<sphalerite_>
Dezgeg: do you have a good reference on the MMC protocol where I could work out how to add range-protecting functionality to mmc-utils?
<erlandsona>
Basically reaper is a daw (digital audio workstation or music production software) and I'm trying to get the "Native Linux" version installed with NixOS.
<boomshroom>
First thing he mentions is libSwell.
<erlandsona>
If you look in the reaper_linux_x86_64/REAPER there's a libSwell pre-packed with the tar.xz now.
<erlandsona>
The video is a touch old.
<Dezgeg>
hmm, IIRC the spec is non-public
<boomshroom>
Well that sounds like a problem
<erlandsona>
I wouldn't know :shrug:
<erlandsona>
Why do you say that?
<boomshroom>
You just said that you gave me outdated information.
<erlandsona>
Apologies but that's all I have to work from too?
<sphalerite_>
Dezgeg: whyyyyyyyyyy :'(
<Dezgeg>
MMC_SEND_WRITE_PROT is the command to query write protection status so that should be safe to try
<erlandsona>
So basically there's the two binaries that get copied to the $out in builder.sh `libSwell` and `reaper5` and whatever else is under the REAPER directory. But from the video the guy seems to be able to just run the reaper5 bin and it just works?
<erlandsona>
I figured I should be able to utilize buildFHSUserEnv for the derivation and be able to accomplish something similar.
<clever>
erlandsona: if you run `file` on reapter5, what does it say?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 15aa4fc Will Dietz: networkmanagerapplet: 1.8.2 -> 1.8.6...
<NixOS_GitHub>
[nixpkgs] adisbladis pushed 1 new commit to master: https://git.io/vNDI8
<erlandsona>
clever: result/bin/reaper: a /nix/store/65l6hr8snf4v823f974k97jc65i7bhvf-bash-4.4-p12/bin/bash script, ASCII text executable
<clever>
erlandsona: can you gist that bash script?
<erlandsona>
It's all in my github repo `/erlandsona/reaper4nixos.git`
swayf has joined #nixos
<boomshroom>
I'm pretty sure that the files are quite a bit smaller than they need to be.
<clever>
erlandsona: you dont need to use a builder.sh, pkgs.runCommand would do the same thing
<clever>
but that doesnt include the contents of the reaper bash script
<boomshroom>
If the application is just the binary and libSwell, it should be possible to do it with just patchElf.
<clever>
/nix/store/j4vcckdcbim2qqyr0kjw9dcvlcm9as12-reaper-derivation/reaper5: error while loading shared libraries: libasound.so.2: cannot open shared object file: No such file or directory
<sphalerite_>
Dezgeg: won't load in dillo >.> I'll try out that sdcmd tool though
<erlandsona>
clever: can I add you to that repo so you can push up whatever you're running into there?
<clever>
erlandsona: i can just throw it into a gist when i'm done
<sphalerite_>
samueldr: I don't think it's related, I have those boot parts as well but I think they're separate from the main thing
<Dezgeg>
no, boo* is a separate thing
<erlandsona>
libasound.so.2? Is that from alsa? Or would that need to be added in multiPkgs?
<erlandsona>
clever: sweet.
<samueldr>
good, though I'm curious what they are with the veyron platform
<Dezgeg>
it's a separate hardware partition, effectively
<Dezgeg>
outside the usual MBR / GPT partitioning
<sphalerite_>
OH! Actually…
<sphalerite_>
dd bs=4M if=/dev/mmcblk2 skip=1 count=1 | sha256sum and dd bs=4M if=/dev/mmcblk2boot1 count=1 | sha256sum give me the same thing
<Dezgeg>
sounds like a coincidence
<boomshroom>
clever: Hi. I managed to get binutils to build, but it was extremely hacky and ugly. Now I need to pass it to GCC, but it expects bintools and there doesn't seem to be an obvious way to override bintools with the custom target.
<Dezgeg>
are they all zeroes or something?
<sphalerite_>
oh yeah, all zeroes
<sphalerite_>
:(
* samueldr
needs to plug in another keyboard instead of switching around
<petersjt014[m]>
IS there any way with the repl or something to show the result of evaluating/refactoring a nix file?
<samueldr>
yeah, boot0/1 are zeroed one on mine too
<petersjt014[m]>
like so all the imports are pulled in?
<Dezgeg>
on the tegra u-boot goes to mmcblk0boot0 IIRC
<samueldr>
yeah, that's what I assumed reading the kernel doc
<Dezgeg>
and mmcblk0boot1 holds the u-boot variables, apparently
<petersjt014[m]>
because there is some useful driver or something in the virtualbox-demo file and I want to figure out what it is
<samueldr>
but seems unused on veyron, since they allegedly have SPI flash
<clever>
./result/bin/reaper5: ./result/bin/reaper5: no version information available (required by ./result/bin/reaper5)
Rusty1_ has joined #nixos
<samueldr>
(allegedly since I haven't seen the chip on mine yet)
<clever>
dontStrip fixes the above error
fyuuri has quit [Ping timeout: 268 seconds]
<samueldr>
at first glance, looks like zeroing the first few MB worked, sphalerite_, if you dd to mmcblk0, then read back, you get the original contents, right?
<sphalerite_>
samueldr: yeah
<sphalerite_>
samueldr: badblocks -n is a nice way to test it
<clever>
./result/bin/reaper5: symbol lookup error: /nix/store/0xkvp0ngxxr46psic51fp84aqynk362c-reaper/bin/libSwell.so: undefined symbol: gdk_init_check
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
duncan^ has quit [Quit: WeeChat 1.9.1]
duncan^ has joined #nixos
<boomshroom>
A quick google tells me that "GDK" seems to mean GIMP Drawing Kit.
<clever>
the puzzling part, is that it refers to a gdk symbol, but doesnt link to any gdk libs
<MichaelRaskin>
Might expect they are loaded via GTK?
<samueldr>
sphalerite_: sorry to report that I can't help with read-only blocks :(
<samueldr>
wrote /dev/zero to first 16MB, re-read back 16MB of zeroes
<sphalerite_>
samueldr: huh. Maybe it's all about the kernel, and you'll find the same thing as me with a mainline kernel
<sphalerite_>
samueldr: it shows up as mmcblk2 for me as well, while with the stock kernel it's mmcblk0
<samueldr>
maybe, that was done in chromeos, using nothing else than what's on chromeos 64
<Dezgeg>
I doubt it
<sphalerite_>
best way for me to test this would be checking if badblocks fails on chromeos too
<sphalerite_>
only issue with that is that I don't have chromeos on here anymore x)
<samueldr>
sphalerite_: badblocks isn't on chromeos :(
<samueldr>
though, I could probably load crouton if it works with arm
<sphalerite_>
samueldr: can use it from a chroot or whatever
<samueldr>
yeah, the point being *I need to do something* :)
<erlandsona>
clever: hmmm... I've got the same undefined symbol error you're getting.
nDuff has joined #nixos
<clever>
erlandsona: yeah, thats next on our todo list
<sphalerite_>
samueldr: or I could do something ;)
<samueldr>
:D, you know how to get the recovery img on linux?
<petersjt014[m]>
I'm in the repl and have bound /etc/nixos/configuration.nix to a variable x. calling `:l x` tells me that I '`cannot auto-call a function that has an argument without a default value (‘config’)`'. What kind of value is it expecting?
<samueldr>
(on nixos you need to fix google's script, they did a dumb and reference /bin/echo)
<clever>
petersjt014[m]: that loads all of nixos, which loads the configuration.nix to configure itself
<nDuff>
Howdy -- I'm trying to build some tools for an utterly ancient system (to allow data to be migrated off of it). It's running a kernel that precedes the 2.6.32 release which nix's packages are compiled to support. Is building a stack with an older ABI likely to be a straightforward/feasible approach?
marusich has quit [Quit: Leaving]
aarvar has joined #nixos
<petersjt014[m]>
Alright. How can I see what the five variables are it added to the scope?
<clever>
petersjt014[m]: try to eval config and options, i believe those are 2 of the main ones it adds
mkoenig_ has quit [Ping timeout: 264 seconds]
<petersjt014[m]>
adding `:t` doesnt seem to be
<clever>
petersjt014[m]: you can also just spam tab
<erlandsona>
clever: from the reaper forums: reaper_5352_developer_linux_x86_64.tar.xz libSwell-x86_64-debian-stretch-gtk3.tar.xz
<petersjt014[m]>
k, I'll try that
fendor has quit [Read error: Connection reset by peer]
thc202 has quit [Ping timeout: 260 seconds]
<erlandsona>
I'm wondering if there's some FHS stuff from debian-stretch or ubuntu that libSwell is using.? GTK3... you think we might need to pull in GNOME deps?
<boomshroom>
clever: Based on how hacky my solution was to build a working cross-binutils, is there really a simple way to obtain a build of GCC that can target arbitrary triples, or should I just build it the old fashioned way?
<clever>
boomshroom: dont know
<Profpatsch>
where my bois at
<clever>
erlandsona: its not attempting to link against gtk, thats the weird part
<petersjt014[m]>
See what It looks like when the imports are pulled in, etc--basically right before it's converted into a derivation
<petersjt014[m]>
one of the imports has something I'm looking for
<clever>
petersjt014[m]: then just eval config in the repl and youll see every config option in all of nixos
<buhman>
I'm reading these wonderful pills. Nix's language seems pretty similar with languages that have the same features, like haskell; why was it decided to write a completely new language instead of writing maybe some dsl thing on top of another existing language?
stephenjudkins has quit [Remote host closed the connection]
<licebmi>
I'm trying to set up mopidy, no weird configurations (at least I think that's the case), but I encounter some issues with collisions.
<licebmi>
I'm still wraping my head around the manual, so I'm searching for docs and I only find relevant information about nix-env but I'm just using the declarative configuration, so I don't think it's applicable.
<erlandsona>
licebmi: if you follow the directory path of the collisions it should tell you which libs you have installed that are conflicting. Usually I've fixed those errors by removing one or the other from my config.
<licebmi>
erlandsona: I identified them, but how should I remove them, AFAIK those are dependencies of the package I'm trying to install
<licebmi>
gst-python and pygobject
<licebmi>
I didn't explicitly installed them or listed on my configuration
<erlandsona>
So chances are either of those dependencies are being installed by something else in your config... like gnome or kde or a language pack or something. So once you figure out which 'parent' packages are installing the conflicting deps you can uninstall the parent.
<erlandsona>
That said, clever or MichaelRaskin might have more insight than myself, as I'm also noobish with Nix stuff...
<erlandsona>
learning slowly but surely.
<licebmi>
oh, OK. Still that seems to me like a bug right?
<licebmi>
So I was checking out the meta.priority attribute on the package declaration
<licebmi>
am I doing something absurd by checking this out?
ottidmes has quit [Ping timeout: 264 seconds]
coot has joined #nixos
stephenjudkins has quit [Remote host closed the connection]
stephenjudkins has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 2 new commits to gcc-7: https://git.io/vNDmc
<NixOS_GitHub>
nixpkgs/gcc-7 3d24ae9 Tuomas Tynkkynen: ncurses5: Use v6 sources but compiled with --with-abi-version=5...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/gcc-7 11e98ea Tuomas Tynkkynen: gcc7: More resync from gcc6...
<erlandsona>
licebmi: looks like gst-python relates to GStreamer and pygobject has something to do with gtk3? I know gnome installs gtk3 so maybe you're getting conflicts trying to install GStreamer if it's already installed by the gnome3 stuff?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] adisbladis pushed 1 new commit to staging: https://git.io/vNDmg
<ivanivan>
I'm not that familiar with Makefiles or kernel modules, just looked at lots of examples.
<boomshroom>
Lets see how well this works when completely bypassing nixpkgs and treating the cross compiler like a normal derivation that just happens to share a version and source with another.
jtojnar has quit [Quit: jtojnar]
jtojnar has joined #nixos
<boomshroom>
binutils was over with successfully way faster than the other method.
<boomshroom>
`../.././libcpp/expr.c:770:18: error: format not a string literal and no format arguments [-Werror=format-security]`
sigmundv__ has quit [Ping timeout: 264 seconds]
Supersonic112 has quit [Disconnected by services]
<boomshroom>
Is Werror=format-security set by Nix, and if so, how do I disable it?
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
<erlandsona>
MichaelRaskin: just added gtk3 to the makeLibraryPath list and still same gdk_init_check error?
<MichaelRaskin>
Maybe they actually link against v2 stuff? And you need gtk2?
<erlandsona>
clever: MichaelRaskin, well well well... lol after downloading the repo with wget to look around a bit more, I realized I skipped over the top level readme where I found "Requirements: libc6, libstdc++ for gcc 4.x or later"
<erlandsona>
Also libgdk-3
markus1189 has joined #nixos
<palo>
I have a problem with the /nix/store/ it is growing and growing, I have for example multiple intellij versions in there. but nix-collect-garbage does not pick them up
<palo>
There are also other software I had installed once (I only use nix-shell -p or configuration.nix)
nocoolnametom_ has joined #nixos
<clever>
palo: run nix-store --query --roots on each path to intellij
<clever>
palo: what differs between each of them?
markus1199 has quit [Ping timeout: 256 seconds]
<palo>
nix-store --query --roots /nix/store/*idea-ult* showed some result files and /run/current-system /run/booted-system
<clever>
those cant be GC'd
<clever>
there should also be some numbered entries
<palo>
so I delete /root/result for example ?
<palo>
than it can be GC'd ?
<clever>
yeah, deleting that symlink will let it GC some more things
<erlandsona>
so where do I get libgdk-3 from in Nix?
<boomshroom>
palo: The result links generated by nix-build act as GC roots. Why did you build you run nix-build in /root?
<clever>
nix-locate libgdk-3
<clever>
gnome3.gtk.out 1,166,488 x /nix/store/rqsnb5fv5rwkih74ygb5ar092p82wcl9-gtk+3-3.22.11/lib/libgdk-3.so.0.2200.11
nocoolnametom_ has quit [Ping timeout: 252 seconds]
<palo>
boomshroom: I don't know must be a while ago when I started and fiddled around a bit, and did not know that result files are generated
<palo>
thanks everybody
<erlandsona>
ugg...
<palo>
one thing maybe
silver has quit [Read error: Connection reset by peer]
<palo>
the /nix/store/.links folder is pretty big. should it be there ?
<clever>
palo: that is how nix deduplicates things
<boomshroom>
I better question would probably be why you were in /root at all rather than your unprivileged home directory.
<palo>
clever: thanks
<palo>
so removing all the result links freed 50% of my /nix/store
<palo>
nice
<palo>
thanks
<clever>
palo: you can also: ls -lh /nix/var/nix/gcroots/auto/
<clever>
that will show symlinks pointing to ever result you have forgoten about
ivanivan has quit [Quit: WeeChat 1.9.1]
<erlandsona>
Alright I need to check out for the night, but if anyone get's passed that gdk_init_check issue, please please please hit me up with a github issue or ping me on slack @erlandsona or something.
<boomshroom>
wow. I did an ls -l in there and it gave me 700. Most apear to be dead links.
<clever>
boomshroom: yeah, nix will gc the dead links at the start of any gc
<clever>
boomshroom: nix-collect-garbage --max-freed 1 will tell it to stop after deleting 1 byte
<palo>
clever: nice this freed another 4GB :D
erlandsona has left #nixos ["See y'all later"]
<boomshroom>
aparently a garbage collection of 1 byte frees 4 GB. XD
<hodapp>
wut
<clever>
boomshroom: it probably deleted a single file, that was 4gig
<clever>
and it cant delete half a file
elninja44 has joined #nixos
<boomshroom>
it seems deleting /nix/store/trash either take really long, or it actually hanged.
ertes has joined #nixos
<boomshroom>
I suddenly remember why I haven't done a GC in months on this machine.
<hodapp>
that super-important environment that takes 72 hours to build?
d4g_ has joined #nixos
iyzsong has joined #nixos
boomshroom_ has joined #nixos
<boomshroom_>
Right. nix-collect-garbage freezes my computer.
boomshroom has quit [Ping timeout: 260 seconds]
coot_ has quit [Quit: coot_]
<boomshroom_>
It seems as though IRC waited until I rebooted and logged in again before disconnecting my old session.
d4g has quit [Ping timeout: 256 seconds]
boomshroom_ is now known as boomshroom
<adisbladis[m]>
boomshroom_: Deleting /nix/store/trash does indeed take very long..
s33se_ has joined #nixos
<boomshroom>
A quick du reveals it's only 316M. That doesn't seem big enough for its deletion to completely freeze my computer.
<boomshroom>
Should it be safe to rm -rf /nix/store/trash?
s33se has quit [Ping timeout: 256 seconds]
<MichaelRaskin>
Should fail, actually
<boomshroom>
MichaelRaskin: Why, other than the lack of sudo?
<MichaelRaskin>
read-only bind mount of /nix/store
<boomshroom>
how do I view the permisions of a directory without seeing everything inside the directory?
alexteves has quit [Read error: Connection reset by peer]
<adisbladis[m]>
fearlessKim: I'll take a look :)
babyflakes has joined #nixos
nocoolnametom_ has joined #nixos
nocoolnametom_ has quit [Remote host closed the connection]
nocoolnametom_ has joined #nixos
Rusty1_ has quit [Quit: Konversation terminated!]
jb55 has quit [Ping timeout: 268 seconds]
johnsonav has quit [Remote host closed the connection]
rogue_koder has quit [Quit: Konversation terminated!]
nocoolnametom_ has quit [Remote host closed the connection]
nocoolnametom_ has joined #nixos
aneeshusa has joined #nixos
aarvar has quit [Ping timeout: 256 seconds]
<fearlessKim[m]>
adisbladis: ty. I usually switch between nixpkgs worktrees but I have many heavy PRs in the works and it has become impractical. Also the PR is IMO ready so I chose to nag for once ^^
<infinisil>
fearlessKim[m]: you mean you're using separate git worktrees for different prs?
<infinisil>
(As in adding worktrees with `git worktree`)
swayf has joined #nixos
<boomshroom>
Since GCC was giving me so much trouble, now I'm trying Rust. When using a rust platform using the nightly channel from mozilla-nixpkgs, I get "can't find crate for `std`"
<boomshroom>
When I use the default rust platform, the project builds, but then the test complains about it not being nightly.
freeman42x[NixOS has joined #nixos
<fearlessKim[m]>
infinisil: yep, because switching branches is not practical when working on several PRs. I typically have 3 worktrees.
swayf has quit [Ping timeout: 265 seconds]
<infinisil>
Nice
<boomshroom>
I remember there being a problem with rust's stdlib a while back.
<boomshroom>
The alternate solution is forcing nix to not run the tests.
<infinisil>
fearlessKim[m]: Oh, i just thought of a cool idea: Have a fuse filesystem that has a dynamic directory with a git worktree for every nixpkgs pr :O a la nixpkgs-prs/<pr number>
<boomshroom>
doCheck = false; # It builds now.
johnsonav has joined #nixos
<fearlessKim[m]>
infinisil: I am sorry but I won't let you kill my productivity with this kind of idea, I 've grown fond of my worktree workflow :p
<boomshroom>
And then it panics the moment I try running it.
<infinisil>
fearlessKim[m]: :P
<adisbladis[m]>
fearlessKim: Getting some build errors :/ Some of them could be resolved by bumping libpinyin and ibus-libpinyin but fcitx-engines.cloudpinyin is still not building after that
<boomshroom>
And then it panics the moment I try running it. `failed to execute `RUSTC -vV`: No such file or directory (os error 2)`
johnsonav has quit [Remote host closed the connection]
<adisbladis[m]>
fearlessKim: No config. I just did a nox-review
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] aneeshusa opened pull request #34348: Enable Kerberos by default for OpenSSH (master...enable-kerberos-by-default-for-openssh) https://git.io/vNDnx
NixOS_GitHub has left #nixos [#nixos]
<boomshroom>
Just checking, wrapProgram --prefix is what I should use if I want to add additional values to an environment variable of a binary?
<adisbladis[m]>
boomshroom: Yes, like adding to $PATH
<fearlessKim[m]>
adisbladis: so I am in my PR working tree, I run 'nox-review wip' and it tells me no uncommit changes. It suggests an 'against' flag that doesn't appear in help
<boomshroom>
adisbladis[m]: Thank you... except "wrapProgram: command not found"
<adisbladis[m]>
boomshroom: You need to add makeWrapper to nativeBuildInputs
<adisbladis[m]>
fearlessKim: Where rev is a commit-ish like origin/master or a commit-sha
<boomshroom>
Thanks. One more thing, what phase is it recomended to put wrapProgram?
<adisbladis[m]>
boomshroom: Typically postInstall
<boomshroom>
That's where I had it. Thank you
mrkgnao has quit [Ping timeout: 256 seconds]
juhe has joined #nixos
<boomshroom>
And it works! 😄
<boomshroom>
(Yes, I felt the need to use an emoji.)
<adisbladis[m]>
\😄/
johnsonav has joined #nixos
<Ralith>
🎊
<boomshroom>
Fun fact: "Emoji" is Japanese for "picture character" and it's resemblance to "emotion" is purely coincidental.
Ivanych has quit [Ping timeout: 260 seconds]
duckwho has joined #nixos
mrkgnao has joined #nixos
duckwho has quit [Client Quit]
nocoolnametom_ has quit [Remote host closed the connection]
<boomshroom>
Has anyone tried building Nix on RedoxOS?
Itkovian has joined #nixos
<fearlessKim[m]>
adisbladis: when I tried nox-review previsouly it would crash, I guess now I will systematically use it thanks (I am testing your changes). What does nox-update ?
<fearlessKim[m]>
nox-update --help is not too helpful
nocoolnametom_ has joined #nixos
<fearlessKim[m]>
adisbladis: did you manage to have nox-review succeed ? I have with your patch (and certainly without) an include error
<adisbladis[m]>
fearlessKim: No that include error is what I ended up with. Dont know how to fix that.
<fearlessKim[m]>
y right. Will work on it thanks
<fearlessKim[m]>
I think it might be related to the fact I disable the pinyin module in fcitx with -DENABLE_PINYIN=OFF
nDuff has quit [Quit: Page closed]
fragamus has joined #nixos
lonokhov has joined #nixos
verite has joined #nixos
siel has quit [Ping timeout: 255 seconds]
siel has joined #nixos
nocoolnametom_ has quit [Remote host closed the connection]
<fearlessKim[m]>
nox-review is freaking slow
nocoolnametom_ has joined #nixos
nocoolnametom_ has quit [Ping timeout: 268 seconds]
<hyper_ch>
Mic92: thx, it's building now... so maybe I managed to do it correctly
verite has quit [Ping timeout: 240 seconds]
leat has quit [Ping timeout: 248 seconds]
lonokhov has quit [Ping timeout: 264 seconds]
<typetetris>
https://gist.github.com/typetetris/abd0f6abd0dcf9f799de6c618e4aeed6 This now worked, but I arrived at it through trial and error. Maybe one could explain to me, why I need both overrides, why can't I replace the override at `ghc822` with a set union `//` and remove the `overrides = ` line ?
justanotheruser has quit [Ping timeout: 252 seconds]
<typetetris>
doCheck isn't an attribute of the function used with callPackage to create the derivation, so it isn't directly overridable, but mkDerivations usually has it, but uses a default argument for it.
justanotheruser has joined #nixos
iyzsong-x has joined #nixos
boomshroom has quit [Quit: Page closed]
iyzsong has quit [Ping timeout: 246 seconds]
lonokhov has joined #nixos
leat has joined #nixos
<typetetris>
In `let a = "uae"; b = { "uae" = 4545 }; in b.a` does that work?
<sphalerite_>
Using master or nixpkgs-unstable for your nixos system is a Bad Idea™ as it hasn't passed the basic tests of staying bootable. nixos-unstable is fine. But you can also just run 17.09 and install individual packages from nixos-unstable, nixpkgs-unstable or master
<sphalerite_>
yes (but https, not http), or keep a checkout, or just use -I nixpkgs=https://nixos.org/channels/nixos-unstable/nixexprs.tar.xz on a one-off basis
binh has quit [Quit: Leaving]
<sphalerite_>
you'll probably want to do that as root as well, because user channels are confusing and weir
<typetetris>
But how do I get access to nixos-unstable as my normal desktop user, after doing the nix-channel command as root?
<typetetris>
I need to manipulate my `NIX_PATH`?
<sphalerite_>
nix-env -iA nixos-unstable.foo
<sphalerite_>
(also make sure to nix-channel --update)
<typetetris>
and of course it downloads a second nixos now :) (Of course, that was the point of it. It is just, that my internet is painfully slow and I need to whine a little bit about it for no avail.)
<sphalerite_>
you can also keep a checkout of 17.09 and cherry-pick individual commits onto it
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ma27 opened pull request #34350: nixos/powerManagement: set `cpuFreqGovernor` with `mkOptionDefault` (master...fix-powermanagement-default) https://git.io/vND8Y
NixOS_GitHub has left #nixos [#nixos]
<sphalerite_>
that results in having to build the changed stuff locally, but means you can use existing dependencies, reducing the closure size a great deal
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
fyuuri has joined #nixos
<typetetris>
Can I somehow get that hackage2nix that is used to update the hackage packages in nixpkgs?
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<dhess>
typetetris: what are you trying to do with hackage2nix?
<typetetris>
I would try to get up to date hackage packages with nixos-1709, so only replace `pkgs/development/haskell-modules/hackage-packages.nix` I presume?
<dhess>
typetetris: hmm that's gonna be tricky. For one thing, *tons* of Haskell stuff has changed since 17.09 because of all the cross-compiler support added to master recently.
ssmike has quit [Remote host closed the connection]
ma27 has quit [Ping timeout: 256 seconds]
<dhess>
typetetris: I would recommend that you run something stable as your system-level channel, something like nixos-17.09 or nixos-17.09-small, and then use the nixpkgs-unstable (or nixos-unstable) channel for your Haskell development
<typetetris>
I will do that.
<typetetris>
using nixos-unstable for haskell development.
<dhess>
typetetris: so the root channel will be something like nixos-17.09 and when you run `nixos-rebuild switch` it will use that channel; and then in your user account you can use the nixos-unstable channel , so that when you use nix-build or nix-env from that account, you'll get the Haskell packages etc. from that channel
<typetetris>
If I do `nix-env -q` what other option would I need to see, which channels the packages come from?
<dhess>
that would almost certainly be more stable and less work than trying to mix haskell packages from an unstable channel with a stable channel
<typetetris>
ok
<typetetris>
sounds reasonable
<typetetris>
The stuff in my `.config/nixpkgs/config.nix` will apply only to the `nixos` channel or to the `nixos-unstable` channel also?
<dhess>
run that from your user account, not the root account!
ciil has joined #nixos
<typetetris>
of course :)
<dhess>
you can always switch back if something breaks. I find that nixos-unstable is pretty good.
<fearlessKim[m]>
$ nix-channel --list not showing the root channels is really treacherous
<dhess>
vaibhavsagar: around?
<hyper_ch>
hmmm, samba on master seems to be broken
ma27 has joined #nixos
fuyuuri has joined #nixos
ambro718 has joined #nixos
<ambro718>
I just upgraded a machine via nixops and I get ssh "WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED!" (once at the end of deploy, and further in nixops ssh). There is no error, deploy and ssh connection are still successful. What is going on?
glenn_ has quit [Remote host closed the connection]
glenn_ has joined #nixos
glenn_ has quit [Ping timeout: 246 seconds]
ericsagnes has quit [Ping timeout: 268 seconds]
glenn_ has joined #nixos
knupfer has joined #nixos
MP2E has quit [Remote host closed the connection]
ericsagnes has joined #nixos
Itkovian has joined #nixos
ottidmes has joined #nixos
glenn_ has quit [Ping timeout: 252 seconds]
<fearlessKim[m]>
ambro718: it means the ssh key of the VM has been changed I guess
<ambro718>
fearlessKim[m]: I didn't change it
<fearlessKim[m]>
nixops must have changed it I suppose
Drakeson has quit [Ping timeout: 240 seconds]
asuryawanshi has joined #nixos
glenn_ has joined #nixos
pie_ has joined #nixos
asuryawanshi has quit [Ping timeout: 248 seconds]
pie__ has joined #nixos
Turion has joined #nixos
pie_ has quit [Ping timeout: 260 seconds]
dbe_ has joined #nixos
FareTower has quit [Ping timeout: 256 seconds]
swayf has joined #nixos
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos
hotfuzz_ has joined #nixos
hotfuzz has quit [Ping timeout: 240 seconds]
primeos has quit [Quit: WeeChat 1.4]
i-am-the-slime has joined #nixos
KingOfThePirates has joined #nixos
<KingOfThePirates>
anybody awake at this hour?
<hyper_ch>
well, it's almost noon
<hyper_ch>
so I guess there's billions of people awake
<KingOfThePirates>
oh snap I forgot there's other lands
CapsAdmin has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] teto opened pull request #34351: [WIP] buildLinux: Add more overrides (master...kernel_wip) https://git.io/vND0c
NixOS_GitHub has left #nixos [#nixos]
Swant is now known as IKEA
<MichaelRaskin>
I would say you forgot that the default land for #nixos is Europe.
<hyper_ch>
Europe is a land? oO
<KingOfThePirates>
I would agree
<MichaelRaskin>
Well, a landmass
<MichaelRaskin>
And in landmass sense it even includes Switzerland!
<sphalerite_>
hyper_ch: this is the 3rd time it's happened for me
<KingOfThePirates>
btw I'm looking in my search history to find that error. that's why I haven't responded to the help responses yet
<hyper_ch>
sphalerite_: why do you create a dataset only to destroy it?
<KingOfThePirates>
practice
<sphalerite_>
hyper_ch: for the example's sake :)
<hyper_ch>
and wouldn't you need to set mountpoint=legacy?
<sphalerite_>
hyper_ch: in practice, I'll have used the dataset and no longer need it
<sphalerite_>
hyper_ch: I set mountpoint=legacy on tank so all the datasets inherit it
<hyper_ch>
smart :)
rogue_koder has joined #nixos
<hyper_ch>
but why will that kill x server
<hyper_ch>
and why even first do it on the seperate dataset
<sphalerite_>
systemd notices there's a mount missing, and goes into emergency mode, killing everything that's not part of emergency mode
<sphalerite_>
it's really stupid
<hyper_ch>
:)
<sphalerite_>
I just need to stop using nixos-generate-config
<KingOfThePirates>
do I need systemd to boot from nixos?
<srhb>
KingOfThePirates: Yes, NixOS is based on systemd.
glenn_ has quit [Ping timeout: 246 seconds]
<KingOfThePirates>
Ok. I didn't see that mentioned in the install part of the manual
<hyper_ch>
or just alias nixos-rebuild to "zfs snapshot tank/nixos@now; nixos-rebuild"
<srhb>
(Needing it sounds a bit weird, you can't really avoid it. I wonder what prompted the question)
randomstrangerb has quit [Ping timeout: 240 seconds]
<srhb>
KingOfThePirates: You don't *need* anything but the iso to boot that. What's the problem at hand?
elninja44 has quit [Remote host closed the connection]
quam6246 has joined #nixos
<KingOfThePirates>
It says it can't find /dev/something
<srhb>
KingOfThePirates: When you do what?
<KingOfThePirates>
after the lvm part
<srhb>
KingOfThePirates: Exact commands and error messages really help. :)
randomstrangerb has joined #nixos
<sphalerite_>
KingOfThePirates: yeah you're probably creating the usb wrong
<KingOfThePirates>
I put a FAT32 partition. And I put a nixos iso on it
<KingOfThePirates>
Then boot
<sphalerite_>
yeah no you want to write the iso directly to the usb
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sphalerite_>
cp nixos-17.09*.iso /dev/sdc
<srhb>
(that is, /dev/sdx. not /dev/sdxN)
<sphalerite_>
(substitute /dev/sdc as appopriate)
<srhb>
Note that it will nuke everything on the drive.
<KingOfThePirates>
ok so I've been trying to do it on a hard drive. over the past couple days I did it successfully with ubuntu multiple times. working out the quirks and learning some basic kernel knowledge. i'm still new
<srhb>
That's fine. :)
glenn_ has joined #nixos
<KingOfThePirates>
like make a 5 GB hard drive and put an iso on it
<srhb>
It's a bit more difficult to not use the entire device.
<srhb>
I'm honestly not sure how to do it.
<KingOfThePirates>
you mean 1 partition for 1 physical hdd/ssd?
<srhb>
I mean write the iso directly to the usb device, taking up all partitions.
<srhb>
I don't believe the manual covers this case.
<srhb>
Unless unetbootin knows how to do it.
<sphalerite_>
no, unetbootin is garbage
<KingOfThePirates>
I think the dd part should do it for me. But I'm lacking knowledge to even use it
<srhb>
sphalerite: Alas, it is the recommended way to do it according to the manual
<KingOfThePirates>
I'm using ubuntu right now so let me pull it up
<srhb>
KingOfThePirates: Use sudo dd if=/path/to/iso/file of=/dev/usbdevice
<sphalerite_>
KingOfThePirates: on your existing ubuntu or whatever system, run lsblk, plug in the USB stick, run lsblk again. The entry that wasn't there before is your usb stick, and that's the thing you want to put in the of= for dd
<srhb>
KingOfThePirates: Where usbdevice is something like sdc. You can get the actual name by plugging in the stick and taking a look in the bottom of dmesg or something :)
<sphalerite_>
srhb: ugh
<srhb>
Or lsblk
<sphalerite_>
I don't think unetbootin is even still maintained
<srhb>
sphalerite: Well, I agree... :P
<KingOfThePirates>
Ok. and when I make my pseudo usb. Is FAT32 good? I can make it anyting
<KingOfThePirates>
thing
<srhb>
KingOfThePirates: You should not format it at all
<srhb>
KingOfThePirates: The iso contains the partition table and everything
<srhb>
KingOfThePirates: This is why you should write it to the raw device.
<srhb>
KingOfThePirates: It will simply take over the entire disk.
<srhb>
Wait, "pseudo usb" ?
<KingOfThePirates>
Oh lol maybe I understand now
<srhb>
What do you mean?
<KingOfThePirates>
I was tripping over myself
stranger___ has joined #nixos
<KingOfThePirates>
I make a partition and I pretend it's a usb
<KingOfThePirates>
Then I boot from it
<srhb>
I've... No idea how to do that.
<KingOfThePirates>
works with ubuntu
oida has quit [Quit: WeeChat 1.6]
<KingOfThePirates>
did it like 5 times
<srhb>
That doesn't mean it works with our iso :)
<KingOfThePirates>
over the past week
<sphalerite_>
KingOfThePirates: if you're going to fiddle around with stuff like that it's probably easier to just install nix on ubuntu, then use that to install nixos without booting an installer image at all
<KingOfThePirates>
do i need a vm? I'm trying to avoid that
<KingOfThePirates>
but if it can live in harmony or on top
<KingOfThePirates>
i'll do it
<sphalerite_>
KingOfThePirates: in any case, that solution won't work because of how our installer finds its data
ssmike has quit [Remote host closed the connection]
<sphalerite_>
there's no VM necessary, nix should be fairly unintrusive and you can uninstall it again afterwards using rm -r /nix ~/.nix-*
<KingOfThePirates>
i'm not stuck on any method. i'd just like to avoid vm. and i don't have a usb
<KingOfThePirates>
but if there's any way to make it work. i'm happy to experiment
<srhb>
I'm inclined to agree with the method proposed by sphalerite_ if and only if you can't scrounge up a raw device to boot from, which would be much simpler
<hyper_ch>
sphalerite_: can nix be installed and then kexec be built?
<srhb>
But it *is* more involved.
<sphalerite_>
hyper_ch: yes. but you don't need to boot into the installer at all if you have nix
<hyper_ch>
never used just nix
dan_b has quit [Ping timeout: 252 seconds]
<srhb>
hyper_ch: Nonsense, you use just nix all the time ;-)
<sphalerite_>
hyper_ch: the kexec method is more for building a tarball you can just stick on a linux system without nix and boot into that, there's not usually any point in using it on a system that already has nix
simpson has joined #nixos
asuryawanshi has joined #nixos
<KingOfThePirates>
hold on let me read all the response
<KingOfThePirates>
responses
<hyper_ch>
using nix != just nix
<sphalerite_>
hyper_ch: or rather, clever's version of it. kexec does still make sense if you need to mess about with the partition table and are currently on a system living in those same partitions
<sphalerite_>
but then you don't need the tarball
<srhb>
sphalerite_: What does that version do? ramfs all the things?
<sphalerite_>
srhb: yes, the kexec installer does that generally
<srhb>
I haven't tried it.
<srhb>
My first NixOS install was the only funky one (direct from Arch)
<sphalerite_>
but you don't need to use the tarball if you have nix, since you can just build the script and use that directly
<sphalerite_>
haha, pretty much all my nixos installs since my third or so have been funky
<sphalerite_>
and I've done… many
<srhb>
At work we do some weirdness with self installing isos, but that's all.
<srhb>
(And that really does launch the regular installer with a systemd job that does everything)
<sphalerite_>
pff, that's what netboot is for :)
<srhb>
netboot requires dhcp!
<srhb>
And I can't convince our networking infrastructure that that would be A Good Idea
<sphalerite_>
really? wow
<MichaelRaskin>
Well, if you have a laptop and physical access…
asuryawanshi has quit [Ping timeout: 256 seconds]
<srhb>
There's an abundance of things we could do. I'm trying not to cause too many waves since they're already accepting the NixOS thing ;)
<srhb>
"Don't worry about it!"
* sphalerite_
starts wondering what the smallest viable netboot server might look like
Turion has quit [Remote host closed the connection]
<sphalerite_>
probably a pi zero?
<MichaelRaskin>
Well, I guess booting from flash also needs physical access to the machine, so there is little externally observable difference
<sphalerite_>
exactly
<MichaelRaskin>
sphalerite_: hm, the smallest smartphone with USB OTG, I would guess?
<srhb>
We load the isos via the RMC (which is a bit messy, admittedly...)
<sphalerite_>
MichaelRaskin: why USB OTG?
<MichaelRaskin>
That still means you get to sit in a relatively quiet room, so better than manual netboot.
<KingOfThePirates>
when you guys are saying "just nix" and "a system that already has nix", do you guys mean nix-env or nixos?
<KingOfThePirates>
just trying to stay on the same page
<sphalerite_>
KingOfThePirates: nix. nix is the package manager, nix-env is one tool that's part of the package manage
<MichaelRaskin>
sphalerite: because then you plug a small converter, plug USB-Ethernet, put a large microSD, and here you have you server batteries literally included
<KingOfThePirates>
yeah that's what i meant to refer to
<KingOfThePirates>
i have nix
<KingOfThePirates>
but not nixos
<srhb>
KingOfThePirates: Then you can use that to install NixOS without leaving your current system.
<srhb>
If you have spare partitions at least
<KingOfThePirates>
can i use it to install to it's own partition?
<KingOfThePirates>
I have lots of space
<KingOfThePirates>
unallocated
<srhb>
Yes, that's what sphalerite_ was suggesting.
<srhb>
(I think)
glenn_ has quit [Read error: Connection timed out]
<KingOfThePirates>
still reading the rest
<sphalerite_>
KingOfThePirates: then you can get the tools necessary to install nixos using `echo {} > /tmp/configuration.nix ; nix-env -f '<nixpkgs/nixos>' -I nixos-config=/tmp/configuration.nix -iA config.system.build.nixos-{install,generate-config}`
zzamboni has joined #nixos
<sphalerite_>
iirc
<sphalerite_>
then you can follow the instructions in the manual on ubuntu rather than in a booted nixos installer
<KingOfThePirates>
as long as i don't wipe all partitions i'll be fine. i'll keep trying with fresh starts until things are stable
<sphalerite_>
bar anything that requires fiddling with currently mounted partitions. If you already have the partition where you want nixos installed set up you're good to go
<srhb>
Be mindful of keeping your current system bootable though
<srhb>
You'll need to take extra steps during the installer to achieve that.
<KingOfThePirates>
define "current system bootable"
<KingOfThePirates>
I have many bootables on this machine
<srhb>
You're booted into some system. Once you run the installer, I don't think the default options will keep the bootloader intact. You'll need to make it bootable in the new (NixOS-managed) bootloader
<srhb>
s/it/them then
<KingOfThePirates>
windows, refind, ubuntu, 3 backups, another ubunutu
<KingOfThePirates>
no usbs though
<srhb>
If you have NixOS take over the bootloader (which is the usual case) you'll need to at least enable autodiscovery of other bootable systems to have them included in the boot menu.
<srhb>
But that's a later step.
<srhb>
(just be sure you give it some thought or ask here before running nixos-install)
<KingOfThePirates>
"the kexec method is more for building a tarball you can just stick on a linux system without nix and boot into that" why would a linux system not be able to get nixpkgs?
<sphalerite_>
srhb: well it should be easily reversible by just doing update-grub+grub-install on the ubuntu system afterwards :)
<sphalerite_>
KingOfThePirates: it's more about the effort of getting nixpkgs on it vs just copying a tarball over
<sphalerite_>
KingOfThePirates: especially if you'll wipe it right afterwards
<KingOfThePirates>
also i only recently heard of the word "systemd". do i need this before i do anything?
<KingOfThePirates>
it seems really useful when reading their website
<srhb>
sphalerite_: Right, but since they have no USB stick, it might suck to end up with no bootable system :)
<srhb>
KingOfThePirates: Nope, you do not.
<srhb>
KingOfThePirates: It is the "service manager" that NixOS is based on
<KingOfThePirates>
mmk
i-am-the-slime has quit [Ping timeout: 256 seconds]
<sphalerite_>
KingOfThePirates: most linux distros, including ubuntu, use systemd :)
<KingOfThePirates>
oh cool
<hyper_ch>
I liked upstart
glenn_ has joined #nixos
<srhb>
That's a first. :)
<sphalerite_>
hyper_ch: switch to chrome os, that uses upstart ;)
<sphalerite_>
what did you like about it? That whole business was before I started caring about init systems
<hyper_ch>
sphalerite_: I just liked it
justanotheruser has quit [Quit: WeeChat 1.9.1]
<MichaelRaskin>
sphalerite_: the same, and this is exactly why I liked upstart.
<MichaelRaskin>
I mean, systemd made me care about using _anything_ but systemd.
justanotheruser has joined #nixos
rauno has quit [Remote host closed the connection]
<MichaelRaskin>
NixOS is better than many systems at abstracting systemd away
<MichaelRaskin>
Still not perfect.
<MichaelRaskin>
(which is why I no longer use NixOS mainline)
<sphalerite_>
what do you use for init again?
<MichaelRaskin>
For PID1? sinit, because… because I don't want it to actually do anything.
<hyper_ch>
upstart was just sysv init with improvements and backward capabilites IMHO
<sphalerite_>
MichaelRaskin: neat. and service management?
<KingOfThePirates>
when you say "You're booted into some system. Once you run the installer..." what exactly do you mean by "system"? the machine? the os desktop? the bootloader? or something else?
<sphalerite_>
unrelated, I'm building webkitgtk on my chromebook... It's taking a while…
<sphalerite_>
KingOfThePirates: sorry, what are you citing?
<MichaelRaskin>
I don't actually need any service management. I want, like, 10 services, I export independent runner-scripts from NixOS (NixOS supports it, which is nice), I have my bootscripts spawn these service scripts as needed. They are all safe to stop by sigterm, and they don't fail in a sense that service manager would catch, like ever.
<sphalerite_>
my guess would be the OS from context, but I'm not sure
<KingOfThePirates>
[05:11] <srhb> You're booted into some system. Once you run the installer, I don't think the default options will keep the bootloader intact. You'll need to make it bootable in the new (NixOS-managed) bootloader [05:11] <srhb> s/it/them then
seanparsons has quit [Ping timeout: 240 seconds]
<srhb>
KingOfThePirates: The system you're (probably) writing this message from now.
ma27 has quit [Ping timeout: 256 seconds]
<srhb>
KingOfThePirates: Once you reach the step where you're running nixos-install, the default behaviour is to install a bootloader to where you'll be specifying it in your configuration.nix. If you're not cautious here, you might end up with only NixOS being bootable on that system, not your Ubuntu, not your Windows, not anything else. And if things go wrong, perhaps even NixOS is not bootable.
glenn_ has quit [Ping timeout: 246 seconds]
<srhb>
KingOfThePirates: So I'm just urging caution. :)
<KingOfThePirates>
ah i see i see
<srhb>
I'm trying to be extra explicit about this since you seem to not have any emergency bootable device.
<MichaelRaskin>
sphalerite: For the record: I don't even have a global system dbus instance, although I do spawn dbus instances around LibreOffice that somehow requires this; but they have the same life extent as LO instance itself.
<sphalerite_>
MichaelRaskin: neat
<srhb>
gchristensen: Does ofborg respect meta.platforms when issued build commands? :)
seanparsons has joined #nixos
<MichaelRaskin>
And I have moved a lot of the system-management logic into a Common Lisp system management daemon. I think this captures more of Lisp OS dreams than spending a heroic effort into reimplementing IDE driver, again…
glenn_ has joined #nixos
<srhb>
MichaelRaskin: You should write a blog post about your system setup. It seems to come up quite often :)
<MichaelRaskin>
I am currently refactoring it into something that can be submitted as a ELS talk.
<MichaelRaskin>
The system per se, not just the writeup.
<sphalerite_>
so glad I have a graphical browser on my chromebook now…
<sphalerite_>
no javascript or correct rendering yet but it's already so much better than w3m
zzamboni has quit [Quit: Leaving.]
<MichaelRaskin>
(For a million of reasons, I want to let the Nix part be used with system management daemons in whatever language, but I personally will develop the Common Lisp toolset for it)
zzamboni has joined #nixos
zarel has joined #nixos
glenn_ has quit [Ping timeout: 252 seconds]
zzamboni has quit [Client Quit]
<sphalerite_>
sounds good. also why did you pick a github username that's so hard to remember :p
IKEA is now known as Swant
<MichaelRaskin>
Well, why would you need to remember it? I mean '7c6' is easy to remember/recognise, the rest can be grepped…
fuyuuri has quit [Ping timeout: 260 seconds]
<MichaelRaskin>
I have multiple logins at many services, I separate them by topic (well, more or less) and I don't want to take anything someone else would meaningfully want
zzamboni has joined #nixos
<sphalerite_>
so what about your IRC name? :p
thc202 has joined #nixos
<MichaelRaskin>
Around getting the commit access to Nixpkgs there was (back then) the idea of using real names for contributors to be clear about partial copyright etc. Then everyone gave up
glenn_ has joined #nixos
<sphalerite_>
why I would want to remember it: one of these days I'm going to think "didn't MichaelRaskin have a cool boot setup?" and won't just be able to go to github.com/michaelraskin to find it. I'll have to grep the IRC logs or something instead, which is obviously an earth-shattering inconvenience ;)
licebmi has joined #nixos
<MichaelRaskin>
Actually, my nick at IRC is not registered (at some point nickserv decided to kill the registration, I didn't care enough to re-register), so maybe someone will actually snatch it at some point…
<gchristensen>
srhb: no, it is currently too dumb to respect platforms
<srhb>
gchristensen: OK, just curious :)
glenn_ has quit [Ping timeout: 252 seconds]
fendor has joined #nixos
<sphalerite_>
I've installed nix on a foreign system, but can't nix-copy-closure to it because nix-store is only in the PATH once .profile has been sourced. Any idea how I could make sure it's available to non-login SSH sessions as well?
<qwebirc442554>
so it seems i need to: (1) make a partition. don't format. (2) nix-env -f '<nixpkgs/nixos>' --arg configuration {} -iA config.system.build.nixos-install
glenn_ has joined #nixos
<srhb>
qwebirc442554: No no, you're mixing up instructions now
<srhb>
qwebirc442554: The "don't format" was in response to "how do I make the iso bootable"
<srhb>
qwebirc442554: If you're installing from your current system, you don't need to do any of that.
<qwebirc442554>
i compartmentalized the instructions in my head so it won't be hard to rearrange
<srhb>
qwebirc442554: Just follow the regular instructions after installing nixos-install and nixos-generate-config
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
justan0theruser has quit [Client Quit]
peacememories has joined #nixos
justanotheruser has joined #nixos
<tilpner>
gchristensen - Hey, if your #hetzner question was "Can I speak English in here?": There isn't much talk in there, but most of it was English in the last few days
<gchristensen>
:) no worries, I was wanting to ask about PXE support for server provisioning. I went the kexec route and got the job done :)
<tilpner>
gchristensen - I messaged support a few days ago. In the future, support will upload project-local ISOs on request, which I assume is what you mean PXE support
<gchristensen>
interesting
<tilpner>
*+by
<tilpner>
gchristensen - Do you just want one server, or are you looking into nixops support?
<gchristensen>
ideally I'd be able to do what I can do with Packet.net (I'mnotafanboyIpromise) where I just say "boot from https://mydomain/nixos.ipxe" and they do it, no upload required, but uploading an ISO isn't so bad :)
spear2 has quit [Remote host closed the connection]
<gchristensen>
tilpner: I'm not sure yet :) depends what is possible
glenn_ has quit [Ping timeout: 252 seconds]
<srhb>
Does it ever make sense to have a broadcast address that doesn't correspond to what can be computed from the address and prefix? Or that is unset? Or is 0.0.0.0?
<tilpner>
gchristensen - I modified clevers kexec_tarball into a self-extracting executable that requires no interaction. After scp-ing and executing it, it will repartition (though not zfs anymore), install NixOS, and then reboot. It's not quite "boot from here", but it's easier to automate now
<gchristensen>
nice
tertleeltret has quit [Quit: Connection closed for inactivity]
i-am-the-slime has quit [Ping timeout: 256 seconds]
rauno has joined #nixos
oida has joined #nixos
Jackneill has joined #nixos
glenn_ has quit [Ping timeout: 252 seconds]
yenzenz has quit [Ping timeout: 240 seconds]
jensens has joined #nixos
quam6246 has quit [Quit: My computer has gone to sleep. ZZZzzz…]
glenn_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dotlambda opened pull request #34353: pythonPackages.discordpy: Use websockets==3.4 (master...discordpy) https://git.io/vNDw0
NixOS_GitHub has left #nixos [#nixos]
i-am-the-slime has joined #nixos
glenn_ has quit [Ping timeout: 246 seconds]
<qwebirc442554>
i'm embarrassed to say it took me awhile to realize "nix-env -f '<nixpkgs/nixos>' --arg configuration {} -iA config.system.build.nixos-install" is what installs nixos-install
glenn_ has joined #nixos
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
i-am-the-slime has quit [Ping timeout: 256 seconds]
<srhb>
qwebirc442554: Don't be. :-)
<srhb>
qwebirc442554: But good that you realized!
<pmeunier>
Hi! I'm getting a stack overflow (possibly infinite recursion), but I'm pretty sure it's not an infinite recursion. How can I increase stack size?
swayf has quit []
swayf has joined #nixos
zzamboni has joined #nixos
<sphalerite_>
qwebirc442554: oh yes isn't it blindingly obvious? ;)
asuryawanshi has joined #nixos
vidbina has joined #nixos
Isorkin has joined #nixos
asuryawanshi has quit [Ping timeout: 256 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] yesbox opened pull request #34354: Enable use of the network.useHostResolvConf option (master...optional_host_resolvconf) https://git.io/vNDrZ
NixOS_GitHub has left #nixos [#nixos]
ma27 has quit [Ping timeout: 240 seconds]
ma27 has joined #nixos
va1entin has joined #nixos
<va1entin>
When I do a nixos-rebuild I get this message a lot of times: perl: warning: Setting locale failed.
<va1entin>
according to the output of "locale" my locale is en_DE.UTF-8, even though I have defaultLocale = "en_US.UTF-8" in my config
<va1entin>
I have consoleKeyMap = "de" in my config aswell
jekor has quit [Read error: Connection reset by peer]
<va1entin>
Does anynone have a hint what might be causing this error? I also can't start "Zim" because it says "unsupported locale setting"
<NixOS_GitHub>
[nixpkgs] fpletz closed pull request #34354: nixos/containers: Enable use of the network.useHostResolvConf option (master...optional_host_resolvconf) https://git.io/vNDrZ
NixOS_GitHub has left #nixos [#nixos]
<fearlessKim[m]>
full logs oh yeah \o/
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 2f39b19 Tuomas Tynkkynen: mdadm: Make 4.0 the default...
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vNDKs
<NixOS_GitHub>
nixpkgs/master e305380 Michael Raskin: Merge pull request #34356 from dezgeg/mdadm4...
NixOS_GitHub has left #nixos [#nixos]
<gchristensen>
tilpner: I think these are good ideas, would you mind sending a PR?
<tilpner>
Yes, I'm on it
<tilpner>
Just testing in GH preview
<gchristensen>
cool :D
i-am-the-slime has joined #nixos
<gchristensen>
one day I'll replace these last two pieces with something other than PHP...
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 8ed3a90 Maximilian Bosch: nixos/powerManagement: set `cpuFreqGovernor` with `mkOptionDefault`...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 50dda06 Franz Pletz: Merge pull request #34350 from Ma27/fix-powermanagement-default...
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vNDKE
elurin has joined #nixos
elurin has left #nixos [#nixos]
<tilpner>
gchristensen - How would you like the string to be formatted? I have no experience with PHP, and SO says there's heredocs in PHP too
<gchristensen>
heredoc would be finee
<fearlessKim[m]>
Regarding the layout loading order, I noticed that fcitx messed with that, for instance I have layout ="fr,us" but when I enable fcitx the order is reversed
mkoenig has quit [Ping timeout: 240 seconds]
<fearlessKim[m]>
(just in case that helps va1entin)
<tilpner>
gchristensen - Heredoc looks really weird, because the EOD can't have indentation :/
<gchristensen>
tilpner: its an 89-line-long file that I'd like to delete soon :P I'm not real fussed
mkoenig has joined #nixos
i-am-the-slime has quit [Ping timeout: 256 seconds]
<qwebirc442554>
how come guides use directories for drives like /dev/sdx but i have /dev/nvme0nx?
<GiGa|Laptop>
It's a Samsung printer, so I can't imagine it's in hplip
ylwghst has joined #nixos
stepcut has quit [Remote host closed the connection]
swayf has joined #nixos
Rusty1_ has joined #nixos
qwebirc442554 has quit [Ping timeout: 260 seconds]
<hodapp>
weird... I defined a Python package with buildPythonPackage (cvxpy), and had to do the same for several dependencies, but in the resultant environment I couldn't 'import cvxpy' due to it missing various things defined in buildInputs, even basic stuff like 'multiprocess'
<hodapp>
and making them nativeBuildInputs didn't fix anything; I had to just add them to the top-level list in python36.withPackages that I'm using in the environment
<hodapp>
that is not a thing I've seen happen before
<tilpner>
GiGa|Laptop - I think that should be services.printing.drivers = [ pkgs.gutenprint ];
i-am-the-slime has quit [Ping timeout: 256 seconds]
<hodapp>
Yes. I have violently destroyed around 3 or 4 printers by now - at least 2 using a claw hammer, and one using a foot or two of steel pipe.
<symphorien>
GiGa|Laptop: I have drivers=[ samsung-unified-linux-driver_1_00_37 ];
<hodapp>
for the first two, it was because I needed the ink cartridges, and the printer refused to power on fully because one of the ink cartridges was supposedly empty, and if it wouldn't power on, it wouldn't slide the head out where I could actually get it.
<symphorien>
maybe that the one you need
glenn_ has quit [Ping timeout: 276 seconds]
<MichaelRaskin>
Some printers actually work with Generic Postscript. These are not completely hell.
<makefu>
symphorien: ah yes, i have the same in my config. i've added it to the wiki
glenn_ has joined #nixos
sridii has quit [Remote host closed the connection]
michaelpj_ has joined #nixos
<symphorien>
thanks :) I was going to do that :)
<GiGa|Laptop>
symphorien: not seeing it in CUPS management, I'm expecting a universal driver or something I presume?
<GiGa|Laptop>
symphorien, found it I think - "Samsung M262x M282x Series (en)"
srid_ii has joined #nixos
<symphorien>
beware that maybe you need an older version of the driver -- newer ones drop support for old printers
ylwghst has joined #nixos
<GiGa|Laptop>
THanks for your help :)
glenn_ has quit [Ping timeout: 252 seconds]
alunduil has joined #nixos
<GiGa|Laptop>
It's certainly working :)
jensens has joined #nixos
glenn_ has joined #nixos
pie__ has quit [Ping timeout: 240 seconds]
michaelpj_ has quit [Ping timeout: 240 seconds]
glenn_ has quit [Ping timeout: 246 seconds]
asuryawanshi has joined #nixos
ssmike has quit [Remote host closed the connection]
glenn_ has joined #nixos
asuryawanshi has quit [Ping timeout: 240 seconds]
michaelpj_ has joined #nixos
fuyuuri has joined #nixos
xcmw has joined #nixos
xcmw has quit [Client Quit]
simendsjo has joined #nixos
i-am-the-slime has joined #nixos
juhe has left #nixos [#nixos]
fenedor has joined #nixos
zzamboni has joined #nixos
zzamboni has quit [Client Quit]
zzamboni has joined #nixos
glines has joined #nixos
fendor has quit [Ping timeout: 256 seconds]
zzamboni has quit [Client Quit]
<glines>
for a lot of development tasks, I use something like "pkgs = import <nixpkgs>" and then "pkgs.callPackage ./default.nix {}"
ottidmes has quit [Ping timeout: 246 seconds]
<glines>
however, I'm usually using the unstable nixos channel, which occasionally breaks... is there a way to "import <nixpkgs>" for some stable version of nixos that I know will continue to work?
<clever>
glines: put the channel name between the < and >
<clever>
the name from `nix-channel --list`
<clever>
and i think that only works for channels root has added
<GiGa|Laptop>
is there a way I can say in the main packages file "for buildMonoPackages use mono50"?
knupfer has quit [Ping timeout: 252 seconds]
<srhb>
GiGa|Laptop: Assuming buildMonoPackages takes mono as an argument, you'd override that.
zzamboni has joined #nixos
i-am-the-slime has quit [Ping timeout: 256 seconds]
<GiGa|Laptop>
don't believe so, buildMonoPackages requires mono, and mono is an alias for mono40
stackshadow[m] has left #nixos ["User left"]
glenn_ has quit [Read error: Connection timed out]
<GiGa|Laptop>
it is possible I have no idea what I'm takling about srhb, still new to this.
glenn_ has quit [Read error: Connection timed out]
justanotheruser has quit [Ping timeout: 256 seconds]
rauno has quit [Remote host closed the connection]
<sudoreboot[m]>
I have a problem where I'm trying to build a project that is trying to include a C header file which produces the error `expected specifier-qualifier-list before ‘bool’`.
<sudoreboot[m]>
I'm not sure if it's an issue with the nix expression not including some needed packages.
<sudoreboot[m]>
Anyone know what's going on there?
glenn_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joncojonathan opened pull request #34358: keepass: Change from mono40 to mono50 (master...keepass-mono50) https://git.io/vNDD2
NixOS_GitHub has left #nixos [#nixos]
mlin has joined #nixos
<GiGa|Laptop>
Ooooh, I've never seen a PR come up like that before
<sudoreboot[m]>
Btw the reason I'm writing here is that it builds on Ubuntu but not on NixOS
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/staging ac3b8fe Vladimír Čunát: kwin: fixup build with cmake-3.10 via upstream patch
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vNDDP
<mlin>
Hello all, how do I get network manager tools? I have tried `networking.networkmanager.enable = true;`, and ran `nixos-rebuild switch`. I don't see the new configuration in my EFI boot screen, nor are the nm tools installed
<clever>
mlin: is /boot mounted correctly?
<GiGa|Laptop>
If I take the checkout I've already worked on, do a "git checkout -b update-keepass238" can I work on the files to make a further change and then do another push and PR for something else?
<mlin>
clever: well I would assume so, since my system is booting
<mlin>
It shows all of the builds I performed during install
<clever>
mlin: changes being reverted like that can happen if /boot is mounted
<clever>
mlin: double-check that /boot is mounted before the nixos-rebuild
justan0theruser has quit [Quit: WeeChat 1.9.1]
glenn_ has quit [Ping timeout: 252 seconds]
vidbina has quit [Ping timeout: 260 seconds]
justanotheruser has joined #nixos
<mlin>
clever: hmm, okay. How can I check? I see that /boot contains efi/ and loader/
<clever>
mlin: is it listed when you run `mount` ?
<GiGa|Laptop>
I don't want the Push I've just done to be included in the second PR?
ssmike has joined #nixos
<mlin>
clever: ah I see that it is not
vidbina has joined #nixos
<mlin>
clever: is this a hardware-configuration.nix issue?
<clever>
mlin: yeah
michaelpj_ has joined #nixos
<clever>
mlin: youll want to manualy mount it, and fix your config, then nixos-rebuild switch
<clever>
nixos-generate-config will regenerate hardware-configuration.nix based on what is currently mounted
zzamboni has quit [Quit: Leaving.]
<mlin>
clever: I see, how is my system booting without /boot mounted then?
jluttine has joined #nixos
glenn_ has joined #nixos
<clever>
mlin: your /boot is read-only, the bootloader reads the kernel and boots, but the OS never touches /boot
<GiGa|Laptop>
I'm trying to make a pull request but it's still including my changes to all-packages.nix, even though I didn't think it was in this branch. Any advice please?
asuryawanshi has joined #nixos
mkoenig has quit [Ping timeout: 240 seconds]
mkoenig has joined #nixos
quam6246 has quit [Quit: My computer has gone to sleep. ZZZzzz…]
<CapsAdmin>
clever, i'll try that, thanks
asuryawanshi has quit [Ping timeout: 246 seconds]
stepcut has joined #nixos
zzamboni has quit [Quit: Leaving.]
stepcut_ has joined #nixos
stepcut has quit [Ping timeout: 246 seconds]
michaelpj_ has quit [Ping timeout: 240 seconds]
dan_b has quit [Ping timeout: 276 seconds]
zzamboni has joined #nixos
zzamboni has quit [Client Quit]
<simpson>
Hm, does haskellPackages.llvm-general work for anybody? It appears to not work on GHC 8, and `haskell.packages.ghc7103.ghcWithPackages (p: with p; [ llvm-general ])` doesn't work either.
ssmike has quit [Quit: ssmike]
quam6246 has joined #nixos
orivej has quit [Ping timeout: 256 seconds]
ssmike has joined #nixos
ssmike1 has joined #nixos
zzamboni has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to staging: https://git.io/vNDHV
<NixOS_GitHub>
nixpkgs/staging 8ebba20 Franz Pletz: libogg: 1.3.2 -> 1.3.3
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 9c6a17c Franz Pletz: uhd: fix build due to boost upgrade
<NixOS_GitHub>
[nixpkgs] fpletz pushed 58 new commits to master: https://git.io/vNDHw
<NixOS_GitHub>
nixpkgs/master ae939c2 Franz Pletz: sslh: 1.18 -> 1.19
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master b5737de Franz Pletz: lldpd: 0.9.8 -> 0.9.9
ssmike has quit [Ping timeout: 256 seconds]
ssmike1 is now known as ssmike
swayf has quit [Read error: Connection reset by peer]
swayf_ has joined #nixos
Myrl-saki has quit [Ping timeout: 240 seconds]
ssmike has quit [Remote host closed the connection]
glenn_ has quit [Remote host closed the connection]
randomstrangerb has quit [Ping timeout: 240 seconds]
randomstrangerb has joined #nixos
elcid has joined #nixos
quam6246 has quit [Quit: My computer has gone to sleep. ZZZzzz…]
rauno has joined #nixos
tmaekawa has joined #nixos
zzamboni has quit [Quit: Leaving.]
tmaekawa has quit [Client Quit]
stranger___ has joined #nixos
swayf_ has quit [Read error: Connection reset by peer]
<i-am-the-slime>
How might I get bluetooth to work?
<vcunat>
stranger___: I know about it (I do watch for @mentions). I even know before it happened actually :-) but it's difficult to find time for larger changes (or more risky ones).
<stranger___>
vcunat: why not just let it go through the staging?
<vcunat>
Staging isn't meant for testing.
<vcunat>
I haven't read details of that PR yet, so I can't even assess the risks of breaking X with various drivers.
<stranger___>
vcunat: the thing that it's an option, i don't enable it by default, nor i want it
<gchristensen>
it still requires holistic thinking about how it interacts with the rest of the systems
ottidmes has joined #nixos
<vcunat>
I'll try to check that it doesn't really change anything risky without enabling the option.
i-am-the-slime has quit [Remote host closed the connection]
<vcunat>
(but not today; I'm sorry)
<stranger___>
sure
amir is now known as RHEA2013
nallar has joined #nixos
Ross has quit [Ping timeout: 265 seconds]
nallar is now known as Ross
stepcut_ has quit [Remote host closed the connection]
<stranger___>
gchristensen: mesa should stay the same, mesa-glvnd is built separately. then when you set hardware.opengl.useGLVND = true, mesa-glvnd is used for default filling /run/opengl-driver/lib/ that's basically the whole change
<defau1t>
That's if I want it to compile at nixos-rebuild time. I'm looking for how to use one that I already made
<stranger___>
vcunat: do you use intel video?
chaker has joined #nixos
<defau1t>
tilpner: yeah I used that to build the kernel
randomstrangerb has quit [Ping timeout: 256 seconds]
<vcunat>
stranger___: not on this machine (where I do most nix stuff)
<vcunat>
Here I only have a Nvidia GPU.
randomstrangerb has joined #nixos
fuyuuri has quit [Quit: WeeChat 2.0.1]
<vcunat>
stranger___: assuming you're the author of this PR, the intention is that this PR is an in-between state? This seems like it still requires the LD_LIBRARY_PATH thing that we have.
<chaker>
Hey teozkr, How repo.nix differ from project/repo.nix? I can see that both file got generated by sbtix, and they have different scala version (but they have the same sbtVersion).
<vcunat>
So that first we try on this that our driver setup with GLVND is OK, and then we use libglvnd stubs as the build-time providers for libGL and friends?
orivej has joined #nixos
<teozkr>
chaker: repo.nix are the dependencies of your project, project/repo.nix are for your plugins and their dependencies
<teozkr>
SBT plugins*
<stranger___>
vcunat: it's simpler. everything builds against usual mesa as before, we only poke with mesa_drivers filling /run/opengl-driver/lib
<chaker>
Aa, okay. But is the difference in the versions reasonable?
<chaker>
I think it may lead to some problem?
<stranger___>
vcunat: i use this change for optimus setup, so i can now put intel+nvidia drivers together. libglvnd libgl dynamically figures what drivers to use by looking at what xorg modules are used
jb55 has quit [Ping timeout: 256 seconds]
dieggsy has joined #nixos
<vcunat>
I forgot this could help the mess around switchable GPUs.
<teozkr>
chaker: are you running SBT 0.13.x or did you manage to port to 1.x?
<teozkr>
if you're running 0.13.x then that's normal since all 0.13 plugins (as well as SBT itself) run Scala 2.10
<stranger___>
vcunat: if only single driver set, than there is no difference to user whether this option is enabled or not
<teozkr>
while SBT 1.x is compiled against Scala 2.12
boomshroom has joined #nixos
<boomshroom>
Hello!
CapsAdmin has quit [Ping timeout: 260 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] bgamari opened pull request #1824: configure: Use $CC instead of cpp directly (master...patch-2) https://git.io/vNDFu
NixOS_GitHub has left #nixos [#nixos]
<chaker>
I'm still using 0.13.15. So if I understand correctly, SBT build specification (and its plugins) will be compiled using Scala 2.10. But than the project will be compiled with whatever version is present in repo.nix
<boomshroom>
I trying to build the RedoxOS toolchain and when building binutils, I get an error of "bison: m4 subprocess failed: No such file or directory". I have both bison and gnum4 in the shell path and I've tried with both bison and bison2. Does anyone have any ideas?
<chaker>
I think I will try to invest some time trying to use 1.x because I still having problem that seems to be coming from the inconsistencies in the Scala version between both files.
CapsAdmin has joined #nixos
knupfer has quit [Ping timeout: 265 seconds]
ixxie has joined #nixos
tusj has joined #nixos
<stranger___>
vcunat: > This seems like it still requires the LD_LIBRARY_PATH thing that we have <- yes it currently does. is it planned to drop LD_LIBRARY_PATH thing? and what is planned instead?
<CapsAdmin>
after i built nix on msys2 i get "warning: unable to download 'https://google.com': Problem with the SSL CA cert (path? access rights?) (77); retrying in 564 ms"
<CapsAdmin>
when running for instance nix-prefetch-url https://google.com
<CapsAdmin>
or for instance nix-channel --update
<stranger___>
vcunat: rely on glvnd for picking correct drivers? i believe it's still not ready for that :\ at least i heard amd is not supported
dbe_ has joined #nixos
swayf has quit [Read error: Connection reset by peer]
swayf has joined #nixos
<vcunat>
hmm, yess, it's possible that AMD (in some driver variants)
<stranger___>
vcunat: thanks for talk on that! i've changed pr description to match what it really does
asuryawanshi has joined #nixos
fyuuri has joined #nixos
<stranger___>
rebuilding everything against mesa-glvnd is indeed to early
asuryawanshi has quit [Ping timeout: 256 seconds]
<hyper_ch>
infinisil: so, sneding incremental snapshots (non-raw mode) with new format works... I altered my script now to send the snapshots created by auto-snapshot :)
<tilpner>
gchristensen - I prefer my format for when entries get merged (it's a little more compact), but yours is better while they aren't. Now that the partial log is *free*, you could even make it a little longer
pie_ has quit [Ping timeout: 248 seconds]
<simpson>
Hi! What's the best way to bind LLVM from a memory-safe language in nixpkgs? I tried Haskell's bindings but they don't build with either GHC 8 or 7.10.3.
<tilpner>
gchristensen - That was responding to the first link. Your second link is better, I tested that exact layout too (and then got carried away in making it even smaller)
chaker has left #nixos ["ERC (IRC client for Emacs 25.3.1)"]
<cinimod>
How do install thunderbird? If I do `nix-env -qa thunderbird` then I am told `error: selector ‘thunderbird’ matches no derivations`
<boomshroom>
simpson: Since C++, and by extension LLVM, you would have to enforce the safety yourself.
<stranger___>
going to sleep. bye folks
<tilpner>
cinimod - nix-env -iA thunderbird should work
stranger___ has left #nixos [#nixos]
<gchristensen>
tilpner: cool :) thanks. I've found people don't know those sections can be expanded without being told so, hence the note
<tilpner>
Uhh, nixpkgs.thunderbird
<simpson>
boomshroom: Yes, I know. Lemme rephrase: I ain't writing C++ today. Where are working non-painful LLVM bindings in nixpkgs?
tusj has quit [Remote host closed the connection]
<boomshroom>
simpson: The alternative is to bypass the C/C++ bindings and write a library that outputs bitcode or ir directly.
knupfer has joined #nixos
<vcunat>
or nixos.thunderbird (typical on NixOS). It depends on what names you give to your channels.
<cinimod>
not supported on ‘x86_64-darwin’
<cinimod>
meep
<vcunat>
Ah, Darwin :-)
<gchristensen>
most gui stuff on darwin is best fetched the old fashion way, cinimod
<cinimod>
Oh well I am sure the thunderbird folks have a dmg
<boomshroom>
simpson: If you're looking at memory safe languages, then you're better off looking at the package managers for the specific languages rather than the system package manager.
<simpson>
boomshroom: Everybody suggests that LLVM's low-level ABI is even more unstable than its high-level bindings...
<cinimod>
gchristensen: but I have been bitten by the nix bug
<simpson>
boomshroom: It's okay to just say "I don't know", y'know.
<gchristensen>
cinimod: alas, it sounds like NixOS is in your future... :D
<cinimod>
Not unless it has excel
<tilpner>
Uhm, Libreoffice Calc?
<vcunat>
cinimod: it's possible that it will just build and work - it's just that we haven't tested it yet
<vcunat>
(but probably it will need some adjustments)
<boomshroom>
simpson: That's probably accurate, but it's the only way to avoid using the bindings directly. If you are using the bindings, then you just need to have one of the llvm packages in your path.
<cinimod>
I am not skilled enough to make adjustments (yet)
<vcunat>
cinimod: you can just try it out by env NIXPKGS_ALLOW_BROKEN=1 nix-env -iA nixpkgs.thunderbird
<cinimod>
sure
<simpson>
boomshroom: Sorry, that doesn't seem to be why `haskell.packages.ghc7103.ghcWithPackages (p: with p; [ llvm-general ])` isn't working for me. Does it work for you?
<boomshroom>
I'm not as familiar with Haskell's bindings, but from my understanding, they're divided into "pure" which output ir text directly, and "impure" which interacts with the bindings.
<simpson>
It doesn't even build for me.
<cinimod>
tilpner: my experience with excel substitutes has not been great
<tilpner>
cinimod - Yeah, I wasn't quite serious. I don't use either :/
<simpson>
`pythonPackages.llvmlite` builds, though, so I guess that Haskell's failed me for the last time once again.
<boomshroom>
When using Haskell and nix, I usually use stack with the nix flag. Then in stack.yaml, you would have a packages field of nix where you specify the native packages you're using.
zzamboni has joined #nixos
<cinimod>
That was from using R and (r)gdal via nix yesterday
<cinimod>
> you can just try it out by env NIXPKGS_ALLOW_BROKEN=1 nix-env -iA nixpkgs.thunderbird
<cinimod>
Sadly it failed
<boomshroom>
simpson: When I was playing with Haskell and LLVM, I was mostly using llvm-pretty which has a very nice API and bypasses the bindings (for better or for worse).
<cinimod>
vcunat: see above
<hyper_ch>
hi vcunat
<vcunat>
hyper_ch: hi :-)
<simpson>
boomshroom: Well, that one builds, at least, so maybe it'll be useful. Thanks.
<cocreature>
simpson, boomshroom: hey, I’m one of the maintainers of the llvm-hs family of packages. if you’re having build problems, feel free to open issues!
<cocreature>
8.0 and 7.10 should definitely be supported
sud0x3 has joined #nixos
<simpson>
cocreature: Should I be using llvm-general or llvm-hs? I've been reading http://www.stephendiehl.com/llvm/ which assumes llvm-general.
<vcunat>
cinimod: the package support for Darwin in nixpkgs just isn't on par with x86_64-linux. (And it's complicated by various stuff, like just access to such virtual machines.)
<boomshroom>
simpson: No problem. I was even playing with operator overloading and was able to make wrappers to use operators like (+) on arbitrarily sized llvm integers.
<cocreature>
simpson: obviously I’m biased given that I created llvm-hs but I think it’s fair to say that llvm-general is dead at this point and you should be using llvm-hs.
<boomshroom>
simpson: Probably. I want to write a compiler for a haskell-like language. I've already made something that mostly works in Rust, but the parser is broken and the type checker was incredibly difficult to write and unreadable.
glenn_ has joined #nixos
ssmike has joined #nixos
freeman42x[NixOS has quit [Ping timeout: 252 seconds]
<boomshroom>
The idea is that it would have some form of substructural type system like Rust so I could use it without allocating on the heap. (Because in my domain, we don't have one.)
<hyper_ch>
so, the zfs backup script seems to work nice.. now all I have to do is nixos-option-ify it :)
mlin has quit [Ping timeout: 260 seconds]
fresheyeball has joined #nixos
<boomshroom>
I'm still confused by the "bison: m4 subprocess failed: No such file or directory" that I got when trying to compile the redox toolchain.
vidbina has quit [Ping timeout: 256 seconds]
<fresheyeball>
hey out there
<hyper_ch>
I doubt anyone is "out there" in that cold
<fresheyeball>
I am having a hell of a time getting hakyll-sass to work
mrkgnao has quit [Ping timeout: 240 seconds]
<fresheyeball>
It depends indirectly on static linking with libsass
<bgamari>
Is there any way to get nix to somehow identify what phase a particular command is coming from?
<bgamari>
I am seeing a derivation fail with an erroneous mkdir during or after fixupPhase
<bgamari>
but I can't for the life of me figure out where it is coming from
<clever>
bgamari: can you paste the error here?
<bgamari>
checking for references to /tmp/nix-build-linux-config-4.9-armv7l-unknown-linux-gnueabihf.drv-3 in /nix/store/fhvjhvii4bh18xracfb916lhgacns5k7-linux-config-4.9-armv7l-unknown-linux-gnueabihf...
<bgamari>
waiting for children
<bgamari>
building of '/nix/store/ikxn76fbpwsnl3r3jvwsd438bx3fm3my-linux-config-4.9-armv7l-unknown-linux-gnueabihf.drv': read 142 bytes
<bgamari>
mkdir: cannot create directory '/nix/store/fhvjhvii4bh18xracfb916lhgacns5k7-linux-config-4.9-armv7l-unknown-linux-gnueabihf': Not a directory
<bgamari>
(that is with -vvvvv, which unfortunately wasn't particularly helpful)
<bgamari>
clever, note the "checking for references to..."
<clever>
bgamari: that sounds like it failed after the build finished
macslayer has joined #nixos
<bgamari>
which suggests that it's somewhere near fixupPhase
<bgamari>
clever, yes, the build is very nearly done
<bgamari>
it makes it through installPhase and most, if not all, of the way through fixupPhase
<bgamari>
I don't know where else this could come from?
<macslayer>
Hey, I have a quick `fontconfig` question
quam6246 has joined #nixos
Jackneill has quit [Ping timeout: 240 seconds]
<clever>
bgamari: try running "set -x" in the postInstall region
<macslayer>
I want to use fontconfig to configure my `monospace` font on my NixOS system, and for some reason, even after installing `fontconfig` and `hack-font`, I can't seem to find the Hack font.
freeman42x[NixOS has joined #nixos
<macslayer>
For example, doing `fc-cache -fv; fc-list | grep hack` returns nothing
<bgamari>
okay, so this raises a question that I have long been wondering about
<bgamari>
there seems to be a pervasive assumption that $out will be a directory
<bgamari>
but that isn't always the case, as we see here
<clever>
very rarely will it be a file
<gchristensen>
holy crap fpletz
ma27 has joined #nixos
<clever>
bgamari: propagated build inputs work via $out/nix-support/something
<clever>
bgamari: so you need to ensure you dont have those if $out is a file
<bgamari>
clever, but there are a variety of places which assume that it will be a directory or a script
<gchristensen>
fpletz: you've been on a tear
<bgamari>
e.g., stdenv will attempt to source buildInputs that are bound to be files
<bgamari>
found*
<fpletz>
gchristensen: 18.03 is coming up, gotta get everything in shape :>
<macslayer>
clever: Ah, okay
<macslayer>
I'll do that then
<gchristensen>
fpletz: you're too good for us
<clever>
bgamari: i think only for setup hooks will the stdenv source things
<bgamari>
yes, that is possible
macslayer has quit [Quit: Konversation terminated!]
y0no has joined #nixos
dan_b has joined #nixos
<clever>
bgamari: can you gist the derivation that is having this failure?
pie__ has joined #nixos
<tmplt>
Speaking of fonts, I have one which don't exist in nixpkgs, what is the NixOS equivalent to `mkfontdir; mkfontscale; xset +fp .`? `+fp` complains about incorrect server address
stephenjudkins has quit [Remote host closed the connection]
<bgamari>
clever, it's actually just pkgs.linux
stephenjudkins has joined #nixos
<bgamari>
clever, but under cross-compilation
<bgamari>
I found the problematic code
<bgamari>
it's pkgs/stdenv/generic/setup.sh, line 1089
sud0x3 has quit [Quit: WeeChat 2.0.1]
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever>
bgamari: ah, so the setupHook attribute was defined on the derivation, and its trying to create $out/nix-support/setup-hook, for future sourcing
stepcut has joined #nixos
<bgamari>
hmm
__monty__ has joined #nixos
<clever>
bgamari: if you eval it in nix-repl, can you confirm the presense of that attribute?
<clever>
and also via running `env` around postInstall
<__monty__>
How can I compare the versions of all packages in two or more generations?
<bgamari>
clever, linux-config doesn't appear to have a setupHook attribute
<clever>
the cross-compile may add things
<bgamari>
mm
zzamboni has joined #nixos
lexsomething has joined #nixos
xcmw has joined #nixos
<bgamari>
well that is problematic
<lexsomething>
hello guys :) how r u doin?
<boomshroom>
lexsomething: Fine. I have a few things that I want to work on but are stuck for various reasons.
hamishmack has quit [Quit: hamishmack]
<lexsomething>
yeah, buddy, same issue
<lexsomething>
so much plans, so many obstacles, so little time.
<bgamari>
clever, strongly enough I don't see a setupHook
<boomshroom>
Speaking of which, I should probably work on 1 of the 2 homework assignments that are due tomorrow.
<fresheyeball>
clever executableSystemDepends
vcunat has quit [Quit: Leaving.]
<bgamari>
hmm, although the hash strangely doesn't match what it was building previously
Intensity has joined #nixos
<fresheyeball>
clever: I just tried putting them in both executableSystemDepends, and librarySystemDepends, no dice, wont link
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 36103e9 Franz Pletz: nixos/powerManagement: remove duplicate definition...
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vNDxj
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 8417649 Franz Pletz: hwloc: 1.11.8 -> 1.11.9
<lexsomething>
meanwhile, let's say I have a list of strings in format : libatk-1.0.so.0 => not found , libatk-1.0.so.0 => not found , libatk-1.0.so.0 => not found , etc, and so on. What will be the best way to generate a list of NIX packages based on required file names??? I would like to query something like https://rpmfind.net but for nixpkgs
vidbina has joined #nixos
<buhman>
is the nix-pills references to pp-aterm out of date?
<buhman>
it seems like that doesn't exist and it really means to say to use "indent"
<symphorien>
lexsomething: there is nix-index / nix-locate if I understand correctly what you want
<bgamari>
clever, well, /nix/store/sqiflknbarqvkqsvjkrcr421rqdgvavl-gcc-wrapper-6.4.0 is indeed in propagatedBuildInputs
gilligan_ has joined #nixos
<bgamari>
so that is presumably where the setupHook is coming from
<__monty__>
Is there no way to find out what versions you'd be at if you did a rollback?
ylwghst has joined #nixos
<chisui>
I'm currently looking at https://github.com/NixOS/nix/blob/master/src/libutil/archive.cc unfortunatly it's an uncommented blob of spaghetti. Can someone tell me what this `caseHackSuffix` String is used for? It seems to be messing something up in `ncurses`
civodul has joined #nixos
raynold has quit [Quit: Connection closed for inactivity]
stepcut has quit [Remote host closed the connection]
<lexsomething>
symphorien, thank you, really appreciate your help <3 , I definitely will look on it. In my particular usecase I do not have packages installed that provide, let's say libatk-1.0.so.0, so I expect from tool to query online\offline database and tell me that I need to install particular _package_name_ that provide libsomething.so
a6a3uh has quit [Quit: Bye]
<lexsomething>
I maybe wrong, but i'm afraind nix-index will query /nix/store only for installed packages. Doesn't it?
stepcut has joined #nixos
<symphorien>
no it also works for not installed package
<symphorien>
+s
<samueldr>
you're seeding it with something that's loaded from the cache
<Rovanion>
So I'm building avr firmware and am missing the inttypes.h header. Which package should I install, or rather how do I figure this out on my own?
<tilpner>
chisui - AFAICT, the case hack is a workaround for when macOS uses HFS+ in case-insensitive mode
<tilpner>
chisui - I think you may have been focusing on the wrong issue :/
ylwghst has quit [Ping timeout: 248 seconds]
<tilpner>
chisui - Hmm, but you're not on macOS, right?
<Rovanion>
tmplt: Hmm, didn't seem to help I'm afraid...
MP2E has joined #nixos
<tmplt>
You can find it in avr/include/inttypes.h
<Rovanion>
I'm not sure where I find the avr folder to begin with.
<clever>
Rovanion: nix-build '<nixpkgs>' -A avrlibc
<tmplt>
I'm not sure if you must do anything special on NixOS to reach it from avr-gcc; I have my avr headers in a local `include`
ma27 has quit [Ping timeout: 265 seconds]
statusfailed has joined #nixos
ma27 has joined #nixos
<tmplt>
Additionally, I've had to pass -L"<path-to-avrlibc>/avr/lib/avr5` to the compiler and copy an object file into the same directory I'm working in. I don't know why that is
<Rovanion>
clever: Yes there's the header. Do I need to do anything special for the make avr-gcc look there for headers?
<clever>
Rovanion: try adding avrlibc to the buildInputs
<chisui>
tilpner: This is the issue: https://github.com/NixOS/nixpkgs/issues/34364 I have files on my system with that suffix and archive.cc is the only file I could find in the nix sources that contains that string
<Rovanion>
Thank you for being so patient with me!
<tilpner>
chisui - I'll get the ISO and see if I can reproduce
<__monty__>
Can I have a confirmation that it's not possible to compare the versions of packages across generations?
<17SAAM33Z>
[nixpkgs] fpletz pushed 1 new commit to staging: https://git.io/vNDhq
<17SAAM33Z>
nixpkgs/staging 192f30f Will Dietz: nettools: 1.60_p20161110235919 -> 1.60_p20170221182432...
17SAAM33Z has left #nixos [#nixos]
* lexsomething
still compiling rust dependecies at 80c for 15th minute ;`3 , guess, it's lex's fault the he decided to install package rootless in user environment.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vNDhO
<NixOS_GitHub>
nixpkgs/master 103cf02 Franz Pletz: libewf: 20140608 -> 20171104
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 5f57b61 Franz Pletz: sleuthkit: 4.2.0 -> 4.5.0
fresheyeball has quit [Quit: WeeChat 1.9.1]
srdqty has joined #nixos
knupfer has quit [Ping timeout: 246 seconds]
kreetx has quit [Ping timeout: 256 seconds]
<lexsomething>
at least I'm calm that it will compiles fine without any questions, not like gentoo, thanks to nix pure functions.
<lexsomething>
but i'm not an expert, defau1t, just a newbie user
<MichaelRaskin>
Hm, maybe make a Nixpkgs overlay with it and host it on the development site?
<MichaelRaskin>
I mean, making such a Nix package is easy, better reasons are only about inclusion in Nixpkgs.
hiratara has quit [Ping timeout: 252 seconds]
<defau1t>
Mm, how does one use overrideAttrs... lexsomething
stephenjudkins has quit [Remote host closed the connection]
<gchristensen>
MichaelRaskin: for packages with few rebuilds, I've considered automatically sending build jobs per attribute changed during the evaluation phase
fresheyeball has joined #nixos
<__monty__>
MichaelRaskin: Not familiar with overlays. Would that work if you want package-master versions of multiple projects?
<MichaelRaskin>
You need to add something writeable by all the build users to the directories accessible by builds. Then you write a fetcher that creates an empty file (m-m-m, hash known in advance!) with the name derived from builtins.currentTime (m-m-m, always new output path!), but actually updates the shared-storage clone. Then you need to write a derivation that is not fixed-output, that looks at the shared-storage clone and checks out the master branch.
hiroshi has quit [Ping timeout: 240 seconds]
quam6246 has quit [Quit: My computer has gone to sleep. ZZZzzz…]
silver_ has joined #nixos
fresheyeball has quit [Ping timeout: 256 seconds]
<MichaelRaskin>
Well, you can always just switch off the sandboxing…
<MichaelRaskin>
I think something like noChroot allows you to switch off the sandboxing for a single derivation (source fetching — don't forget to depend on current time, though)
concatime has quit [Ping timeout: 260 seconds]
silver has quit [Ping timeout: 240 seconds]
jtojnar has quit [Remote host closed the connection]
jtojnar has joined #nixos
<winem_>
Hi, I tried to build one of our applications with aarch64 as dependency but I run into this error "error: a ‘aarch64-linux’ is required to build ‘/nix/store/rw4jfzwdgvpdzxxnh6dwrvc7lnkwfy14-crypto-d585fd2.drv’, but I am a ‘x86_64-linux’". we use the packages from the nixpkgs-unstable channel. I'm a bit confused since the error message is quite describtive but I don't know how to fix that
<MichaelRaskin>
That doesn't break/work around any Nix assumptions
stephenjudkins has quit [Read error: Connection reset by peer]
<lexsomething>
and, I'm sorry to say team, my nix-index installation\compilation failed, after 30 min of rust compilation . out of stack error https://pastebin.ca/3965443 . It is kinda ironycal in contrast with my words from 00:25 . well I'll go sleep and have a look at it tomorrow, cya . bennofs[m]
<__monty__>
MichaelRaskin: That'd always be a local solution though, right? Ideally anyone interested in developing this project could just nix-shell -p package-git.
gilligan_ has quit [Remote host closed the connection]
stephenjudkins has joined #nixos
dan_b has quit [Ping timeout: 256 seconds]
<MichaelRaskin>
__monty__: automatically generate the updated master reference, and publish as an overlay.
<gchristensen>
tilpner: hmm I'm liking the approach of detecting the support based on the build output, in the comment posting step for now. it keeps the front and build phases pretty simple, which I think is good
<__monty__>
MichaelRaskin: How does one use overlays? Is it like channels?
<boomshroom>
__monty__: I personally use overlays to add custom packages without modifying nixpkgs.
<gchristensen>
tilpner: the trouble there is then I'd be saddling you with porting the PHP to Rust which is a bit unfair :)
<MichaelRaskin>
Well, the manual tells the details (I happen not to use any), but basycally you checkout the corresponding repos into a specific place (there are options on how to configure this place), and then all of them modify what is seen on importing Nixpkgs
hiroshi has joined #nixos
<boomshroom>
gchristensen: Any oportunity to post software to Rust makes me interested. Tell me more?
<gchristensen>
boomshroom: oh hello =)
<tilpner>
gchristensen - I might be okay with that, I worry more about being able to test changes. I assume there's a lot of initial effort required for a dev setup
<gchristensen>
tilpner: I admit I've played fast and loose w.r.t. testing, and have only started adding ways to test the code more recently
<__monty__>
boomshroom: Tell me why rustaceans are obsessed with this? Often it's not the best way forward imo.
<tilpner>
gchristensen - That may be acceptable for you, because you have access to the infrastructure running ofborg. I don't, and that would make attempted ports very painful
freeman42x[NixOS has quit [Quit: Leaving]
stephenjudkins has quit [Remote host closed the connection]
stephenjudkins has joined #nixos
<gchristensen>
tilpner: I know :( it makes testing newer thing spooky, too, because it is important I don't accidentally do things visibly
<boomshroom>
__monty__: I personally don't advocate for the "Rewrite everything in rust," I do enjoy atempting to port stuff myself. The main reasons is Rust is generally safer than C/C++ and faster than most interpreted languages and languages with a garbage collector. That said, it can be quite painful trying to learn how to write rust in the first place.
cinimod has quit [Quit: Connection closed for inactivity]
<boomshroom>
And asking someone to rewrite their code in rust is asking them to throw out all their hard work and start from scratch.
<gchristensen>
I'd say it isn't acceptable to me, tilpner, though, long term. in the beginning it was a race to provide value to the community. now that it does, the testing becomes more important
defau1t has quit [Ping timeout: 260 seconds]
<tilpner>
gchristensen - I'm okay with you saying "Sorry, I don't have ofborg setup for easy contributions yet, I'll port this myself (eventually)", I just want to make sure this feature doesn't bottleneck on lack of volunteering :)
<__monty__>
boomshroom: I just don't see the value in making many things faster especially considering the slower development that often entails.
zzamboni has quit [Read error: Connection reset by peer]
zzamboni1 has joined #nixos
<__monty__>
boomshroom: My perspective is shaped by the haskell community where instead of rewriting stuff people come up with better ways to do things usually. That's way more valuable imo.
<gchristensen>
__monty__, boomshroom may I suggest ##nixos-discusses-rust for this? :P
<tilpner>
gchristensen - Maybe we/you can discuss/document a testing setup, probably not in #nixos though
randomstrangerb has quit [Ping timeout: 256 seconds]
<boomshroom>
__monty__: That's why I don't go around asking people to rewrite things in rust. I just start trying to port stuff as a personal exercise.
<MichaelRaskin>
tilpner: #nixos-borg does exist
* tilpner
was not aware
<tilpner>
Thanks!
<MichaelRaskin>
(or are you there already)
<gchristensen>
boomshroom: if you'd like to try porting something, I can suggest one in #nixos-borg
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to gcc-7: https://git.io/vNyfu
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/gcc-7 e2c1c76 Franz Pletz: Merge remote-tracking branch 'origin/master' into gcc-7
randomstrangerb has joined #nixos
<__monty__>
gchristensen: I'll stop discussing it. Just wondering why rustaceans are especially obsessed with rewriting things. Especially things which weren't written in C/C++.
hiratara has joined #nixos
raynold has joined #nixos
<gchristensen>
__monty__: I suspect part of it is just it can be a fun thing to do and provide some challenges to help them get better... and perhaps more viable when doing it in rust than in, say, NodeJS or Ruby
<lexsomething>
I'm sorry to say, C u mentioned rust here, i'm gonna kick some rust'asses tommorow cuz, in my perfect pure purr world some small rust package 1)takes forewere to compile 2) even fails to finish compilation after that. I bet C analogue wouldn't have such issues on my systems. It's not like I want to start a discussion, just preparing you and myself mentally for the challenges of tomorrow
<simpson>
lexsomething: Don't worry about it; all software is terrible.
<NixOS_GitHub>
[nixpkgs] fpletz pushed 4 new commits to master: https://git.io/vNyf9
<NixOS_GitHub>
nixpkgs/master 93d7267 Symphorien Gibol: bombono: add symphorien as maintainer
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 4b0dc51 Symphorien Gibol: bombono: import patches from AUR to fix gcc7 build
<__monty__>
gchristensen: Quoting a terrible language and a very slow language doesn't exactly make for a convincing argument : ) Don't get me wrong I really appreciate some projects, like ripgrep for example it's great. Most of the RiiR sentiment is just counterproductive though.
lexsomething has quit [Ping timeout: 260 seconds]
* gchristensen
shrugs
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vNyfp
<NixOS_GitHub>
nixpkgs/master 5ba11ca Linus Heckemann: pmidi: init at 1.7.1
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 2b49df1 Franz Pletz: Merge pull request #34344 from lheckemann/pmidi...
<tmplt>
Can nix-shell merge shell.nix with the current env instead of replacing it? i.e. so that one can use an IDE within a nix-shell
thblt has quit [Ping timeout: 240 seconds]
<boomshroom>
__monty__: It is.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master f9238bf Will Dietz: ipvsadm: init at 1.29...
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vNyJI
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 4bb42d0 Franz Pletz: Merge pull request #34235 from dtzWill/feature/ipvsadm...
jtojnar has joined #nixos
<Ralith>
tmplt: nix-shell does not clobber the environment you run it in by default
<Ralith>
except for nix-specific things anyway
<__monty__>
boomshroom: Why is it a thing though? Is it because rust attracts a huge number of hipsters or something? Why is this not a thing (or hardly a thing) in languages like haskell, clojure, other languages that have ~recently gained popularity?
<Ralith>
if you want an IDE environment derived but not equal to from your build environment, you should just define it that way
<tmplt>
I should define the new env as a derivation of the one I'm calling nix-shell from?
<__monty__>
Is there a reason to use package overrides rather than overlays?
<boomshroom>
__monty__: Maybe? It is my favourite language personally because of its lack of a garbage collector and all the languages you mentioned use one. Clojure in particular runs on the JVM which can turn some people off. From what I can tell, the fanboys see it has having all the benefits of C/C++ and of higher level languages and ignore the downsides of being exceedingly hard to learn and being very verbose.
<MichaelRaskin>
I think that Rust, unlike languages like Haskell, tries to change fast and needs to make sure it is suitable for randomly selected tasks in the low-level land. RiiR serves as random sampling, testing and feedback.
<__monty__>
boomshroom: My issue with that is that garbage collectors are a non-issue in most cases. Some people make way to big a deal of having a GC.
<__monty__>
MichaelRaskin: I disagree about haskell not changing fast. But this is at least a better motivation for rewrites than I've seen so far.
<MichaelRaskin>
(I am sometimes afraid that maybe RiiR people look at the recent history, see how Mozilla basically annoyed the world into rewriting half the web into semi-decent HTML and want their Mozilla advocacy — for Rust — to match that; I hope these fears are unfounded)
<boomshroom>
__monty__: For the most part they are fast enough. Some people see any detraction to performance as a bad thing (despite the fact that Rust often performs on par with Go) and want maximum performance.
<MichaelRaskin>
__monty__: I have been told not to use significant parts of Prelude in Haskell. Rust wants not to have such a situation with std.
<boomshroom>
__monty__: My perfered domain has much harder requirements, namely that I don't have a heap to collect or allocate from until I set one up myself, but it's extremely rare that you'll find someone operating in that domain.
<MichaelRaskin>
And when I tried to use Haskell (and when I actually used Clojure) it is acceptable to say that you need to try and formulate your task in a way that matches. Rust tries to be able to just take a spec which is good enough for a C program.
<__monty__>
boomshroom: People who value performance above developer convenience are nearly always wrong imo. Computers are plenty fast, people are not.
jb55 has quit [Ping timeout: 260 seconds]
<__monty__>
MichaelRaskin: Afaik the borrowing system pretty much forces your hand in many cases.
<__monty__>
(Maybe we should actually move off-channel, wasn't planning on going into it this much.)
<MichaelRaskin>
__monty__: it is a layer deeper than the spec, though. But yes, there is a need for unsafe.
<MichaelRaskin>
#nixos-discusses-rust ?
<boomshroom>
If you need something done 2 billion times a second, you're spending half of your time doing that thing. It's unlikely you'll ever write that extreme, but it's very common to only have 16 milliseconds to perform all your logic in time to render.
<gchristensen>
boomshroom, __monty__ please drop it from this channel.
Acou_Bass has quit [Quit: byeeeeeeeeeeeeeee]
<boomshroom>
OK. I'm a fan of the language, but RiiR people have become a vocal minority.
Acou_Bass has joined #nixos
simukis has quit [Quit: simukis]
<gchristensen>
that isn't dropping it from this channel :)
shabius has quit [Remote host closed the connection]
stepcut has quit [Remote host closed the connection]
jtojnar has quit [Remote host closed the connection]
jtojnar has joined #nixos
ma27 has quit [Ping timeout: 276 seconds]
mkoenig_ has joined #nixos
petar has quit [Remote host closed the connection]
mkoenig has quit [Ping timeout: 248 seconds]
petar has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
zzamboni1 has quit [Quit: Leaving.]
chisui has joined #nixos
<chisui>
I'm trying to set my password using the hashedPassword option. I run mkpasswd -m sha-512. Enter my password and past the output into the hashedPassword option and run nixos-rebuild switch. but I can't login with that user. This is the User entry in my configuration.nix file https://gist.github.com/chisui/069ce1df6d48d65cec71c18a1bac5fbb
<chisui>
Am I missing something?
__monty__ has left #nixos [#nixos]
<clever>
chisui: one min
<chisui>
clever: thanks
<tilpner>
chisui - "If the option users.mutableUsers is true, the password defined in one of the three options will only be set when the user is created for the first time."
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<chisui>
tilpner: so if I want to use hashedPassword I have to disable mutableUsers?
<tilpner>
chisui - And users.mutableUsers is true by default
fyuuri has quit [Ping timeout: 268 seconds]
<tilpner>
chisui - That's what the manpage says, yes
<clever>
chisui: if you had the hashedPassword in the config before the user had been made, it would have taken effect
nix-gsc-io`bot has joined #nixos
<clever>
chisui: but since the user is pre-existing, its defering to the existing password hash