<pie_>
i was running nix-shell on the directory not the default.nix file
<pie_>
and i have a shell.nix in there
<pie_>
so it was running the shell.nix instead
<samueldr>
nix-shell with a file is like when it has a shell.nix, it's using that derivation as the input
<pie_>
wait thats still weird though
<dminuoso>
Yaniel: Oh man I have been chasing a ghost. Naming my deployment "unifiController" was not the smartest idea.
<samueldr>
it's really -A which is a speical case AFAIUI
<dminuoso>
Yaniel: The package has been gone so long...
<Yaniel>
lol
<pie_>
samueldr i dont follow
<gchristensen>
hrm. the iso from the website (64 bit, headless) doesn't manage to import my zfs pool
<samueldr>
your system runs the "same" zfs? (not zfsUnstable, 19.03 to 19.03?)
<samueldr>
no overlays either?
<gchristensen>
oh!
<gchristensen>
luks!
<gchristensen>
thank you for the thinking!
itorres1 has joined #nixos
<pie_>
wait nix-shell doesnt even have a -A
itorres has quit [Ping timeout: 272 seconds]
<pie_>
oh yeah it does im blind
* pie_
needs to slow down
<pie_>
samueldr after a few rereads, yeah thats what i expected, -A is not just a "do the same thing but on this attribute"? it does something else...
<samueldr>
AFAIUI, that's it
<pie_>
does nix-build do something weird too
jgt has joined #nixos
<samueldr>
a bit late to change the past, but it could have been a better UX to instead have -A act like -p, but on the given attrset; always getting the derivation, and use something like --give-me-a-build-environment to get that environment
<pie_>
or is nix-shell some kind of degenerate case
dl4mfo has joined #nixos
<samueldr>
not sure, but I think it's nix-shell being special cased
<samueldr>
(the source is surprisingly easy to deal with if you want to see)
<pie_>
ok thats pretty annoying, id like a proper -A for nix-shell too :p
LogicAside has quit [Remote host closed the connection]
<pie_>
ok so i made another proxying shell.nix so im not using -A but i still get result :I
dl4mfo_ has quit [Ping timeout: 268 seconds]
ThatOtherGuy has joined #nixos
<pie_>
*get the built environment
ertw has joined #nixos
<pie_>
wait a minute. i wonder if theres some shellhook magic from the python infra going on here or something
<pie_>
why is everything so confusing xD
wildtrees has quit [Quit: Leaving]
Fahrradkette has quit [Quit: Leaving]
<pie_>
im increasingly feeling the urge to read source code because nothing ever makes sense. problem is there is an infeasible amount of source code.
<pie_>
this isnt my day job xD
<pie_>
well...ill just bear with this for the moment
<pie_>
"If we create a shell.nix file which calls buildPythonPackage, and if src is a local source, and if the local source has a setup.py, then development mode is activated."
<pie_>
"and if src is a local source,"
<pie_>
aha! thats why i didnt run into it earlier!
<exarkun>
everybody loves magic
<pie_>
especially if it makes sense
<pie_>
this would be less of a problem if things were consistent and knowledge transferred
<pie_>
samueldr were you perhaps thinking of https://github.com/NixOS/nix/issues/726 ? it seems to me like -A *is* consistent after all, I was just confounded by that fancy python stuff ^ (i think)
<{^_^}>
nix#726 (by copumpkin, 3 years ago, open): nix-shell -p is too magical
<DjDingo>
hi all. I am kind of confused why I have added "nixpkgs.config.allowUnfree = true" to my configuration.nix but still cannot install unfree packages
<DjDingo>
I have already rebuilt
mkoenig_ has quit [Remote host closed the connection]
<pie_>
does sudo -E not work on nixos or something?
<ivan>
DjDingo: are you using nix-env to install them?
<DjDingo>
ivan: yes
<ivan>
alias nix-env='NIXPKGS_ALLOW_UNFREE=1 nix-env'
<pie_>
i think you have to set it in nix-env's config separately
<pie_>
the nix env error message describes how to do it
markus1189 has joined #nixos
<DjDingo>
ivan: so after I add the line to my configuration.nix and rebuild I still need to put in the "allowUnfree" argument to install?
<DjDingo>
i thought it is only for one time
<DjDingo>
oh I understood now lol. What a fool am I lol
endformationage has quit [Quit: WeeChat 2.5]
domogled has joined #nixos
growpotk- has joined #nixos
jgt has joined #nixos
dansho has quit [Ping timeout: 246 seconds]
johnw has joined #nixos
jgt has quit [Ping timeout: 264 seconds]
johnw has quit [Ping timeout: 276 seconds]
kleisli_ has quit [Ping timeout: 245 seconds]
doyougnu has joined #nixos
abathur has quit [Ping timeout: 245 seconds]
sudoforge has joined #nixos
thePirateKing has joined #nixos
<sudoforge>
okay; i feel like a dummy. i've tried several times to get nixos-minimal-64 running in virtualbox to no avail. after selecting to start the installer, it sits at "running udev..." for eternity. any thoughts?
<thePirateKing>
hey guys does anybody use cachix? It is down and i cant do literally anything on my computer because of it
<thePirateKing>
like cant rebuild config cant install any packages I intend on using
<thePirateKing>
do i need to figure out which packages its trying to pull from cachix and fix that so i can rebuild without cachix?
revtintin has joined #nixos
kleisli_ has joined #nixos
<thePirateKing>
yeah no thats not solution i might have to literally reinstall
<thePirateKing>
if it stays down too long
Okinan has quit [Quit: Konversation terminated!]
mexisme_ has quit [Ping timeout: 248 seconds]
dl4mfo_ has joined #nixos
doyougnu has quit [Ping timeout: 250 seconds]
dl4mfo has quit [Ping timeout: 268 seconds]
jgt has joined #nixos
drakonis_ has joined #nixos
zmlww[m] has joined #nixos
growpotk- has quit [Ping timeout: 248 seconds]
drakonis has quit [Ping timeout: 276 seconds]
jgt has quit [Ping timeout: 264 seconds]
drakonis2 has quit [Ping timeout: 272 seconds]
<ivan>
sudoforge: try adding nomodeset to the boot options?
<ivan>
try removing hardware from the VM?
daveyjones01 has joined #nixos
thePirateKing has quit [Ping timeout: 276 seconds]
johnw has joined #nixos
johnw has quit [Ping timeout: 276 seconds]
atlas- is now known as atlas_
atlas_ is now known as atlas-
atlas- is now known as atlas_
Yakulu has joined #nixos
lambda-11235 has quit [Quit: Bye]
Yakulu has left #nixos ["Disconnected: Replaced by new connection"]
Yakulu has joined #nixos
dl4mfo_ has quit [Read error: Connection reset by peer]
jgt has joined #nixos
jgt has quit [Ping timeout: 276 seconds]
xkapastel has joined #nixos
orivej has joined #nixos
johnw has joined #nixos
szicari has quit [Ping timeout: 250 seconds]
daveyjones01 has quit [Ping timeout: 272 seconds]
karswell has quit [Read error: Connection reset by peer]
johnw has quit [Ping timeout: 250 seconds]
<nix-build>
[nixpkgs] @Provessor opened pull request #66462 → hunspell-dicts: add support for Australian dictionary → https://git.io/fj7Q8
szicari has joined #nixos
lord| has quit [Read error: Connection reset by peer]
sudoforge has quit [Ping timeout: 264 seconds]
lord| has joined #nixos
sudoforge has joined #nixos
parseval has quit [Quit: Connection closed for inactivity]
<nix-build>
[nixpkgs] @thefloweringash opened pull request #66466 → ec2-instance-connect-cli: init at 1.0.0 → https://git.io/fj7Qx
daveyjones01 has joined #nixos
thePirateKing has quit [Ping timeout: 276 seconds]
asymptotically has joined #nixos
mix has joined #nixos
mkoenig has joined #nixos
<mix>
Hello! Maybe someone can help me out... I have a build of emacs with some packages and it works fine when I build it and run it ./results/bin/emacs. I have a line `(import /etc/nixos/emacs.nix {inherit pkgs;})` in my configuration.nix but when I switch the system emacs doesn't have my packages :/ I don't understand what am I doing wrong.
Jackdk__ has joined #nixos
<Jackdk__>
Does anyone know whether the default network manager config in nixos does uncommon things with DNS? I can't connect through a captive portal because I get host not found errors, but I can get online on a tablet
<pie_>
samueldr: i just realized criu is short for crazy russians in userspace
<mix>
@Jackdk I can only confirm that I very often have problem with DNS when switching networks. But usually they get resolved just by restarting firefox.
jgt has joined #nixos
<Jackdk__>
Mix: thanks, no luck, #63754 seems related?
<nix-build>
[nixpkgs] @peti pushed 30 commits to haskell-updates: https://git.io/fj77O
gxt has quit [Quit: WeeChat 2.5]
<colemickens>
Is this with NM? I've noticed it has started adding edns0 as well...
zupo has joined #nixos
mix has quit [Read error: Connection reset by peer]
steshaw_ has quit []
steshaw_ has joined #nixos
<colemickens>
Ive had problems where o have to restart Firefox when changing networks too, that seems possibly distinct from the edns0 issue...
<colemickens>
Just last night I opened my laptop, same network as when I'd closed it, removed edns0 from /etc/resolv.conf and Firefox still didn't work until a restart.
<colemickens>
I'd tried using DoH but that causes more problems with captive wifis, some...
<steshaw_>
I've been really surprised that I have priority issue when attempting to install tools for Rust development but when using nix-shell to bring those same packages into scope, there is no problem. Any ideas? https://gist.github.com/steshaw/fb156b7a2ad460f4326d95b949a82cb0
daveyjones01 has quit [Remote host closed the connection]
daveyjones01 has joined #nixos
boxscape has joined #nixos
boxscape is now known as boxscapeR
o1lo01ol1o has joined #nixos
ottidmes has joined #nixos
<boxscapeR>
If I want to switch to a new channel to get updated software (say, the newest jdk), is it advisable to also to also do `nixos-rebuild --upgrade`?
<eyJhb>
boxscapeR: --upgrade only updates the channel, so if you switch channels to a new one, it should already be up-to-date
<eyJhb>
` Fetch the latest version of NixOS from the NixOS channel. ` as far as I know. So it shouldn't hurt :)
sudoforge has quit [Ping timeout: 244 seconds]
<boxscapeR>
hm... I was under the impression that was you're describing is done by `nix-channel --update`?
<boxscapeR>
I might be misunderstanding what you mean though
<boxscapeR>
^ eyJhb
xkapastel has quit [Quit: Connection closed for inactivity]
<eyJhb>
boxscapeR: There are 100% someone my qualified than me to answer this, but `Fetch the latest version of NixOS from the NixOS channel.` basically describes, that it will update the `nixos` channel from `sudo nix-channel --list`. That is at least, what I would expect from it. :)
<eyJhb>
So I am guessing it will run the command you describe underneath :)
o1lo01ol1o has joined #nixos
dispanser has quit [Read error: Connection reset by peer]
dispanser has joined #nixos
jtojnar has joined #nixos
chloekek has quit [Ping timeout: 276 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
Ariakenom has quit [Read error: Connection reset by peer]
Shouou has joined #nixos
meck has quit [Quit: WeeChat 2.5]
<boxscapeR>
huh, strangely enough doing nixos-rebuild --upgrade didn't install anything new after adding a new channel, but once I did nix-channel --update and then nixos-rebuild again, it's now reinstalling a bunch of stuff
<boxscapeR>
(technically `nixos-rebuild --upgrade boot` in both cases)
shibboleth has joined #nixos
<infinisil>
gchristensen: Huh, why the new nick for {^_^}?
<dminuoso>
Mmm. So I have the following situation. My laptop runs nixos with zsh inside kitty, and my server is a relatively raw nixos installation. When I ssh into the server, then things like backspace dont work correctly. Why is that?
<ivan>
maybe wrong TERM or zsh needs to be configured with more keybindings
<clever>
the one in nixpkgs throws an error at eval time, lol
<tokudan>
works-for-me®
<tokudan>
i'm on 19.03
<clever>
unstable here
<tokudan>
what's this line from the example? with (import /home/maarten/code/nixos/yarn2nix { inherit pkgs; });
<clever>
clone yarn2nix, and then put the path of your clone there
npmccallum has joined #nixos
<clever>
or use fetchFromGithub
<tokudan>
hmm... node2nix was a lot simpler
<clever>
yarn2nix has 2 ways to run, either with IFD, or where you pre-generate the nix once
<clever>
i think that readme is showing the IFD route
<joepie91>
yarn2nix docs are a bit confusing, but usage is easier imo
<tokudan>
IFD?
<clever>
,ifd
<joepie91>
hold on
<nix-build>
import-from-derivation (IFD) is when you evaluate nix from a derivation result, for example `import (pkgs.writeText "n" "1 + 1")` will evaluate to 2. This is sometimes problematic because it requires evaluating some, building some, and then evaluating the build result. It has been described as "such a nice footgun."
<tokudan>
joepie91, maybe once you've understood it
<joepie91>
this presumes that the repo contains a yarn.lock
<joepie91>
(I'm using a patched yarn2nix that removes the --no-patch flag because it broke shit)
<tokudan>
wow... complicated
<clever>
a patch to remove patching? lol
<joepie91>
clever: a patch to patch out the non-patching!
<clever>
tokudan: most of the complex stuff in my example is to deal with cross-compiling to windows
<joepie91>
tokudan: there's a bit of indirection in my example due to having to insert a config file that's not part of the repo, but other than that it's fairly straightforward, and most importantly does not require a separate "generate a default.nix" step
<joepie91>
the IFD approach takes any source that contains a yarn.lock
<joepie91>
and generates the relevant Nix code on build time
<clever>
joepie91: the source doesnt even need to contain a yarn.lock, as i confirmed a few days ago
<joepie91>
clever: I think it does if you sandbox builds?
<clever>
joepie91: you can use yarnLock = ./yarn.lock; to use an out-of-tree lock file
<tokudan>
what makes this so complicated for me is that I have absolutely no experience with yarn, node or any other related tool
<joepie91>
oh in that sense
<clever>
then you dont have to PR yarn support to the project, or fork it
<joepie91>
yeah
<joepie91>
tokudan: ah, right. tl;dr of Node package management = you have a package.json which specifies dependencies as semver ranges, then (depending on package manager used) you have a package-lock.json or yarn.lock which contains the exact package versions (and URLs and hashes!) that were installed with the package manager
<joepie91>
the lockfile is to reproduce your dependency set, but the semver specifications in package.json specify what you could update your deps to safely
<joepie91>
you give yarn2nix a yarn.lock because that means it doesn't have to resolve any deps itself based on package.json semver ranges, it can just take the work previously done by Yarn that it put into the lockfile
<joepie91>
and directly translate it to Nix
<joepie91>
and so the resulting Nix-produced dependency tree matches what Yarn would have installed upon a `yarn install` in that same repo / with that same lockfile
<tokudan>
that sounds good so far
<joepie91>
any part in there that's specifically unclear?
<tokudan>
it's not unclear, there's just so many details
<clever>
in my case, electron-packager wanted to download a windows build of electron at build time
<clever>
so i have to generate a compatible cache dir
<clever>
export ELECTRON_CACHE=${electron-cache}
<clever>
and then tell electron to use it
hmpffff has joined #nixos
<joepie91>
tokudan: grunt has no special role in package management. it's possible for a package.json to specify tasks to run after installing dependencies (called a 'postinstall script'), but I don't know whether yarn2nix honors these.
<joepie91>
I /think/ it would, but I almost never use postinstall scripts, so I don't know :)
shibboleth has joined #nixos
<joepie91>
grunt itself is ultimately just a binary, one of many on your system
<joepie91>
if nothing calls it somehow, it does not get executed
<clever>
this video points out a number of "quirks" with JS
<tokudan>
no, not welcome to it. going to have a bare-minimum contact with it
MmeQuignon has joined #nixos
<manveru>
i still consider it an achievement that can build almost all js apps these days with nix :)
<tokudan>
it may be possible, but the tooling for nix (yarn2nix, npm2nix) is somewhat complicated and seriously lacking documentation. and outside of nix it looks even worse
vika_nezrimaya has joined #nixos
<manveru>
yeah
<Shyim[m]>
Hey, i am really new to Nixos. PhpStorm is currently with version 2019.1 in nixpkgs. A pr is already open to update it. Is there a way to include that update in my configuration.nix?
<tokudan>
i mean for the common ./configure; make ; make install i just point nix to the source
<manveru>
Shyim[m]: sure, you just need to know the commit hash of that PR
<nexgen>
so your distro is mostly suitable only for open source software
ThatOtherGuy has quit [Ping timeout: 272 seconds]
<nexgen>
which is available say on github
<nexgen>
like Gentoo
<nexgen>
but with a more modern package manager
<edef>
we have skype, steam, etc packaged
<maralorn>
any hints on how I can test a substituter config? I have setup my server with the nix.serveNix option. I can succesfully connect to "ssh nix-ssh@myhost nix-serve". But nix says "cannot connect to 'nix-ssh@myhost'.
<manveru>
Shyim[m]: sorry, i didn't look up the name of the package, it should be `jetbrains.phpstorm` apparently
<maralorn>
clever: I just tested by login in on another tty as root. Am know configuring my sshServe to accept also my root-ssh-key.
steshaw_ has quit [Quit: Connection closed for inactivity]
<Shyim[m]>
same error after renaming 🤔
<tokudan>
nexgen, if you use nix only as a package manager you only get the benefits (rollbacks, purity, etc) of nix for the packages you install that way and not for the whole system
<manveru>
Shyim[m]: guess i'll actually have to try it myself ;)
<Shyim[m]>
its ok. But thanks ❤️
<nexgen>
I have ZFS for the whole system
<nexgen>
though missing reproducible builds
<nexgen>
I use Devuan
<tokudan>
nexgen, you're also missing a mostly reproducible system
<nexgen>
what do you think about GUIX as a package manager and as a distro?
dansho has joined #nixos
<simpson>
nexgen: Still wrong channel for GUIX. What are you *actually* trying to figure out?
<tokudan>
reinstall for nixos is basically: copy /etc/nixos from the old system or a backup and run nixos-install and you're done
<nexgen>
I am trying to understand whether I need NixOS
<simpson>
nexgen: Nope, you don't. In fact, you likely don't need computers at all.
<tokudan>
once you've seen the benefits you just wont want to miss it
<nexgen>
simpson, why do you think so?
ThatOtherGuy has joined #nixos
<simpson>
nexgen: Historically, people have not needed computers. Do you have some specific package that you're trying to get working with Nix?
<infinisil>
People have survived without Nix for thousands of years, so probably you don't need it. I'm personally not sure how they managed that, but it must be possible somehow
<nexgen>
I am not interested how people used to do it earlier
asymmetric has quit [Quit: Peace.]
<nexgen>
I am thinking about a convenience in my case
asymmetric has joined #nixos
<tokudan>
that's the point. do you see any benefits in using puppet or ansible?
<nexgen>
me?
<tokudan>
yes
stepcut has joined #nixos
<nexgen>
ansible - yes
<tokudan>
nixos is the next step
<nexgen>
nix does the same thing about configs?
<Shyim[m]>
manveru: i love you. Its working ❤️
<Shyim[m]>
Thanks!
<nexgen>
config deltas?
<tokudan>
puppet and ansible are legacy from our viewpoint
<tokudan>
nexgen, no deltas. nix builds a whole new system everytime without any legalcy
<manveru>
Shyim[m]: cool :)
<tokudan>
nexgen, no deltas. nix builds a whole new system everytime without any legacy
<nexgen>
not binary, but something like declarative deltas?
<nexgen>
or states
<clever>
nexgen: basically, you just define how to build the final setup, and nix will use the hash of those directions to see if something is already built
<clever>
nexgen: so if a sub-step of the build has already been done before, it gets reused
<clever>
no need to compute a delta from the previous
<nexgen>
I see your channel is very active
<srhb>
Lots of nice people in here. :)
<tokudan>
nexgen, puppet and ansible define part of a system. if you later remove a definition for one part of the system, it's left in an unclear state. it wont be removed, but it's just unmaintained. it's a legacy.
stepcut has quit [Remote host closed the connection]
<tokudan>
nexgen, nix and nixos do not have that problem as the whole system is exactly defined at any time
<erba>
I just want to thank everyone working in the best distro I've used (especially for haskell) could not be happier!
<erba>
*on the best
<nexgen>
very good
<nexgen>
I came here through a Gentoo
<nexgen>
needing to compiles a grsec kernel
<imdoor>
hi, i want to pin my current nixpkgs in a project? how do i get the sha256 and rev values i'm currenlty using? (can't use nix-prefetch-git, cause that'd pull something newer)
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<Shyim[m]>
Yes. I am installed yesterday first time nixos in a vm. Now I want to switch everything to nixos ❤️
<nexgen>
do you have a package definition for a minipli grsec?
<imdoor>
*the first sentence wasn't ment to be a question
<tokudan>
nexgen, you basically have to tell nix how exactly you want your kernel compiled. then you can use that kernel
chloekek has quit [Ping timeout: 248 seconds]
<simpson>
nexgen: However, you might want to re-evaluate what grsec is actually providing for you, as it's not a very good patchset. Introduces new vulns, poorly tested, hilariously poor communication with downstream, and isn't even clearly meeting its own goals.
<srhb>
nexgen: Partially at least, yes.
<nexgen>
simpson, I have read a lot of good things about it though
<nexgen>
would like to try it at least
<nexgen>
and see if it helps e
<nexgen>
*me
<nexgen>
building of dotnet core 2x actually works?
<nexgen>
it is generally mentioned as a not easy to build
<srhb>
nexgen: No idea. But it looks like an at least partial binary deployment, considering the use of patchelf.
mix has quit [Ping timeout: 276 seconds]
<clever>
nexgen: thats also just the binary blog being patched
<nexgen>
oh
<nexgen>
:(
<clever>
nexgen: some things like electron are still blob based, even though they are open source
<nexgen>
backdoored by Microsoft
<clever>
nexgen: the source for electron, once you run gclient, is over 30gig, so iterating on a proper package takes hours per attempt
<nexgen>
well, I can use mono yet
<nexgen>
two packages could be presented: 1) bin 2) source
<dminuoso>
Hi. Im a bit confused, I installed 'nix-env -iA nixos.ncurses6.dev' which has bin/infocmp - but that binary does not end up in my profile.
<nexgen>
anyone could do a desirable choice
<srhb>
nexgen: You're very welcome to contribute a working package. :)
<clever>
dminuoso: files from .dev are usually ignored by nix-env
<clever>
dminuoso: you probably want `nix-shell -p ncurses6
<nexgen>
I still to not know how to do it, though I would be glad
<srhb>
nexgen: I think most other people feel the same way.
<dminuoso>
clever: Mmm, well I do want that binary in order for certain kitty functionality to work.
<nexgen>
if it even builds at all, I have read there are some difficulties
<dminuoso>
clever: And Id prefer to have that in my systemPackages.
<imdoor>
srhb: nixpkgs.lib.version gives me "19.09pre184803.d567c486ca5". nix complains if i try to pass either of the strings seperated by '.' as the sha256 value when calling fetchFromGitHub (which i'm using for pinning)
<srhb>
nexgen: I bet.
<clever>
dminuoso: it might work better in systemPackages
<srhb>
imdoor: Oh, don't pass the sha, just use the ,tofu one
<srhb>
imdoor: But that last bit is the rev
<clever>
imdoor: the rev is d567c486ca5 and tofu the sha256
<srhb>
imdoor: So use that as your rev for fetchFromGitHub, then pass 52 zeros as the sha256.
<srhb>
imdoor: The error message from the hash mismatch will give you the correct sha.
<imdoor>
srnb, clever: ok, thanks :)
<nexgen>
can NixOS base be rebuild completely from source? reproducibly?
<srhb>
nexgen: Depends what you mean by reproducibly
<dminuoso>
clever: huh so systemPackages is treated differently from installing via nix-env?
<simpson>
nexgen: Nix is as reproducible as the toolchains you choose to use with Nix. As we make toolchains better and better, Nix automatically exhibits their reproducibilities.
<clever>
infinisil: nix's purity doesnt prove that the binary cache isnt malicious, bit-per-bit reproducible prooves that the binary cache hasnt done anything nasty
<clever>
dminuoso: yeah
iyzsong has quit [Ping timeout: 258 seconds]
<dminuoso>
clever: Mmm, is this covered in any of the manuals?
<clever>
dminuoso: probably
<dminuoso>
Heh.
gerschtli has quit [Quit: WeeChat 2.4]
<clever>
dminuoso: ive not re-read them in a while
maralorn has joined #nixos
<ajs124>
what does nix verify mean when it says that a path is untrusted?
<nexgen>
how hardened profile helps for security?
gerschtli has joined #nixos
<ajs124>
or more general, is there any documentation on the 2.x nix utils?
stepcut has joined #nixos
<dminuoso>
Oh no..
<infinisil>
clever: Yeah, it can't do that, but for problems like "works on my machine", Nix's purity gives you pretty good reproducibility for that
<simpson>
nexgen: Checks boxes, gives feeling of comfort, etc. The usual.
<nexgen>
)
<clever>
infinisil: as long as no funny cpuid stuff is going on
<nix-build>
[nixpkgs] @turboMaCk opened pull request #66484 → acpid: fix escaping in handlers doc example → https://git.io/fj7bm
madhukar93 has joined #nixos
<simpson>
clever: Although, of course, many folks desire such things. I was reading about Spack and EasyBuild yesterday; they both care about getting CPU tuning correct because their users are HPC.
<nexgen>
if configs are so easily replicated in NixOS
<nexgen>
may be there is something comparable to OpenBSD config?
<clever>
simpson: mplayer solves that by having runtime cpu detection, it just builds multiple variants of key functions, and selects the right one at runtime
<Athas>
infinisil: regarding reproducibility, doesn't Nix really boil down to reproducibly constructing a shell script (the builder), the reproducibility of which we then take on faith?
<Athas>
How often has this been a problem in practice?
<simpson>
clever: Right, and AFAIK runtime CPU detection is reasonable in terms of the runtime costs on all of the popular CPUs, so the problem (as usual!) is the piles of existing incorrectly-written C.
<infinisil>
Athas: Nix's sandbox restricts heavily what builders can do
<clever>
simpson: the only time ive seen runtime detecting be a problem, was on an avr, where the cycle count of an `if` statement in an irq was too much, lol
<Athas>
infinisil: oh cool. What mechanisms are used for that?
<Athas>
No network access?
<infinisil>
Yeah that's the major one, no network access, only filesystem access to a specific set of paths
<clever>
Athas: its basically spinning up a dynamicly created container for every build
mix has quit [Ping timeout: 248 seconds]
<Athas>
I mean, a builder can still be nondeterministic through various sources of randomness, but I'm curious whether this is ever a problem for non-contrived non-malicious derivations.
<infinisil>
The sandbox leaves Nix builders with only a limited set of non-deterministic inputs, such as /dev/random or file ordering in directories
<clever>
Athas: some of the manual's in nixos used rng to generate links in the html, to make the links unique within the document
<Athas>
Or process scheduling order, surely?
<infinisil>
clever: That better be deterministic though
<clever>
Athas: yeah, parallel builds have caused issues before, which is why parallelish is off by default
madhukar93 has quit [Remote host closed the connection]
<nexgen>
Athas, are there any configs to make Linux, say NixOS, have similar to OpenBSD locked down configs?
<Athas>
nexgen: probably, but that's missing the point. OpenBSD is good because it's holistically developed. It loses its advantages when you install too much third party software.
<Athas>
NixOS, on the other hand, accepts a heterogeneous world of third-party software, and just imposes some structure on top.
<Athas>
But that does not make a mess of third-party software any more cohesive in documentation, configuration style, code quality, etc.
<infinisil>
nexgen: But yeah, my point is that you can't infer "comparisons aren't welcome" when nobody replies to your question about a comparison, not sure why you thought so
<nexgen>
So NixOS is good as an application platform for many programs
<nexgen>
like Windows
<nexgen>
but more manageable
<Athas>
I'm not sure Windows is particularly "good" for that, except that a lot of software happens to exist for it.
<nexgen>
And for security it is better to use still OpenBSD
<nexgen>
that is what I meant, many software titles
<Athas>
OpenBSD is good if you can accept that it does not scale to beefy hardware, and you need very few features (mostly the base system).
<nexgen>
as an opposite to OpenBSD
<Athas>
NixOS is good if you need all the Linux advantages (beefy or obscure hardware, heterogeneous software).
<nexgen>
I think about using OpenBSD and a hypervisor for NixOS guest
<nexgen>
is it a good idea?
szicari has quit [Ping timeout: 264 seconds]
<Athas>
OpenBSD is slow.
<nexgen>
but secure
<nexgen>
how much slower is it compared to KVM?
<Athas>
Sure, but I'm not certain it makes a good hypervisor for Linux.
<nexgen>
say its vmm
<DigitalKiwi>
infinisil: #fedora used to (they still might) ban people for even mentioning other distros
<nexgen>
if 10-30% slower it does not matter
<nexgen>
if 2-3 times slower than it is already bad
<infinisil>
DigitalKiwi: Damn
<clever>
DigitalKiwi: i once got banned from #debian for asking ubuntu questions
<Athas>
At this point you should go to #openbsd instead, since it's not really a comparison anymore.
<DigitalKiwi>
infinisil: so maybe experiences like that are the source of "comaprisons unwelcome"
<clever>
DigitalKiwi: in my favor, it was an apt question, so it applied to either distro
* sphalerite
is tempted to join #fedora, say "nixos", then leave again
<nexgen>
I am present in #openbsd
<infinisil>
I think comparisons are really valuable
<infinisil>
Lots of people are familiar with *one* system
<nexgen>
without comparison I would not get here
<dminuoso>
DigitalKiwi: See, this is part of why I enjoy the Haskell community (much of which is also found here), where you might not even get any comments for discussing Rust, Prolog or category theory in #haskell
<infinisil>
But to be able to do a comparison you need to be familiar with multiple
<infinisil>
And that takes time
lavalike has joined #nixos
<nexgen>
comparisons are good for better things like your os
<clever>
sphalerite: reminds me of the old days when people joined christian channels, then made the bot quote a verse that had a curse word in it, causing the bot to get auto-banned :P
<nexgen>
and bad for others
<nexgen>
as they bring more people from other distros to yours
<infinisil>
That reminds me, I should clean my computer by getting rid of the windows installation on it
<infinisil>
Haven't used anything other than NixOS on it for like a year now, I think it's time
<simpson>
nexgen: It's not a popularity contest. There are more important considerations, like the fact that Guix forked from Nix; if you deliberately ignore these facets of history, then you're only going to sound anachronistic.
<Yaniel>
ugh, grub failed to install on my efi partition
szicari has joined #nixos
drakonis has joined #nixos
<DigitalKiwi>
infinisil: I'm familiar with nixos, freebsd, macos, arch linux, fedora, centos, various debian/ubuntu derivations...that's without thinking too hard or obscure ones that I haven't used in years/much at all
<infinisil>
DigitalKiwi: Neat, like more than a year with all of them?
<Yaniel>
guess it's trying to tell me to stick to trying out void in a vm
<DigitalKiwi>
yes though some more recently than others
<nexgen>
can I build things like monodevelop IDE without any graphical DE for use with another local host X server running in another distro
<infinisil>
DigitalKiwi: How long do you estimate to stay with NixOS?
<DigitalKiwi>
i'm pretty committed to nixos at this point
<nexgen>
DigitalKiwi, did you evaluate any alternatives?
<nexgen>
comparable if they exist even
<DigitalKiwi>
I've used centos and freebsd on servers for years...like 4-6, I've been an arch user (though a lot less lately) for 11 and a half years, nixos for about a year, full time since december, most of my computers including my servers are on it now.
<infinisil>
nexgen: NixOS can run X server and you can use X port forwarding like normal, so there shouldn't be anything preventing you from doing that
<infinisil>
Ah, yeah but you asked to use it as an X client
<infinisil>
Yeah no problem either, nixpkgs packages are distro-agnostic
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos
<DigitalKiwi>
4-6 is actually low now that I think about it probably since 2010
<DigitalKiwi>
for centos, freebsd is probably closer to 4-6
<DigitalKiwi>
so far like nixos the most
<infinisil>
You probably wouldn't be here if that wasn't the case :P
<DigitalKiwi>
well i'm also in #archlinux
<averell>
me too. i'd wager the overlap is largest for all the distros.
<DigitalKiwi>
lots of recovering arch users here
<clever>
prior to nixos, i ran gentoo
<petercommand>
clever: how's gentoo compared to nixos?
<petercommand>
clever: like, how does gentoo's package manager work?
<clever>
petercommand: similar, its just a giant heap of shell scripts that build everything, and define what the deps are
dispanser has joined #nixos
<simpson>
petercommand: It's alright. Gentoo rebuilds a lot of stuff. There's a lot of features in Gentoo that look extremely baroque when compared to the ability of NixOS to configure anything by writing some Nix.
<clever>
petercommand: the main difference, is the lack of sandboxing, its not able to have conflicting versions co-exist, so rollbacks are missing, and no official binary cache, so you must build everything from source
<petercommand>
hmm..
freeman42y has joined #nixos
doyougnu has joined #nixos
chloekek has joined #nixos
freeman42x has quit [Ping timeout: 264 seconds]
<petercommand>
do you guys think it would be a good idea to have an equivalent of "nix-env -i" for deb/rpm packages that are hard to install via nix where such commands would automatically try to resolve deb/rpm dependencies and install them (whatever that means)
<simpson>
No, also-ran techniques are usually harmful to the ecosystem. Further, Nix itself doesn't really have a concept of difficulty of installation; rather, the difficulty is in writing the packaging.
<simpson>
To imagine packaging a uniform wrapper for *every* RPM or Debian package... Wouldn't that be harder than packaging any one single package? We do have universal techniques, but they involve putting bad software into playground environments.
<petercommand>
well yes, I mean packages that are hard to package for nix
chloekek has quit [Ping timeout: 248 seconds]
<simpson>
We can patch source. We can patch binaries, too, if really necessary. No package is worth sacrificing Nix's build isolation.
<petercommand>
of course it would be harder, I am not saying that it is not
<simpson>
Do you have examples of such packages? Anything that you need, in particular?
Soo_Slow has joined #nixos
<petercommand>
simpson: I was trying to package Zettlr, but I just don't know how to properly package it
<petercommand>
simpson: it uses yarn and electron-builder
<joepie91>
most Electron things you can package by just completely ignoring the included Electron binary and using the nixpkgs version of Electron, running it as `electron /path/to/application`
qqlq has quit [Ping timeout: 272 seconds]
<petercommand>
I tried yarn2nix, but didn't get the whole build process to work properly
<nix-build>
[nixpkgs] @danbst pushed commit from @pacien to master « nixos/postgresql-wal-receiver: add module (#63799) »: https://git.io/fj7AT
__monty__ has joined #nixos
ertw has quit [Remote host closed the connection]
<clever>
rnhmjoj: `mount --bind /foo /mnt/`
<clever>
rnhmjoj: then nixos-install as usual
<rnhmjoj>
clever: thank you, using `/foo` directly is expected to fail?
<clever>
rnhmjoj: yeah, nixos-install does some sanity checks to see if its a dir
daveyjones01 has quit [Remote host closed the connection]
<clever>
if its a mount point*
daveyjones01 has joined #nixos
<rnhmjoj>
i'll try it, thank you again
fendor_ has joined #nixos
reanimus has quit [Quit: Logging off...]
<clever>
rnhmjoj: what are you trying to do?
<rnhmjoj>
clever: i'm testing a change to the nixos-install script
<clever>
rnhmjoj: ah
<clever>
then all the methods of skipping nixos-install dont apply!
shibboleth has joined #nixos
<clever>
youll want to be carefull with efi vars, and boot.loader.grub.device
reanimus has joined #nixos
fendor_ has quit [Client Quit]
fendor__ has quit [Ping timeout: 244 seconds]
<rnhmjoj>
clever: i had to disable all the bootloaders and nix seemed happy but then the installation failed with some error about the filesystems being read-only
<drakonis>
DigitalKiwi: i could've sworn i have seen you elsewhere
<Shyim[m]>
I am trying to install nixos in a vm on vmware. The entire vm is crashing on boot. Has someone an idea? https://i.imgur.com/CdATufj.png 😂
<clever>
Shyim[m]: that seems weird, to fail that early on
<Shyim[m]>
yep 😂
<DigitalKiwi>
drakonis: #archlinux
<drakonis>
no #freebsd?
<DigitalKiwi>
i don't think i've been there, but i have logs of your name in #archlinux
<drakonis>
ah i see okay
<DigitalKiwi>
but i have been on freenode for 11 and a half years so :)
<clever>
nexgen: git clone nixos-configs, add the (relative|absolute) path of qemu.nix to your imports list, and then set qemu-user.arm = true;
<clever>
currently, it doesnt build on the latest nixos-unstable, so youll need to be on a stable channel
<rnhmjoj>
i have tried passing the closure to nixos-install and it again fails with that error
<worldofpeace>
,locate libgmp.so.10
<stepcut>
I'd love to get an RPi4 once they fix the (minor) issue with USB-C power
<{^_^}>
Found in packages: gmp5, gmpxx, julia, julia_07, julia_11
<slabity>
Hey guys, I just installed NixOS on a new Ryzen3+x570 system, but my ethernet controller isn't being seen. Only the wireless adapter. Any ideas how to find out what's wrong?
<sphalerite>
slabity: is it known to work with linux generally? Maybe try using linuxPackages_latest?
<slabity>
spahlerite: Not sure. It's an Intel chip, so I'd assume so, but it's not even showing up in lspci. I am already using linuxPackages_latest
shibboleth has quit [Remote host closed the connection]
mexisme_ has joined #nixos
<worldofpeace>
,locate libtinfo.so.5
<{^_^}>
Found in packages: tlf, arduino, ncurses5, arduino_core, jetbrains.clion
mexisme_ has quit [Ping timeout: 245 seconds]
qqlq has joined #nixos
s94a84d2e has joined #nixos
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<exarkun>
clever: Hi. I am trying to use nixos-configs/qemu.nix and I have coreutils test suite failures with it on armv6l due to coreutils picking up the full path to the executable as "program_name", instead of merely the basename (eg coreutils du test suite wants du to report itself as "du" not "/build/.../du"). Does that sound familiar at all? I'm not sure which piece here is at fault...
<clever>
exarkun: that might be a bug in the argv0 handling of the qemu stuff, i would just disable the tests
<exarkun>
I wonder if there are other packages later that will have similar problems
domogled has quit [Quit: domogled]
<exarkun>
It takes 12 hours just to get coreutils built
<slabity>
That's weird. The NIC isn't even being detected in dmesg
<exarkun>
Long edit/test cycle :/
<clever>
slabity: what about lspci?
<slabity>
clever: Nope. Only the wireless controller shows up
<clever>
slabity: double-check that the pci device is seated properly?
<slabity>
clever: Well it's soldered to the motherboard, so...
qqlq has joined #nixos
<clever>
slabity: just once, the ethernet in my laptop failed hard, dmesg spewed errors, and after a soft reboot, lspci claimed the device didnt even exist
<clever>
slabity: after a full power cycle, the card returned from beyond, and never did it again
<slabity>
Unfortunately power cycling doesn't seem to be doing much. Let me see if there's something I need to enable in BIOS
<slabity>
Maybe it's disabled by default for some reason
endformationage has joined #nixos
o1lo01ol1o has joined #nixos
<slabity>
Strange. My BIOS says the LAN's MAC Address is 'N/A' - I really hope that doesn't mean it's borked
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-http-download: disable test suite to fix the build »: https://git.io/fj7xG
<exarkun>
is `nix.package = fetchTarball { url = <path to nixpkgs tarball>; sha256 = ...;}` the right way to use nix.package? if so, how do I also specify an overlay (so I can disable coreutils checkPhase)?
chloekek has joined #nixos
<exarkun>
I can't tell why you would prefer `nix.package` over `nixpkgs.pkgs` or vice versa and what the interaction between `nix.package` and various other `nixpkgs.*` options might be
<srhb>
exarkun: No, nix.package is the package for Nix itself.
<exarkun>
aaa
<clever>
exarkun: nix.package must be the result of building it, but fetchTarball returns a directory, in this case, of nixpkgs which isnt a even nix (the package manager)
karetsu has joined #nixos
<srhb>
exarkun: you probably want nixpkgs.pkgs and nixpkgs.overlays. Also read the wiki article on overlays.
<exarkun>
right ... the docs say "Nix package", not "nixpkgs" :/ argh
<srhb>
exarkun: Right, nixpkgs, nix, pkgs are all different idioms. :-P
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-pantry: disable test suite to fix the build »: https://git.io/fj7xl
o1lo01ol1o has quit [Remote host closed the connection]
szicari has quit [Ping timeout: 276 seconds]
mix has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage2nix: update list of broken builds »: https://git.io/fj7xu
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « Merge pull request #66468 from NixOS/haskell-updates »: https://git.io/fj7xQ
sudoforge has quit [Ping timeout: 264 seconds]
gentauro has quit [Quit: leaving]
growpotk- has joined #nixos
qqlq has joined #nixos
sudoforge has joined #nixos
mkoenig has quit [Ping timeout: 268 seconds]
growpotkin has quit [Ping timeout: 276 seconds]
mkoenig has joined #nixos
tnks has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @philandstuff opened pull request #66487 → amazon image: fetch meta-data everytime in case if something changes → https://git.io/fj7xj
growpotkin has joined #nixos
tnks has joined #nixos
domenkozar[m] has joined #nixos
domenkozar[m] has quit [Changing host]
fendor has quit [Ping timeout: 272 seconds]
hmpffff has joined #nixos
hmpffff has quit [Client Quit]
leotaku has joined #nixos
bdesham has joined #nixos
mix has quit [Read error: Connection reset by peer]
<exarkun>
I'm failing to get coreutils to not run checkPhase
<stepcut>
I got Idris2 built, but I am not sure how to get it to build something, I created a Hello.idr and hello.ipkg and ran `~/.idris2/bin/idris2 --build hello.ipkg`. It produced Hello.ttc and Hello.ttm but no executable
o1lo01ol1o has quit [Remote host closed the connection]
<Shyim[m]>
I have added environment.noXlibs = true; to my configuration and started rebuilding. It compiles now much pages. Is that normal? :D i hoped it will just remove some packages
<lngd>
it may be recompiling packages that link have been linked with x shared libs
mexisme_ has joined #nixos
lngd has quit [Client Quit]
<Shyim[m]>
ahh :)
<stepcut>
ugh, I need to find a different irc client, I constantly post things in the wrong channel with this one
imdoor has quit [Quit: imdoor]
imdoor has joined #nixos
imdoor has quit [Remote host closed the connection]
imdoor has joined #nixos
drakonis has quit [Quit: WeeChat 2.4]
sudoforge has quit [Ping timeout: 272 seconds]
growpotkin has quit [Ping timeout: 276 seconds]
imdoor has quit [Client Quit]
imdoor has joined #nixos
<{^_^}>
[nixpkgs] @nh2 opened pull request #66490 → zlib: Properly clean up static/shared distinction → https://git.io/fj7hZ
<nh2>
clever: thanks; this allowed me to solve my problem and motivated me to make https://github.com/NixOS/nixpkgs/pull/66490. CC matthewbauer, this should address the concerns in your last zlib commit message
imdoor has quit [Remote host closed the connection]
imdoor has joined #nixos
lambda-11235 has quit [Quit: Bye]
gentauro has joined #nixos
selfsymmetric-mu has quit [Remote host closed the connection]
selfsymmetric-mu has joined #nixos
xorAxAx has quit [Quit: Gateway shutdown]
imdoor has quit [Quit: imdoor]
imdoor has joined #nixos
xorAxAx has joined #nixos
<infinisil>
Shyim[m]: I can't recommend that option at all..
<Shyim[m]>
i have removed it already
<Shyim[m]>
it fails to built
<infinisil>
It rebuilds a ton and when I tried it, it even increased the closure size
<Shyim[m]>
some pkg-config issues
xkapastel has quit [Quit: Connection closed for inactivity]
<mexisme_>
hi all, i think everyone in @nixos-darwin is afk, but just wondered if anyone in her is on Nix Darwin, and getting a few unusual compilation errors in the last week-ish?
<mexisme_>
i.e. since the last macOS patch release
leotaku has joined #nixos
imdoor has quit [Client Quit]
imdoor has joined #nixos
imdoor has quit [Remote host closed the connection]
imdoor has joined #nixos
___laika has quit [Ping timeout: 268 seconds]
levdub has joined #nixos
<samueldr>
hey, peeps, just a quick reminder about how there's only 4 weeks left to submit to the CFP for nixcon 2019 :) https://cfp.nixcon.org/nixcon2019/cfp
<vaibhavsagar>
how do people decrease the cycle time of working with `dockerTools`? Building, adding, and running are all incredibly slow
lambda-11235 has joined #nixos
<{^_^}>
[nixpkgs] @grahamc pushed 0 commits to revert-66468-haskell-updates: https://git.io/fj7jF
<selfsymmetric-mu>
I'm getting these on rebuild: `"error: getting attributes of path /nix/store/<some_hash>/xvidcore-1.3.5": No such file or directory`
<rnhmjoj>
infinisil: thank you
<selfsymmetric-mu>
It's true. The file does not exist. But why is Nix looking for it?
<selfsymmetric-mu>
I am running `sudo nixos-rebuild switch --keep-going`.
<selfsymmetric-mu>
And before I run that I run `sudo nix-channel --update nixpkgs unstable`.
Soo_Slow has quit [Remote host closed the connection]
tokudan_ has joined #nixos
<{^_^}>
[nixpkgs] @Infinisil pushed to haskell-updates « hackage2nix: Unmark pantry and http-download as broken »: https://git.io/fj7jb
<{^_^}>
nix#2673 (by trofi, 26 weeks ago, open): 'nix-channel --update' stopped working on nix-2.2.1: error: while setting up the build environment: getting attributes of path '': No such file or directory
<selfsymmetric-mu>
Wait no, that one just looks like a bad path.
<selfsymmetric-mu>
I have no idea what's going on, but it means I'm blocked on this machine.
<selfsymmetric-mu>
Oddly enough I do not have this issue with the same configuration on other machines.
<selfsymmetric-mu>
This issue persists across attempts and across reboots.
<selfsymmetric-mu>
However, something interesting is that when I removed some old generations the breaking path _changed_.
<drakonis>
i'm running out of space the moment i install systemd-loader
<drakonis>
aint that just wonderful
<drakonis>
since it does not allow relocating the kernel images to a different location
<selfsymmetric-mu>
Has anyone seen a problem like this or should I report a new bug? I can't find this one on nixpkgs.
<selfsymmetric-mu>
Actually, in this case, to which repository should I report a bug? `NixOS/nix` or `NixOS/nixpkgs`?
<drakonis>
i'm mildly confused right now
<drakonis>
does systemd-boot offer the ability to decide which location the kernel images are stored?
<gchristensen>
systemd-boot uses .conf files where you can provide a path to the file
<drakonis>
hm, but its not set up to store the kernel images on /boot instead of /boot/efi, right?
<zmlww>
Creating my first package. The package in question uses Gradle to build: can anyone recommend another package that uses Gradle that would serve as a good reference?
<drakonis>
i'm griping that its filling up my efi partition and preventing me from booting
tgamblin has joined #nixos
<danderson>
systemd-boot can only boot stuff in the EFI partition, to my knowledge
<danderson>
though maybe I'm wrong about that, the only time I've used systemd-boot in detail is when I made secure boot work on my laptop
<drakonis>
and grub is hosed up on nixos-unstable
<drakonis>
and above
<danderson>
systemd-boot is very good for that, because it supports booting a blob that's the combo of kernel+modules+initrd
<selfsymmetric-mu>
I can't really do without `xvidcore` since that's what Chromium and Steam use. So this laptop is no longer usable for me.
tgamblin has left #nixos [#nixos]
<selfsymmetric-mu>
I just wish I could know why Nix is looking for a file that's not there! It doesn't even provide any justification, even with tracing on!
tgamblin has joined #nixos
<drakonis>
danderson: no clues on how to do that?
phreedom has quit [Ping timeout: 260 seconds]
___laika has quit [Ping timeout: 246 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
___laika has joined #nixos
tgamblin has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<danderson>
sorry, I got nothing. Quite new to nix/nixos.
daveyjones01 has quit [Remote host closed the connection]
phreedom has joined #nixos
___laika has quit [Ping timeout: 248 seconds]
jtojnar has quit [Quit: jtojnar]
lord| has quit [Quit: WeeChat 2.5]
jtojnar has joined #nixos
xorAxAx has quit [Quit: Gateway shutdown]
lord| has joined #nixos
xorAxAx has joined #nixos
mbrgm_ has joined #nixos
___laika has joined #nixos
justanotheruser has quit [Ping timeout: 244 seconds]
mbrgm has quit [Ping timeout: 252 seconds]
mbrgm_ is now known as mbrgm
tgamblin has joined #nixos
<tgamblin>
.join nixos
jtojnar has quit [Client Quit]
tgamblin has quit [Remote host closed the connection]
<zmlww>
Is it discouraged to use a .deb or .rpm instead of the source for a package?
drakonis has quit [Quit: WeeChat 2.5]
<gchristensen>
it is strongly preferred to build from source
<zmlww>
Sorta figured, alright. Guess I'll have to figure out this Gradle mess after all :)
drakonis has joined #nixos
<gchristensen>
ouch.
<gchristensen>
I admit even I would just reach for the deb/rpm.
<pie_>
:DDD
<zmlww>
Oh, I don't know how bad it is. It's my first package.
<pie_>
oh no. :D
<zmlww>
Maybe I should just do the deb/rpm at first?
<{^_^}>
[nixpkgs] @emilazy opened pull request #66497 → stm32loader: init at 0.5.0 → https://git.io/fj5eG
<zmlww>
I'd like it to be accepted into nixpkgs in the end.
<pie_>
also, wait, we can handle deb/rpm?
___laika has quit [Ping timeout: 245 seconds]
hmpffff has joined #nixos
<zmlww>
pie_: I saw a blog post about deb, doesn't seem so bad.
<pie_>
im not saying we cant i just didnt know we can, if we ca
<exarkun>
nix doesn't like my buildMachines because host key mismatch. Which known_hosts file does nix look at when connecting to build machines?
<gchristensen>
exarkun: the root user's and the host system
<selfsymmetric-mu>
Oh! `nix-store --verify --check-contents` to the rescue. It reports to me that my xvidcore path has disappeared but still has valid referrers, which is true.a
<selfsymmetric-mu>
I wonder if it'll suggest a fix...
<exarkun>
gchristensen: ty
<selfsymmetric-mu>
I'll try running it with `--repair` after this.
<clever>
selfsymmetric-mu: you can also `nix-store --query --roots /nix/store/foo` to see what is rooting it, and `nix-store --delete /nix/store/foo` to just remove it
<pie_>
and openstore() is called without arguments in nix-build.cc
<pie_>
though im not actually sure how that argument gets treated either
<infinisil>
pie_: And I recommend not asking specific people, this makes a lot of people not even look at the question, and clever might not be here or doesn't have time
<pie_>
i guess thats what i should find out first somehow
<pie_>
infinisil fair enough
<NickHu>
Ah I guess my channel was slightly out of date - wine is probably a big derivation and hydra doesn't want to keep lots of them I suppose
<gchristensen>
the cache has never been GC'd :)
<pie_>
maybe it was a checkout that never got build properly?
<pie_>
(idk)
<Ashy>
gchristensen: how big is the hydra cache?
<gchristensen>
180t
* Ashy
whistles
<Ashy>
nice
<pie_>
hm so thats about 7200$ of some random 10TB i googled, not counting redundancy
<pie_>
*10TB hdd
<pie_>
except this is amazon
<pie_>
on that note
<pie_>
is amazon storage actually reliable?
thc202 has quit [Quit: thc202]
<pie_>
i wonder if all that 180TB is still good
<gchristensen>
s3?
<gchristensen>
their SLA on durability is 99.9999999%
<pie_>
hm, one would hope thats true, if yes, sounds good
<gchristensen>
"Amazon S3 Standard, S3 Standard–IA, S3 One Zone-IA, and S3 Glacier are all designed to provide 99.999999999% durability of objects over a given year. This durability level corresponds to an average annual expected loss of 0.000000001% of objects. For example, if you store 10,000,000 objects with Amazon S3, you can on average expect to incur a loss of a single object once every 10,000 years. In addition,
<gchristensen>
Amazon S3 Standard, S3 Standard-IA, and S3 Glacier are all designed to sustain data in the event of an entire S3 Availability Zone loss."
<pie_>
nice
<exarkun>
So statistically 1979.12109375 bytes of the cache are bad now.
<exarkun>
:)
<gchristensen>
if each byte were an object
<gchristensen>
but they're not talking bytes
<exarkun>
I said "statistically" :)
not_a_robot has joined #nixos
<exarkun>
everyone knows that means "I'm about to talk about of my ass" right
<pie_>
statistically is a dangerous word
<pie_>
yeah as i was sayin xD
<gchristensen>
I think we have ~250,000,000 objects
phreedom has quit [Remote host closed the connection]
szicari has quit [Ping timeout: 250 seconds]
<exarkun>
Then it's barely 500 years before one of them is lost
<gchristensen>
oh dear
kleisli_ has quit [Ping timeout: 248 seconds]
<exarkun>
Do you think S3 will refund the project when WW3 destroys all of their storage in 3 years
o1lo01ol1o has quit [Remote host closed the connection]
<exarkun>
will it be a complete refund or a prorated refund
<gchristensen>
prorated on the 500yr?
<pie_>
read that as perforated at first
<not_a_robot>
Does anyone have experience with mkWrapper and neovim? Specifically in getting neovim able to find python3
<exarkun>
gchristensen: right
<gchristensen>
:D
o1lo01ol1o has joined #nixos
<exarkun>
pie_: I take it as a given there will be perforation at that point.
<emily>
S3's SLA is nonsense
freeman42y has quit [Ping timeout: 250 seconds]
<emily>
an affront to statistics and actuarial science :p
szicari has joined #nixos
<gchristensen>
is it?
<emily>
exarkun: actually, when us-east1 gets wiped out by WW3, it'll just be reported as an elevated error rate in some regions
<emily>
gchristensen: there is no way to be 99.9999999% confident in the reliability of data in S3. "jeff bezos goes insane and personally deletes every single bit of data" is more than 0.00...01% likely
<emily>
unknown unknowns and all
orivej has joined #nixos
___laika has joined #nixos
<exarkun>
I have certainly experienced AWS being less than ... completely transparent in their admission of service degradation.
<emily>
you can't just extrapolate "we've been doing okay for 13 years" to "the universe will end before we renege on our promise to keep your data safe and sound"
tgamblin has joined #nixos
<exarkun>
emily: But of course the SLA isn't a promise of what they will actually provide
<exarkun>
emily: It is part of a contract for getting $7 refunded when they lose your entire enterprise's dataset.
<emily>
right
<pie_>
i have someone that told me they had data corruption and amazon told them nope everthing is fine *shrug* who to trustttt
<emily>
in practice keeping copies in a couple of clouds probably gives you enough 9s.
<emily>
but AWS would never suggest that, of course :p
* pie_
wonders how to debug nix with gdb so that variables actually show up...
<pie_>
(the c++)
<exarkun>
did you rebuild it with debug symbols?
<pie_>
probably not, but i think i tried that before and still didnt manage it
o1lo01ol1o has quit [Ping timeout: 244 seconds]
<pie_>
i wish we have a DEBUG ALL THE THINGS flag >.>
<pie_>
(that actually worked (?))
<exarkun>
keen feature for nixpkgs would be a semi-automatic "debug" output for all derivations w/ debugging enabled for the whole closure
* pie_
tries to rebuild with debug symbols
<nh2>
pie_: I'm pretty sure I did that once with gdb. Double-check whether -O is passed or something similar that destroys the debugging experience. For example, some of the hardening flags do that; you need to disable that