated expressions until we get proper import-from-derivation or whatever
<heyho>
hallo ich habe folgendes problem
<heyho>
Hello, I have the following problem when i run nixos-rebuild the sha256 hash from teamviewer is different from what was expected.am i doing something wrong?
<aristid>
Ralith: i think i'd prefer that. i've just spent half an hour or so waiting for rq to get to the stage where it builds its v8-rs dependency module, just to find that I made a typo while specifying the buildInput for the native system v8 library. and that's not the first such problem
<aristid>
heyho: it could happen that teamviewer changed its binaries without bumping the version number
<Ralith>
it sounds like rq is an exceptionally heavyweight project I guess
<aristid>
Ralith: is it really?
<Ralith>
certainly I've never waited half an hour for a build
<Ralith>
you might want to do your testing/debugging in a nix-shell of the package so you don't have to start again from scratch
<heyho>
any suggestions to fix itt?
<aristid>
Ralith: i don't actually want to learn rust, i just want to try out rq
<aristid>
and i'm really stubborn it seems :D
<aristid>
heyho: well, yeah, do you have a nixpkgs checkout already?
<Ralith>
yes, I was talking about testing/debugging of the nix package
<aristid>
Ralith: also it always feels like i'm almost done anyways :D
<Ralith>
yay!
<aristid>
lol, now it must be using the wrong version of v8 or sth
<aristid>
heyho: then, find a directory "teamviewer" or so
<aristid>
heyho: in there, there should be a default.nix. you will find a line sha256 = "..."
<heyho>
ok thanks will try it
<aristid>
heyho: change it to the sha256 from the error message you got earlier
<aristid>
heyho: then, do nix-env -f . -iA pkgs.teamviewer
<aristid>
from the root of the checkout
<aristid>
heyho: however, i stopped using teamviewer myself because i don't want to give a closed-source component root access to my computer.
<clever>
aristid: also, after the latest update, the connect button just vanished
<clever>
so it doesnt do anything
<heyho>
aristid do you have a open source alternative?
<clever>
heyho: ssh or vnc are my main choices
goibhniu has joined #nixos
<heyho>
clever for both i need port forwarding. or do i misunderstand something?
<clever>
for vnc, the forwarding can go at either the client or server, depending on your needs
<clever>
for ssh, it has to be at the server
<clever>
there are also vpn options to bypass the port forwarding
<heyho>
thanks you
pstn has joined #nixos
<heyho>
one more question, i get "argument list too long" when i try to start qutebrowser. cant find a wayto fix it
vikstrous has quit [(Quit: -)]
<clever>
heyho: it sounds like makeWrapper was ran on makeWrapper, so its an infinite loop
Maxdamantus has quit [(Ping timeout: 260 seconds)]
<aristid>
heyho: just to make sure you don't think nixos is something it isn't: nixos is a fairly niche distribution still, so if you have any kind of "unusual" requirements, it will commonly require you to fix them directly in the nixpkgs git repo
Maxdamantus has joined #nixos
msd has quit [(Remote host closed the connection)]
<aristid>
Ralith: uuh this is getting ridiculous: v8-rs requires a newer version of v8 than nix has (although both are in the 5.4 release series it seems?!)
<Ralith>
geez
<Ralith>
you could try an older version of v8-rs
<Ralith>
compiling a new v8 doesn't sound like fun
<Ralith>
at least if it's anything like the rest of the chromium project
pstn has quit [(Remote host closed the connection)]
<Ralith>
you should say "ffs switch to stable dependencies" to the upstream maintainer ^^
<aristid>
Ralith: well, it's v8 5.4.232 vs 5.4.whatevs that changes the public interface in v8.h
<aristid>
v8 is at 5.7 now
<aristid>
so they.. changed the interface in a stable release or something
hiratara has joined #nixos
<aristid>
Ralith: i think i'm just going to mark the file as broken and commit the package to nixpkgs anyways, because most of the things i did _will_ be useful to whoever wants to package rq and has more patience than I am
<Ralith>
IANAcommitter but that seems reasonable
<NixOS_GitHub>
[nixpkgs] aristidb pushed 1 new commit to master: https://git.io/v1WQx
<NixOS_GitHub>
nixpkgs/master e3dcf5d Aristid Breitkreuz: rq: init at 0.9.2 (broken because our v8 is too old, and I'm too weak to update v8)
<aristid>
pity.
mizu_no_oto has joined #nixos
jgertm_ has quit [(Ping timeout: 246 seconds)]
mizu_no_oto has quit [(Client Quit)]
thc202 has quit [(Quit: thc202)]
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
glines has joined #nixos
<viric>
clever: I have the pi2 rebuilding nixos to match release-16.09 (I had a dumb change in *coreutils*)
<viric>
it takes days.
<clever>
yeah, i need to get the rpi3 up to double my speed
<viric>
didn't you put it at work yet?
<viric>
why would it double the speed? clock? instruction set? cores?
<clever>
an rpi2 + rpi3
<viric>
Haha.
<clever>
2 slaves
<clever>
but i cant get the kernel from nixos to boot on the rpi
<viric>
(I had old dtbs and my machine didn't boot)
<viric>
and yes - they had the same names
<clever>
ah
<clever>
i'm using the dtb from the firmware repo, let me check the kernel build
<viric>
the kernel store path should have them
<viric>
- yes, it's a shame nixos does not copy them. :)
<viric>
I could have saved me enough troubles by adding that 'cp'
<clever>
i thought that the entire point of DT, is that the kernel can take a DT in and configure itself correctly
<viric>
clever: the DTs are per-kernel. DTBs are to make a single kernel binary to boot in different boards
<viric>
what is in the DT and what not varies from version to version of linux
<Dezgeg>
no, DTs are _supposed_ to be independent of the kernel version, it's just that apparently the raspberry people failed to adhere to that
<viric>
aaaah
<viric>
:)
<viric>
let's see if that helps clever
<Dezgeg>
in the mainline world you do get yelled at if you try to break old dtbs (except if you're an important person enough)
<viric>
ok
<clever>
heh
<viric>
I should raise my expectations.
<clever>
i read the source code, and the area that is failing appears to get the GPIO address by asking the firmware over the mailbox
<clever>
which sort of defeats the entire point of DT
<clever>
and shouldnt be capable of failing this badly
mizu_no_oto has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<kier>
i'd be happy to donate my pi to a community build farm, if only I could get nixos running on it in the first place :p
<clever>
viric: yep, i found the dtbs, let me copy them
<gchristensen>
kier: we typically don't trust hardware not hosted centrally, but a non-official project is possible
<kier>
fair enough
<Dezgeg>
that brcmvirt_gpio stuff is indeed interesting... it looks like the raw gpio access is actually done in firmware and there's just a shared memory channel to do that
<Dezgeg>
so I guess the firmware version also needs to match
<clever>
viric: and nope, the DTB inside the linux derivation is an exact match for the DTB from the firmware repo
Sonarpulse has joined #nixos
<clever>
Dezgeg: any other ideas?
<Dezgeg>
does the firmware release match the kernel?
<viric>
ok
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<clever>
Dezgeg: it doesnt say which version it is, let me try booting it
<Dezgeg>
strings start.elf | grep 2016 seems to "work"
<clever>
VC_BUILD_ID_TIME: Nov 19 2016
<clever>
not really a kernel version though
<Dezgeg>
the tagged releases in the repo have timestamps
<clever>
but git log does reveal kernel: Bump to 4.4.27
<Dezgeg>
i.e. 1.20160620-1
<clever>
Linux (none) 4.4.33-v7+ #928 SMP Sat Nov 19 19:39:23 GMT 2016 armv7l GNU/Linux
<clever>
i booted kernel7.img and its not what git log said it was
<clever>
it is also not matching the source you mentioned
<Dezgeg>
interesting
<clever>
let me double-check its the right kernel7
<clever>
it is
<clever>
bootcode: Adds support for Pi 3 LEDs that get controlled through a gpio expander
<clever>
thats strange, and possibly related
<NixOS_GitHub>
[nixpkgs] zimbatm closed pull request #19582: Fetch one of (master...fetchOneOf) https://git.io/vP6ar
Maxdamantus has quit [(Ping timeout: 248 seconds)]
<eqyiel[m]>
apart from grepping through nixpkgs, is there a way to find out what package or module controls a file? I want to modify /etc/pam.d/passwd but I'm having a hard time finding the nix expression that writes to it