<iqubic>
It is in default.pa, but it should be in daemon.conf
<iqubic>
IDK how to tell nixos to put it there.
<clever>
iqubic: look on the nixos option search, what options does pulse have?
<iqubic>
Got it.
<iqubic>
I'll just change this to use daemon.config here.
doyougnu has quit [Ping timeout: 268 seconds]
<clever>
and daemon.config doesnt take a string, but a { flat-volumes = "no"; }
<iqubic>
Interesting.
<iqubic>
So it takes an attribute set?
<clever>
yeah
<iqubic>
Cool, seems to be working now.
<iqubic>
not sure if this solves my initial problem, but we'll see.
<iqubic>
Initial problem meaning, the problem that lead me to change my PA config.
perique has joined #nixos
vk3wtf has quit [Ping timeout: 250 seconds]
<iqubic>
well, that just made my issues worse.
drakonis has quit [Quit: WeeChat 2.3]
<clever>
what was the issue?
Guest18885 has quit [Ping timeout: 245 seconds]
drakonis has joined #nixos
<iqubic>
Firefox uses pulseaudio to play sound. When watching YouTube I get this weird issue where pausing the video, or even just skipping a few seconds ahead by using the video scrubber causes my sounds to get set to 100%, which is way too loud.
<iqubic>
clever: What exactly does the flat audio thing do?
<drakonis>
but have you tried reading the explanation
bfrog has joined #nixos
bfrog has quit [Client Quit]
halfbit has quit [Quit: WeeChat 2.3]
<drakonis>
hmm
halfbit has joined #nixos
<drakonis>
i have a problem now
<iqubic>
Yes.
praetorg has quit [Remote host closed the connection]
<iqubic>
I'd like to know the problem.
<drakonis>
error: value is a set while a list was expected, at /nix/store/bdjgj1dawnir67r25hak53kbvgfm2rq6-nixpkgs-19.03pre166987.bc41317e243/nixpkgs/pkgs/top-level/stage.nix:208:8
praetorg has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fhPdA
jluttine has joined #nixos
<iqubic>
Without context I'd say that you gave a set when a list was expected.
<CMCDragonkai>
Anybody find that when installing fonts on NixOS, while other apps are able to use them, Chrome doesn't seem to realise that the fonts exist?
<CMCDragonkai>
I mean installing them in the HOMEDIR.
<_deepfire>
ah, so if something isn't needed by Cabal, it shouldn't need to be null'd?
<_deepfire>
regardless of presence in ghc://libraries/
<clever>
_deepfire: yeah, but it also shouldnt be baked into the compiler
<clever>
_deepfire: i think you want to check what ghc-pkg list gives, when using just the bare ghc derivation
<clever>
and thats what should be null
<_deepfire>
yes, I see..
<_deepfire>
makes sense
<hodapp>
ugh, found yet another thing that's broken in Cura
<clever>
of note, stack2nix undoes the compiler-configuration
revtintin has joined #nixos
<clever>
_deepfire: so you wind up with duplicate packages, and sometimes nix-shell grabs the wrong one, resulting in errors like expected Data.String, got Data.String
n0qqe has joined #nixos
<_deepfire>
clever: that's because of overridden GHC's package?
<_deepfire>
*GHC's own package
<clever>
stack2nix undoes the fix, and then defines duplicates that arent null'd out
<clever>
and then its somewhat random which version gets picked
<_deepfire>
yuk
<_deepfire>
so stack2nix should check for 'null'-ness
<_deepfire>
well, that combined with unattended upgrades + regular reboots would certainly make any attacker's life less pleasant and reduce their ability to propagate further.. esp. if they had to use a time-costly exploit to get in in the first place
<_deepfire>
but one'd have to make the reboots rather frequent, too, to make it effective, I guess..
eadwu has quit [Quit: WeeChat 2.3]
halfbit has joined #nixos
jackjennings has quit [Quit: Connection closed for inactivity]
sigmundv_ has quit [Ping timeout: 240 seconds]
maximiliantagher has joined #nixos
lassulus has quit [Ping timeout: 240 seconds]
hellrazo1 has quit [Ping timeout: 250 seconds]
maximiliantagher has quit [Ping timeout: 245 seconds]
maximiliantagher has quit [Ping timeout: 250 seconds]
<ottidmes>
gchristensen: how accurate is that, because I just opened a PR of which I am a maintainer (but I was not the first to publish the package, maybe that is the reason)
<wedens>
angerman: btw, is it available in current cabal version or does it still require HEAD?
<elvishjerricco>
Is there a way to use nightly rustc with carnix right now?
doyougnu has quit [Ping timeout: 245 seconds]
maximiliantagher has quit [Remote host closed the connection]
ericsagnes has joined #nixos
sicklork1n has quit [Ping timeout: 272 seconds]
Rusty1 has quit [Quit: Konversation terminated!]
rauno has quit [Ping timeout: 268 seconds]
maximiliantagher has joined #nixos
maximiliantagher has quit [Ping timeout: 240 seconds]
vk3wtf has quit [Ping timeout: 240 seconds]
sicklork1n has joined #nixos
<wedens>
I haven't used nixops, so I have a newbish question: if I deploy a system using nixops, can I use it as regular nixos (doing nixos-rebuild, etc) or do I have to always use nixops to change something?
<clever>
wedens: you will always need to use nixops, nixos-rebuild ignores the nixops config
<clever>
wedens: but if your config is simple enough, you could copy it back into the configuration.nix and go backwards
<wedens>
clever: makes sense. thanks. am I right that nixops builds the closure locally and copies it to nixops-managed machine?
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @r-ryantm to master « duply: 2.1 -> 2.2 »: https://git.io/fhXvy
<wedens>
clever: no sudo? also can I switch existing nixos installation to nixops? will it just use configuration provided by nixops without collisions?
<clever>
wedens: it will switch over to whatever nixops provides, and if you get some bits like fileSystems or boot.loader wrong, it may not boot right, but it should still work with the old generations in the bootloader
<clever>
wedens: i have switched my nas and router from nixos-rebuild to nixops
palo has quit [Ping timeout: 244 seconds]
palo1 is now known as palo
<wedens>
are there any drawbacks to managing localhost via nixops instead of nixos-rebuild?
<clever>
wedens: cant really thing of anything
<wedens>
good. thanks clever
jtojnar has quit [Quit: jtojnar]
<Izorkin>
Please check this PR #53245 #54475 #51902
xkapastel has quit [Quit: Connection closed for inactivity]
<Myrl-saki>
[myrl@myrl:~]$ du -chs /nix/store/*
<Myrl-saki>
-bash: /run/current-system/sw/bin/du: Argument list too long
<Myrl-saki>
Well, crap.
<clever>
Myrl-saki: du recurses for you, omit the * and change --max=
<Myrl-saki>
Yeah, -d seems to work.
<Myrl-saki>
[myrl@myrl:/nix/store]$ du -hd 1 ./ | sort -h
<Myrl-saki>
Can you get file length from Nix?
<Myrl-saki>
Oh, apparently readFile also requires the file to be a Nix string, so `builtins.length (builtins.readFile ...)` is not only inefficient, it won't even work.
<clever>
Myrl-saki: why do you need file length in nix?
<Myrl-saki>
clever: Basically, build an image for {1,2,4}GB.
<clever>
Myrl-saki: cant you check the size at bash time, and adjust it there?
sinner has joined #nixos
sinner is now known as Guest64196
<Myrl-saki>
clever: Yeah, that's what I'm doing right now, and it requires me to populate the file using `dd`, because `fallocate` won't work under runInLinuxVM, so I have to dd if=/dev/zero, which takes 50 seconds, then dd if=file, which takes another 50 seconds.
<clever>
Myrl-saki: what about the preVM script?
<Myrl-saki>
Hm.
<Myrl-saki>
That might actually work.
<Myrl-saki>
clever: OTOH,it still might not work if /nix/store is under FS.
<Myrl-saki>
ZFS
<Myrl-saki>
[myrl@myrl:~]$ fallocate -l 1G ./test
<Myrl-saki>
fallocate: fallocate failed: Operation not supported
<clever>
try truncate, rather then fallocate
<clever>
fallocate doesnt mean much on zfs
<Myrl-saki>
Hm. Will do.
<Myrl-saki>
Oh cool.
<Myrl-saki>
Oh wait.
<Myrl-saki>
Oh yeah.
<Myrl-saki>
[myrl@myrl:~]$ truncate -s 2G ./test
<Myrl-saki>
[myrl@myrl:~]$ du --apparent-size -chs test
<Myrl-saki>
2.0G test
<Myrl-saki>
Works! \o/
<clever>
truncate just sets the size, without allocating any blocks
<Myrl-saki>
Right.
<clever>
so its basically instant, on FS's that support space files
<clever>
Myrl-saki: in the case of zfs, all blocks on disk are immutable, so pre-allocating null blocks with fallocate is pointless, since your doing to be getting even more blocks when you have real contents
<Myrl-saki>
Right.
<Myrl-saki>
Ah. This one also works.
<Myrl-saki>
[myrl@myrl:~]$ fallocate -d -l 2G test
<Myrl-saki>
Should be equivalent, I believe.
<Myrl-saki>
Okay, I'm stupid.
<Myrl-saki>
[myrl@myrl:~]$ fallocate -d -l 2G test
<Myrl-saki>
fallocate: cannot open test: No such file or directory
Makaveli7 has quit [Quit: Leaving]
<Myrl-saki>
I just have to optimize dd now.
<teto>
dtz: I did a fetch all on your nixpkgs repo, so many branches xD !
<EffSquared>
Hi folks. Could anybody with access to the build bot run the two build command I proposed for my pull request here: https://github.com/NixOS/nixpkgs/pull/54626
<{^_^}>
#54626 (by FlorianFranzen, 2 days ago, open): buid-support: Add Ubuntu bionic base images
<EffSquared>
Much appreciated. Thanks
viric_ has joined #nixos
elgoosy_ has quit [Quit: Leaving]
<Myrl-saki>
Oh wow, this is actually kinda funny.
<Myrl-saki>
I have my squashfs.img dependency checking on /nix/store, because I disabled compression.
<Myrl-saki>
Yep, I didn't realize that this actually "messes" with the /nix/store.
<clever>
nix checks for deps, by serializing the entire $out (via nix-store --dump $out), and then just doing a dumb string search for the hash from every input
<elvishjerricco>
I've always been a little dissatisfied by that. I don't have a better solution, and it certainly works great most of the time, but it is kinda hacky
<Myrl-saki>
I guess the nice thing about this is that propagated-build-inputs is just a stdenv thing, I believe?
civodul has joined #nixos
<elvishjerricco>
Yea
<clever>
Myrl-saki: yeah, as the stdenv processes every buildInputs entry, it checks the nix-support on them, and can recursively follow the propagated ones
And has joined #nixos
<Myrl-saki>
OTOH, it's also techncially possible to add a dependency that's not part of *inputs, right?
<elvishjerricco>
Don't think so. At least not if you include transitive *inputs.
<Myrl-saki>
Though that should be catchable in buildtime, and you can't really do that randomly.
<clever>
Myrl-saki: nix requires that it be part of your input closure
<Myrl-saki>
Ah.
<Myrl-saki>
It will error if there's a spurious dependency on the output?
<clever>
nope
<clever>
but the sandbox will also not provide a copy at buildtime
<elvishjerricco>
at least with sandboxing on, I doubt it's even possible to get a runtime dep that isn't a build time dep
<Myrl-saki>
Right.
<clever>
and it wont be counted as an input, so anything down the road also wont get it at buildtime
<elvishjerricco>
It's not visible at all, and you can't generate the hash
<clever>
elvishjerricco: ive seen people quote old storepaths in nix expressions
<elvishjerricco>
Ah, that's a good point
<Myrl-saki>
sphalerite: Ping. What do you mean by "will allow accessing the encrypted data even after the passphrase has been changed, simply by using the old header"?
<clever>
Myrl-saki: yeah, if you have a copy of the old header, you can undo a password change
<Myrl-saki>
Aw
<Myrl-saki>
So use the old header on the new data?
<elvishjerricco>
Rewriting the whole disk would be the only theoretical way to fix that one, right?
<Myrl-saki>
And you can do it even if the data has been changed after the password change?
<clever>
Myrl-saki: there are backup and restore commands
<clever>
Myrl-saki: yes
<clever>
elvishjerricco: no
<clever>
luks has a single master key, used for all disk encryption, that can never change (enless you want to re-encrypt the entire disk)
<elvishjerricco>
right
<clever>
that master key, is then encrypted with your password, to fill a keyslot
<elvishjerricco>
by "rewriting the whole disk", I meant re-encrypt
<clever>
and you can potentially have multiple copies of the same master key, encrypted by different passwords
<clever>
the normal commands for changing a pw, just change the keyslot that protects the master key, to enrypt it with a new pw
bananchick_pasha has joined #nixos
<Myrl-saki>
Right. The key used for encryption is still the same.
<clever>
elvishjerricco: yeah, at that point, your better off just making a new luks device, and copying the contents over
<Myrl-saki>
IIUC?
<clever>
Myrl-saki: yep
<Myrl-saki>
I should have expected that.
<elvishjerricco>
Sidenote: I was looking at the LUKS spec to try and learn how it avoids leaking which blocks are duplicates of each other, but the spec doesn't actually specify how the configured cipher is applied to the actual data
<clever>
elvishjerricco: i suspect it uses the block# as a nonce
<elvishjerricco>
An easy way to avoid leaking duplicates is to just xor the address of the block onto it before encrypting, but I have no idea if that's actually secure, so I wanted to see what LUKS does
phry has quit [Quit: WeeChat 2.1]
<clever>
elvishjerricco: so the real encryption key, is a combination of the masterkey, and the block#
<elvishjerricco>
Yea that could do it
<elvishjerricco>
Though
<clever>
one simple (but probably not the safest) way to do such a thing, is just sha256(realkey+block#)
phry has joined #nixos
<clever>
concat, not add
<clever>
so you have no way to undo it, and no way to relate the keys from 2 blocks
<elvishjerricco>
(I'm reading a lot about crypto lately, so I still do not know what I'm talking about :P)
And has quit [Remote host closed the connection]
<Myrl-saki>
Hm.
Makaveli7 has joined #nixos
<clever>
elvishjerricco: go over the libsodium docs and try playing with their functions
<elvishjerricco>
But the definition of a secure block cipher only specifies security for the same key
<elvishjerricco>
It doesn't specify that a key can't be leaked if it's reused with modifications
<Myrl-saki>
What Nix module should I use for a mutable flash drive LiveCD?
<clever>
elvishjerricco: oh, that reminds me of the days when i was trying to get tv capture working on my hdtv setup
<elvishjerricco>
course we don't technically have provably secure block ciphers yet AFAIK
<Myrl-saki>
sd-image?
<clever>
elvishjerricco: after a bit of research, i discovered a flaw in mpegts crypto
<clever>
elvishjerricco: it doesnt use a nonce, and h264 has a lot of padding in the form of nulls
<Myrl-saki>
Technically, it should now be live flash drive and not liveCD. :P
<clever>
the paper i found, says that you can just search the mpegts stream, for 2 blocks with identical ciphertext
<clever>
and because the key is only 16bits, you can then lookup the key, using a rainbow table
<elvishjerricco>
lol
<elvishjerricco>
that's pretty bad
<clever>
the problem, is that the lookup could take up to 3 minutes (with "modern" hardware at the time of the paper)
<clever>
and the upstream source can be rotating the key as often as every 5 seconds if they wanted to
bananchick_pasha has quit [Quit: Quit]
<clever>
mpegts crypto, involves the sending a black-box bytestream to the CAM, which will then spit out the current 16bit key
<clever>
that bytesteam can then do anything, from just giving it the protected key, to giving the CAM firmware updates
<clever>
in the case of satelite systems, the CAM is on a smartcard
<clever>
elvishjerricco: ive even heard stories of hacker level tricks being pulled by sat companies
<clever>
shipping out firmware updates in 20 pieces, that all look harmless on their own
<clever>
then the 21st piece assembles them all, and boom, bricks all the illegal CAM's
<MasseR>
Can I give priority to a cache? I have a local cache server and I would like nix to fetch from there primarily and fall back to cache.nixos.org
<wedens>
arianvp: there is also fonts.fontconfig.confPackages
o1lo01ol1o has joined #nixos
jasongrossman has joined #nixos
seanparsons has quit [Ping timeout: 245 seconds]
maximiliantagher has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
regina has joined #nixos
<Aleksejs>
Hello, I've run rebuild switch --upgrade yesterday and now teamviewer is broken - I don't see my computers, it asks to login, but both login and password boxes are not available
moredhel_ has joined #nixos
maximiliantagher has quit [Ping timeout: 240 seconds]
<regina>
i'd like to open a nix-shell with the androidsdk, how can accept it's license? the help message tells me to use the config option android_sdk.accept_license, however the package seems to use licenseAccepted?
seanparsons has joined #nixos
<EffSquared>
Aleksejs: As a first step I would use nixos-rebuild --rollback switch to see if that brings back a working teamviewer.
<jasongrossman>
regina: I'm not positive, but I'd try nixpkgs.config.android_sdk.accept_license = true; first.
<jasongrossman>
regina: I use that with programs.adb.enable = true;
<azazel>
how it's possible to test failures on derivations? I've a situation where I want to test if a "checkPhase" of a derivation actually works so I've added a test for it using "tryEval", but the evaluation fails hard because of the dependency failure between the the tested derivation and the tests one... any idea
<azazel>
?
evck has quit [Ping timeout: 244 seconds]
<regina>
jasongrossman: thanks
brejoc has joined #nixos
sicklork1n has joined #nixos
PeterHK has joined #nixos
<PeterHK>
someone has a dockerTool container build for django app?
<teto>
timokau[m]: mbsync is faster than offlineimap I think so stay with it
maximiliantagher has quit [Ping timeout: 246 seconds]
Makaveli7 has quit [Quit: Leaving]
<jboy>
teto: Yes, that's what I want to do. But I'm getting the same "GSSAPI Error" that timokau[m] reported getting. Googling for that error turned up the IRC logs.
<timokau[m]>
The issue is actually that cyrus-sasl will try to use gssapi support when its available (which it is on nixos). That will fail for some reason.
<timokau[m]>
If its not available, it will fall back to plain (e.g. communicate unencrypted). That is not nice for security, but at least it works.
<timokau[m]>
So what I've been doing for now is override mbsync to disable gssapi support
<jboy>
timokau[m]: oh, I see. Thanks for pointing that out!
<timokau[m]>
jboy: You're welcome! I spent way too much time debugging this, so I'm happy if I can save you that work
<jboy>
timokau[m]: Well thanks to you I may get some actual work done today
<timokau[m]>
jboy: Still be careful with that "solution". I've settled on it because I'm not willing to invest even more time in it, the relevant mail account is "just" my uni mail and I mostly sync in secure networks. Still not a nice solution.
<jboy>
Yes, I see. Does your university use Exchange?
<jboy>
ah. It only fails with my university account, which is an Exchange server; fastmail works fine.
<timokau[m]>
jboy: Maybe leave a message at the upstream issue I liked, mentioning that you have the same problem. Might increase the chances of someone with the relevant knowledge looking at it.
maximiliantagher has quit [Ping timeout: 240 seconds]
<symphorien>
what does nix-info say ?
brejoc has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
brejoc has joined #nixos
<{^_^}>
[nixpkgs] @dtzWill opened pull request #54885 → chelf: init at 0.2.2 → https://git.io/fhXzf
xkapastel has joined #nixos
oida has quit [Remote host closed the connection]
avn has quit [Remote host closed the connection]
fusion809 has joined #nixos
lokado has joined #nixos
<lokado>
Can I use write a script with nix-shell shebang and make it executed in a FHS environment created by buildFHSEnv?
<{^_^}>
[nixpkgs] @markuskowa opened pull request #54886 → openblas: make the optimization target overridable → https://git.io/fhXgr
<kai_w>
how can I get extra files into /run/opengl-driver/lib (or otherwise onto LD_LIBRARY_PATH)? Specifically I want to use libcuda for compute but integrated graphics for X11
brejoc has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<noonien>
i always wondered, what does configuration.nix change on the os
ng0 has quit [Quit: Alexa, when is the end of world?]
<tilpner>
etu: Can I add your repo to search.tx0.co for indexing?
<noonien>
i know different modules behave differently, but
drewr has quit [Remote host closed the connection]
<etu>
tilpner: Sure
endformationage has joined #nixos
<noonien>
what happens if i add a file to /etc, it's kept, right?
<jasongrossman>
noonien: configuration.nix is a function than NixOS uses to build everything it needs to boot.
<etu>
tilpner: My repo is a bit weird because I have my own modules, but I find it pretty well organized
<jasongrossman>
noonien: Maybe, but it could get overridden, so it's best to let NixOS manage the whole of /etc. At least at first.
<noonien>
ah, i see!
<noonien>
so there's no magic
<jasongrossman>
There's lots of magic!
<tilpner>
etu: Those are the most interesting ones :)
drewr has joined #nixos
<inf>
hey all! I'm trying to change boot.plymouth.logo, but it doesn't seem to work. /etc/plymouth/logo.png → /etc/static/plymouth/logo.png → /nix/store/*-custom.png links are properly created, but breeze-plymouth still displays default nixos logo.
<noonien>
i'm not sure how they're selected on install
<jasongrossman>
noonien: Are you actually worrying about multiple machines simultaneously? I'm not sure people are answering the right question for you.
<jasongrossman>
noonien: Even if you are, it might not be the best way for you to start.
<noonien>
No, I'm not doing anything of the sorts, i just stumbled opon this repo and was curious
<jasongrossman>
noonien: OK, just go back to walking instead of running if you get frustrated!
ryantrinkle has quit [Ping timeout: 268 seconds]
init_6 has quit []
<noonien>
Of course, i'm not actually going to do multiple machine setups atm, but i find NixOS/Nix fascinating and don't mind learning extra stuff
<Dagger2>
there should be no good reason to do so -- if you've got a problem with v6, fix the problem rather than immediately abandon it
<noonien>
hmm, is there any way to install go packages as binaries from the nixos configuration?
<inf>
Dagger2: yeah... Sadly I use nodejs which, when sees ipv6 network only, keeps on asking DNS for AAAA records only, even when ipv4 shows up :>
<noonien>
basically, `go get` then copy the binary
<noonien>
i know i can create a nixpkg, was wondering if there's an easier way
<inf>
Dagger2: on the other hand, I don't want to wait for network presence, because *resons* (we use network only for management purposes, and software should work without any network connection ~just fine)
<inf>
inb4 migrate backend service to ipv6. yeah, that might be a plan... :P
<inf>
oh, and in the end I can't guarantee ipv6 network will actually work in places where this will be deployed anyway.
agander has joined #nixos
<Dagger2>
"a working network" seems like a reasonable thing to add to the requirements list
dermetfan1 has quit [Quit: WeeChat 2.3]
<Dagger2>
that nodejs behavior does sound pretty broken though :/
<inf>
it no longer is, when you deploy it on someone's elses premises/network.
<inf>
and, yeah, nodejs is pretty broken in that regard, have to research it further, but that was a simplest fix i could get when i had to get it working *now* :)
vidbina_ has quit [Ping timeout: 240 seconds]
<inf>
ie. by fixing one issue, i stumbled upon a different one, but not in my software this time.
<inf>
s/fixing/working around/g
<jasongrossman>
noonien: Are there not Nix derivations for the packages you want?
<noonien>
i just searched, and there are!
<jasongrossman>
noonien: Yay!
dermetfan has joined #nixos
dermetfan has quit [Client Quit]
dermetfan has joined #nixos
<noonien>
now i'm really looking forward to creating my own derivations
vidbina_ has joined #nixos
<Dagger2>
inf: what part of nodejs is it that's doing this? dns.lookup() looks like it *ought* to work...
Khetzal has quit [Remote host closed the connection]
<inf>
Dagger2: i don't really remember, have to sit down and research it again. we are specifically using https://www.npmjs.com/package/ws in electron backend process :P
reinhardt has quit [Quit: Leaving]
maximiliantagher has joined #nixos
<kenogo>
I'm currently packaging a printer driver, and it needs cups to be able to call a shell script. Right now the shell script is located in the same /nix/store/.../ directory as the other driver files. It starts with a #!/bin/sh shebang. However, cups fails running it with an exec format error
<andi->
yeah, if you choose to use nixpkgs for that... I would hope that in order to move on we would have a community maintained shim-layer that bootstraps a new enough nix in case we want to break semantics.
<gchristensen>
not latest, just enough of an old one
<apaul1729>
so i'm currently writing a nixpkg for darwin that requires xcode. i have xcode 10.1 installed, and i've added it with `nix-store --add-fixed --recursive sha256 /Applications/Xcode.app` . however, when I try to enter a shell with `nix-shell -p darwin.xcode` it complains about not having specifically the Xcode_8.2.xip file
<andi->
while we are at the topic: I recently tried to install nix from nixos.org on Debian.. it was less then comfy.. very bumpy road :/ I should have opened issues back then but I only wittnessed it on other peoples computers.
<apaul1729>
any way i can work around this? either by specifying the /nix/store path with nix-shell? the darwin/xcode/default.nix in current master even has a line for version 10.1
Leo` has joined #nixos
<timokau[m]>
gchristensen: But if users would need to go through 10 bootstrapping steps until they can use the latest nixpkgs, that wouldn't be much of an improvement
<andi->
timokau[m]: I suspect you'd only need to compile one - if at all. It could already be cached by hydra at that point. You just need a compatible expression that points you to the correct NARs
<timokau[m]>
andi-: Yeah if there is an expression that maintains backwards compatibility that would be fine. But we'd probably need to think about that *before* the inclusion, wouldn't we? Otherwise the packaged nix wouldn't know where to look for such an expression.
<gchristensen>
its okay, we have about 50 years before it hits stable
<andi->
timokau[m]: IMO that is the job of the guy maintaining the Debian package. He is not in this alone but I hope he knows what it means..
<andi->
gchristensen: 6ish :P
<timokau[m]>
gchristensen, andi- : Good point :D
<andi->
also it doesn't mean that the version that is there will be in stable.. There can be more updates before that..
<andi->
It just should be updated before the unstable -> testing / testing -> debian11 freeze/migration
<andi->
iirc experimental -> testing is automated (these days)
<timokau[m]>
But still, landing in unstable would trickle down to ubuntu right? I suspect not many people are running debian stable on end-user pcs anyway (probably being naive there)
<andi->
if I ask people around me you would be suprised ;)
<andi->
(more like buster but yeah)
ryantrinkle has joined #nixos
erasmas has joined #nixos
linkrage has joined #nixos
<ejpcmac>
Hi! If someone is merging here, can you please merge the PR #53959 to update Elixir to 1.8.0? It’s been 15 days since I submitted it.
Leo` has left #nixos ["ERC (IRC client for Emacs 26.1)"]
deni has left #nixos ["WeeChat 2.3"]
hoijui has joined #nixos
Ariakenom has quit [Quit: Leaving]
maximiliantagher has joined #nixos
hyper_ch2 has quit [Quit: Page closed]
maximiliantagher has quit [Ping timeout: 240 seconds]
<Mic92>
nobody has packaged eclim by accident?
<Taneb>
It'd be quite impressive to package something by accident, I think
<apaul1729>
woops i slipped and packaged clang, sorry about that
<Taneb>
Mic92: what is eclim?
<Taneb>
(I have enough trouble packaging things intentionally)
<Mic92>
lsp for java
<Mic92>
mhm does not really implements the lsp protocol so
<Taneb>
There seems to be packaged an emacs mode called eclim
agander has quit [Ping timeout: 246 seconds]
oskar has joined #nixos
hoijui has quit [Ping timeout: 264 seconds]
<oskar>
Hi, Nix newbie question: Whenever I do sudo whatever for a nix-env package, the command is not found. I guess my .profile is not read (or parsed differently) for root. How do I sudo correctly with nixpkgs (on debian on crostini on chromeos)?
<wedens>
Mic92: eclipse jdt does
<symphorien>
oskar: you can see the modifications nix needs in your .bashrc iirc
<symphorien>
copy this in root's .bashrc
<oskar>
thanks, will try that
<wedens>
Mic92: also there was some lsp server based on intellij
<pie__>
egh. to tired for it right now but i guess i will have to try packaging it
<pie__>
unless someone is super bored :[
<pie__>
* :p
<tilpner>
If you don't need the GUI, maybe you can get it working with services.xserver.wacom.enable and environment.etc."X11/xorg.conf.d/70-wacom.conf"
<{^_^}>
[nixpkgs] @kalbasit pushed to master « nixos/pam: refactor U2F, docs about u2f_keys path (#54756) »: https://git.io/fhX7f
<tilpner>
What's wrong with your wacomScript, infinisil? c.c
<pie__>
tilpner, yeah i did notice those
<pie__>
tilpner, what i dont understand, is that the tablet works, so whats this other stugg?
<pie__>
stuff
<pie__>
it was plug n play
<linkrage>
lewo, I'm starting to test podman today. Installed latest podman from the unstable channel. I'll keep you up to date with my testing/findings etc. :)
<{^_^}>
[nixpkgs] @grahamc opened pull request #54892 → elixer: remember where a package is defined → https://git.io/fhX7l
<tilpner>
I don't know, I don't use drawing tablets
<goibhniu>
pie__: with pressure sensitivity and all?
vidbina_ has quit [Ping timeout: 246 seconds]
<pie__>
goibhniu, yup
<goibhniu>
beautiful
<pie__>
and xsetwacom list devices lists nothing
dbmikus has quit [Ping timeout: 246 seconds]
dbmikus has joined #nixos
`_ has quit [Ping timeout: 240 seconds]
<dmj`>
Is there something like a postLibraryBuild phase in haskell nixpkgs infra
hexagoxel has quit [Quit: ZNC 1.6.5+deb1+deb9u1 - http://znc.in]
dbmikus has quit [Ping timeout: 240 seconds]
hexagoxel has joined #nixos
dbmikus has joined #nixos
ottidmes has joined #nixos
chessai has joined #nixos
<gchristensen>
I need a person who: (1) uses Nix (2) on Linux (not NixOS) (3) with a Linux Kernel of 4.19 or older. is there anyone available? we think we've found a bad kernel regression, and need to see if it replicates off NixSO
<tilpner>
Spin up a quick VPS?
<gchristensen>
we could, but I'd rather find someone else who already has it :P
<aszlig>
best would be 4.18 *and* 4.19
<tilpner>
Oh, dply.co is dead
<aszlig>
tilpner: we'd need nested virtualisation then
Ariakenom has joined #nixos
<Taneb>
gchristensen: I'm currently on Ubuntu 4.15 on my work computer
<ottidmes>
I have two questions about Rust on NixOS, I build my project fine on the mozilla overlay nightly, but when I tried to build it from nixos-unstable nightly I got an error that "no matching package named `dialoguer` found". It said it searched the registry https://github.com/rust-lang/crates.io-index but when I search that registry myself I can find serveral packages mentioning different versions of it, how come
<tilpner>
aszlig: Ah. DigitalOcean has that, if you end up needing to go there
<ottidmes>
that it still is reported as not found? I made sure it was not a version issue by using an older version mentioned on the registry
apaul1729 has quit [Quit: Ping timeout (120 seconds)]
<aszlig>
it will need to recompile the kernel, so it will take a while
<aszlig>
immae: oh, very good
<gchristensen>
immae: please also run that test expression ^
<aszlig>
so having this tested on 4.15 and 4.19 is even better
<immae>
do I have to be up to date with nix too?
<immae>
(current:
<Taneb>
aszlig: just with nix-build, yeah? Doing that now
<immae>
(current version: 2.1.3)
<aszlig>
immae: that shouldn't matter
<immae>
ok
<aszlig>
Taneb: thanks
<immae>
running on my side too
<Taneb>
aszlig: how long do you expect this to take?
elgoosy has joined #nixos
<aszlig>
Taneb: depends on your machine, because as said, it needs to compile kernel 4.19
<ottidmes>
my second Rust question is why clippy seems to have dissapeared from the mozilla overlay, I just copied the example shell.nix and it has cargo-clippy in /bin but when I ran nix-shell again today it rebuild it, because I had it pointed to master.tar.gz rather than a specific commit as per the example, but now its missing, I tried downgrading and pinning it, but its unclear to me how to get back to the old
<ottidmes>
situation. I just rooted the old rust build with cargo-clippy for now, but I want to be able to reproduce this env in the future
<immae>
aszlig: I’m curious, what does nixosTest do? Do you have documentations about it?
<aszlig>
immae: it's basically similar to <nixpkgs/nixos/tests/make-test.nix>
<vdemeester>
I think I may had that upstream (aka on Nixos/nixpkgs) :)
<linkrage>
vdemeester, I see thank you!
<immae>
aszlig: I added the sandbox=relaxed rcently, and I assure you I had to rewrite a lot of my custom packages because of this, so it seems to be sandboxed :)
<aszlig>
immae: okay, so the sandbox works...
<vdemeester>
linkrage: interesting… you installed it using `nix-env -i podman` right ?
<linkrage>
vdemeester, exactly
<vdemeester>
linkrage: right, then you need to install `conmon` manually too
<linkrage>
vdemeester, I see I still don't know how to install it in this case, as nix-env -i conmon does not find it :)
<aszlig>
immae: do you have anything in the "sandbox-paths" option?
<immae>
if you have a better one to suggest I can put it
<aszlig>
immae: so the build uses the ar command from busybox
<linkrage>
vdemeester, btw since podman depends on runc, I had to install that one as well and symlink just for the test but I'm sure there is a better way. Perhaps you're configuration. I just don't understand hot to use your.nix file
<aszlig>
because it's a builtin of busybox
Makaveli7 has joined #nixos
<vdemeester>
linkrage: ah.. right, I may need to update the package to patch the default path podman looks for conmon & co then
hellrazor has quit [Ping timeout: 240 seconds]
<ar>
i'm not builtin :(
<immae>
yes, but it’s not named ar so I’m surprised
<vdemeester>
linkrage: for the nix file, if you download it, import it in your config and set `programs.podman = true;` it should work
<vdemeester>
:P
<immae>
But again aszlig it’s because I didn’t find a better option, so I’m open to use something else if you have a suggestion
<linkrage>
vdemeester, thanks!
<aszlig>
immae: it doesn't need to be, because when it's running in busybox, the latter will just invoke ar internally without looking up PATH
<immae>
ah ok
<immae>
This is why
<aszlig>
immae: well, you could use something like a statically linked dash if you want to keep it minimal
sigmundv_ has quit [Ping timeout: 250 seconds]
<linkrage>
vdemeester, I don't need docker daemon to run containers right (as that's the idea of using podman/runc)
<aszlig>
immae: i also do have the finished build cached on my hydra, if you're not afraid to fetch binaries
<Twey>
If I use `nix copy --to file://cache` to put the closure of a package into a cache, and I transfer that cache to another machine, can I import all those NARs back into the store somehow?
<Taneb>
aszlig: need anything else from me before I head off?
<aszlig>
Taneb: thanks! so it *is* reproducible on non-nixos systems with <4.19, that at least narrows it down
<aszlig>
Taneb: nope, thanks a lot again :-)
<Taneb>
:) happy to help! Hope you figure it out
<tilpner>
Twey: Did you try --from?
<linkrage>
vdemeester, thanks. How do I "import" your podman.nix ? I'm unable to find info on the topic
<tilpner>
(Also note that nix can copy to ssh targets directly, if that's the transfer method you just used)
<inf>
hmm, going back to broken systemd-networkd-wait-online when `networking.useNetworkd = true; networking.enableIPv6 = false;` are set... it turns out reenabling ipv6 solves a problem on one of my boxes... another one has two network interfaces, and it seems like networkd expects dhcp to be present on both of these with default config?
<vdemeester>
linkrage: once you got the file in you configuration, `imports [ ./podman.nix ];`
<linkrage>
vdemeester, thank you!
<vdemeester>
(then you can use `programs.podman.enable = true`)
<vdemeester>
linkrage: np ;)
aleph- has quit [Ping timeout: 246 seconds]
aleph- has joined #nixos
<Twey>
tilpner: Yes, but it wants a path — there's no ‘install everything in this cache’ option
<Twey>
tilpner: I guess I just need to record the root of the closure and give it that
<linkrage>
vdemeester, I get: attribute 'nur' missing, at /root/podman.nix:27:19
<aszlig>
immae: so, did you have any success with the binary cache? otherwise i could also send you the NAR directly
<vdemeester>
linkrage: ah right
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<vdemeester>
linkrage: removing `nur.repos.vdemeester` should make the thing work
<linkrage>
vdemeester, I see, thanks!
<aleph->
Hmm is there an issue with tarballs.nixos.org?
<vdemeester>
(I did that module before having the package upstream, so I refer to mine :P)
<gchristensen>
aleph-: shouldn't be2, what're you seeing?
<das_j>
Is it possible to extend an existing string? Something like systemd.services.x.execStart = "${pkgs.mkdir}...; ${config.systemd.services.x.execStart}" ?
<vdemeester>
linkrage: there is also some stuff you can clean (like the `[registries.insecure]` as it's related to my home network :D )
<linkrage>
vdemeester, yes I did that already, I just have to see why my other VM with the local registry is not up yet :D
<linkrage>
vdemeester, do you feel podman+kubernetes on nixos combination is ready for production use?
aleph- has quit [Read error: Connection reset by peer]
aleph- has joined #nixos
<aleph->
Okay better.
<aleph->
So yeah is there a pkg besides pkgs.steam I should install?
<aleph->
Since that seems to fail building.
<immae>
aszlig: it’s building again with a dash, thanks (this will help me for my own packages too :) ))
<aszlig>
immae: ah, cool :-)
<immae>
It was easier to recompile it finally :p
<aszlig>
immae: yeah, should at least prevent surprises like this, as busybox actually has a lot of builtins
maximiliantagher has joined #nixos
<aleph->
Hmm doesn't even seem to be a package for steam.
<aleph->
,locate steam
nadley has quit [Quit: No Ping reply in 180 seconds.]
<{^_^}>
Found in packages: metasploit
<aleph->
Well that's unhelpful.
<clever>
aleph-: what does `nix-build '<nixpkgs>' -A steam` say?
<aleph->
Sec
<immae>
aszlig: yes, I was surprised that it could find busybox ar while not having it in the path
<immae>
feels better now :p
nadley has joined #nixos
hakujin has joined #nixos
apaul1729 has quit [Ping timeout: 245 seconds]
<aleph->
clever: Huh that works now.
<aleph->
Or... not.
<hakujin>
attempting to build hydra from source and get a compilation error re: "state.hh:13:10: fatal error: parsed-derivations.hh: No such file or directory"
maximiliantagher has quit [Remote host closed the connection]
maximiliantagher has joined #nixos
_kwstas has joined #nixos
maximiliantagher has quit [Remote host closed the connection]
dsiypl4_ has joined #nixos
* _deepfire
found haskell-modules/generic-builder.nix do a 'mv /nix/store/1gps7cqrr93s5v9clq4slfgrdddxyh5c-data-default-instances-containers-0.0.1/lib/ghc-8.7.20190115/package.conf.d/data-default-instances-containers-0.0.1.conf /nix/store/1gps7cqrr93s5v9clq4slfgrdddxyh5c-data-default-instances-containers-0.0.1/lib/ghc-8.7.20190115/package.conf.d/.conf'
xkapastel has quit [Quit: Connection closed for inactivity]
<_deepfire>
"$pkgId" = ""
<_deepfire>
..and that's due to local pkgId=$( ${gnused}/bin/sed -n -e 's|^id:[ ]\+||p' $packageConfFile )
<_deepfire>
..which, again, I guess, doesn't handle too many spaces
<truthadjustr>
hi i need help to make teamviewer work in my nixos18.09. Followed https://github.com/NixOS/nixpkgs/issues/51320, teamviewer installed fine, it launched the GUI but when I connect, message is "this teamviewer version is incompatible ..."
<{^_^}>
#51320 (by johnalotoski, 8 weeks ago, open): Teamviewer 13 doesn't accept connections on NixOS on 18.09 with KDE Plasma/sddm while logged in
<gchristensen>
immae: how did it go?
<{^_^}>
[nixpkgs] @peti pushed 4 commits to haskell-updates: https://git.io/fh1Tf
<aszlig>
immae: no, that's fine, so i can send the report, thanks a lot
<gchristensen>
yay!
tmaekawa has quit [Client Quit]
<immae>
Good :)
<gchristensen>
sort of a good-news/bad-news yay, but yay!
<johnw>
what is the -j equivalent option for 2.x "nix build"?
<gchristensen>
--max-jobs probably
<johnw>
no, rejects that option
<johnw>
--option max-jobs N?
<gchristensen>
maybe?
<johnw>
i can't seem to find via google the mapping from nix-build to nix build options
phizzz has joined #nixos
freeman42x has quit [Read error: Connection reset by peer]
q6AA4FD has quit [Read error: Connection reset by peer]
freeman42x has joined #nixos
<phizzz>
how do i install glibc-2.27? when i try to rebuild i get this error: undefined variable 'glibc-2' at /etc/nixos/configuration.nix:58:5
q6AA4FD has joined #nixos
<phizzz>
i've tried quoting "glibc-2.27
<phizzz>
and that doesn't work either
agander has joined #nixos
<srhb>
phizzz: You should not be installing glibc in any manual fashion to begin with, probably. What are you trying to achieve?
maximiliantagher has quit [Remote host closed the connection]
<phizzz>
i'm trying to do a source install of ns-3
<srhb>
,libraries phizzz
<{^_^}>
phizzz: Don't install libraries through nix-env or systemPackages, use nix-shell instead. See https://nixos.wiki/wiki/FAQ/Libraries for details.
<phizzz>
srhb: ok i'm struggling here. i'm trying to build this thing that needs ns-3 and waf, i have installed both via systemPackages but now can't figure out how to build it. so i was trying the olde fashioned way
<srhb>
phizzz: Yeah, it's a common point of confusion. The short version is that the "olde fashioned way" usually does not work very well in nix, but using nix-shell or even writing a proper nix derivation for your piece of software does.
snajpa- has joined #nixos
snajpa has quit [Read error: Connection reset by peer]
<srhb>
phizzz: If you can elaborate on the piece of software you're trying to build, someone may be able to assist.
aither has quit [Read error: Connection reset by peer]
<phizzz>
srhb it's some visualization thing, but i'm only the sysadmin.
<gchristensen>
sphalerite: ping?
<sphalerite>
gchristensen: pong
<gchristensen>
sphalerite: mind sending out calendar invites for our meeting tomorrow?
<phizzz>
srhb: i know it needs waf and ns-3, so would i make a derivation that has something like: buildInputs = [ pkgconfig ns-3 waf ]; ?
<srhb>
phizzz: Something like that (ignoring complexities with nativeBuildInputs which you can probably do safely for now) -- there's also wafHook, not sure how well that works, but you may want to grep nixpkgs for that to see examples of usingit.
<phizzz>
man why is this so complicated
<srhb>
It's mostly just different. :) once you've gathered the deps it's usually just like writing a build "the olde fashioned way" but a bit more structured.
aither has joined #nixos
<spacekookie>
Is there some elixir packaged software/module that I can look at to figure out how to deal with the mix dependencies?
<phizzz>
did the nixos devs just miss the days when an os install actually took weeks to complete?
<qyliss^work>
spacekookie: I've definitely seen some in Nixpkgs
<srhb>
phizzz: I'll race you :-P
<gchristensen>
phizzz: my impression is we've miscommunicated something, because it probably isn't as hard as you expect
<gchristensen>
phizzz: just a few days ago I reinstalled my laptop in about 2hrs and was back working, with everything configured
<phizzz>
i might still have some debian slink floppies around
<srhb>
phizzz: I'd say the learning curve _is_ a quite bit steeper when you actually need to build custom software.
<typetetris>
Hi there! Why do I need a special client for cachix? (If it is off topic, what is the right channel for that?)
<spacekookie>
qyliss^work: hmm, I saw that and I'm not quite sure yet
nefix has joined #nixos
<srhb>
typetetris: I think it's on topic. I don't think there's a better answer than "it's not affiliated with NixOS as such"
<nefix>
hi! I'm trying to add a blocklist to my /etc/hosts. How can I read the file from the main configuraiton? I think there's a readFile method, but I can't find anything about it
<phizzz>
ok so if i make this derivation with all the deps i need. then run nix-shell, but how do i pass the default.nix derivation to the shell?
<typetetris>
hmm, it's for distributing nix store paths, isn't it? Why not just use nix-serve or some such?
<qyliss^work>
buildMix doesn't seem to actually be used in nixpkgs at all
<tilpner>
typetetris: The machine that calls the cachix CLI will not do the serving
<tilpner>
typetetris: I think you can use cachix caches without the cachix CLI. Not sure about pushing though
<apaul1729>
phizzz you can do `nix-shell ./default.nix`
<qyliss^work>
nix-shell should use default.nix by default
<qyliss^work>
at least if there's no shell.nix
<typetetris>
tilpner: Okay. I just want to add it as an additional substituter and not install some cli stuff.
ng0 has joined #nixos
<tilpner>
typetetris: You can find the public key on https://foo.cachix.org. Just add that to your nix.conf
<phizzz>
so how do i get waf on the path in this nix-shell?
<phizzz>
symphorien: i'm not sure what you mean about waf not in $out/bin
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @alyssais to master « defaultGemConfig.tzinfo: fix for >=2.0 (#54881) »: https://git.io/fh1mv
<judson>
Thanks
<symphorien>
phizzz: usually executables are in /nix/store/somehash-name/bin/foo
mmlb08 has joined #nixos
<symphorien>
$out is the variable which contains /nix/store/somehash-name when you write the derivation
<symphorien>
for some reason waf was the file $out instead of being a file in the directory $out/bin
slack1256 has joined #nixos
<symphorien>
the command I gave creates a new derivation which symlinks waf to $out/bin/waf
q6AA4FD has joined #nixos
<symphorien>
this way, waf is in the conventional place
<matthewbauer[m]>
you need waf to be in the same directory as you are building
<matthewbauer[m]>
that's why it's in $out
<matthewbauer[m]>
so you should just cp to your directory
<symphorien>
ah ok
<gchristensen>
sphalerite: you intended for it to meet at 1900UTC, yeah?
jonreeve has joined #nixos
agander has quit [Quit: Leaving]
<sphalerite>
gchristensen: yep
<gchristensen>
gerat
<sphalerite>
2000CET :)
<gchristensen>
"timezones are hard"
<sphalerite>
woooooo it worked
<gchristensen>
it did work! it arrived intact!
<sphalerite>
except you can't be sure since the sig won't verify :o
worldofpeace has joined #nixos
<gchristensen>
eh
<gchristensen>
you're right, we should probably cancel
<judson>
Bah. I'm trying to translate my fish config to home-manager, and I'm getting `cannot coerce a list to a string, at /nix/store/x3c0khyy25bqd9339pak7gpsrg9y5q2g-nixpkgs-18.09.1985.749a3a0d00b/nixpkgs/lib/types.nix:200:26`
<ramses_>
Hey guys, is there any way to do something like this in configuration.nix "if hardware-config.nix defines fileSystems.X.device, then set fileSystems.X.options to this value, otherwise don't set fileSystems.X"?
<jonreeve>
symphorien: awesome, that seems to have worked. Thanks!
<ramses_>
It doesn't really matter where the thing is defined, but I'm not sure about the order in which imports are parsed and in my case, it will always be defined in hardware-config
azzknot has joined #nixos
jonreeve has quit [Quit: WeeChat 2.3]
<azzknot>
can anyone verify pmx2pdf is missing from texlive full?
<devalot>
ramses_: The order doesn't really matter since all the config sections get merged at the end and then applied as a whole set.
maximiliantagher has joined #nixos
<devalot>
But an answer to your question would be something like: mkIf config.filesystems...
<ramses_>
devalot: so is there any way to add something to a section only if that section exists? Like, I don't want to create it if it didn't exist yet
cantstanya has quit [Ping timeout: 256 seconds]
<devalot>
ramses_: Probably a better way of doing it is to create a new option that you can set, then only apply the config change if that option is set. Then update all your modules to respect the option.
<ramses_>
devalot: the thing is, that I would prefer to rely on the detection mechanism that creates hardware-configuration.nix, instead of having another option
apaul1729 has quit [Ping timeout: 250 seconds]
hellrazor has joined #nixos
<azzknot>
never mind it was in the scripts they just forgot to link it... I'll do it myself. bye
azzknot has quit [Client Quit]
<devalot>
ramses_: Then I think playing with mkIf and looking the setting up in config might work for you.
<{^_^}>
infinite recursion encountered, at /var/lib/nixbot/nixpkgs/master/repo/lib/modules.nix:336:9
<infinisil>
Yeah that doesn't work
<infinisil>
And that's not what I showed
<infinisil>
The fact that you assign to `options` directly is important
sinner has joined #nixos
sinner is now known as Guest62896
<ramses_>
infinisil: I tried that too, but then I get the error that no device is defined for the mount point... Meaning that he creates the fileSystem."/boot" thingy
<{^_^}>
[nixpkgs] @vcunat pushed to staging « syslinux: fix build with new binutils »: https://git.io/fh1s1
<infinisil>
Ah darn, I didn't test that..
<infinisil>
Sorry
<{^_^}>
[nixpkgs] @veprbl opened pull request #54899 → pythonPackages.pyarrow: redo fix on darwin → https://git.io/fh1s9
<infinisil>
Ah, it will only error when evaluating the system
<infinisil>
Hmm..
<infinisil>
It might actually be impossible with how the module system currently works
<infinisil>
Oh, I have an idea though, *tries*
fendoer has joined #nixos
elgoosy has joined #nixos
fendoer has quit [Remote host closed the connection]
aanderse_ has joined #nixos
<ottidmes>
Just to suggest something that ought to work, but might not be the prettiest, you could always first extract the information you want with a call to import <nixpkgs/config> ... and then call it again supplying the gathered information, since its lazy it should not be that costly, this way you can still leverage the module system to generate the data, but work around the potential circular stuff, maybe instead
<ottidmes>
of <nixpkgs/nixos> you can even call evalModules directly
<ottidmes>
infinisil: so mkOption's do stack, I was wondering about that, I did use the mkOption apply trick a few times, but apply was always undefined, I had not tested whether it would work if it already was defined
<infinisil>
ottidmes: I don't think I'm using any apply here?
<{^_^}>
[nixpkgs] @ElvishJerricco opened pull request #54900 → shellFor: Don't suck in src to compare to deps. [Fixes #51079] → https://git.io/fh1ZJ
<{^_^}>
[nixpkgs] @Mic92 pushed 3 commits to release-18.09: https://git.io/fh1Zk
<ottidmes>
infinisil: Huh, I never said you did, "I did use the mkOption apply trick", as in, I used the trick of setting mkOption for an option that already has an existing mkOption, in my case to define an apply for it
<{^_^}>
[nixpkgs] @Mic92 closed pull request #54657 → Update to Go 1.11.5 → https://git.io/fh6Z8
<ramses_>
infinisil: I'm not sure I understand how to use this... I'm a bit new to nix still
<infinisil>
Sometimes I feel like I'm especially gifted at misunderstanding people, not only in text, but also when talking to people
phizzz has quit [Quit: leaving]
<ottidmes>
is there no way to build a Rust project in Nix without going through a nix-shell? carnix fails on stable and unstable for me, buildRustPackage seems to have a useless registry since even packages that are more than a year old are missing
<infinisil>
ramses_: Ah, I guess you probably have a problem assigning the `options` attribute. You need to put all your standard nixos configuration under the `config` attribute, and then put the `options = ...` part next to it
<infinisil>
ramses_: Or alternatively, split it up into another file
<symphorien>
ottidmes: buildRustPackage has no registry
<symphorien>
change the cargoSha256 and it will refetch everything
<ottidmes>
symphorien: really, is that the cause? if so, some bad UX there
<symphorien>
same trap as changing the url of fetchurl without changing the sha
worldofpeace has quit [Remote host closed the connection]
<ottidmes>
symphorien: any idea how to deal with needing edition 2018 with buildRustPackage?
<symphorien>
you need a recent enough rustc, that's all
apaul1729 has quit [Ping timeout: 240 seconds]
<symphorien>
the edition is declared in Cargo.toml irrc
<ottidmes>
symphorien: hmm, I might be trying to build from stable, let me see if thats the cause
<symphorien>
the mimimum is 1.31 iirc
shibboleth has joined #nixos
ramses_ has quit [Ping timeout: 256 seconds]
<ottidmes>
symphorien: it worked! but if version 1.31 is the minimum then my unstable (less than a month old) has just about support for it since is at 1.31
<ekleog>
Does anyone have documentation about `newScope` and the like? I'm looking for a way to add a patch to LLVM, and it looks… fun
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
henrytill has left #nixos ["ERC (IRC client for Emacs 27.0.50)"]
vk3wtf has quit [Quit: WeeChat 2.3]
vk3wtf has joined #nixos
<ekleog>
wth when I try to `:l <nixpkgs>` then `newScope` in `nix repl` it's showing me it's coming from `pkgs/top-level/splice.nix`, which looks amazingly xorg-specific
<ekleog>
how the hell is that working for llvm?
<{^_^}>
[nixpkgs] @andir pushed 56 commits to staging-18.09: https://git.io/fh1BY
<ekleog>
hmm ok so I guess it's just that xorg gets injected into llvm, because what could possibly go wrong
<ekleog>
Ericson2314: Would you happen to know about the above? you appear to have made quite a few changes that relate to `newScope`