<samueldr>
lambda-11235: looked into it quickly, it is fixed on master, looks like the next unstable bump would have that fix
<bourbon>
I remember reading about a nixos fork that used clojure data structures instead of the nix dsl, but I can't seem to find it. Anybody remember it?
<samueldr>
maybe not the one you're thinking about, not exactly a _nixos_ fork, and not clojure, but guile; that sounds like guix
<bourbon>
actually.... that sounds about right
<bourbon>
thanks, samueldr!
Chiliparrot has quit [Ping timeout: 264 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fj9aV
<mankyKitty>
okay different question .. using `pkgs.haskellPackages.shellFor` .. how does one add a binary to the path for that shell?
<{^_^}>
[nixpkgs] @peterhoeg merged pull request #30712 → systemd user services shouldn't run as root and other "non-interactive" users → https://git.io/vdAdT
<ashkitten>
ugh, i have to pass in overlays to the home-manager config because it doesn't seem to use the system overlays
<ashkitten>
that's really annoying
<ashkitten>
so now i've got `home-manager.users.ash = import ./home-manager config.nixpkgs.overlays`
justanotheruser has quit [Ping timeout: 258 seconds]
<ashkitten>
and in `home-manager/default.nix` i have `overlays: { pkgs, ... }: nixpkgs = { inherit overlays; };`
DariusTheMede has joined #nixos
<DigitalKiwi>
worldofpeace: they did a new release of bitwarden desktop and i updated the nix expression to work with electron_5...still opens the devtools on launch i don't know if that's normal or not though
Izorkin has joined #nixos
DariusTheMede has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @Kiwi opened pull request #65764 → bitwarden: init at 1.15.2 → https://git.io/fj9wm
<DigitalKiwi>
hopefully I don't have any dumb mistakes
arcnmx has quit [Remote host closed the connection]
wfranzini has quit [Remote host closed the connection]
mexisme has joined #nixos
alan2 has joined #nixos
<kandinski>
I asked earlier about binaryCaches preferences, and I think I've found a bug in it. In my case, I had [ "http://dymaxion.local" "https://cache.nixos.org" ] and it fails because I'm no longer in the network where host dymaxion is. It does work and move to the next substituter if I write "http://dymaxion.com" because the failure is when the host cannot be resolved. The domain name esolver enters
<kandinski>
exponential backoff.
alan2 has quit [Client Quit]
alan1 has quit [Read error: Connection reset by peer]
doublex_ has quit [Quit: Quit]
Edes has quit [Ping timeout: 272 seconds]
day|flip has quit [Remote host closed the connection]
DariusTheMede has joined #nixos
<DigitalKiwi>
i end up using --option a lot to get around problems like that (usually i want to change substituters and/or builders)
DariusTheMede has quit [Ping timeout: 244 seconds]
<mbrgm>
hey! what's the preferred way of running nixos off a local nixpkgs clone? I did this for a long time for my nixOPS deployments by setting nixpkgs= accordingly, but now I started using nixos on my desktop and I'm not quite sure what's the best way, getting everything (like search unknown command etc.) right?
<tilpner>
mbrgm: You can set NIX_PATH=nixpkgs=path/to/your/checkout, but that will not get you command-not-found
<tilpner>
I'm not sure how to recreate the same index that Hydra builds, but you can use nix-index/nix-locate for more functionality than command-not-found would give you
<mbrgm>
tilpner: yeah, that's what I did for nixops. now I thought about moving this definition to configuration.nix? there's several ways to achieve what I'm trying to do, so I was interested in about how other people are doing it...
<{^_^}>
[nixpkgs] @FRidh pushed 108 commits to staging-next: https://git.io/fj9rw
<tilpner>
You can set nix.nixPath
<tilpner>
But that's an annoying two-stage thing, you need to pass -I to nixos-rebuild every time you change nixPath
<tilpner>
As for single-stage solutions... they are possible, but more complicated (AFAIK)
DariusTheMede has quit [Ping timeout: 268 seconds]
<ctp>
Hi. I have some trouble to use ZFS encryption while bootstrapping a machine from the minimal install image. Kernel/userland mismatch. I guess there is a simple solution how to reload the kernel modules at this stage without rebooting? Here's my issue https://stackoverflow.com/questions/57321726/nixos-zfs-encryption-in-installer
knupfer has joined #nixos
<ctp>
Does someone have a hint for me? :D I am documenting the entire procedure for other who might run into same issue
veske has joined #nixos
nhey_ has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed commit from @ivan to master « gnupatch: avoid rebuild from c4ce832e PR #65687 »: https://git.io/fj9rD
MmeQuignon has joined #nixos
weedloser has quit [Ping timeout: 248 seconds]
DariusTheMede has joined #nixos
<tilpner>
Uh oh
<tilpner>
ctp: Don't just install zfs user tools with nix-env
<{^_^}>
[nixpkgs] @vcunat pushed commit from @ivan to release-19.03 « gnupatch: avoid rebuild from 2af7aac7 PR #65687 »: https://git.io/fj9rH
<ctp>
tilpner: I did the configuration.nix part as well
<tilpner>
One solution is to build your own installation image with zfs unstable enabled
<tilpner>
ctp: You may also be interested in building the entire disk image of the finished installation, though that's a bit more uncommon
<tilpner>
And may not work well with encryption
justanotheruser has quit [Ping timeout: 245 seconds]
orivej has joined #nixos
<ctp>
Yeah, I was thinking of it as a source for reproducibility of the entire build, but since it's ZFS, snapshot is fine enough as image of entire disk. OK, I have to run now, but will try the custom image thing later. Thank you so much!
<tilpner>
ctp: With this method, you can customise your image to automatically start installation of your configuration. Just boot, wait, reboot, and you're in your finished NixOS system. Theoretically...
<ashkitten>
> gnome-shell for auto-detecting the GNOME Shell version
<{^_^}>
undefined variable 'gnome-shell' at (string):255:1
<ashkitten>
...
fendor has joined #nixos
<ashkitten>
so instead of supplying the gnome-shell version from nixpkgs which we can acquire without building gnome-shell, we're just putting it in the build inputs
ottidmes has joined #nixos
<abbe>
New NixOS (release-19.03 nixpkgs channel) user. If I see a package that's just got some bugfix updates (no version bump or anything) in release-19.03 branch, how do I install them ?
<abbe>
I tried 'nix-channel --update ...' and did 'nixos-rebuild switch', and 'nix-env -i ...'
<abbe>
but I still get same buggy package, whereas i see the corresponding fixes being present in nixpkgs directory on the host
<srhb>
abbe: Before nixpkgs/release-19.03 gets promoted to nixpkgs-channels/nixos-19.03 it must pass certain test sets, perhaps that's what you're seeing.
<srhb>
The latest commit on nixos-19.03 is currently 4cc5592fe2d264577ca8fb0b4a42f1768030011e -- it might be easier to talk about the package in question rather than in hypotheticals :)
<abbe>
sure, just a second
DariusTheMede has quit [Ping timeout: 245 seconds]
<srhb>
As in, they count as two channels, which means if you update one, the other won't be.
<abbe>
oh! so I should remove my user's ?
<srhb>
YEs.
<srhb>
At least, that's the default setup :)
<srhb>
There's other ways to do it of course, including just not using channels..
ng0 has joined #nixos
<Izorkin>
Mic92: please run command build mariadb.connector-c_3_1 mariadb.client mariadb.server
domogled has joined #nixos
<abbe>
srhb: okay, i purged nixpkgs channel from abbe, but not sure about channels(root): "nixpkgs-19.03.173214.4cc5592fe2d, nixos-19.09pre186563.b5f5c97f7d6". https://hasteb.in/ciqufora.js
psy3497 has quit [Ping timeout: 246 seconds]
<srhb>
abbe: can you sudo nix-channel --update ?
<srhb>
I'm not sure how you would up with the nixpkgs one..
<abbe>
srhb: did that, still two channels
<srhb>
BUt not in nix-channel --list, right?
<srhb>
Now you just want to name your one root channel nixos for clarity, update and you should be good to go.
<srhb>
Rename probably means --remove and --add
<abbe>
right, two channels were nerver in nix-channel --list
<abbe>
it's only in output of "nix-shell -p nix-info --run nix-info" where I noticed
lsix has quit [Ping timeout: 272 seconds]
<srhb>
Well.. There must have been two at some point, or something really broke.
erictapen has joined #nixos
<srhb>
But the channels mechanism is so confusing that who knows anyway.
<srhb>
Anyway: End up with 1 channel, as root only, called nixos, pointing to nixos-19.03, and you're good to go.
<ivan>
fgaz: I'm seeing this too on my own nix-serve's, it might be a bug in nix
is_null has joined #nixos
<ivan>
fgaz: or maybe a dependency because I don't see anything relevant in 2.2.2
o1lo01ol1o has quit [Remote host closed the connection]
<abbe>
srhb: after doing nix-channel --remove nixpkgs i don't see any channels in nix-channel-list but /root/.nix-defexpr/channels/nixos symlink is still present
<srhb>
abbe: It's only relevant for nix-env, ignore it..
<abbe>
okay
<srhb>
(You generally won't be using nix-env as root anyway)
<abbe>
so with no channels I guess I won't be able to install any packages ?
<abbe>
or will I ?
o1lo01ol1o has joined #nixos
<abbe>
using 'nix-env -i ...'
<srhb>
abbe: Your user should have a symlink channels_root iirc
<srhb>
So you're saying nix-channel and sudo nix-channel is empty now
<abbe>
correct
<srhb>
OK, so just add nixos-19.03 with the name nixos to root now
<abbe>
my question is with empty channel list for both users should I be able to install a package using 'nix-env -i ...' basically just to verify my system is consistent with general assumption
<srhb>
No.
<abbe>
okay
<abbe>
but I tried installing ffmpeg (which I have not installed before ever) using 'nix-env -i ffmpeg'
<abbe>
and it starts getting installed
<srhb>
Yargh.
<abbe>
yes, so I guess something is inconsistent in my setup, some hidden channel or something ?
<srhb>
ls /nix/var/nix/channels/{root,abbe} ?
<srhb>
Oops sorry
<srhb>
ls /nix/var/nix/profiles/per-user/{root,abbe}/channels ?
<srhb>
abbe: Clearly you still have a nixos channel in root, did you update?
<Izorkin>
Mic92: rollback PR, added test patch
<srhb>
abbe: When you sudo nix-channel --update is when the actual profile gets created, in order to facilitate rollbacks to earlier versions
<abbe>
srhb: so while you were gone, i added a new channel with url release-19.03 with nixos alias, and then removed it, after that all of my channels were gone for real, now I re-added that same channel, and things are rebuilding
<abbe>
and i only see one channel in nix-info
<srhb>
abbe: OK. fwiw you should not use release-19.03 unless you're willing to risk breakage. That's the untested version of nixos-19.03
<srhb>
So it could eat your system in funny ways.
<azazel>
funny... it depends :-)
<srhb>
s/funny/interesting then ;-)
<azazel>
:-)
<abbe>
hmm
<abbe>
sorry that is what I meant, didn't use release-19.03, but nixos-19.03. sorry for confusing
<srhb>
OK great, sounds like you're in good shape :)
<{^_^}>
[nixpkgs] @prusnak opened pull request #65782 → runwayml: init at 0.8.0 → https://git.io/fj9il
<Mic92>
Izorkin: mind sending your patches also to https://github.com/MariaDB/server. I mean pkgs/servers/sql/mariadb/cmake-without-client.patch and pkgs/servers/sql/mariadb/cmake-fix-crypt-libs.patch
<Mic92>
Then it won't become a burden in future.
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<emilsp>
is there a way I can turn a shell.nix into a tarball that I can _enter_ at any time? getting a bit tired of rebuilding my build env on every reboot.
<{^_^}>
[nixpkgs] @Ma27 pushed commit from @angristan to release-19.03 « vagrant: 2.2.3 -> 2.2.5 »: https://git.io/fj9iF
psyanticy has joined #nixos
<emilsp>
thanks manveru and DigitalKiwi
<__monty__>
+1 for direnv/lorri
<emilsp>
do I have to use direnv? last time I looked at it, it left a bad taste security wise.
<manveru>
how so?
<manveru>
and yeah, i don't think lorri will work without direnv
<emilsp>
ideally, I'd always be the one executing a thing rather than relying on a side effect of changing my cwd to execute things.
<manveru>
but i'd be really interested how direnv is supposed to be insecure (given you don't blindly run `direnv allow` in every directory you go to)
<rawtaz>
emilsp: terribly sorry for a silly question, but what was the problem you were having with shell.nix?
<emilsp>
rawtaz: building android-sdk often fails intermittently and then typescript langserver takes ages to build and doesn't work the way I want it to work.
<rawtaz>
hmm. im probably missing too much context :) thanks though!
<emilsp>
rawtaz: I want something that I can shell into immediately rather than rebuilding the whole derivation that my shell.nix declares. Usually after the first build, it will be fine, nix-shell my-project/shell.nix will return in less than 30 seconds, but the first build is always very painful. As such, I was hoping to find out a way to have my shell.nix derviation evaluated once and then saved somewhere on disk in
<emilsp>
a way that I can easily _shell into_ when I need to.
<manveru>
emilsp: other option is to run something like `mkdir .nix-gc-roots; nix-shell . --pure --indirect --add-root .nix-gc-roots/dep`
<manveru>
and not use <nixpkgs> in shell.nix
<manveru>
otherwise the shell will depend on your channel, which might change, and your shell will require updating
<emilsp>
I can replace nixpkgs channel with a specific commit that doesn't change all the time, right?
<manveru>
yes
<emilsp>
I'll have to look into that then :)
<emilsp>
Gah, there's always too many _other_ things I need to do so I never get around to fixing my nix scripts :/
<__monty__>
emilsp: You could also use `myEnvFun` instead of nix-shell.
veske2 has quit [Quit: This computer has gone to sleep]
niso has joined #nixos
ghasshee has joined #nixos
<manveru>
anw, you can ask more in #lorri :)
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<kandinski>
I'm setting up a couple of NixOS systems (my own laptop and a network proxy cache for binary substitutions) and the snag I just found is that iohk pakages don't have a valid signature. What's the standard way to add public keys so nix can recognise packages? The word "key" is a bit too ambiguous to be a good search term.
ng0 has quit [Quit: Alexa, when is the end of world?]
mexisme has quit [Quit: WeeChat 2.5]
<Taneb>
kandinski: you want to add a "trusted-public-keys" to your nix.conf file
<Miyu-chan>
Ugh, does anyone know how to use foreign keys with inner join on Nixops's sqlite3 database?
<Miyu-chan>
foreign key(machine) references Resources(id) on delete cascade
<yorick>
Miyu-chan: oh no
<Miyu-chan>
Where's the name in this? ; ;
<Miyu-chan>
I hhaven't really used SQL directly.
noudle has joined #nixos
<Miyu-chan>
It seems idk. unnamed?
<yorick>
Miyu-chan: foreign keys are just constraints, this one is called 'machine'
<Miyu-chan>
Oh.
<yorick>
you still have to inner join ON machine=Resources.id
<Miyu-chan>
Thank youuu
zupo has joined #nixos
<Miyu-chan>
Finally.
drakonis has joined #nixos
<Miyu-chan>
SELECT value FROM Resources r INNER JOIN ResourceAttrs a ON r.id=machine WHERE r.name="my-key-pair" and a.name="privateKey";
<Miyu-chan>
Does this seem sane?
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
<Miyu-chan>
...I'm stupid. I finally understand what's happening lol. The line that I sent is saying that the name column constrains or osmething of the sort.
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
<averell>
also you should use ' quotes for strings
drakonis_ has quit [Ping timeout: 272 seconds]
erictapen has joined #nixos
noudle has quit []
DariusTheMede has joined #nixos
<Miyu-chan>
SELECT * FROM DeploymentAttrs INNER JOIN Resources USING (deployment) INNER JOIN ResourceAttrs ON Resources.id=ResourceAttrs.machine WHERE DeploymentAttrs.name='name' and DeploymentAttrs.value='deployment_name' and Resources.name='my-key-pair' and ResourceAttrs.name='privateKey';
<Miyu-chan>
Is this sane?
<gchristensen>
no, it is not typically a good idea to directly query an application's database when you're not the application
<gchristensen>
but yeah that looks right
noudle has joined #nixos
<Miyu-chan>
I forgot how to extract the privkey lmao
<Miyu-chan>
IIRC, my lst way of doing that was using the --export and passing it through jq.
<Miyu-chan>
Now that I think about it.
<gchristensen>
yeah that is a good way
<Miyu-chan>
Welps, either way, was a good crash course to SQL TBH.
<gchristensen>
:) SQL is such a good tool to know
DariusTheMede has quit [Ping timeout: 258 seconds]
<CMCDragonkai>
i've got the new gopro hero 7 black
<CMCDragonkai>
i connect it to USB, the screen says it is now usb connected
<CMCDragonkai>
however i cant seem to see any extra disks to mount
<CMCDragonkai>
i can see that the usb device is there with usb-devices
<clever>
CMCDragonkai: modern cameras dont register as a mass-storage device, they use something else, that directly serves files, not blocks with an FS
<maralorn>
When I "nix.nixServe.enable = true" on my Server, enter correct authorizedKeys and use "binaryCaches = [nix-ssh@my-server ...]" on my Laptop. Any reasons this shouldn‘t work?
<__monty__>
Passwordless key? And does the server sign stuff?
iyzsong has quit [Ping timeout: 252 seconds]
iyzsong has joined #nixos
work_ has joined #nixos
joehh has joined #nixos
fendor_ has quit [Ping timeout: 244 seconds]
drakonis1 has joined #nixos
ZeDestructor has joined #nixos
paulaneds has joined #nixos
<paulaneds>
hey all, how do I install a python package locally?
<paulaneds>
I have added pythonPackages.requests to systemPackages, but when I do import requests in my script Iget
<paulaneds>
ImportError: No module named requests
<paulaneds>
but if I run it from nix-shell -p pythonPackages.requests
<gchristensen>
^ this creates a special python just for you which has requests "built in"
<{^_^}>
[nixpkgs] @rawkode opened pull request #65787 → nheko: use qt5's mkDerivation → https://git.io/fj9ML
<{^_^}>
[nixpkgs] @rawkode opened pull request #65788 → vokoscreen: use qt5's mkDerivation → https://git.io/fj9Mq
ottidmes has quit [Ping timeout: 272 seconds]
<__monty__>
gchristensen: What does nix-shell do differently btw? Afaik with ghc for example you need to do withPackages even in a nix-shell?
<gchristensen>
nix-shell runs a setup hook
<joehh>
hello all, excited about recent nixos accomplishment
<joehh>
had a client visiting a remote site at short notice (1000km+ from them, 3000km from me) which required one of our nixos appliances. created a custom installer image, which connected back to our networks via openvpn. Got client IT to burn to usb, boot random piece of hardware with it, so we could then install nixos on device and then takeover with nixops. client then took device to remote site for
<joehh>
installation. No idea how one would accomplish this without nix/nixos/nixops.
<iqubic>
What's the latest version of Emacs available on the unstable branch?
paulaneds has quit [Remote host closed the connection]
<iqubic>
To whomever maintains the emacs derivation in nixpkgs, I would like to point out that this bug currently exists in emacs 26.2, which is the current latest version in unstable. https://debbugs.gnu.org/36810
<__monty__>
johnw: ^
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos
<gchristensen>
iqubic: IRC is not a bug tracker, if you'd like to report a bug, please open an issue on https://github.com/nixos/nixpkgs -- or even better, send a PR :)
<iqubic>
I know. I will do that.
<iqubic>
I'm not sure how to fix up the emacs derivation, but I'll try.
<iqubic>
Should I open a new branch for my PR?
<gchristensen>
yes
selfsymmetric-pa has joined #nixos
<iqubic>
What should I call this new branch?
meteo has joined #nixos
<gchristensen>
anything you want
<__monty__>
iqubic: The branchname doesn't matter as long as it's in your fork.
<iqubic>
Ah. I see.
<averell>
i'm thinking it's not there because 26.2.90 is a pretest and the real release is 26.3 (which would be in the mirrors then, unlike this release)
<iqubic>
Ah. I see.
<iqubic>
Makes sense.
<iqubic>
If that is true, then I'm not sure what I should do.
<averell>
you can still make a emacs-pretest version or something if the need is there. we have that for many packages.
<iqubic>
Oh, I see.
<__monty__>
Or backport the patch?
<gchristensen>
at any rate, whatever happens, we should cc the current maintainer and get their feedback
<iqubic>
backporting the patch sounds like the harder way to solve this.
<averell>
but it'll fix the issue for more people :)
<iqubic>
I'm going to make a bug report on the Nixpkg github.
<iqubic>
Just to make sure I'm not going to be posting to the wrong place.
<__monty__>
iqubic: Yep.
ramses_ has joined #nixos
<anton-latukha[m]>
I am currently ignorant:
<anton-latukha[m]>
Couple of times looked before to find the cause of the state of master trunk.
<anton-latukha[m]>
So, what changes broke that 54776 packages and then healed 55100+?
<ramses_>
Hi guys, I am having some issue with a mkIf expression that does not seem to be evaluated lazily, like I have an expression "mkIf cond expr" and eventhough cond evaluates to false, nix throws an exception from with expr. Is this intended behaviour?
<anton-latukha[m]>
In what component the issue was.
<gchristensen>
anton-latukha[m]: link to where you see that?
<samueldr>
anton-latukha[m]: cancelled jobs
<gchristensen>
samueldr++
<{^_^}>
samueldr's karma got increased to 101
<samueldr>
there's no distinction about the reason
<anton-latukha[m]>
<samueldr "anton-latukha: cancelled jobs"> Got it. Yes, also a logical probable variant. Also current delta shows there is some major improvement (>+300). So also logically there were some big change during that swing. Canseled jobs can be related to the change, example: saw the errors, and canseled and fixed the change. I got really satisfied with responce not going to pursue the quest.
<Izorkin>
Mic92: please check next update - #65222
<adisbladis>
iqubic: Nope, but I don't really use package.el
<iqubic>
Right. That's what I figured.
<iqubic>
How do you install emacs packages?
<adisbladis>
emacsWithPackages <3
<iqubic>
I see.
<adisbladis>
This morning we actually had a big change in how we generate those packages, so it's likely that we will be _much_ more up to date going forward
<Miyu-chan>
Anyone else getting Fastly issuse? :/
<adisbladis>
I just created #nixos-emacs in case anyone wants to join in on working more on emacs packaging
<adisbladis>
I have some big plans that will hopefully materialise this weekend
<iqubic>
I hope you don't mind, but I've been using your exwm-overlay like this: http://dpaste.com/0CM7S3B And I've been loading this by putting in the imports section of my configuration.nix
<hyperfekt>
Can someone help me understand overlays - I thought if I use super I can't get recursion from something being indirectly dependent on itself? https://pastebin.com/4C55pETb
<Miyu-chan>
gchristensen: SHould I comment under the gist, or?
<gchristensen>
no, please gist the result tome
<thoughtpolice>
Miyu-chan: If we could get a .drv or something to try and reproduce that would be nice, too.
<thoughtpolice>
Though I may not be able to get to it immediately this morning...
asymptotically has quit [Remote host closed the connection]
asymptotically has joined #nixos
<thoughtpolice>
Miyu-chan: Sorry, that was just an example in the comment, `pwih6g12kymmk9nccajz7a934cp6n63h.narinfo` is completely random. You'll want to replace it with the problematic artifact you're trying to fetch.
<Miyu-chan>
Does this say anything? I just copied the command in the gist, since I'm nto so sure what to have as an argument.
<thoughtpolice>
Unfortunately, nix build really hides this :(
<thoughtpolice>
Miyu-chan: You'll actually want the .nar itself, since that's what you're trying to fetch, in this case. (Though it could be hung on the narinfo, as well)
<Miyu-chan>
gchristensen: That works for me.
<thoughtpolice>
FWIW, you can luckily just take the hash of the actual thing you want to build, in this case, `1vf92zjw7zgrvblr343dgr4930c9g2mk`, and just add `.narinfo` on the end
<Miyu-chan>
thoughtpolice: And that seems to be right. I have good connection to HK and SG.
meteo has quit [Ping timeout: 245 seconds]
<Miyu-chan>
FWIW, even AWS is not working.
<thoughtpolice>
That's an odd one. I don't have any server access in Asia, etc to test at the moment... So it seems you're getting interfered with at some other level, perhaps...
<thoughtpolice>
Ugh. That's so annoying. :( What's strange though is you aren't unilaterally blocked out, it seems, since you can fetch the narinfo files...
<thoughtpolice>
I wonder if it's some shithead gateway proxy or something that's trying to like, unzip the .xz files or something. I've had some nightmares with that before.
<Miyu-chan>
And, well, the EC2 instance is also located in Singapore, if that's any indication of anything.
<Miyu-chan>
Hm. I wonder.
<Miyu-chan>
Yeah. Had a guess.
DariusTheMede has quit [Ping timeout: 258 seconds]
<Miyu-chan>
[myrl@myrl:/tmp/tmp.bqeMefzSaP]$ nix run nixpkgs.bc
<Miyu-chan>
[2 copied (1.4 MiB), 0.3 MiB DL]
<Miyu-chan>
So smaller packages seem to work.
<thoughtpolice>
Hmm. Here's a big package: `nix run nixpkgs.foundationdb61 -c fdbcli --version`
<thoughtpolice>
200mb . I think? Try that maybe?
Mr_Keyser_Soze has quit [Ping timeout: 245 seconds]
<Miyu-chan>
I'll try foundationdb60, since that's what's available in my local nipkgs.
<Miyu-chan>
And seems to be not working too.
<Miyu-chan>
[myrl@myrl:/tmp/tmp.bqeMefzSaP]$ nix run nixpkgs.foundationdb60 -c fdbcli --version
o1lo01ol1o has quit [Remote host closed the connection]
<Miyu-chan>
Anyways, it seems alive now. :)
<thoughtpolice>
Miyu-chan: You mean it works?
<averell>
hm, maybe it's cached now... :)
<Miyu-chan>
Yep.
<Miyu-chan>
So I can't provide any more debugging details. :/
<thoughtpolice>
I'll ask around internally but I think if you're not even getting response headers from the cache, something might be screwing you up between you and the POP. Which yeah, is going to be very hard to debug...
<thoughtpolice>
It shouldn't take *that* long to actually prime the cache even for large things....
<thoughtpolice>
Oh. We can also purge it
<adisbladis>
aria: I created #nixos-emacs to discuss emacs packaging without spamming here
<averell>
i was thinking the (shithead gateyway proxy) in between might be done caching
<thoughtpolice>
Miyu-chan: Try `./fastly-cache-debug.sh nar/02pk0mmxj3rkda9nm7m90c4jfwj6yk8791c69682l18a5mq31m7c.nar.xz` again now
<thoughtpolice>
If it still works (it may take a second or two longer now, since it needs to prime a cold object), then yeah I'm guessing it's either some dumb proxy or bizarre network issues between you and the POP which are... not easy to control.
<Miyu-chan>
Hm.
<niso>
is anyone aware of a nixos option for executing a script after each system-boot?
<Miyu-chan>
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0* Connected to cache.nixos.org (151.101.2.217) port 443 (#0)
<Miyu-chan>
*
jophish has joined #nixos
<Miyu-chan>
Let me try with fastly-cache-debug on ghc's nar.xz.
<thoughtpolice>
Interesting. So what might be happening from what I can tell is that: when you contact Singapore (SIN), it has to contact Ashburn (IAD). This forms two layers to the cache. I wonder if the fact IAD is "cold" has anything to do with it.
jophish has quit [Client Quit]
<thoughtpolice>
When you said it worked, I tried to download that same NAR, and I saw it cached in *both* places.
<thoughtpolice>
Right now I'm trying to populate the object in IAD, again
<niso>
gchristensen: temporary patching something in a hacky way. a) i want to start the nix-daemon b) i want to execute a build script
<Miyu-chan>
Maybe it's just my slow internet acting up?
<Miyu-chan>
Give me a few, I might have false positived.
<thoughtpolice>
That's also possible.
jophish has joined #nixos
<Miyu-chan>
I think ghc.doc works, but ghc itself does not.
<thoughtpolice>
What is the .nar file for ghc.doc, if you can grab it?
<niso>
gchristensen: thanks, guess i'll wrap it in systemd services
MmeQuignon has joined #nixos
* niso
wonders why nix-daemon doesn't start automatically
<Miyu-chan>
Oh wait. Cool.
<thoughtpolice>
gchristensen: I also realized something yesterday, which is that I think I can point my debug/custom version of the Fastly Cache configuration to point right to our S3. This means we can actually test the real cache with my own custom changes, with no interference to cache.nixos.org
<Miyu-chan>
thoughtpolice: The headers for the first command just displayed itself. Good thing I didn't C-c C-c it.
<gchristensen>
thoughtpolice: oh cool!
<gchristensen>
yeah
<thoughtpolice>
Interesting. So there may be some kind of inter-POP problem going on, I wonder...
<thoughtpolice>
Two things: it was in neither cache at the time you fetched. But it should now be in SIN. Can you run that command again? Just immediately?
o1lo01ol1o has joined #nixos
<Miyu-chan>
So, around a minute and 30 seconds after. Same thing.
<Miyu-chan>
Err, same behavior as the first run.
<Miyu-chan>
Only shows me > Host: cache.nixos.org
<thoughtpolice>
Hmmmm. So it will probably come through eventually just... it's taking a long time either A) from SIN <-> IAD, or from B) IAD <-> AWS
<thoughtpolice>
I would *think* the problem is A) considering, like. It works fine for a lot of other stuff!
<thoughtpolice>
But the backbone between SIN/IAD shouldn't have major interferences like that. I hope.
<thoughtpolice>
But what's more interesting is you were able to get it work once and then it stopped after my PURGE.
<thoughtpolice>
What we really want to see is, if it starts working and is cached, if you can fetch quickly, and repeatedly.
<thoughtpolice>
That way we can actually confirm whether the PURGE was related or not.
<thoughtpolice>
A bit whack-a-mole, unfortunately.
<Miyu-chan>
Hm.
<Miyu-chan>
I'm testing with this right now, and it's working as of the moment. nar/0hq55n0il2r8n2inznyapda6b3xgmjcvyyzwpmq329vhkg71hpc8.nar.xz
<zimbatm>
next, add some sort of on-disk cache to not destroy cache.nixos.org
<averell>
pretty cool. is dir-listing a security thing?
<thoughtpolice>
Miyu-chan: The interesting thing is that, the ghc.doc deriv is cached in IAD now. FoundationDB was not cached in IAD.
<thoughtpolice>
I really wish I could just purge from one POP but I'm not sure (OTTOMH) if that's possible for debugging...
<thoughtpolice>
zimbatm: neato
<Miyu-chan>
thoughtpolice: Yeah, all seems good so far. :)
<Miyu-chan>
My home connection is 5 Mbps, so it's taking a bit more than seconds. :P
<thoughtpolice>
Okay, so that is very weird. My best guess is this: when you request objects from SIN, it must ask for them from IAD. If the objects are not in the cache at IAD, it takes a *very* long time, seemingly, to get them to you
<thoughtpolice>
But if they are in IAD, then the transfer is fast. I'm.... not sure what to make of this.
__monty__ has quit [Ping timeout: 245 seconds]
orivej has quit [Ping timeout: 268 seconds]
<thoughtpolice>
This would seem to imply that for some bizarre reason, when you go through SIN, it takes a long time for IAD to fetch from AWS. But if I go through another POP, say, Dallas to IAD -- it doesn't take long at all, even if it's uncached!
<zimbatm>
averell: it's not a security issue but it means unpacking the whole NAR file to get all the files. NAR is just a concatenation of files
<thoughtpolice>
Which seems contradictory...
meteo has joined #nixos
DariusTheMede has joined #nixos
<Miyu-chan>
Yeah, I think I get what you're saying.
<thoughtpolice>
I have logging configured so perhaps I can get more insights
<Miyu-chan>
So there's no connection issues between SIN and IAD, nor IAD and AWS, but if IAD doesn't have it cached, and SIN requests it from IAD, it goes wonky?
rouma has joined #nixos
<thoughtpolice>
Yeah, that's what it seems like.
<thoughtpolice>
It would be nice if some other user in Asia (who would go through Singapore) could confirm this...
<thoughtpolice>
I'll ask around internally about this too, but who knows... This is all unfortunately speculation on my part at this exact moment.
DariusTheMede has quit [Ping timeout: 246 seconds]
<Miyu-chan>
Mkk. Thanks for helping!
<adisbladis>
thoughtpolice: I have a machine in Hong Kong
<adisbladis>
If that helps
orivej has joined #nixos
<thoughtpolice>
adisbladis: try that `for` loop from earlier
<selfsymmetric-pa>
What command can I run to see all reverse dependencies of `binutils-2.30` on my system?
<selfsymmetric-pa>
I tried out `vulnix --system` and I'm getting `70 derivations with active advisories`. Yikes.
<thoughtpolice>
adisbladis: If you're going through singapore, you should see something like `cache-sinXXXXX-SIN` in the `fastly-debug-path` field
domogled has quit [Quit: domogled]
<Miyu-chan>
I don't have any experience in what you're doing, but is it possible that IAD ignores SIN for miss events, and the reason why it works after some time is because another POP hits it?
<infinisil>
Prints all paths that have /nix/store/... as a transitive dependency
<thoughtpolice>
Miyu-chan: Possibly but if that were the case I'd think we'd be seeing many, many errors and problems internally (and a lot of irritated customers, too)
<Miyu-chan>
Also, my 19.03 is ~1 month outdated, so that might contribute to the cache misses.
<thoughtpolice>
Okay, so that one actually had a HIT on HKG on the first request, which means it wasn't purged (maybe someone requested it)
jgt has quit [Ping timeout: 250 seconds]
orivej has quit [Ping timeout: 245 seconds]
<hyperfekt>
samueldr: how do you prevent your overlay-compat thingy from picking itself up when evaluating the config and thus leading to infinite recursion? :o
<samueldr>
I uh... don't?
<samueldr>
I mean, I don't do anything particular
<hyperfekt>
weird, my rebuilds got stuck in a loop evaling config over and over when i tried it
<samueldr>
the way it works is that the only location it's defined is in NIX_PATH, e.g. nixpkgs-overlays=/etc/nixos/modules/overlays-compat/, and then in the nixos configuration, which I think might not follow NIX_PATH?
<samueldr>
though I may be wrong
dansho has joined #nixos
<samueldr>
I also could have been lucky I guess?
<hyperfekt>
nixos-rebuild definitely picks up the NIX_PATH for me 🤔
<samueldr>
I think I just don't know at that point :)
<samueldr>
I wrote that thing up when I was learning about nix and nixos, and haven't touched it or updated it since
<samueldr>
though I'm using it
<hyperfekt>
haha that's okay, we're all just scribbling summoning circles after all :b
<samueldr>
one of the worst thing about nixos: everything just works, you're never faced with the need to just re-do it from scratch
JonReed has quit [Remote host closed the connection]
<lordcirth>
lol
<samueldr>
and since it just all works, you don't know really what you did, months later, but you don't really need to either!
ramses_ has quit [Ping timeout: 260 seconds]
<hyperfekt>
i mean i'd rather not know what i did and have it written down that not knowing what i did and not having it written down anythere as would be the case if not declarative :b
<nurelin[m]>
Hello, how can i override the cargo-vendor version in nixos-19.03 ?
<__monty__>
,overlays
<{^_^}>
__monty__: Did you mean overlay?
<{^_^}>
Overlays look like `self: super: { foo = ...; }`. Use the self argument to get dependencies, super for overriding things and library functions (including callPackage). More info: https://nixos.org/nixpkgs/manual/#sec-overlays-install
<__monty__>
nurelin[m]: ^
<__monty__>
How can I load a configuration into nix repl? I'd like to troubleshoot my home.nix this way. It expects arguments { config, pkgs, ... }.
<nurelin[m]>
it has been updated in nixos-unstable
<symphorien>
what about using unstable's buildRustPackage then ?
jgt has quit [Ping timeout: 264 seconds]
<__monty__>
nurelin[m]: Why is that a problem? I can overlay individual haskell packages just fine.
<nurelin[m]>
symphorien: I will try that, thanks
__monty__ has quit [Ping timeout: 258 seconds]
<nurelin[m]>
__monty__: I just don't know how to declare the overlay since cargo-vendor is in a not common place in the nixpkgs tree
_kwstas has joined #nixos
<ghasshee>
Good Morning guys, I have another basical question. What 'a single derivation' denotes for ? and How can we make a double derivation ?
bennofs has joined #nixos
<sphalerite>
ghasshee: there's no such thing as a double derivation, your text is probably referring to a single derivation as opposed to more than one.
<ghasshee>
sphalerite: yes, sir. I got an error: nix-shell requires a singlle derivation
<sphalerite>
ghasshee: right, so you're probably passing multiple derivations or no derivations to it
drakonis_ has joined #nixos
<ghasshee>
sphalerite: I see that it might mean no derivation. Thanks, I going to try.
logzet has quit [Ping timeout: 264 seconds]
drakonis has quit [Ping timeout: 272 seconds]
orivej has joined #nixos
<fgaz>
is there a way to manually add a nar to the store?
_kwstas has quit [Quit: _kwstas]
<aminechikhaoui>
if I have an expression that takes many arguments and then needs to inherit those and pass them to another expression, how can I achieve that without having to rewrite the list of args
<aminechikhaoui>
something like http://ix.io/1QnJ, tried callPackage but I think what I'm doing is wrong
<ivegotasthma>
hello, any idea why I cannot run a java binary? It's giving me no such file or directory found, when I'm right there with bash. I suspect it's something related to the linker, but file is giving me correct output.
<gchristensen>
is it a jar?
<ivegotasthma>
no, it's a binary
<ivegotasthma>
java: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked, interpreter /lib64/ld-linux-x86-64.so.2, for GNU/Linux 2.6.9, BuildID[sha1]=b82a586842f6da3df8a61093daadf04180e85c16, not stripped
thc202 has quit [Ping timeout: 250 seconds]
drakonis_ has joined #nixos
<gchristensen>
/lib64/ld-linux-x86-64.so.2 is sensible output, but that path probably doesn't exist
<ivegotasthma>
it doesn't
<ivegotasthma>
is it possible to patch it somehow?
<ivegotasthma>
nvm, found the wiki page
srid has quit [Ping timeout: 268 seconds]
<catern>
is there a way to list a closure with only derivations, no output paths?
<catern>
fixed-outputs would also be acceptable
<catern>
the reason I want to do this is to copy a tree of derivations to another host and build it again there, because that host doesn't trust the signature of artifacts built on my host
<catern>
I tried nix-store --query --requisites $(nix-store --query --deriver mypath), but that seems to include output paths, not just derivations
<{^_^}>
[nixpkgs] @JohnAZoidberg opened pull request #65806 → xorg.libXTrap: init at 1.0.1 (Fix build of xorg.xtrap) → https://git.io/fj99k
x24 has quit [Remote host closed the connection]
x24 has joined #nixos
ngirard has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @flatulation opened pull request #65807 → Fix/firefox bin update → https://git.io/fj99L
sdemos-ish has joined #nixos
<catern>
although hm I don't just want low-level derivations because some of those don't exist on my source host
<{^_^}>
[nixpkgs] @JohnAZoidberg opened pull request #65808 → gdcm: Fix on darwin (include ApplicationServices) → https://git.io/fj99q
lambda-11235 has joined #nixos
<srhb>
catern: Won't importing drvs have the same trust restrictions as the artifacts anyway?
<catern>
srhb: I don't believe so, you can add an arbitrary drv
<Miyu-chan>
AaaaAAa. I can't run ./create-ami.sh in EC2?
<Miyu-chan>
error: a 'x86_64-linux' with features {kvm} is required to build '/nix/store/mhkywh7j4jq6nxrcdqrvnc4ysjk0zhb0-nixos-disk-image.drv', but I am a 'x86_64-linux
<Miyu-chan>
' with features {benchmark, big-parallel, nixos-test}
<catern>
(well... I don't remember exactly, I guess the output paths need to be right, but at least it *should* work)
<Miyu-chan>
Based on what I understand, it's because of the qemu.
<srhb>
Miyu-chan: You just need to add the feature. At least, if you have a kvm device node available.
<catern>
srhb: another reason it has to work is build offloading - the farm gets the low level drvs from untrusted users
<srhb>
(I guess qemu will run, but slowly, without)
<Miyu-chan>
Gah, yeah. :/
<noudle>
i just installed "nomachine-client" but no i dont find the executable, how is it called? o.O
o1lo01ol1o has quit [Read error: Connection reset by peer]
<Miyu-chan>
It will take a few hours for me to upload an AMI from my home. Welps.
drakonis_ has quit [Ping timeout: 264 seconds]
<Miyu-chan>
I have no other choice.
<ivegotasthma>
gchristensen: is it possible to patch a dynamic library the same way?
<hyperfekt>
i have a script I'm using in a derivation and it's in my config directory. Using it just like that worked! Nix is so neat *-*
<{^_^}>
[nixpkgs] @sikmir opened pull request #65811 → qmapshack: use Qt mkDerivation → https://git.io/fj99S
drewr has joined #nixos
noudle has quit []
veske has quit [Quit: This computer has gone to sleep]
dftxbs3e has quit [Ping timeout: 250 seconds]
_kwstas has joined #nixos
<__monty__>
Hmm, hit a snag. I'm trying to write a function for my home.nix that goes through all the installed packages and extracts any .app's, so I can link them into ~/Applications.
<__monty__>
But apparently you're not allowed to run `pathExists` on outPaths of other derivations?
chloekek has joined #nixos
<srhb>
__monty__: This seems like the wrong approach. Instead, I would map a function over them that either makes a new package that puts every .app into a known location within the new package, and then ensure that that's properly joined into the profile buildenv.
<srhb>
__monty__: Then all you have to do is symlink that location in your profile into ~/Applications
<srhb>
s/either//
<srhb>
(Apparently I had a bad idea too and scrapped it, JIT)
<__monty__>
srhb: Hmm, how to go about the last step? I'll just assume the apps are in package/Applications for now.
tindvik[m] has joined #nixos
<Church->
Hmm my kernel module is broken
* Church-
grumbles
<srhb>
__monty__: Can't you make ~/Applications a symlink to ~/.nix-profile/Applications?
<__monty__>
.nix-profile/Applications doesn't exist. That's the problem I'm trying to solve.
<__monty__>
nix-darwin has a hook for that I believe but I guess home-manager doesn't.
<__monty__>
But I want to install apps from home.nix.
<__monty__>
/nix/store/package/Library *is* linked into .nix-profile otoh.
<__monty__>
Though not everything in package/Library is present there.
<srhb>
__monty__: Add it to pathsToLink?
<__monty__>
-_-
<srhb>
So then it just becomes a buildEnv over all the .app packages you want..
<__monty__>
srhb: Oh, it's a buildEnv feature? Probably why I didn't think of it.
o1lo01ol1o has joined #nixos
<davidtwco>
Is it more idiomatic when using patchelf to --set-interpreter to "${stdenv.glibc.out}/lib/ld-linux.so.2" or "$(cat $NIX_CC/nix-support/dynamic-linker)"?
<__monty__>
srhb: Any other dirs you link? (Applications, Library, bin)
o1lo01ol1o has quit [Ping timeout: 245 seconds]
selfsymmetric-pa has quit [Remote host closed the connection]
<srhb>
__monty__: I don't use Mac, so no, I generally don't use this at all. :)
<x24>
How would I go about adding additional python packages to FreeCAD?
dansho has quit [Ping timeout: 248 seconds]
dansho has joined #nixos
_kwstas has quit [Quit: _kwstas]
_kwstas has joined #nixos
ngirard has joined #nixos
linarcx has joined #nixos
linarcx has quit [Client Quit]
rouma has quit [Remote host closed the connection]
linarcx has joined #nixos
<__monty__>
srhb-- Haha, very funny, I got pwned.
<__monty__>
"dumping very large path >256MiB; this *may* run out of memory.
<__monty__>
srhb: I figured simply including the buildEnv in home.programs would be fine but apparently it's not. What am I doing wrong?
rouma has joined #nixos
<Church->
What is trying to build this dang pycaption lib...
___laika has joined #nixos
<Church->
Ah there we go, it was gnomecast
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
johnny101m has joined #nixos
___laika has quit [Ping timeout: 246 seconds]
<__monty__>
x24: You'd have to override the package's buildInputs but I'm not sure that can be done easily in an overlay. I'd recommend changing the expression so that becomes possible and submit a PR to nixpkgs.
<x24>
Thanks!
Jackneill has quit [Remote host closed the connection]
x24 has quit [Remote host closed the connection]
hyperfekt has quit [Remote host closed the connection]
fpob has quit [Ping timeout: 244 seconds]
selfsymmetric-pa has joined #nixos
fusion809 has quit [Remote host closed the connection]
zupo has joined #nixos
zupo has quit [Client Quit]
srid has joined #nixos
srid has quit [Changing host]
fpob has joined #nixos
<__monty__>
srhb: Hmm, I tried moving it to a regular overlay but I still get the very large path warning. How can I get around this?
<__monty__>
A different problem, this buildEnv would have to refer to other things from my overlays is the only solution making sure it comes alphabetically last?
eof has quit [Quit: eof]
<{^_^}>
[nixpkgs] @davidtwco opened pull request #65813 → cc-wrapper: Allow specifying extra compilers to wrap. → https://git.io/fj9Ha
<LnL>
I don't think propagated checkInputs are a thing
<kisik21>
Then why it doesn't throw an error on invalid argument? :c
<LnL>
since test dependencies are only needed for the current pacakge
<LnL>
attributes end up as environment variables so it's free form
<averell>
ooh, twitch restream. very good.
<kisik21>
right, tests are not recursive
<LnL>
separating variables from special attributes like buildInputs would indeed make sense and allow better validation, but that's currently not the case
kreisys has joined #nixos
<__monty__>
If only I had the bandwidth : /
selfsymmetric-pa has joined #nixos
ghasshee has quit [Ping timeout: 248 seconds]
linarcx has quit [Quit: WeeChat 2.5]
linarcx has joined #nixos
_kwstas has quit [Quit: _kwstas]
_kwstas has joined #nixos
zupo has joined #nixos
<{^_^}>
[nixpkgs] @petabyteboy opened pull request #65815 → cbmem: init at 4.9 → https://git.io/fj9Hp
linarcx has quit [Quit: WeeChat 2.5]
linarcx has joined #nixos
<kisik21>
do I need to build flask applications as, well, applications or libraries? especially if they are to be deployed with something like uwsgi
<kisik21>
its so confusing when you blur the line between those by going web development
veske has joined #nixos
bennofs has quit [Quit: WeeChat 2.3]
zupo has quit [Ping timeout: 248 seconds]
pietranera has quit [Quit: Leaving.]
<Izorkin>
Mic92: forced update and rebase PR 65222
<ToxicFrog>
Argh. I'm having an issue with mvn2nix -- it generates the json fine, but when I try to build it, six dependencies (out of hundreds) produce a "Non-resolvable import POM: Cannot access 4thline-repo (http://4thline.org/m2) in offline mode and the artifact <name> has not been downloaded from it before"
<ToxicFrog>
But looking at the build logs, it was downloaded, and is present in the nix store!
kim0 has joined #nixos
ottidmes has joined #nixos
erasmas has quit [Quit: leaving]
srid has quit [Ping timeout: 244 seconds]
xkapastel has quit [Quit: Connection closed for inactivity]
psyanticy has quit [Quit: Connection closed for inactivity]
<fendor_>
can I just install the nix executable and use it by supplying a custom store location?
<pie_>
fendor_: kind of
<pie_>
depends on if you want sandboxing, and you generally want sanboxing
<pie_>
also dpends on if youre in single user or multi used mode i guess
<pie_>
hm
<pie_>
not sure what difference that makes other than needing the nix daemon for multi user mode i think
srid has quit [Changing host]
srid has joined #nixos
<pie_>
i think you also need a statically compiled /bin/sh or somesuch because youll get weird errors if you use sandboxing and dont have that (it gets mounted into the sandbox for some reason)
<pie_>
i might be wrong about this stuff though
<pie_>
actually why _does_ nix mount a sh instead of having its own, wtf?
<thoughtpolice>
nix run --store $HOME/my-test-store nixpkgs.hello -c hello
<thoughtpolice>
It kind of depends on what you want to do, really. The --store argument lets you bind mount a random directory, as if it was /nix/store. But it doesn't just magically let you use Nix-compiled packages without 'nix' installed itself, since something has to perform the bind mount step
DRW77 has joined #nixos
<thoughtpolice>
There are some various utilities floating around -- basically all the exact same code -- that do this. Nix's "--store" argument is one such implementation
<thoughtpolice>
But there are some others that do not depend on --store at all
<thoughtpolice>
(Or 'nix' the command)
fendor__ has joined #nixos
<fendor__>
thoughtpolice, alright, thx! I think I killed the vm, but i will try that again!
DRW77 has left #nixos [#nixos]
<catern>
fendor_: ignore everyone else, you can use that static-nix thing and --store to build something reproducibly in a store from a random directory
fendor_ has quit [Ping timeout: 246 seconds]
<catern>
alas! he's gone
<catern>
(and you get binary caching too, you don't lose that...)
<{^_^}>
[nixpkgs] @bgamari opened pull request #65817 → build-support: Add p11_kit to appimage dependency set → https://git.io/fj97q
___laika has joined #nixos
jgt has joined #nixos
<fendor__>
catern, not gone, just disconnected for a second :)
<camsbury>
well this is all relevant to what I asked before
<camsbury>
copying to a cache using `nix copy --to` and getting hash mismatches
wfranzini has quit [Remote host closed the connection]
rtjure has joined #nixos
<camsbury>
when I garbage collected all the things on the cache
<camsbury>
so it's like... why would that happen
wfranzini has joined #nixos
<catern>
fendor__: oh, I didn't see that your tail got longer :)
<fendor__>
catern, I will try the static-nix appraoch once I have fixed the vm! some issue with mounting the tmp folder
___laika has quit [Ping timeout: 245 seconds]
fendor__ has quit [Ping timeout: 244 seconds]
veske has quit [Quit: This computer has gone to sleep]
drakonis_ has joined #nixos
camsbury has quit [Remote host closed the connection]
drakonis has joined #nixos
o1lo01ol1o has joined #nixos
drakonis_ has quit [Ping timeout: 250 seconds]
<worldofpeace>
DigitalKiwi: Probably should submit the PR for bitwarden and see if we could come up with something. could be a bug in the application. have you tried searching if other applications had that issue?
linarcx has quit [Read error: Connection reset by peer]
<worldofpeace>
DigitalKiwi: and that issue with qt was fixed in master
<{^_^}>
#65756 (by Kiwi, 22 hours ago, open): Updated bitwarden-cli with ./generate.sh so that it builds. Fixes #63158
<DigitalKiwi>
did i make the commit message bad on that second one
<jlv>
I'm trying to fix a broken package, but the only feedback I'm getting is "builder failed with exit code 1". Is there any way to get more information about why it failed?
<rawtaz>
DigitalKiwi: are you using it for yourself or more users? locally or their online one?
<DigitalKiwi>
just me and the online one right now
<rawtaz>
ok cool. does it work well? any problems/downsides? im thinking of trying it out for a bunch of users, maybe 20 or so. probably self-hosted though
<rawtaz>
and most importantly - does it feel secure? :-)
<{^_^}>
[nixpkgs] @FRidh pushed 70 commits to staging-next: https://git.io/fj95v
fendor__ has joined #nixos
fendor__ has quit [Remote host closed the connection]
<DigitalKiwi>
they have had outside security audits and everything is on github i've only been using it a few days but highly impressed not sure i like that so much of it is electron but maybe that's not a big problem idk
<{^_^}>
[nixpkgs] @grahamc opened pull request #65818 → shotcut: use qt's mkDerivation → https://git.io/fj95O
<rawtaz>
thanks for your opinion.
jfroche91 has quit [Ping timeout: 245 seconds]
<DigitalKiwi>
you're welcome, always glad to give opinions ;)
<DigitalKiwi>
worldofpeace: i'm fixing the things it's building a lot of stuff :<
abathur has joined #nixos
<DigitalKiwi>
worldofpeace: should i fix the period in bitwarden-cli while i'm at it?
<worldofpeace>
DigitalKiwi: sorry I don't know what you mean about the period
<worldofpeace>
DigitalKiwi: also it may be buliding lots of stuff because there was a little bloop on nixpkgs. perhaps use an older revision base.
<DigitalKiwi>
See contributing guidelines, no periods at the end of description.
<DigitalKiwi>
and you know what when i copied it i said to myself they're not supposed to have periods but i left it for consistency with the other one...that i could have changed (but can't remember when i did which)
<gchristensen>
might as well
<DigitalKiwi>
can i still change it in the PR that i did that fixed it or since it's merged do i need a new one
<rawtaz>
not much point then, cuz your period commit message will still be in the history :)
<rawtaz>
you broke the entire history
<gchristensen>
eh?
<DigitalKiwi>
how do i do the revision base thing i only know how to do it if checkout the the channel and then create the branch i don't know how to do the existing branch onto it after i rebased onto master
<rawtaz>
gchristensen: never mind :)
<DigitalKiwi>
make the borgbot reject anything that has a period at the end of description :D
<gchristensen>
nice idea
<DigitalKiwi>
and that doesn't quote homepage urls
<gchristensen>
homepage URLs don't need to be strings, URL syntax is ok
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<DigitalKiwi>
noooo must be consistent! and quoted is the best (not quoted, click a link, the semicolon is parsed as part of url, doesn't work) (and also someone was wanting to deprecate url syntax anyway)
<exarkun>
is there some simple way to get any kind of networking between a host and a nixos vm built w/ nixos-build-vms?
<infinisil>
alienpirate5: Ah I see (should've posted that from the beginning), but unfortunately there's no way to remove assertions in NixOS, you can only remove the module fully (disabledModules) and replace it with one that doesn't have the assertion
<averell>
what if you nix-shell --pure? maybe it's just that the installed clang is on the PATH first?
<m5n>
averell: still the same issue
<m5n>
Running `env` shows that clang 7 is used for an env variable called `NIX_CC`
<m5n>
It's always set to that even if I try other versions of clang or use pure/sandbox
rouma has quit [Remote host closed the connection]
<m5n>
Looking at $PATH, I can see that both clang7 and clang8 are there, but clang7 is there first, as you suggested is the case
ericsagnes has quit [Ping timeout: 272 seconds]
rouma has joined #nixos
<averell>
in what paths? if i do this locally i get 7/8 depending on what i provided to nix-shell, but i'm on master.
<m5n>
I get those paths: /nix/store/qvagarl5ghs07v39zhszm5x9rn08k231-clang-wrapper-7.1.0/bin and /nix/store/i2nf6hslid0ak8fzgpj2m082sprihi92-clang-7.1.0/bin no matter what I build
<m5n>
I think that Mac uses clang as the default CC and Linux uses gcc as the default one
MarcWeber has quit [Ping timeout: 272 seconds]
<m5n>
averell: maybe try using a different version of GCC? I suspect that will trigger the same behavior on Linux for you
dftxbs3e has joined #nixos
dftxbs3e has quit [Read error: Connection reset by peer]
hmpffff has quit [Quit: nchrrrr…]
o1lo01ol1o has joined #nixos
dftxbs3e has joined #nixos
<averell>
yes, you are right. also about teh standard CC :)