<gchristensen>
"rofi-pass comes with a tiny helper script, which makes it easier to create new pass entries. Just run it with --name "My new Site" +user "zeltak" +branch "branch" +custom "foobar" +autotype "branch :tab user :tab pass"" (lol)
<alphor>
I'm installing on an old raspberry pi, not one of the fancy new ones.
<gchristensen>
the "just" there is funny :)
<benley>
I wish I could figure out how to convince nixos to just run freaking "zfs mount -a" at boot
<benley>
there is a zfs-mount service that's supposed to do it
<benley>
and it just ... isn't
filterfish has quit [(Ping timeout: 268 seconds)]
indi_ has quit [(Remote host closed the connection)]
indi_ has joined #nixos
m0rphism has joined #nixos
bennofs has quit [(Ping timeout: 245 seconds)]
<alphor>
or am I meant to edit configuration.nix in /etc/nixos. That seems to make more sense.
<benley>
You can edit it in /etc/nixos
<benley>
I have it come from my home directory on my home machine, but the default is for it to be mutabale in /etc/nixos
<alphor>
benley: mk, that sounds like a good idea.
<alphor>
thanks :)
indi_ has quit [(Ping timeout: 258 seconds)]
AllanEsp1nosa has joined #nixos
mizu_no_oto has joined #nixos
indi_ has joined #nixos
<ToxicFrog>
Hmm. kdirstat/qdirstat don't seem to be available :/
slack1256 has quit [(Remote host closed the connection)]
hugod has joined #nixos
SovereignBleak has quit [(Quit: Connection closed for inactivity)]
adlaiduncan9 is now known as adlaiduncan10
<manveru>
hmm, enpass is kinda nice... can use it with seafile, but the form input is pretty finicky
ambro718 has quit [(Ping timeout: 264 seconds)]
indi_ has quit [(Remote host closed the connection)]
indi_ has joined #nixos
cjhowe has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
AllanEsp1nosa has quit [(Ping timeout: 256 seconds)]
edvorg has quit [(Ping timeout: 248 seconds)]
<NixOS_GitHub>
[nixpkgs] chris-martin opened pull request #21515: ppl-address-book: init at 2.4.0 (master...ppl) https://git.io/vMI5a
filterfish has joined #nixos
<eacameron>
How do I figure out what "sysconfdir" is set to for nix-copy-closure?
<eacameron>
The docs refer to it but I'm not sure what it's set to.
zraexy has joined #nixos
indi_ has joined #nixos
bitchecker has quit [(Ping timeout: 258 seconds)]
indi_ has quit [(Remote host closed the connection)]
indi_ has joined #nixos
vtomole has quit [(Ping timeout: 260 seconds)]
<clever>
eacameron: i think its /etc or /etc/nix
<eacameron>
clever: Thanks. I found that answer in the IRC logs.
<eacameron>
We really need to get that into the manual or something.
<eacameron>
clever: Also, the format of the file used by NIX_REMOTE_SYSTEMS is somewhat opaque. Your examples have "1 1" at the end, but I've seen others that just have "1" at the end.
<eacameron>
I can't find any docs on that file format. It would be super nice if it were a readable format, or even a nix file.
<clever>
eacameron: ah, i believe one is the max-jobs, and the other is the speed factor
<eacameron>
speed factor?
<clever>
how 2 systems with an equal number of cores compare to eachother in terms of speed
<eacameron>
clever: Ah hmm...
<clever>
let me pull up a nix file...
tvon has quit [(Remote host closed the connection)]
<clever>
eacameron: once chromium starts responding and lets me, lol
<eacameron>
clever: can't you install google-chrome?
sdothum has quit [(Quit: ZNC - 1.6.0 - http://znc.in)]
<clever>
the problem isnt the browser, its the number of open tabs
<clever>
eacameron: it is up to 21gig of swap used so far...
<eacameron>
clever: heh...tabs-outliner is a wonderful extension
<clever>
i'm testing one-tab right now, and its already getting buggy
<clever>
it consumes whatever window is active, when it gets the button event
<clever>
even it i changed windows after clicking the button
<eacameron>
yowza
<clever>
it still saves them to the master list, but its anoying when it saves the wrong window
<clever>
which is usualy the one i was just using
<eacameron>
clever: You need to go to ebay and get a 72gb ram 16 core (dual xeon) compy for $700 like I just did :P
<clever>
lol
<clever>
i alredy have 16gig of ram and 8 cores
<eacameron>
16gb ram is so 2010s
<clever>
yeah, chrome burns thru it so fast
<eacameron>
LOL
<clever>
id need 30 or 40gig of ram to stop it from swapping
<clever>
up to 70% swap usage!
katyucha_ is now known as katyucha
<eacameron>
clever: yah tabs-outliner
<NixOS_GitHub>
[nixpkgs] timbertson opened pull request #21516: gup: 0.5.5 -> 0.6.0 and extract src into JSON (master...gup) https://git.io/vMIbG
<eacameron>
clever: Lets you keep track of tabs much better and you can feel better closing some of them without losing your place
indi_ has quit [(Remote host closed the connection)]
tvon has quit [(Ping timeout: 258 seconds)]
indi_ has joined #nixos
filterfish_ has joined #nixos
filterfish has quit [(Remote host closed the connection)]
<sphalerite>
clever: gchristensen pinged you yesterday when I had some questions about PXE booting, but then I went to sleep
<sphalerite>
https://gist.github.com/lheckemann/78e4cb38ae3cc280b6d89f8f13e7727f here's what I use to run my PXE server with which I'm trying to boot a machine so I can install nixos on it. Any idea what I might be doing wrong so that the nix store is mounted read-only and nixos-install doesn't work?
indi_ has quit [(Remote host closed the connection)]
systemfault has quit [(Quit: Bye!)]
rly has joined #nixos
<clever>
sphalerite: are you using nixos release or unstable?
<sphalerite>
release
shanemikel has joined #nixos
<clever>
sphalerite: i fixed the store being read-only, but that fix has only made it into nixos-unstable
<sphalerite>
aaah right
<clever>
sphalerite: you can always switch nix-channel back to release after it netboots, if you wish to install release into the target machine
<sphalerite>
so I just replace pkgs in default.nix with an unstable nixpkgs?
<clever>
or build it with -I nixpkgs=/path/to/nixos-unstable
<sphalerite>
Cool, thanks!
<clever>
with -I nixpkgs will cover both lines 5 and 6 at once, so you cant miss one of them
ixxie has joined #nixos
<sphalerite>
huh, I tried doing mkdir /tmp/nixpkgs-unstable && cd /tmp/nixpkgs-unstable && GIT_DIR=~/nixpkgs/.git GIT_WORK_TREE=. git checkout nixos-unstable, and there's no default.nix..?
<clever>
what files did it make?
<sphalerite>
Also, is there any work on making the download output prettier? I do seem to remember that it was on someone's todo list somewhere, but is it actualy happening or is that something I could take a look at?
<clever>
sphalerite: ah, i think i know what happened, 'git checkout' knows what the state of files is (via .git state), and only checked out files that differ between the current state, and nixos-unstable
<clever>
so now the state i .git os "corrupt", and your lacking the files that matched up
MichaelRaskin has quit [(Quit: MichaelRaskin)]
<sphalerite>
aaah right
<clever>
i would just do "git checkout" in the nixpkgs dir as normal, and then switch it back to master when your done
systemfault has joined #nixos
ebzzry has joined #nixos
tvon has quit [(Ping timeout: 258 seconds)]
derjohn_mobi has quit [(Ping timeout: 248 seconds)]
justbeingglad has joined #nixos
justbeingglad has left #nixos []
deepfire has quit [(Ping timeout: 246 seconds)]
alx741 has joined #nixos
alx741_ has quit [(Ping timeout: 256 seconds)]
indi_ has joined #nixos
deepfire has joined #nixos
kampfschlaefer has joined #nixos
mpraglowski has joined #nixos
thc202 has joined #nixos
shanemikel has quit [(Quit: My MacBook Air has gone to sleep. ZZZzzz…)]
<jluttine>
has anyone used kdeconnect (in kde) successfully in nixos?
<jluttine>
i can't find my device, not sure where the problem is.. router, cellphone, laptop, nixos, kde, ...
<viric>
LnL: it iterates lots of symlinks in the fs
<viric>
LnL: I see about 15 readlink/s sometimes, and 200readlink/s other times
<viric>
But it's unrelated to ram cache, I understand, because I have seen both situations right after boot
<LnL>
do you have an l2 cache?
<viric>
it's a usual zpool create, zfs create
<viric>
single fs
<viric>
(that means "I don't know")
indi_ has quit [(Remote host closed the connection)]
<LnL>
no then, and also just a single disk?
<nekroze>
viric: is it a ssd?
<viric>
no
<viric>
single disk. no ssd
indi_ has joined #nixos
<nekroze>
zfs will run the l2arc on the same disk when you give it no other
<nekroze>
you might try playing with the options to change syncing to passthrough... I can't remember the option name right now but there is something there.
<nekroze>
also you should make sure your ashift matches the size of your disk sectors. zfs is pretty good at guessing but some drives lie
<nekroze>
unfortunately ashift can only be set on zpool creation
<viric>
ashift=12 I added on creation, yes
<nekroze>
so 4k sectors
<viric>
what is that lssomething command that gives the block sizes?
<nekroze>
fdisk is a decent representation
<viric>
lsblk -t
<nekroze>
hdparm I think is the truth
<viric>
not lsblk -t? :)
<nekroze>
oh yeah that exists too. you learn something new every blah blah blah
<nekroze>
:)
<viric>
:)
<viric>
my usb disk says WSAME=32M, and my local disk says WSAME=0B
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
m` has joined #nixos
eacameron has quit [(Ping timeout: 260 seconds)]
berce has joined #nixos
Shou has joined #nixos
<m`>
Hi! I was trying to deploy a simple config to a virtual machine via nixops. works fine. I'm having trouble to add a custom nix-expression to it. Here's my config: http://codepad.org/QF9l93Mw
<m`>
Would be super nice, if someone could chekc this and tell me how to properly pass the necessary arguments to 'my-hello'
eacameron has joined #nixos
<rly>
peti: Hydra seems rather unstable. Is anyone actually fixing all the problems encountered?
<rly>
peti: i.e., I would expect everything to run under cgroups such that this stuff wouldn't happen.
<viric>
rly: I'm worried on git-annex times. Similar approach to nix, yes
kampfschlaefer has quit [(Ping timeout: 258 seconds)]
<rly>
viric: in the case of git-annex does N depend on the time one has been using the tool?
ixxie_ has quit [(Ping timeout: 245 seconds)]
<rly>
viric: in the case of Nix N is sort of limited to system complexity.
<viric>
rly: by N you mean what?
<rly>
viric: the number of symbolic links created in bulk.
<viric>
rly: it's up to the number of files you have in the annex
<rly>
viric: I don't actually use annex, but what do you typically store there?
<viric>
rly: recordings (sound and video) and scanned books
eacameron has quit [(Ping timeout: 264 seconds)]
shanemikel has quit [(Quit: My MacBook Air has gone to sleep. ZZZzzz…)]
<ixxie>
consoleKeyMap determines the keyboard layout right?
ebzzry has quit [(Ping timeout: 260 seconds)]
<manveru>
ixxie: for the ttys at least
<manveru>
not for X
<niksnut>
peti: well, there hasn't been a commit in a day
<niksnut>
the latest evaluation appears to be current
ertes has quit [(Ping timeout: 258 seconds)]
ertesx is now known as ertes
<NixOS_GitHub>
[nixpkgs] viric pushed 3 new commits to release-16.09: https://git.io/vMLtw
<NixOS_GitHub>
nixpkgs/release-16.09 bd9cd79 Franz Pletz: nixos: compress squashfs with xz...
<NixOS_GitHub>
nixpkgs/release-16.09 64eaa7e Lluís Batlle i Rossell: Tune mksquashfs for smaller isos....
<NixOS_GitHub>
nixpkgs/release-16.09 0ae95d8 Lluís Batlle i Rossell: Make the minimal iso not use profile/minimal, +vim...
<niksnut>
peti: I think the problem here is that when hydra-evaluator checks a jobset and there are no changes, it doesn't clear the previous error message
xadi has quit [(Ping timeout: 265 seconds)]
<manveru>
"14474.86 MiB freed by hard-linking 1474031 files" ...
<manveru>
really wish it'd do that when installing stuff already
<manveru>
gchristensen: tested locally, but may you can try it too?
<manveru>
basically when you enter stuff in the search box, it'll set the document hash too, and fill in the search field from the document hash when you first open the page
ryantrinkle has quit [(Ping timeout: 258 seconds)]
eacameron has joined #nixos
indi_ has joined #nixos
alx741_ has joined #nixos
alx741 has quit [(Ping timeout: 246 seconds)]
eacameron has quit [(Ping timeout: 268 seconds)]
<zarel>
I can't understand what's the status of gnome-screensaver. Is it still used in gnome3?
mizu_no_oto has joined #nixos
Taneb is now known as nvd
nvd is now known as Taneb
mudri has quit [(Ping timeout: 260 seconds)]
mpraglowski has quit [(Ping timeout: 248 seconds)]
danza has joined #nixos
danza has quit [(Quit: Leaving)]
<alphor>
btw before rebuilding the config I checked, networking.firewall.allowedTCPPorts isn't auto added nor does it show up on nixos-option. rebuilding now.
roberth has joined #nixos
<bennofs>
note that you need to run nixos-option as root if any file belonging to your configuration is only readable as root
latk has quit [(Ping timeout: 246 seconds)]
alx741_ has quit [(Quit: alx741_)]
<alphor>
oops, nvm. there is a port defined by sshd.nix in one of the profiles, and now I have a duplicitous 22 entry
wangoe has joined #nixos
simendsjo has joined #nixos
mudri has joined #nixos
eacameron has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
eacameron has quit [(Remote host closed the connection)]
ryantrinkle has joined #nixos
mguentner2 has joined #nixos
mpraglowski has joined #nixos
alx741 has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vMLV3
<NixOS_GitHub>
nixpkgs/master 3dbf6b4 Vladimír Čunát: font-manager: fix missing icons...
derjohn_mob has quit [(Ping timeout: 258 seconds)]
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to release-16.09: https://git.io/vMLVB
<NixOS_GitHub>
nixpkgs/release-16.09 4817e84 Vladimír Čunát: font-manager: fix missing icons...
ixxie has quit [(Ping timeout: 246 seconds)]
jgertm_ has quit [(Ping timeout: 252 seconds)]
dgn has quit [(Ping timeout: 258 seconds)]
msd has joined #nixos
<msd>
LnL: got wxwidgets 2.9 compiling and working on Darwin without downloading any extraneous packages. Pushing now.
<joncfoo>
is there a wiki entry or instructions somplace that show how to get icons & gtk themes working correctly with NixOS?
<joncfoo>
e.g. whether to specify themes & icons in system packages or user; switching themes/icons; etc
<LnL>
msd: great! I'll look at it in a bit
<NixOS_GitHub>
[nixpkgs] mdaiter opened pull request #21527: wxGTK29: fixed Darwin build (master...wxGTK29Darwin) https://git.io/vMLwz
<LnL>
you generally don't install dependencies like that in a user profile
<LnL>
how are you trying to use it?
<phorse>
I'm just gotten a book on idris (programming language), and when I try to compile the sample program in the first chapter gcc complains about not finding -lgmp
<phorse>
so I did the naive thing and installed gmp
<phorse>
simpson: hmm, I don't think I need to package anything, though. I'm just trying to write/compile/run some programs on my own
<gchristensen>
it is amazing that the first 11 pages of pull requests are from the last year, and the last 2 pages are for a few years.
msd has joined #nixos
<gchristensen>
7 pages since October even.
msd has quit [(Remote host closed the connection)]
Shou has joined #nixos
<LnL>
prs or issues?
<gchristensen>
PRs
<LnL>
ah, yes indeed
<LnL>
we're over 320 again
<gchristensen>
"again" not sure when we were over 320 before :P
<LnL>
I think at some point, but just for a short time
<LnL>
what, we can change the target branch for prs?
<gchristensen>
yeah
<LnL>
is that new?
<gchristensen>
a few months old by now, I think
Shou has quit [(Ping timeout: 265 seconds)]
msd has joined #nixos
<LnL>
I was wondering before if we should add a status:staging label or something because I thought we couldn't change it
<msd>
LnL: the Mac output appears as a failure because the file output (build log) is too long to display. The package should build easily on a Darwin system.
<gchristensen>
imagine if all ~90 people triaged a PR today :)
ryanartecona has joined #nixos
<LnL>
I have 3 machines building nix stuff at the moment :p
<LnL>
msd: wxGTK works, poedit just depends on it so nox tries to rebuild that
AllanEsp1nosa has joined #nixos
<msd>
LnL: I believe the poedit error is unrelated to the wxGTK build. The poedit rebuild tries to already pull wxGTK29 and is set to build for UNIX systems. Currently, the Mac package doesn't even have the wxGTK29 package; therefore, the error must have existed prior to wxGTK29's addition to the Mac ecosystem. This seems to just be a by-product of rebuilding an already broken package.
Itkovian has joined #nixos
<LnL>
yes it's also broken at the moment, but it looks like the only thing that depends on it on darwin
<MichaelRaskin>
gchristensen: I am not sure there _are_ 90 truly untriaged PRs
<LnL>
msd: are you trying to get rid of wxGTK28? there are only 2 packages that depend on 2.9
<gchristensen>
MichaelRaskin: I think there are probably ~90 prs with version bumps that could be merged today
<MichaelRaskin>
Most of the PRs have some substantial discussion, and just adding a random person there wouldn't immediately help
<msd>
LnL: not in particular. Just wanted to add a version of wxGTK to the Mac ecosystem, so that we could build more potential packages on the system.
<LnL>
I'm pretty sure there are at least a couple of new packages/updates that are buried under older stuff
<LnL>
allright, just wondering
<msd>
LnL: although 2.8 is a pain in the ass to get working on modern versions of the Darwin kernel/Mac framework -- all of the builds I've come across have suggested directly downloading the 10.6 framework and directly building against that
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 2 new commits to master: https://git.io/vMLy6
<NixOS_GitHub>
nixpkgs/master 10836bd Matthew Daiter: wxGTK29: fixed Darwin build
<MichaelRaskin>
Some of these are WIPs with some meaningful question asked by submitter that should be answered before merging, and I am not sure how to answer it…
<gchristensen>
maybe a no-activity
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 5 new commits to master: https://git.io/vMLSI
<NixOS_GitHub>
nixpkgs/master e2bb0c5 Florent Becker: js_of_ocaml: propagate dependency on ppx_deriving...
<NixOS_GitHub>
nixpkgs/master 51e0173 Florent Becker: ocamlPackages.ocaml_pcre: change name from ocaml-ocaml-pcre to ocaml-pcre...
<MichaelRaskin>
I refuse to have an opinion about that
<gchristensen>
ok, why?
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<gchristensen>
(surely you have a reason, which I'm very interested in, even if you'd like to share in private)
<MichaelRaskin>
Well, I am sometimes guilty of closing PRs without comments after hand-merging them (and I got justly criticised for that).
<gchristensen>
like applying the patch yourself and not saying so in the PR?
<MichaelRaskin>
Cherry-picking with manual conflict resolution an mentioning the submitter in the commit message, but not writing a reasonable PR-closure message, yes
mpraglowski has joined #nixos
ryanartecona has quit [(Quit: ryanartecona)]
<gchristensen>
aye
<MichaelRaskin>
Also, as too often I do not support the policy decisions even after they have been made, I try not to have an opinion if I don't have arguments that I am willing to use for persuasion
dtzWill has joined #nixos
<MichaelRaskin>
I mean, it is reasonable to have strong positions and to try to reconcile them using explanations; it is reasonable to vote on the bikeshed colour and commit to support the result; as I am too often against the result in the first case, I cannot credibly (even for myself) commit to honest support of the outcome in case I don't have any arguments.
<MichaelRaskin>
And then it makes little sense to even form an opinion in cases where I agree that I have no clue.
<LnL>
urhg, I just rebuilt the stdenv without a change I wanted
<flxw>
Hi all. I'm looking for a light-weight way of doing simple edits in /etc (for packages that don't provide options to set). Concrete example: /etc/inputrc. I'm fine with the default bash ships, just one line "a b c" I'd like to replace by "a b d". In the docs I found overrideDerivation and substituteInplace. Is there a easier way of achiving what I want, or is this route about right?
<gchristensen>
gotcha, MichaelRaskin
latk has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<MichaelRaskin>
I have one position: closing PRs by people with commit access can be taken strictly lighter than in case of non-committers, and first three PRs of a newcomer merit the most care.
<MichaelRaskin>
Note that this is only about greater-than contraints: no position on absolute values of the optimal policies.
msd has quit [(Remote host closed the connection)]
msd has joined #nixos
msd has quit [(Remote host closed the connection)]
<gchristensen>
that is an interesting idea
<MichaelRaskin>
I mean, if I make a PR — that happens — that is explicitly because I want to have a place for a reasoned rejection. If someone closes after a couple of weeks with no feedback, I can assume nobody cares and commit myself. An contributor of 10 merged PRs can assume that her contributions are taken seriously, and therefore if some PR is treated in a wrong way, resubmitting with a better explanation is a viable idea. The new
mguentner2 is now known as mguentner
bennofs has quit [(Ping timeout: 245 seconds)]
msd has joined #nixos
<LnL>
why the hell is dd 10x faster on my nixos box the it is on darwin
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vML7f
<NixOS_GitHub>
nixpkgs/master 20c2b45 Vladimír Čunát: Merge branch 'staging'
pstn has quit [(Ping timeout: 246 seconds)]
lhart has joined #nixos
<lhart>
hi! I'm kind of new to nixos and right now, I'm trying for the first time to build a package from source. I'm starting to get it but I have one problem during nuix-build.
<lhart>
I have an output looking like this :
<lhart>
stdenv/setup: line 695: ./configure.ac: Permission denied
roberth has quit [(Ping timeout: 260 seconds)]
<MichaelRaskin>
What are you trying to do?
<MichaelRaskin>
It sounds like you try to execute configure.ac instead of generating configure using autoconf (and maybe automake/libtool/aclocal)
<lhart>
Hi! thanks.
<lhart>
hum
<lhart>
the configure file is already loaded with the source thats why.
<lhart>
but maybe it's not needed?
<MichaelRaskin>
Then why is configure.ac executed?
<Judson>
gnupg doesn't seem to include gpgkey2ssh?
<MichaelRaskin>
No-no, if there is configure, you want to run it. But you do not want to run configure.ac
mpraglowski has quit [(Ping timeout: 252 seconds)]
hiratara has quit [(Ping timeout: 258 seconds)]
<lhart>
Hum so configure.ac isn't the same thing as configure?
<lhart>
maybe a very stuid question
Fare has joined #nixos
<lhart>
*stuîd
<MichaelRaskin>
No. configure.ac is Autoconf template for creating configure
<lhart>
*stuîd
<lhart>
oh ok
<MichaelRaskin>
(and that part is not really Nix-specific)
<lhart>
so i don't even need to give any configure file? nix will take care of it itself, is that right?
hiratara has joined #nixos
<MichaelRaskin>
I think you need to add autoreconfHook to buildInputs
<lhart>
using configure.ac as an argument?
<MichaelRaskin>
No, autoreconf assumes that configure.ac is the template it needs
<MichaelRaskin>
Or you could add autoconf automake libtool to buildInputs
<MichaelRaskin>
And manually write pre
<MichaelRaskin>
preConfigure
<lhart>
ok i get it. thank you! I'll try to debug the rest myself... haha
<lhart>
Have a nive evening.
<NixOS_GitHub>
[nixpkgs] 7c6f434c closed pull request #20767: releaseTools.debBuild: split checkinstall into separate phase (master...debbuild) https://git.io/v1IFc
<Judson>
Does anyone else have the problem that dmenu doesn't track application updates?
<Judson>
e.g. updating Chromium lets me launch the right version from a terminal, but dmenu still launches the old version
lhart has quit [(Quit: Page closed)]
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vML5V
<NixOS_GitHub>
nixpkgs/master b0bfa88 Tom Hinton: Fiddle LD_LIBRARY_PATH to include GDAL...
<NixOS_GitHub>
nixpkgs/master 6003eeb Michael Raskin: Merge pull request #21207 from larkery/patch-1...
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vML5y
<NixOS_GitHub>
nixpkgs/master 2990cd8 Robert Helgesson: clblas-cuda: use system gtest library...
<NixOS_GitHub>
nixpkgs/master 9e1ec53 Michael Raskin: Merge pull request #21456 from rycee/fix/clblas-cuda...
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vML55
<NixOS_GitHub>
nixpkgs/master e243ec4 Jeremy Shaw: netbpm: tell config.mk where to find JPEGLIB and JPEGHDR_DIR or jpegtopnm and friends will not be built.
<NixOS_GitHub>
nixpkgs/master 3b0720d Michael Raskin: Merge pull request #21473 from stepcut/netpbm-fix-libjpeg-paths...
FareTower has joined #nixos
Fare has quit [(Read error: Connection reset by peer)]
fare__ has joined #nixos
katyucha has quit [(Read error: No route to host)]
<NixOS_GitHub>
nixpkgs/master 62adccb Michael Raskin: Merge pull request #21483 from rardiol/freeorion...
<sphalerite>
Judson: does ~/.dmenu_cache exist? If so, does removing it fix it?
<sphalerite>
(I assume you mean dmenu_run)
fare__ has quit [(Ping timeout: 258 seconds)]
<NixOS_GitHub>
[nixpkgs] abbradar opened pull request #21531: releaseTools.debBuild: split checkinstall into separate phase (master...debbuild) https://git.io/vMLd2
<sphalerite>
clever: was https://github.com/NixOS/nix/issues/896 the issue you meant when I asked earlier about prettier output when downloading/building derivations?
Matrixiumn is now known as Xackle
stepcut has quit [(Remote host closed the connection)]
stepcut has joined #nixos
<MichaelRaskin>
gchristensen: and now could you please list the remaining 70 easy-to-process PRs you meantioned earlier?
<MichaelRaskin>
I am slightly at loss about Ruby/Node package addition, because they always include and update to a huge generated pile-of-text…
<MichaelRaskin>
You couldn't help perceiving it as a comment?
<msd>
LnL: Erlang actually requires wxGTK 3 for wx support (and I think observer:start() support, but I'm not certain). Gonna go work on it and make some pushes
msd has quit [(Remote host closed the connection)]
<LnL>
doesn't that use wxmac?
<LnL>
oh
<rly>
gchristensen: version bumps should not require human interaction.
<rly>
gchristensen: if NixOS fails, it will be, because of a lack of automation.
<rly>
gchristensen: the only thing for which there should be human PR interaction is the automated scripts which generate the .nix expressions.
<rly>
gchristensen: ask yourself this question: If NixOS was a company, would it work in this manual way?
<shanemikel>
rly it depends on the package. it looks like a lot of the haskell infrastructure is automatic (with regards to the version bumps), but for that to work you need essentially a CI system (continuous integration) which is something you can use the Hydra farms for... etc
<LnL>
rly: that would work if software made sense
<rly>
shanemikel: the thing is, I can't use Hydra.
fare__ has joined #nixos
<rly>
shanemikel: and many other people also can't.
<rly>
shanemikel: a fundamental piece of infrastructure is missing, which causes inefficiency in the way the whole community works.
<rly>
LnL: software makes sense to me.
<gchristensen>
yeah, software doesn't make sense / work that way
<MichaelRaskin>
Well, among my packages where the updates are automated, trying to verify those updates often yields some interesting results
<shanemikel>
well, the problem with the automatic iteration is you need some powerful system to iterate and try to build a HUGE variety of version combos
<shanemikel>
so the system can create comprehensive reports of which combinations build/don't build
<rly>
shanemikel: I built Emacs for years by automatically building the last 30 git versions.
<rly>
shanemikel: it *always* worked.
<rly>
For our company infrastructure, I will do exactly the same, btw.
<gchristensen>
rly: if it is really truly that easy, please apply that process to as many packages you can manage.
<rly>
I don't have the time nor the interest to care about specific versions.
<shanemikel>
Oh... well that's fine if it's an Application (isn't a library depended on by other packages)
<gchristensen>
and yes: companies typically do operate this way
<shanemikel>
but that's because they have infrastructure (CI systems) to do the iteration and run tests
<LnL>
well that's because a company generally has control over the parts they update
acertain has quit [(Ping timeout: 244 seconds)]
<rly>
gchristensen: and spam github with PRs from automated systems?
<shanemikel>
it's also much more difficult to do that with software written in "dynamic" languages... especially with FOSS where there isn't a common discipline for testing and code coverage (or a bureau to enforce it)
<shanemikel>
*bureaucracy*
<rly>
gchristensen: as I explained, I think NixOS will fail in the end if we don't do this.
<rly>
gchristensen: the promise of NixOS is increased automation with less people.
<rly>
gchristensen: that means doing things in a smarter way.
<rly>
The basic Nix system is smarter.
<gchristensen>
I disagree about it being the failure nor the promise
<rly>
But all the non-creative work should be eliminated.
<rly>
Even hardcoding a URL is a mistake.
<gchristensen>
I just don't agree with the basic premise here
<rly>
Just use a search engine to find where libfoo is.
<shanemikel>
rly what's hardcoding supposed to mean in a language you don't have to compile?
<rly>
shanemikel: uhm, hardcoding has nothing to do with the evaluation model.
<shanemikel>
I think it has everything to do with it...
systemfault has joined #nixos
<gchristensen>
anyway, sure, spam the tracker with PRs. good ones will get merged, bad ones won't, if your system works it'll be adopted in short order I'm sure
msd has joined #nixos
<gchristensen>
nobody who is "part of" nixos needs to buy in to your plan in order for you to do it
<rly>
gchristensen: that's not really true, because I would have to setup infrastructure for it and support it forever.
<gchristensen>
no, you would only have to support it until it is proven good and adopted as standard infra
<rly>
gchristensen: I can setup the infrastructure, but I am not sure whether I want to support it forever.
msd has quit [(Client Quit)]
<LnL>
I think getting hydra to build prs is more important at this point
<rly>
gchristensen: is the current infrastructure even non-proprietary?
<gchristensen>
rly: what infra?
<rly>
gchristensen: the full Hydra configuration including DNS and network configuration, etc.
<rly>
gchristensen: or is it just manually setup?
<gchristensen>
LnL: sure, but rly can go do this without any sort of buy-in.
<LnL>
rly: all the configuration for hydra is on github
<rly>
I do have a usecase for some packages.
<rly>
And asking for upgrades of packages becomes old *really* quick.
<rly>
Although, I also have considered writing a script which just asks for an upgrade automatically ;)
<rly>
Haha
<gchristensen>
why are you asking for upgrades instead of submitting upgrade patches?
<rly>
gchristensen: it was less work.
<gchristensen>
well
<rly>
gchristensen: and like I said, writing upgrade patches by hand is something I refuse to do.
<MichaelRaskin>
When the upgrade is a simple version-bump, checking for a new version is most of the work…
<gchristensen>
so automate it
<rly>
gchristensen: yes, I already did that for some internal packages, which we might open-source.
<gchristensen>
well then we might use it
<rly>
gchristensen: it also took some time for me to mature in Nix development.
<gchristensen>
but we don't owe you anything, and you don't owe us anything either. if you want Nix to be better, that frequently means doing it yourself
<NixOS_GitHub>
[nixpkgs] volth opened pull request #21535: miredo: init at 1.2.6 (master...miredo-init-1.2.6) https://git.io/vMLNc
namor has joined #nixos
rly has quit [(Quit: leaving)]
<shanemikel>
well, if you take a look at how well NixOS is doing OOB usability wise compared to the huge monstrosity of a bureaucracy that is, say, Debian.. and how much power you have compared to a simple and unstable rolling release like Arch (comparably sized project)
<shanemikel>
I'd say NixOS is hardly a failure
<shanemikel>
gchristensen: that guy is an ass
<shanemikel>
asked for my engineering credentials :)
<gchristensen>
??
namor has quit [(Ping timeout: 258 seconds)]
ryantrinkle has quit [(Ping timeout: 245 seconds)]
<shanemikel>
do you feel like nixpkgs and nixos are meant to be huge "successful" distributions.. I always got the impression that they were sort of experimental (usable, sure!) or reference implementations. I always expected to see nixpkgs incarnations tracking various distros (like an ubuntu nixpkgs, redhat nixpkgs, etc) for developers and packagers to easily target a variety of traditional distros from a single box
<shanemikel>
perhaps wishful thinking, though
proteusguy has quit [(Ping timeout: 258 seconds)]
msd has joined #nixos
<simpson>
shanemikel: Why can't we just kill the old traditional distros?
<shanemikel>
I don't think NixOS is perfect for everything.. the disadvantage of the oldies, is exactly the source of their strength.. that is, they enforce some kind of organization and version convergence in fossland.. I think it's essential to get that *somewhere*, and nix is too flexible for that
<simpson>
shanemikel: AFAICT Nix is the only package manager that actually internalizes and enforces POLA. It's valuable on that alone; cap-safe semantics are good.
<shanemikel>
it's clear the old distros are gonna lose out to these systems on servers and in larger organizations, but I think they're still pretty nice for single user desktops.. and super small IT
<simpson>
What's nice about them? FHS?
<shanemikel>
well, yes..
danharaj has quit [(Ping timeout: 246 seconds)]
<simpson>
Then for those few packages that need it, nixpkgs has a FHS wrapper.
<shanemikel>
in a kind of perverse way, that's what I'm saying
<simpson>
The only thing "nice" about RHEL or Debian is that many of us already know how to manage those systems. That doesn't necessarily mean that they're Doing It Right.
<NixOS_GitHub>
[nixpkgs] volth opened pull request #21536: virt-top: init at 1.0.8 (master...virt-top-1.0.8) https://git.io/vMLAp
<shanemikel>
no, certainly not. But It's nice to have an authority.. and somehow I see that necessity as a direct consequence of technical inferiority
<dash>
(it's not a consequence of technical anything, it's a consequence of multiple people/groups with different goals/desires)
<shanemikel>
I'd love to be wrong about that, believe me
<simpson>
(Conway's Law strikes again!)
msd has quit [(Remote host closed the connection)]
alx741_ has quit [(Client Quit)]
alx741 has quit [(Ping timeout: 248 seconds)]
<dash>
simpson: well, this is even less subtle than conways law
<dash>
simpson: since the entire purpose of an os distribution is to harmonize disparate software packages :)
<simpson>
shanemikel: Anyway, there's no central authority for Debian or RHEL packages. Just like with NixOS, there's one large open-source group that produces the lion's share of packages, and then a bunch of private and corporate players that produce small handfuls of proprietary packages.
<shanemikel>
the authority is the product, not the organization in this case
<simpson>
Oh, the value is in "stable Debian"?
<shanemikel>
not any in particular.. personally, I've never been able to use stalbe
<simpson>
I guess I don't grok your point, sorry.
<dash>
i think it's a little much to call unstable debian a "product"