<clever>
jasongrossman: are you able to ssh into the machine?
<jasongrossman>
clever: Hm, yes, but I'd have to set up another machine first, so I'd rather try other solutions first.
<jasongrossman>
clever: I mean I'd have to set up another machine before I could even attempt it.
Ariakenom has quit [Read error: Connection reset by peer]
<jasongrossman>
clever: But now I'm intrigued about what you were going to suggest. :-)
<clever>
watch the logs over ssh and see if its actually hanging
<clever>
its possible that only the gui end is having issues
<jasongrossman>
clever: Yes, good. I'll try my other things first, because physically lazy and also busy with other bits of life, but that is a very nice suggestion for later. Thanks.
<jasongrossman>
clever: I've already tried getting in on another tty, and that works fine up to a point but I can't, e.g., restart the display manager.
<clever>
check the journal then, `journalctl -b 0` and then scroll to the end with G
<jasongrossman>
clever: (In fact since I can log in on another tty there's probably no point in sshing in, right?)
<clever>
yeah
<jasongrossman>
clever: OK, I'll try that.
<jasongrossman>
clever: Thanks.
<mikky>
for the moment there I tought kexec got new features like spawning a new container with custom kernel or something :)
<jasongrossman>
clever: Trying to rebuild switch without apache first, because I was already in the middle of trying that.
<clever>
infinisil: ah yeah, i think i saw it a while back
<infinisil>
mikky: Maybe check out this guide (it's using clever's kexec thing) ^^
ris has quit [Ping timeout: 250 seconds]
<mikky>
infinisil: just came back from it :)
jasongrossman has quit [Ping timeout: 248 seconds]
jasongrossman has joined #nixos
<mikky>
I take it installing other necessary kernel modules and filesystem tools can be then done at runtime (cryptsetup, dm-crypt and zfs in my case).
<clever>
mikky: the kexec image already has zfs and luks
<mikky>
even better :)
<infinisil>
mikky: Yeah, the hardest part is getting the initial nixos build, then it's smooth sailing whatever you want to do
<infinisil>
Because whatever you fudge up from that point on, you can just rollback
<jasongrossman>
clever: What am I looking for in my journal after the crash, please?
<clever>
jasongrossman: anything that looks like an error around the time of failure
<mikky>
that's one of many reasons I'm giving NixOS a shot. I like the basically stateless declarative approach.
<jasongro`>
clever: Whole bunch of errors. Some of them are in red, but those don't seem important (e.g. failing to mount a volume that isn't used for anything anyway). The final error is failing to start plymouth, which is enabled in my current config but disabled in the one I'm trying to switch to. So that's not much help. Should I paste the output of xjournal -b 0 in a pastebin?
<jasongro`>
clever: Or I could try stable first.
ajs124 has quit [Quit: Gateway shutdown]
<clever>
jasongro`: check to see if any new errors occur when you `systemctl restart display-manager`
jasongrossman has quit [Ping timeout: 260 seconds]
<jasongro`>
clever: That hangs (actually hangs).
rprije has joined #nixos
<clever>
jasongro`: ah, check dmesg then, after things fail
<jasongro`>
clever: Which means it's not easy to see errors at that point.
<mikky>
clever: nice touch with the autoreboot unit
<jasongro`>
clever: OK, cool.
<jasongro`>
BRB
<clever>
mikky: that helps when you forget to setup ssh keys, or your network config isnt right
<clever>
mikky: but once you format the disk, you only have one chance to get things right, and will need to reinstall something with the datacenter tools if nixos doesnt boot
sbdchd has joined #nixos
galaxite has joined #nixos
ajs124 has joined #nixos
jasongro` has quit [Ping timeout: 240 seconds]
<mikky>
clever: actually, in my case I can always go back up until I change the default efi entry and reboot
<mikky>
doesn't make much difference though, I'm going through this, one way or anoter
deepfire has quit [Ping timeout: 244 seconds]
<clever>
mikky: yeah, your undo options depend heavily on the hardware and the tools the datacenter provides
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fNPwg
<clever>
mikky: i have helped people use this process on their own hardware, but they just lacked a keyboard/monitor to install with, lol
<clever>
in that situation, you have no console access and no undo button
<clever>
jasongrossman: nothing of note, has the gui crashed yet?
<jasongrossman>
At that point, yes.
sigmundv has joined #nixos
<clever>
jasongrossman: what about `journalctl -f -u display-manager` ?
<jasongrossman>
clever: That gives me a few instances of this:
<jasongrossman>
lightdm[2183]: Error updating user /org/freedesktop/Accounts/User1000: GDBus.Error:org.freedesktop.DBus.Error.InvalidArgs: No such interface 'org.freedesktop.DisplayManager.AccountsService'
<jasongrossman>
<clever>
sounds like dbus is to blame
<jasongrossman>
Cool.
<mikky>
clever: that could've been fun :)
<jasongrossman>
I'm going to try with stable.
<clever>
jasongrossman: try a `nixos-rebuild boot` then just reboot
<jasongrossman>
clever: Tried that - it hangs during boot.
<clever>
ah, then id try stable next
<jasongrossman>
clever: Actually ... I think I've tried that but maybe I'll just check.
<jasongrossman>
clever: And then I'll try stable.
<clever>
sounds good
<jasongrossman>
:-)
<mikky>
anyway, clever, reading through justdoit.nix - my (very limited) understanding is that this expression just creates a script and prepares configuration for it but doesn't actually *do* anything...right?
wpcarro has quit [Ping timeout: 264 seconds]
<clever>
mikky: correct, you have to run justdoit after logging in over ssh
jasongrossman has quit [Remote host closed the connection]
<clever>
that part is optional, you can also just manually install like you normally would
<clever>
you can also set the options defined in justdoit, using the configuration.nix file
<clever>
and then it bakes those options into the tarball it generates
fiddlerwoaroof has quit [Read error: Connection reset by peer]
<mikky>
I'd prefer to just have a live NixOS with ssh and do everything manually after that.
<clever>
you can just ignore justdoit then
fiddlerwoaroof has joined #nixos
<mikky>
is about 8G enough space for build?
<clever>
plenty of disk space
semilattice has joined #nixos
jasongrossman has joined #nixos
<jasongrossman>
clever: Just confirming that if I do rebuild boot, and then boot, the boot hangs. I'll try with stable now.
<semilattice>
Is there a way to see the attributes for a built derivation? Specifically one in my package overrides. When I look at it in nix-repl, the default print just shows the derivation, when I use :p it shows the same thing, tabbing with x.[tab] seems to not produce anything
<semilattice>
and :a adds 43 atributes, but doesn't list what they are
<clever>
semilattice: is the override in config.nix or configuration.nix?
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mikky>
I see ZFS is quite popular here. Is it just because ZFS is cool or does NixOS itself benefit from it or leverages its features?
<clever>
semilattice: and if you try to eval js-env in `nix repl '<nixpkgs>'` does it print the expected thing?
<jasongrossman>
mikky: I'm pretty sure it's just because ZFS is cool.
<hodapp>
it's because ZFS is freaking awesome
<hodapp>
though I only use it on BSD, not on NixOS
<jasongrossman>
mikky: In fact I feel *less* need for ZFS now that I have rollbacks on my OS. But ZFS is still so cool.
<Ashy>
i wish it could be mainlined with the kernel
<Ashy>
btrfs is getting somewhat close in features though i think
<hodapp>
mlehhh
<semilattice>
is there any benefit of ZFS over btrfs still?
<hodapp>
my luck with btrfs was not good
<jasongrossman>
semilattice: There were still bugs in btrfs not so long ago. Given how important the fs is, I think it's going to be a while before everyone trusts btrfs.
<hodapp>
I was very uninspired by the fact that I managed to get an FS into a state that trying to mount it would panic the kernel
<Ashy>
semilattice: zfs can do encryption without needing a layer underneath
<gchristensen>
it isn't ready for production yet
<semilattice>
@jasongrossman Yeah, I've been using btrfs for at least 3 years personally without issue
<Ashy>
same
<jasongrossman>
semilattice: Excellent!
<semilattice>
or I guess two and a half, close enough :P
<jasongrossman>
semilattice: I imagine everyone who uses linux will switch to btrfs eventually, if no new filesystems get a lot of traction in the meantime.
<semilattice>
The encryption point is pretty good. I'm not using LUKS now but is it any easier/harder in NixOS?
<jasongrossman>
Personally, I never want to go back to a non-COW filesystem, but I don't feel so strongly about the choice between COW systems.
<clever>
jasongrossman: i had btrfs take 6 hours to do a simple hydra based task, and it ultimately gave up and flagged the FS as read-only, forcing a reboot
<clever>
jasongrossman: it then took another 6 hours to delete the .drv files i was making
<mikky>
I've been using btrfs for a couple of years not so long ago but it fell apart for no apparent reason.
<jasongrossman>
semilattice: I don't think NixOS yet has the hooks to let ZFS encrypt the boot volume, whereas it does have the hooks to let LUKS do it.
<jasongrossman>
clever: Ouch!
<hodapp>
mikky: fell apart how?
<clever>
jasongrossman: ive also seen btrfs have problems with the ratio of metadata to data on nixos, and claim its out of space yet df says it isnt
<mikky>
ZFS encryption, or more precisely ZFSonLinux encryption is still experimental I think. At least the on-disk format is not frozen yet.
<jasongrossman>
ZFS encryption of non-boot volumes should be very easy. But it is experimental still, right.
<mikky>
hodapp: it just didn't mount one day
<clever>
jasongrossman: nixos already supports zfs encryption of boot volumes
<clever>
ive seen a few brave souls using it :P
<hodapp>
mikky: kernel panic? or mount just erroring out?
<semilattice>
clever: I've had that happen when I accidentally filled up the disk
<drakonis>
zfs encryption is only experimental outside of linux
<jasongrossman>
clever: Oh cool! But there are no options to tell NixOS how to find the decryption password on boot, that I can see.
<drakonis>
surprisingly enough
<clever>
jasongrossman: it will ask on the console i believe
<jasongrossman>
clever: I'm going to try that! How cool.
<mikky>
hodapp: kernel didn't even recoginzed it as a filesystem. No panics, just your typical "wrong fs type, bad option, bad superblock"
<hodapp>
gaack
<jasongrossman>
gaack indeed
hakujin1 has quit [Ping timeout: 265 seconds]
<hodapp>
in my case, the FS was online and suddenly all attempts to read/write to a specific directory were stalling indefinitely
<{^_^}>
openzfs/openzfs#489 (by lundman, 39 weeks ago, open): 8727 Native data and metadata encryption for zfs
<mikky>
emily: depneds on "use case" I guess
<emily>
and the exact threat it protects against isn't always clear
<emily>
hopefully the best plausible deniability will be the continued ubiquitous adoption of encryption :)
<jasongrossman>
clever: Very nice. Thanks.
<hodapp>
emily: I always viewed it as something like "stick a Windows partition there with 15 GB of horse porn, and make a *really* good act of pretending that that's what you're hiding"
<Ashy>
protection against state level actors in general is fairly theoretical isnt it?
<jasongrossman>
emily: Exactly!
<mikky>
clever: I'll be using it as an install guide
<hodapp>
in order to product your... encrypted Linux partition with absolutely nothing of value on it
fragamus has joined #nixos
<jasongrossman>
Ashy: There are different levels of competence among state-level actors, but I agree that there are plenty of baddies who will torture stuff out of you, if they think it's sufficiently important stuff.
<clever>
mikky: also of note, current nixos zfs requires that you export the pool when your done installing
<clever>
mikky: or the initrd wont be able to import it
<gchristensen>
but my seeecccrrreeeettttsss hodapp
<gchristensen>
my l33t bashrc
<hodapp>
gchristensen: don't worry, I won't tell anyone that you have pirated vinyl rips of Lady Gaga's *entire* discography
<jasongrossman>
gchristensen: LOL
<emily>
Ashy: Edward Snowden isn't dead.
<emily>
Ashy: and he was kind of terrible at opsec.
<Ashy>
jasongrossman: "plenty of baddies" including pretty much all country level security orgs, heh
<Ashy>
emily: his freedom is gone though
<hodapp>
Ashy: I mean he sorta never had that to begin with
<mikky>
clever: thanks for the heads-up
<drakonis>
hm yes his freedom
<emily>
Ashy: I didn't say he did perfectly :p (but, this is probably off-topic)
<Ashy>
yeah heh
<clever>
hodapp: i feel like even mentioning what secrets i am storing may draw in more attackers, lol
<mikky>
clever: isn't lvm a bit extraneous with zfs?
nD5Xjz has joined #nixos
<hodapp>
clever: just tell people it's justin bieber fanfic
<clever>
mikky: swap on zfs isnt stable, and swap on luks + zfs on luks leads to 2 password prompts
<clever>
mikky: so i use lvm to squish zfs+swap into a single luks device
<jasongrossman>
Just out of interest, what goes wrong if you put swap on ZFS?
<clever>
jasongrossman: typically, it just deadlocks
<jasongrossman>
Huh. Is that explicable?
<gchristensen>
zfs needs ram to ... swap ... and it doesn't have any
<drakonis>
wait... then zfs is really badly designed?
<gchristensen>
no, it is designed to use effectively use RAM
<gchristensen>
it is well designed for the "Not swap" use case and if you need the "Swapping" use case, it isn't the tool for that
<jasongrossman>
I wonder whether it's fixable by reducing ZFS's RAM usage then, which is easily done ... although it would have speed side-effects, so it's not a great solution. Just interested.
sir_guy_carleton has joined #nixos
<drakonis>
that makes it sound like the kind of filesystem that should be deployed in servers
<clever>
jasongrossman: it defaults to only using 50% of the ram at most
<jasongrossman>
drakonis: Exactly.
<mikky>
I've actually only seen the "don't swap to zvol" warning in a NixOS wiki page, dated "as of 2014-something"
galaxite has quit [Ping timeout: 265 seconds]
<gchristensen>
well if you want swap, just have a small ext4 partition or whatever
<jasongrossman>
drakonis: Some early versions of ZFS were pretty useless on memory-constrained desktop systems, because it was so tightly designed for servers. It's fine on desktops now, but its history is interesting IMO.
<drakonis>
it still sounds like it will handle different environments than slow storage and fast ram
<hodapp>
does it still have the issues with dedup requiring a ton of RAM?
<emily>
i think that's kind of... inherent to deduplication
<clever>
gchristensen: swap files generally perform worse then dedicated swap block devices
<jasongrossman>
hodapp: I believe so, by design. I think almost nobody uses dedup.
<clever>
gchristensen: and i want the swap to be encrypted, so it needs to be on luks
<gchristensen>
clever: d'oh that is what I meant
<mikky>
hodapp: s/issue/design
rouma7 has quit [Ping timeout: 244 seconds]
<clever>
and then luks asks for 2 passwords on bootup
<mikky>
you don't want to dedup on ZFS unless you have a very, very good reason to
<emily>
clever: you can just encrypt the two separate keys to a third key in a file
<drakonis>
jasongrossman, i'm not sure if it will age well with SSDs and optane/znand becoming commonplace
<clever>
mikky: i limit dedup to the dataset where i keep my virtualbox images and cryptocoin blockchains
<emily>
clever: and write an initramfs script to prompt for the password to decrypt that third key
<emily>
it's all just a pile of shell scripts mounting stuff anyway
<clever>
emily: and that just gets ugly, lvm is a lot simpler
<emily>
oh sure, I was assuming you couldn't use LVM for some reason, sorry.
<gchristensen>
emily: you don't need to go that far even
<mikky>
clever: if you can live without hibernation, use can use /etc/crypttab
<jasongrossman>
drakonis: I don't think it will survive btrfs becoming mainstream anyway, in the long run, because of the licensing issues. But in the long run we're all dead.
<clever>
mikky: i also want hibernation
<Ashy>
clever: i still havent learned to use lots of zfs datsets, sounds like a pretty useful way to structure things
<clever>
Ashy: i typically have a dataset for / /nix /home and then more depending on the use-case of the machine
* emily
doesn't think btrfs is going to become mainstream
<drakonis>
heh how cynical
<mikky>
Ashy: datasets come in very handy as soon as you start to do snapshots on a regular basis
<clever>
Ashy: /nix has snapshots disabled, all others have snapshots enabled
<jasongrossman>
Ashy: I find it simpler NOT to have many ZFS datasets. I'm sure there are pros and cons, but personally I definitely prefer it with just a few. Because otherwise I have to remember two levels of structure: the ZFS structure, and the directory structure.
<drakonis>
btrfs still has about two years until other solutions come up or the world changes significantly
<Ashy>
i think it will, once it's declared "ready for production" distros will start swapping to btrfs by default
<Ashy>
the vast majority of users always use the defaults, linux users included
<jasongrossman>
But yeah, I have separate ZFS datasets for snapshot a lot versus snapshot just a bit versus don't snapshot at all (more or less).
<jasongrossman>
Ashy: Nice pun.
<emily>
but zfs is more practical now, bcachefs seems to have more momentum in terms of being the "next big thing"...
<jasongrossman>
"swapping to btrfs"
orivej has quit [Ping timeout: 240 seconds]
<emily>
and the default is what Ubuntu does... Canonical are the ones who started shipping zfs.ko
<emily>
does it go through the same key derivation a password does?
<clever>
emily: its more of a key to decrypt the master keys
<emily>
right, that should be fine then I guess
<jasongrossman>
mikky: Sorry - I meant NixOS doesn't.
<emily>
I have to note here that you have an entire extra partition just to avoid saving a variable (the inputted password) and using it twice :p
<jasongrossman>
mikky: There are ways round that though.
<gchristensen>
emily: yeah maybe so but one was trivial to implement with existing Nix modules and the other is ?? :)
<clever>
emily: thats why i use lvm :P
<mikky>
jasongrossman: I'm all ears
<emily>
*nods* I wish NixOS's initramfs stuff was a bit cleaner than it seems to be
<jasongrossman>
mikky: I haven't thought about all the possible solutions, but one is to write a little script to add lines to /etc/nixos/hardware-configuration.nix.
<jasongrossman>
mikky: Another is to rerun nixos-generate-config.
<mikky>
both require root though, righ?
<jasongrossman>
mikky: Another is to use the (already existing) NixOS option to force import all pools.
<clever>
mikky, jasongrossman: with the most recent zfs (which nixos has), setting mountpoint= just works
<jasongrossman>
clever: Awesome!
<clever>
so you can `zfs set pool/dataset mountpoint=/foo` and it just does what you want
<mikky>
nice
<clever>
but nixos cant do that for the root or store datasets
<jasongrossman>
\o/
<clever>
so those still need mountpoint=legacy and a configuration.nix entry
<mikky>
store here means /nix?
hakujin1 has quit [Ping timeout: 265 seconds]
<clever>
/nix and /nix/store/
<mikky>
so we need legacy mounts only for / and /nix*
<mikky>
makes sense, I have yet to find a distro which has no problems with zfs mounts for abitrary paths
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
silver_ has quit [Read error: Connection reset by peer]
jasongrossman has quit [Ping timeout: 240 seconds]
n0qqe has joined #nixos
jasongrossman has joined #nixos
<jasongrossman>
clever: Ugh. When I try to rebuild and switch to stable, I get the same problem. And this computer doesn't have an already-installed stable to switch to. I can install one using the installer, of course, but that'll take ages
<jasongrossman>
noqqe has quit [Ping timeout: 240 seconds]
n0qqe is now known as noqqe
<clever>
jasongrossman: try an older stable channel, it may also have the bug? and does `sudo nix-channel --list` show the stable under the name `nixos` exactly?
jmeredith has quit [Quit: Connection closed for inactivity]
<jasongrossman>
clever: Don't understand - sorry. My problem now is trying to switch to ANY other system (without using the installer).
<jasongrossman>
clever: It's not a huge deal if I have to use the installer though. Just take a while.
<clever>
jasongrossman: try switching to 17.09 next?
<jasongrossman>
clever: OK, trying that.
galaxite has joined #nixos
oida has quit [Remote host closed the connection]
oida has joined #nixos
<mikky>
about boot.initrd.network.ssh.hostECDSAKey - how does this one work? is this the path to the key source at runtime or at ramdisk generate time?
<jasongrossman>
I'm trying to downgrade to 17.09 from unstable, and getting this error. Any advice please?
<jasongrossman>
error: The unique option `powerManagement.cpuFreqGovernor' is defined multiple times, in `/etc/nixos/hardware-configuration.nix' and `/nix/var/nix/profiles/per-user/root/channels/nixos/nixos/modules/config/power-management.nix'.
<jasongrossman>
<clever>
jasongrossman: comment that line out in hardware-configuration.nix
<jasongrossman>
I should have mentioned: my hardware-configuration.nix has almost nothing in it apart from the defaults. This is being pulled in from I-know-not-where.
<clever>
mikky: that is an unquoted path on the normal system, nix will copy the key into /nix/store and make it world-readable
<jasongrossman>
Oh wait. that's not true.
<clever>
jasongrossman: nixos-generate-config made that entry, but 17.09 conflicts with that default
<jasongrossman>
That line IS in hardware-configuration.nix. Commenting it now.
<jasongrossman>
Thanks.
<jasongrossman>
Will you please remind me to uncomment it later? kthx
<semilattice>
If it's conflicted by default, surely running nixos-generate-config again would fix it?
<clever>
semilattice: you would have to run the 17.09 version of generate-config
<jasongrossman>
semilattice: :-)
<clever>
which you can only get by successfully finishing nixos-rebuild
^il is now known as vil
<semilattice>
oh true, so nix-shell -I .... -p nix
<jasongrossman>
If I can successfully finish nixos-rebuild then my whole life will be perfect.
<clever>
semilattice: and its not in the nix package
<semilattice>
damn
<semilattice>
is there a nixos package you can get it from?
<clever>
jasongrossman: id also recomend using the boot sub-command when downgrading
<clever>
semilattice: its a bit hidden and simpler to just comment that line out
<jasongrossman>
clever: Yes, thanks.
<semilattice>
clever: So I uh, I got distracted from the issue that I came in here for, but as to "semilattice: and if you try to eval js-env in `nix repl '<nixpkgs>'` does it print the expected thing?", it prints a derivation which is expected
<semilattice>
in this case it pritns «derivation /nix/store/zx3b1fxnzdpgv1x4s3ma18z48faxs2zx-javascript-env.drv»
<clever>
semilattice: and js-env isnt in nixpkgs, so the override appears to be working
<semilattice>
This override is working, but I can't see the attributes for it
<semilattice>
no matter what I try
<clever>
semilattice: oh right
<clever>
semilattice: nix repl cant tab-complete anything with a - in the name
<clever>
only nix-repl can do that
<clever>
nix repl parses that as subtraction and breaks
<semilattice>
clever: oh, wow thanks, that's a silly bug
<clever>
renaming it will be simpler
<semilattice>
yeah, nix-repl is hard for me to install anyway
<semilattice>
since it seems you need to have root? Nix database directory ‘/nix/var/nix/db’ is not writable: Permission denied
<clever>
nope
<clever>
export NIX_REMOTE=daemon
<clever>
nix-repl uses nix1, which cant auto-detect that
<infinisil>
,nix-repl semilattice
<{^_^}>
semilattice: To use nix-repl with Nix 2.0, either use the new `nix repl` or `NIX_REMOTE=daemon nix-repl`. Just using `nix-repl` gives an error because it is linked to Nix 1.x which requires NIX_REMOTE to be set correctly, while Nix 2.0 doesn't (and unfortunately it wasn't kept for backwards compatibility)
<clever>
nix2 can auto-detect it, so the config setting the var was removed
<semilattice>
ahhhh, that clarifies a bunch
<semilattice>
Yeah, renaming it is easier lol `nix-repl> js-env
<semilattice>
error: hash ‘d+nbxBUGKg7Arpsvbnlq61mc12ek3EY8EQldM3GPAhWJ1UVxC6TDGbIvUMNU6obBX3i1+ptCIzV4vq0gFPEGVQ==’ has wrong length for hash type ‘sha512’
<semilattice>
«derivation
<semilattice>
`
<clever>
i dont think nix supports base64 in hashes?
<semilattice>
yeah that's something that I didn't add so no idea where it is
Supersonic has quit [Ping timeout: 255 seconds]
<clever>
semilattice: can you link your pastebin again?
<{^_^}>
[nixos-artwork] @samueldr pushed 2 commits to master: https://git.io/fNPoH
rprije has quit [Ping timeout: 260 seconds]
mizu_no_oto has joined #nixos
johnw has quit [Ping timeout: 260 seconds]
kyren has joined #nixos
oida has quit [Remote host closed the connection]
oida has joined #nixos
mizu_no_oto has quit [Quit: Computer has gone to sleep.]
trylist has joined #nixos
<trylist>
hi guys, does anyone know anything about screensavers/screen sleep. Mine keeps going black after 10 minutes while watching movies on firefox, really annoying to have to constantly move the mouse.
<trylist>
I've tried xset s off
<clever>
[28751:519:0803/233526.020873:ERROR:power_save_blocker_x11.cc(403)] No response to Uninhibit() request!
<clever>
[28751:519:0803/233526.195430:ERROR:object_proxy.cc(617)] Failed to call method: org.freedesktop.ScreenSaver.Inhibit: object_path= /org/freedesktop/ScreenSaver: org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.ScreenSaver was not provided by any .service files
<clever>
[28751:519:0803/233526.195462:ERROR:power_save_blocker_x11.cc(330)] No response to Inhibit() request!
<clever>
trylist: related, i can see chrome trying to turn dpms on/off every time pause/unpause things
<clever>
but in my case, dpms is just plain broken and never goes to sleep
<ivan>
trylist: did you try xset s off -dpms
<trylist>
I'll let you know in 10 minutes or so
<trylist>
thanks
<ivan>
which screensaver thing are you using?
<{^_^}>
[nixpkgs] @aanderse opened pull request #44437 → added the apache auth cookie perl module to nixpkgs → https://git.io/fNPKR
<trylist>
I don't actually know. this is still a pretty basic i3wm install with plain xorg
<trylist>
it's hard to find results in google
<trylist>
I didn't explicitly install anything if that's what you're asking
kyren has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
hakujin1 has joined #nixos
hamishmack has joined #nixos
<init_6>
where to read abourt kernel developing with nixos?
<clever>
init_6: that info is also now in the nixos manual
<clever>
Ashy: ah, thanks, was going to find that
<init_6>
Ashy clever thx
hakujin1 has quit [Ping timeout: 265 seconds]
<Ashy>
init_6: are you doing kernel dev for work?
<trylist>
@clever it looks like that worked, thanks! It feels like this should be a part of firefox and video players in general, I've haven't had that issue on other distros. Would you know of a lead to investigate?
<trylist>
That chrome message you pasted looks like some missed config that might be fixable
<clever>
trylist: try searching its codebase for org.freedesktop.ScreenSaver Uninhibit or Inhibit
<clever>
trylist: and just poking around in dfeet for a minute finds inhibit/uninhibit under org.freedesktop.PowerManagement, via interfaces and org.freedesktop.powermanagement.inhibit
<pie_>
this happens no matter what I pass to -e
<clever>
pie_: nix2 has a bug involving symlinks that point to non-existant files
<pie_>
clever, ugh. wat do
<clever>
pie_: you need to uninstall python
<trylist>
ah, so the browsers are looking for ScreenSaver but nixos sets it up as PowerManagement?
<pie_>
clever, ah ok
<pie_>
clever, sounds kind of bad :p
<pie_>
oh hey that worked.
obadz has joined #nixos
_cyril_ has quit [Ping timeout: 260 seconds]
<clever>
trylist: maybe, or maybe xscreensaver is supposed to forward it over
<trylist>
ok, this is a very good clue. much appreciated
lo_mlatu has joined #nixos
<clever>
[clever@amd-nixos:~]$ xscreensaver
<clever>
xscreensaver: 00:12:59: already running on display :0.0 (window 0x2400001) from process 23295 (clever@amd-nixos).
<clever>
trylist: so i already have xscreensaver running
<lo_mlatu>
help! I jusy installed i3 wm, and copied some .Xresources setting, now I login and only get a xterm, switch back to the gnome desktop showed the same and I have to manually start gnome-shell now, even after deleting .Xresource. What's wrong with my machine? How.can I fix it?
<clever>
trylist: and its set to disabled in the xfce settings
<clever>
lo_mlatu: which display manager?
<lo_mlatu>
in the i3 generation using lightdm, in the gnome generation using gdm, used to work fine
hakujin1 has joined #nixos
<clever>
lo_mlatu: does lightdm give you a selection to choose between several combinaitons of desktop managers?
<trylist>
@clever I'm extremely unfamiliar with X and it's interactions, but it sounds like the browsers are looking for some service waiting to turn off and on the screensaver and not finding it
hakujin1 has quit [Ping timeout: 276 seconds]
<clever>
trylist: yeah, its probably modern changes to stop xscreensaver seperately, so it wont render over things after dpms is shut off
<lo_mlatu>
clever: oh I found it and recovered now, thanks a lot!
<trylist>
@clever what's synergys btw?
fragamus has quit [Ping timeout: 240 seconds]
<clever>
trylist: synergy is a tool for sharing the mouse/keyboard between several machines
<clever>
trylist: the 2 machines behave more like a single machine in a dual-monitor setup
<trylist>
oh synergy, yeah I found that, thought synergys was soemthing different
<clever>
trylist: in my case, the desktop has 3 monitors, and the laptop is beside it, and then it behaves like a 4 monitor setup
<clever>
synergys is the server, and synergyc is the client
<Enzime>
should build fine now that the mpv 0.29.0 macOS fixes were merged
jasongrossman has quit [Ping timeout: 268 seconds]
<samueldr>
Enzime: starting it
<Enzime>
samueldr: do you have permission for it to build on macOS?
<samueldr>
yes :)
<Enzime>
:)
<samueldr>
(I read the thing before blindly starting it)
<Enzime>
samueldr: I was just checking in case you were building it without permission for macOS which might be a bit pointless :p
<samueldr>
yeah, I do understand
<Enzime>
👍
<ryantm>
Is there a way to pass show-trace into nix-env?
<clever>
ryantm: nix-env --show-trace
ArahaelPi has quit [Read error: No route to host]
sir_guy_carleton has quit [Quit: WeeChat 2.0]
<ryantm>
clever: That doesn't seem to have an effect on my particular situation. I'm trying to debug why entering a nix-shell followed by `make nixpkgs/packages-unstable.json` on nixos-homepage repo fails.
<ryantm>
I don't see any difference in the stack trace with that flag
<clever>
ryantm: can you pastebin the whole console output when it fails?, including the commands you ran
<nikki93>
hey all -- so i'm working with a package that has the following line: `preConfigure = "$shell ./platform/unix/automagic";`
ericsagnes has joined #nixos
<clever>
ryantm: that is not a nix failure, so --show-trace will never do anything
<clever>
ryantm: wait correction
<clever>
line 11 is a nix error
<ryantm>
clever: right line 11 is the nix error, but the backtrace seems incomplete.
<clever>
ryantm: try running the nix-env command on 8 directly
<nikki93>
i'm building my own binary that has what that package has as a subdirectory -- how could i edit tha tline to say 'run $shell ./platform/unix/automagic' in the directory 'foo'?
<nikki93>
gonna just try 'cd foo && $shell ./plat...' for now
<maerwald>
and the default display manager doesn't let me choose the session
<{^_^}>
[nixpkgs] @Mic92 merged pull request #44442 → Add MATE to list of desktop managers in manual → https://git.io/fNPPz
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @smithtim to master « nixos/doc: added MATE to list of desktop managers in manual (#44442) »: https://git.io/fNPMt
<{^_^}>
[nixpkgs] @xeji pushed commit from @jfrankenau to master « reaper: 5.93 -> 5.94 (#44445) »: https://git.io/fNPym
Izorkin has quit [Ping timeout: 268 seconds]
Izorkin has joined #nixos
hakujin1 has joined #nixos
logzet_ has quit [Ping timeout: 265 seconds]
logzet has joined #nixos
hakujin1 has quit [Ping timeout: 265 seconds]
<pie_>
I set virt-manager to load an iso file to a cdrom drive, why the heck is it changing the owner/group to root:root and then complaining about permissions??
<{^_^}>
[nixpkgs] @xeji pushed commit from @dasJ to master « scriptaculous: Init at 1.9.0 (#44429) »: https://git.io/fNPya
<typetetris>
bigvalen: nixos manual chapter 4: `nixos-rebuild --upgrade` is an update and a switch.
<typetetris>
sorry `nixos-rebuild switch --upgrade` is the right one.
hakujin1 has joined #nixos
<bigvalen>
Ah, OK. Well, that fails, as above.
<bigvalen>
Some searching seems to indicate it's a problem running a really old nix daemon.
<typetetris>
Hmm, I remember, there was some situation where you had to install nix2 first and do the upgrade with the nix2 tools. But I can't judge, wether it applies here.
<{^_^}>
[nixpkgs] @xeji pushed commit from @dasJ to master « smarty3: Init at 3.1.32 (#44427) »: https://git.io/fNPSw
<typetetris>
bigvalen: Did it work now?
Lisanna has joined #nixos
<Lisanna>
Is there a one-liner or something to delete a store path and all of its referrers?
<clever>
Lisanna: you would need to loop nix-store --delete over its nix-store -qR
<clever>
Lisanna: it has to be a loop, because --delete will stop at the first rooted path
<clever>
if your goal is to get rid of all deps of a given path
<Lisanna>
clever hrm, but those referrers might have their own referrers
<bigvalen>
I'll go try it. Sorry, got distracted when I realised that ZFS didn't work as I expected so I just wasted a load of time with it. Sigh.
<clever>
if you want to get rid of the things depending on X, just normal nix-store --delete X is enough
<Lisanna>
not all deps, but a path + all referrers
<clever>
nix never allows a refernece to something which doesnt exist
<Lisanna>
clever nix doesn't let me delete it if it has referrers.
<Lisanna>
cannot delete path since it is still alive
<bigvalen>
I thought I could go from a zfs mirror to a four-disk raidz2 by pulling a disk from the mirror, creating a three-disk raidz2, do the zfs send, then when it was done, add the disk to the raidz2 pool. Turns out .... no. You can't.
<clever>
Lisanna: thats because of roots, not referrers
<Lisanna>
...oh
<clever>
Lisanna: nix-store -q --roots X
<Lisanna>
clever OK, so just delete all of those and then nix-store --delete will work?
<clever>
bigvalen: i think (havent confirmed) that you can upgrade a lone disk into a mirror between olddisk + raidz2
<clever>
Lisanna: depending on what they are, yeah
<clever>
Lisanna: you need to make sure your not deleting the active generation of any profiles, or the entire profile will be gone
<Lisanna>
yeah I know that
<Lisanna>
these are hydra build products
<clever>
in the hydra root directory?
<clever>
that should only be managed via hydra utils, and you can break hydra if you force it
<bigvalen>
clever maybe I could have been able to do that, had I setup a raidz2 with an offline or missing disk...
<Lisanna>
clever ...good point.......
<clever>
Lisanna: one min
<Lisanna>
*checks if hydra is broken*
<clever>
Lisanna: hydra will believe the path exists, and wont try to rebuild it, because it was rooted and fully protected from gc
<clever>
and if hydra ever needs it ....
<bigvalen>
Seems some people create a raidz2 with 3 disks + sparse file, then mark the sparse file as offline before putting any data on it.
<bigvalen>
This would have been good to research before doing all the other work :)
<clever>
bigvalen: you can also use the file backend for testing, just create a heap of dummy disks and experiment
<Lisanna>
K... guess I'm going to go touch a low-level dependency to make any bad state go away
<clever>
Lisanna: i can explain how to properly gc hydra things, one min
<Lisanna>
clever ok
simukis has quit [Quit: simukis]
<clever>
Lisanna: first, do you know where the keep nr is in your config?
<{^_^}>
[nixpkgs] @rnhmjoj opened pull request #44446 → Make python{2,3} version explicit → https://git.io/fNPSS
<Lisanna>
clever yeah
<clever>
Lisanna: hydra will always root the entire build-time closure for the last $keepnr evals in the jobset
<clever>
Lisanna: how high do you have it set and how many nixpkgs changes have those evals gone thru?
<Lisanna>
clever I have it set to 100 right now, 'cause I never want to have to rebuild these products after they've been GC'd, since that can confuse nix
<Lisanna>
but sometimes I need to purge things
jdnavarro has quit [Ping timeout: 264 seconds]
<clever>
Lisanna: temporarily reduce it, then `systemctl start hydra-update-gc-roots` to tell hydra to prune its own roots
<Lisanna>
clever OK, so that will remove all the GC roots, and then I can nix-store --delete?
<clever>
yeah
<Lisanna>
alright, cool, thanks
<clever>
that will remove the roots that are outside of the keepnr
Anton-Latukha has joined #nixos
rprije has quit [Remote host closed the connection]
<epta>
how to specify a subdir of git repository with `fetchgit`?
Ridout is now known as b
<cocreature>
epta: try something like (fetchgit …) + "/subdir"
averell- has joined #nixos
hakujin1 has joined #nixos
averell- is now known as averell
aarvar has quit [Ping timeout: 264 seconds]
<elvishjerricco>
Have the performance problems with builtins.fetchGit been resolved yet? Last I tried it, the fact that it uses one massive repo caused any new fetches to take ages, due to some expensive algorithm that git runs over the whole repo.
<elvishjerricco>
It was only a big problem if you had multiple large, totally unrelated histories. But that's pretty common with something like fetchGit...
hakujin1 has quit [Ping timeout: 255 seconds]
midchildan_ has quit []
__monty__ has joined #nixos
vmandela has joined #nixos
hakujin1 has joined #nixos
work_ has joined #nixos
alexteves has joined #nixos
hakujin1 has quit [Ping timeout: 265 seconds]
Sonarpulse has joined #nixos
<{^_^}>
[nixpkgs] @pSub pushed 12 commits to add-missing-licenses: https://git.io/fNPhp
pikajude has quit [Remote host closed the connection]
pikajude has joined #nixos
mguex has joined #nixos
vmandela has quit [Quit: Leaving]
ambro718 has quit [Quit: Konversation terminated!]
<juhe>
I have a trouble getting diod network share mounted after network is online. This results in degraded boots because mount was tried before network was online. The systemd mount unit file is generated from /etc/fstab which in turn is generated from config.fileSystems. Is there a way to specify this dependency in fileSystems somehow?
justbeingglad has left #nixos [#nixos]
<ajs124>
juhe, you can add "noauto" "x-systemd.automount" to your fileSystems.<name?>.options
iqubic has joined #nixos
<juhe>
ajs124: would that mean I have to manually mount then?
<ajs124>
That way, the filesystem gets mounted, the first time something tries to access it.
<juhe>
ajs124: that's actually neat... Will try it.
selfsymmetric-mu has joined #nixos
<selfsymmetric-mu>
What's the standard way to set up a ram drive on NixOS?
woffs has joined #nixos
<ajs124>
just use tmpfs? or what exactly do you mean by ram drive?
juhe has quit [Remote host closed the connection]
<iqubic>
selfsymmetric-mu: What is a "ram drive"?
<selfsymmetric-mu>
ajs124: I've never made one before so I'm just looking for basics, nothing fancy. I have a directory that I need to do faster reads from, so I want to move it to ram on boot.
<selfsymmetric-mu>
Maybe there's a way to just tell the OS to cache it? I'm new to the concept.
<Yaniel>
linux does that for you automagically
<iqubic>
Why do you need faster reads from that directory?
<selfsymmetric-mu>
iqubic: To be honest faster reads will probably not help me in this situation. I'm getting weirdly slow org-mode agenda performance even though I've restricted the number of files to a dozen or so. I'm bisected my Emacs configuration with bughunter and confirmed that it's nothing to do with my config.
jperras has joined #nixos
<iqubic>
selfsymmetric-mu: Perhaps you should talk to the people in #emacs and also #org-mode to see if others have ran into this issue as well.
jperras has quit [Client Quit]
<iqubic>
I'm fairly certain that faster reads won't actually help you here.
<selfsymmetric-mu>
iqubic: Yeah, I'm thinking the same thing. Just trying a bunch of tricks. Thanks for the pointers anyway, I may use tmpfs for other things.
<juhe>
ajs124: Hm, unfortunately I had to start the systemd mount service manually after boot, it seems the automount systemd unit is not generated automatically.
<ajs124>
juhe, that's strange. I use this and it works.
<juhe>
ajs124: Maybe diod is special, but... I've found "_netdev" option in mount man page, will try that one too.
juhe has quit [Quit: Leaving.]
hakujin1 has joined #nixos
<ajs124>
Hm, I use it with sshfs and nfs
juhe has joined #nixos
<juhe>
ajs124: It seems that "_netdev" option helped. I'll used that one.
hakujin1 has quit [Ping timeout: 276 seconds]
juhe has quit [Remote host closed the connection]
juhe has joined #nixos
trylist has joined #nixos
drakonis_ has joined #nixos
<tilpner>
Huh, nix copy just grew a file by 12288 bytes (as counted by wc -c)
jdnavarro has joined #nixos
<ivan>
it's a nix worm
judson has joined #nixos
trylist has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @xeji pushed commit from @dtzWill to master « f2fs-tools: 1.10.0 -> 1.11.0 (#43715) »: https://git.io/fNXJB
<{^_^}>
[nixpkgs] @xeji merged pull request #44392 → graphite: make systemd create /run directory for us → https://git.io/fNiBc
<{^_^}>
[nixpkgs] @xeji pushed commit from @basvandijk to master « graphite: make systemd create /run directory for us (#44392) »: https://git.io/fNXJu
<drakonis_>
LnL, /nix/store/dkh7l9a4sx7zqh8riqbj3z21sz25p8xy-bash-4.4-p23/bin/bash: /build/source/lib/elixir/generate_app.escript: /usr/bin/env: bad interpreter: No such file or directory
<infinisil>
I want `nix-doctor` that finds problems and fixes/tells you how to fix them
<infinisil>
drakonis_: You need to call patchShebangs on that file
dcci has joined #nixos
<drakonis_>
yes
<drakonis_>
but i'm first trying to figure out why updating quakespasm is breaking it
<dcci>
LnL: samueldr looks like rebooting "fixed" it
<__monty__>
samueldr: Imo something that works basically everywhere unless someone's (mis)configured a couple very shell specific files is a better default than something that *only* works with two specific shells. Maybe it's just me though.
<LnL>
dcci: hmm well I'm not sure why the daemon wasn't running, unless you didn't reboot since you upgraded from 2.0
<samueldr>
__monty__: open a PR fixing this if you think it is the right solution for macOS installation
<LnL>
dcci: 2.0 had a problem where the service wouldn't restart automatically
work_ has quit [Quit: Connection closed for inactivity]
Fare has quit [Ping timeout: 240 seconds]
<samueldr>
but __monty__, do know that there are multiple conflicting opinions about the right place to put this *so it works* (not nix-specific, I'm talking environment at large)
<judson>
Caught in a dilemma: I think I need a more modern simp_le-client than is in NxOS stable on my NixOps machines. All in all I'd rather not switch to my master checkout of nixpkgs; but the alternative seems to be maintaining a little subtree of python until 18.09 drops.
<LnL>
drakonis_: sounds like I forgot to test the linux build then
f0i has quit [Ping timeout: 256 seconds]
cizra has left #nixos [#nixos]
<samueldr>
and then when a user follows up the advice they deemed right, it probably leaves the other locations to load the environment not working
<{^_^}>
[nixpkgs] @xeji pushed commit from @bignaux to master « caprice32: unstable-2018-02-10 -> unstable-2018-03-05 (#44462) »: https://git.io/fNXkR
<samueldr>
reliably setting the environment programatically for a user is **hard**
<drakonis_>
now, i wonder if i'm doing nix-build wrong
<drakonis_>
it should be run with -A in the repository directory right?
<__monty__>
samueldr: "Do you have a .bash_profile?" Is quite an easy troubleshooting question though. "Let's implement environment setup not just once but once for each shell that's reasonably popular." Sounds like a much higher maintenance burden. I'm not in the know enough to fix this, just trying to influence someone who is. I'm easy to appease though, make fish a first class citizen of the installation process
<__monty__>
and I'll shut up ; )
<samueldr>
__monty__: kinda worked on it for non-nixos linux
<drakonis_>
but i suppose they're not policy compliant
<samueldr>
not at all, but in some ways really easy to deploy
<__monty__>
samueldr: A warning during the installation process would be easy imo, i.e. if exists(.bash_profile...): warn "You might need to read this FAQ:..." Much friendlier than "If you're not using either bash or zsh, figure it out."
<samueldr>
so it's not for upstream $distro use, but for deployment
<judson>
Any thoughs about getting ACME working on NixOps? I've got a cert that expires on the 10th, but I'm getting what looks like "the client is too old" errors
abueide has joined #nixos
<samueldr>
judson: hmmm, let me check my 18.03 things serving acme certs
<samueldr>
» expires on 8 august 2018
<samueldr>
good sign there
<samueldr>
judson: I'll probably have to figure something too
<judson>
samueldr, the acme module seems to use simp_le which is 0.6.0 in 18.03
Fare has quit [Ping timeout: 244 seconds]
Guanin has joined #nixos
<judson>
I'm seriously contemplating a local overlay for enough of pythonModules to build a modern simp_le
<judson>
It's only, what, 20 packages ;(
<samueldr>
oh, in my case it's probably something else
<ajs124>
simp_le on 18.03 is working fine for me
<dhess>
whoa, tha'ts not good
<samueldr>
well, it's possibly
<samueldr>
(the container thing I'm using doesn't have any inode left)
<judson>
It worked to register them, but renewels are failing.
<{^_^}>
[nixpkgs] @xeji pushed commit from @bignaux to master « cadence: init at 0.9.0 (#44281) »: https://git.io/fNXIp
ng0 has joined #nixos
<drakonis_>
i don't get it, how is quakespasm building on hydra but not locally
<dhess>
yay, ghc843 built on aarch64 with my fixes!
<dhess>
However, it took 24 hours on a Jetson TX1. :\
<dhess>
but it did build
<Lycurgus>
apparently instant stuffs make up for the slow
<infinisil>
dhess++
<{^_^}>
dhess's karma got increased to 1
<infinisil>
Nice!
<dhess>
I'm building my Haskell projects with it. If those build successfully, I'll submit a PR to nixpkgs.
<Lycurgus>
1 right now can make up for any number of ass dragons
<dhess>
but it looks pretty good at this point
<dhess>
dhess--
<dhess>
oh well
<samueldr>
dhess: how does the TX1 compare to the community box?
<samueldr>
(you could build ghc on it, and then locally use that build for your projects)
<dhess>
samueldr: it only has 4 cores and 4GB. So: much less beefy. However, the issue with GHC on aarch64 seems to be that there's a memory ordering issue somewhere that causes it to hang on parallel builds.
<dhess>
so: I doubt that individual Haskell packages (including GHC) will build a whole lot faster on the community builder.
<dhess>
but it will be able to build more packages at once, of course.
<dhess>
because my Nixpkgs "fix" is to disable parallel builds for Haskell packages when buildTarget is aarch64
<samueldr>
ah, yeah, single-core performance could cause a reversal here :)
<judson>
How terrible an idea would be to do a nixops deploy with -I nixpkgs=~/nixpkgs, get the cert renewed, and then deploy again without?
<samueldr>
judson: a pretty good test run for the 18.03 to 18.09 upgrade :)
<samueldr>
let's hope you have backups
<samueldr>
(cheekily, there shouldn't be an issues hopefully)
<infinisil>
pie___: Let us debug your haskell problems with nix here perhaps? :)
kyren has joined #nixos
<pie___>
ah crap i tried to do "hoppy-generator" = unst.haskellPackages."hoppy-generator"; but that just understandably breaks on the ghc version i think :o
<pie___>
* :p
<infinisil>
unst??
<infinisil>
Oh unstable
<pie___>
yeah
<pie___>
so how does version checking in hackage-packages work anyway
<pie___>
or does it just use whatever is specified for hackage?
<infinisil>
Packages are either in stackage or not
<pie___>
ah yeah. ok so it does seem that hackage has strict versions set fot these packages
<pie___>
but im not sure why this didnt fail before
<infinisil>
the stackage ones come from the stackage lts versions
Lycurgus has quit [Quit: Exeunt]
<infinisil>
the non-stackage ones are from the latest hackage version
<infinisil>
And there's no version resolution beyond that happening
<pie___>
ok so having looked at this im better off building this package off unstable, but then i get an issue with the cabal versoi
<infinisil>
Breakages need to be fixed up manually
<pie___>
Setup: Encountered missing dependencies:
<pie___>
Cabal >=1.20 && <2.1
hakujin1 has joined #nixos
<pie___>
because i think its cabal 2.2
<infinisil>
Yeah you wanna try jailbreak
<infinisil>
jailbreak makes cabal ignore the version bounds
<samueldr>
except for the website, it's behind a single toggle for the whole stack?
<samueldr>
hah, I was about to say "there's a PR about that, I should probably poke people a bunch more"
<drakonis_>
samueldr, the toggle is hidden though :V
<samueldr>
(it's understandable that bringing in a whole development stack for this feature could cause issues)
<samueldr>
drakonis_: I think I see, while it's in the manual, it's not in the tools documentation
jtojnar has joined #nixos
<samueldr>
though it's a bit hard to actually think about putting that into `man nix*` as it's not part of nix, it's a mechanism entirely implemented into nixpkgs (unless I'm mistaken)
<pie___>
> nix-repl
<pie___>
error: Nix database directory ‘/nix/var/nix/db’ is not writable: Permission denied
<pie___>
so uhh nix-repl should be marked deprecated or something?
<samueldr>
a backwards incompatible change from nix makes it so you need to do *something* (don't remember what off the top of my head) to use nix-repl
<pie___>
btw does nix-repl have a way to find out where an attribute is defined
<samueldr>
I wanted to look into that, see if there was any missing feature
<pie___>
* nix repl
Fare has joined #nixos
<symphorien>
export NIX_REMOTE=daemon
<samueldr>
and if there aren't do a package which aliases nix-repl -> nix repl
<samueldr>
symphorien++
<{^_^}>
symphorien's karma got increased to 2
<pie___>
half the time i just cant figure out how to read the source of something in nixpkgs because i just cant find it
<infinisil>
samueldr: Your new package browser demo messes with the browser "Back" action for me :/
<drakonis_>
reading the pull request, well its ok
<drakonis_>
but i don't get it
<drakonis_>
i don't get something
<samueldr>
infinisil: not only for you, but it shouldn't trap you
<samueldr>
if it does there's a bug, otherwise it's a hard to define "right" behaviour
<infinisil>
,nix-repl
<{^_^}>
To use nix-repl with Nix 2.0, either use the new `nix repl` or `NIX_REMOTE=daemon nix-repl`. Just using `nix-repl` gives an error because it is linked to Nix 1.x which requires NIX_REMOTE to be set correctly, while Nix 2.0 doesn't (and unfortunately it wasn't kept for backwards compatibility)
orivej has quit [Ping timeout: 256 seconds]
<infinisil>
pie___: ^
<pie___>
aha.
kyren has joined #nixos
<infinisil>
samueldr: Ah yeah, not a trap, just that every javascript action is stepped through when going back
marusich has quit [Quit: Leaving]
<samueldr>
infinisil: that's a bit of bikeshedding on the tool, the back button *needs* to still do things in these times of SPAs
<samueldr>
now, what should be considered a navigation and what shouldn't
<samueldr>
it's **hard** when your results set change live
<samueldr>
if you don't update the history with the searchbox location, then change page, then press back, what should happen?
<infinisil>
As long as the page doesn't reload I don't think it needs a new entry
<infinisil>
imo
<samueldr>
infinisil: you see this with your developer eyes, where a "page reload" is something tangible
<samueldr>
but the page doesn't reload when you change channels on the top
<samueldr>
while it would be realistic to expect the back button to go back
kyren has quit [Client Quit]
* samueldr
removes nixer hat, puts on webdev hat
__monty__ has quit [Quit: leaving]
<infinisil>
Hmm.. I personally would prefer it not to do even that
<samueldr>
and there are people that say nix is hard :)
<infinisil>
I never see websites do that
<samueldr>
infinisil: many websites have completely broken the expectations set by the web :/
<samueldr>
and not necessarily for the best
<infinisil>
Clicking a bunch of javascript buttons that instantly transition the page shouldn't clutter the history
<samueldr>
again, you see "javascript buttons" but that's not what the history is about
<samueldr>
it's not about http requests
<infinisil>
I noticed because I needed to click "back" like 30 times to actually get to the last webpage!
<samueldr>
but yeah, it's hard to define the best behaviour in this age where all websites break the expectations from before
<infinisil>
Hmm..
<infinisil>
I'll just leave a comment in the PR I guess
<samueldr>
yes thanks :)
<samueldr>
(even though I know it'll cause disagreements and further bike shedding :))
<samueldr>
(on of my peeve is that, your website shouldn't act differently whether it's an SPA with js actions or completely http-based app)
<samueldr>
(clicking next to get to the next page of result should always push on the history stack, and back should always pop)
<samueldr>
(but then you have the fuzzy lines of live results)
<pie___>
samueldr, i feel like i need to rever to an older haskellPackages or something...
<pie___>
* infinisil
<samueldr>
I was wondering what I could do about that :)
jumblemuddle has joined #nixos
<samueldr>
hmm, can't remember the function to find where a package is defined :(
drakonis_ is now known as Drakonis
<jumblemuddle>
Is there a way to install additional packages into the steam FHS? I'm trying to use steam-run to run a game, but I'm getting a "libxcb.so.1: cannot open shared object file: No such file or directory" error
<{^_^}>
#32881 (by uwap, 32 weeks ago, open): steam / steam-run: make it possible / document how to override steamrt dependencies
<Drakonis>
there are games that use additional libraries not provided by steam
<Drakonis>
alternatively, run steam from a flatpak
<jumblemuddle>
Where's the documentation on steam-run?
<Drakonis>
speaking of which, i should try doing that
fragamus has joined #nixos
<pie___>
maybe language repos should be in subrepos with their own histories...?
<Drakonis>
use submodules for that :V
<pie___>
thats what i meant
<pie___>
i think
<Drakonis>
using flatpak may as well solve most of my headaches wrt proprietary packages
<samueldr>
unless for each commit in $submodule a commit to update $submodule in nixpkgs is made, it becomes a mess of a moving target to build and test
<Drakonis>
automation
<samueldr>
automation doesn't provide a fixed point
<Drakonis>
no lockstep updating though
<samueldr>
(I do agree that a separate repo would be easier to use for development of that environment)