Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<Havvy>
orivej: Right, but I couldn't find an equivalent page on the new wiki.
ssmike has quit [(Remote host closed the connection)]
<orivej>
Havvy: the old wiki was not migrated and would have been lost (for the public) if not for web archive; but I don't know how it became discontinued, except I seem to remember that many articles were perceived obsolete and they were not going to be updated
noam_ has joined #nixos
noam__ has quit [(Read error: Connection reset by peer)]
musteresel has joined #nixos
<musteresel>
Hi. I'm getting ValueError('ZIP does not support timestamps before 1980') in a nix-build .. any ideas how to tackle it? Setting SOURCE_DATE_EPOCH to a recent enough value or unsetting it does not have any effect. I'm building a uzbl.overrideAttrs with only the src changed (to a local git checkout).
digitus has quit [(Quit: digitus)]
crucialrhyme has joined #nixos
<jeaye>
Just a guess, but `export TZ=UTC` in your builder might help.
erasmas has quit [(Quit: leaving)]
jb55 has quit [(Ping timeout: 240 seconds)]
<bgamari>
is there any way to avoid downloading artifacts twice when writing a new expression with stdenv.fetchurl?
<bgamari>
e.g. once manually to hash it, then again with nix to insert it into the store
<bgamari>
I have tried downloading it manually then using `nix add-to-store` to insert it into the store
<bgamari>
but nix tries to download it again regardless
<clever>
bgamari: yeah, there is a special mode, let me find it
<bgamari>
nix-prefetch-url, perhaps?
<musteresel>
jeaye: Didn't help, unfortunately. What confuses me most is that running nix-build --check for the original uzbl package from nixpkgs goes just fine. All I'm doing is changing the src attribute ..
<clever>
bgamari: something better/more powerful
<clever>
prefetch-url only works against urls
<clever>
bgamari: nix-build -A hello.src --hash
<crucialrhyme>
is there like a nixos-beginners channel or is this the place to ask dumb beginner questions?
<bgamari>
hmm, I guess this isn't yet supported in nixUnstable
<clever>
bgamari: that will run the fixed-output derivation for fetching the source, then hash its result, and rename the result in /nix/store/ to match the actual hash, so the data isnt lost
<bgamari>
sounds perfect though
<clever>
bgamari: there is also nix-prefetch-url '<nixpkgs>' -A hello.src, which just evaluates the .src.url and then runs normal prefetch-url on it
<clever>
(but is limited to http like things)
yenzenz has quit [(Ping timeout: 240 seconds)]
<bgamari>
right
<bgamari>
Is it expected that installing nix-env -i nixUnstable breaks nix-build?
<Judson>
Is there a way to easily add postgres support to the current Nix liquibase?
<Judson>
profpatsch?
<Profpatsch>
Judson: yes?
<Judson>
^liquibase
<Profpatsch>
Uh, I might have played with liquibase a while back?
<Profpatsch>
But can’t really remember.
<Judson>
Okay. The existing expression can add mysql support?
<Profpatsch>
There doesn’t seem to be a module.
<Profpatsch>
Oh, blame shows me alright. :)
<Judson>
Even easier: you're a maintainer on the expression :)
<orivej>
bgamari: do you have old nixStable? the current one interoperates with nixUnstable on the same db
<Judson>
50% of this is "can I somehow override the pkgs expression if I can find the postgres driver?"
<Profpatsch>
Judson: Should be easy enough.
<bgamari>
orivej, perhaps?
<Profpatsch>
Add the postgresql jdbc package to the extraJars list.
jb55 has joined #nixos
<orivej>
crucialrhyme: this is the place
<Profpatsch>
Judson: But you have to check if the postgresql_jbdc package is structured the same as the mysql_jdbc package.
<Judson>
I mean, that'd fix it if it were in nixpkgs that way.
<crucialrhyme>
orivej: wonderful, expect some very dumb questions in the near future
<Judson>
And I can start a PR.
<Judson>
For my purposes, I'm wondering if there's a liquibase.override or something?
<orivej>
bgamari: you should be able to update nixStable using the nixUnstable without dumping the db as the cookbook says
<Judson>
That only overrides set fields, though, I kind of understand?
<Profpatsch>
Judson: Seems to be the case.
<Profpatsch>
Judson: I’d do it via PR.
<Judson>
kk
<Judson>
Expect that shortly then :)
<bgamari>
orivej, at this point I seem to be in quite a pickle
<Profpatsch>
Just keep the structure of the mysql, should be fine.
<Profpatsch>
I just checked, both packages seem to function the same way, they have a jar in /share/java/
<bgamari>
orivej, nix eval -f ~/nixpkgs nixUnstable inexplicably fails with
<bgamari>
error: anonymous function at /home/ben/nixpkgs/pkgs/build-support/fetchurl/boot.nix:5:1 called with unexpected argument 'postFetch', at /home/ben/nixpkgs/pkgs/build-support/fetchpatch/default.nix:10:1
<Profpatsch>
So it might work without any debugging if you are lucky.
<Profpatsch>
But maybe write some testing code.
<bgamari>
so i don't even know where my nixUnstable is
<Profpatsch>
Judson: Would be a nice example of a test, compile a minimal java exe for each backend and test it against a running server?
<crucialrhyme>
okay, first dumb question. is it dangerous to initialize a git repo directly in /etc/nixos?
<orivej>
crucialrhyme: no, it's fine
<crucialrhyme>
wow!!! i did not expect that answer. fantastic!
Mateon3 has joined #nixos
<crucialrhyme>
when i installed this i expected everything to be weird and difficult
<crucialrhyme>
but it is actually not at all
Mateon1 has quit [(Ping timeout: 264 seconds)]
Mateon3 is now known as Mateon1
<orivej>
:D
<crucialrhyme>
like even up-to-date unsupported drivers for my rgb keyboard were already in nixpkgs
<crucialrhyme>
i'm still wrapping my head around all the nix pills but even though i don't completely understand everything, it's more usable than ubuntu was for me
<Profpatsch>
crucialrhyme: Universe brain, you can also link some file to /etc/nixos/configuarion.nix from your home directory.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] nyarly opened pull request #32688: liquibase-3.5.3 (update) adding PostgreSQL support (master...postgres-for-liquibase) https://git.io/vbwaw
NixOS_GitHub has left #nixos []
jrolfs has quit [(Quit: WeeChat 1.9.1)]
jrolfs has joined #nixos
tristan0x has joined #nixos
<musteresel>
someone else an idea how to tackle python ZIP errors due to timestamps before 1980 during a nix-build?
<Profpatsch>
musteresel: whaaaat
<Profpatsch>
I’d say patch that pile of crap.
<Profpatsch>
The hell out of it.
<gchristensen>
that is not an uncommon problem in our python tooling
<gchristensen>
there is a fix, but I don't remember what
<musteresel>
Profpatsch: I'm trying to build uzbl ... nix-build <nixpkgs> -A uzbl --check works fine.
tristan0x has quit [(Ping timeout: 240 seconds)]
<musteresel>
But nix-build modified-uzbl.nix where that .nix only contains uzbl.overrideAttrs with src changed to a git repo fails ..
<Profpatsch>
musteresel: See where the error comes from and patch the magic number from 1980 to 1960. :)
<musteresel>
the manual hinted to SOURCE_DATE_EPOCH, but it is completely ignored
<musteresel>
Profpatsch: AFAIK that's a limitation of ZIP itself
<crucialrhyme>
Profpatsch: irc doesn't support gifs but i just did what you suggested, and imagine the gif with the head man and the space explosions
<Profpatsch>
Hrm, 1980 sounds like a fix for FAT/NTFS, where the magic date is 1.1.1980 afaik
<Profpatsch>
crucialrhyme: Awesome!
<musteresel>
gchristensen: The manual says to set SOURCE_DATE_EPOCH .. but that doesn't help. I even replaced the call for python with a ./run-python which sets that env var and then execs python .. still no luck, though
yegortimoshenko has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 2 new commits to master: https://git.io/vbwVB
<Profpatsch>
Why are people still using zip anyway?
<musteresel>
hm ... I could patch python's zip implementation itself, though. Changing that raise [1] into an assignment date_time = (1980,1,1,0,0,0) should solve a few python zip related problems, no? Though it alters also non packaging related behaviour ... [1] https://github.com/python/cpython/blob/master/Lib/zipfile.py#L333
Shados has quit [(Remote host closed the connection)]
tristan0x has quit [(Ping timeout: 265 seconds)]
Shados has joined #nixos
Neo-- has quit [(Ping timeout: 264 seconds)]
radvendii has joined #nixos
<radvendii>
So I just deleted a bunch of things from my systemPackages, but I want it back and I don't remember what they were. Is there a way to compare the packages present in this generation with the ones in the previous generation?
<clever>
radvendii: ls -ld /nix/var/nix/profiles/system*
Supersonic112 has quit [(Disconnected by services)]
thc202 has quit [(Ping timeout: 272 seconds)]
spietz has quit [(Ping timeout: 256 seconds)]
<fearlessKim[m]>
can I see the state for all packages of the current nixos-unstable with https://hydra.nixos.org ? lately I seem to have to recompile a lot more packages on nixos-unstable I am not sure why.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] yegortimoshenko opened pull request #32694: SkypeExport: init at 1.4.0 (master...SkypeExport/init) https://git.io/vbwrM
NixOS_GitHub has left #nixos []
bennofs has quit [(Quit: Connection closed for inactivity)]
phreedom has quit [(Remote host closed the connection)]
mahalel_ has quit [(Remote host closed the connection)]
antonv has joined #nixos
danl1240 has quit [(Quit: My iMac has gone to sleep. ZZZzzz…)]
<antonv>
I've installed nixos on my old good 32 bit laptop
danl1240 has joined #nixos
<antonv>
the stable version doesn't have firefox 57, therefore I tried unstable
<antonv>
But in the unstable the firefox doesn't install on 32 bit
<antonv>
I hope the will update the stable to FF 57 (without dropping support for 32bit)
tristan0x has joined #nixos
aarvar has quit [(Ping timeout: 260 seconds)]
markus1189 has joined #nixos
<kini>
I've noticed a pattern in nixpkgs, where if some software has a disjunctive dependency on foo or bar, both foo and bar are set as inputs to the nix expression, and then when it's instantiated in all-packages.nix it looks like "baz = callPackage baz { foo = null; }" when using bar or "baz = callPackage baz { bar = null; }" when using foo
<kini>
does this mean that nix is not "lazy enough" to just pass in both foo and bar and then decide which one to use within the nix expression?
tristan0x has quit [(Ping timeout: 264 seconds)]
<kini>
that would make all-packages.nix look cleaner, like maybe something like ''foo = callPackage baz { fooOrBar = "foo"; };'' which is easier to read at a glance
<yegortimoshenko>
kini: it needs a priority, it's not about laziness. if you pass both, it would try to use both
<yegortimoshenko>
kini: fooOrBar is less semantically transparent
markus1199 has quit [(Ping timeout: 256 seconds)]
<yegortimoshenko>
you either pass inputs, or you don't
<kini>
I guess that's true; without reading the code, you'd notice that it's passing in both foo and bar as inputs and you couldn't tell which one was actually going to be used
<yegortimoshenko>
fooOrBar involves creating an ad-hoc language that has to be parsed to infer the meaning
<kini>
maybe what makes me think it looks "ugly" is the callPackages mechanism, which makes you have to explicitly declare which inputs you're *not* passing, rather than which inputs you *are* passing
<kini>
yeah, fooOrBar is a bad idea, you're right
<yegortimoshenko>
the idea behind callPackage is to use the convention of having top-level attribute names match attrset attribute names
<yegortimoshenko>
if all inputs were to be explicitly passed, it would take a lot of boilerplate
<kini>
true, true
aarvar has joined #nixos
<radvendii>
clever: thanks. I ended up using `dif -bur` on the two folders
Mateon1 has quit [(Remote host closed the connection)]
Mateon1 has joined #nixos
<Nol>
So, I'm on unstable and I'm having an issue with steam where I can't run 64 bit games due to them not finding libGL.so.1 but can run 32-bit games fine
<radvendii>
Nol: I had a similar problem a while ago. This might be solveable using patchelf
etu has quit [(Read error: Connection reset by peer)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] yegortimoshenko opened pull request #32695: clojure: 1.8.0 -> 1.9.0.273, new tools! (master...clojure/new-tool) https://git.io/vbwKv
NixOS_GitHub has left #nixos []
<radvendii>
Nol: it was kind of a nightmare to do though
<Nol>
:(
<radvendii>
Nol: yeah, i might be wrong though, it might not be the same issue. also you might have an easier time with it than I did
ssmike has joined #nixos
<Nol>
what was the root cause?
<yegortimoshenko>
Nol: are you on unstable?
etu has joined #nixos
<yegortimoshenko>
what's the game that you're trying to run
<Nol>
yes, steam is currently completely broken outside unstable
<Nol>
Opus Magnum
<Nol>
(outside of unstable the library paths no longer get set correctly as of a steam update yesterday)
<yegortimoshenko>
i don't have that one, but i have a few 64-bit games (The Talos Principle), i'll check if i can reproduce this
nuncanada has quit [(Ping timeout: 240 seconds)]
<Yaniel>
even on unstable, only the library view of steam works :(
<Yaniel>
complains something about an outdated libnss in the console
musteresel has quit [(Ping timeout: 272 seconds)]
<Nol>
I also have no inkling of how I would go about debugging this since I'm a new NixOS user as for two days ago :P
<yegortimoshenko>
Yaniel: i'm not sure how i've missed this. true. could you open an issue?
ssmike has quit [(Ping timeout: 240 seconds)]
<Nol>
yegortimoshenko, as an aside if you play games and run NixOS you'll probably enjoy Opus Magnum ;)
<yegortimoshenko>
Nol: i don't play puzzles! hehe
<yegortimoshenko>
or is it about NixOS being a puzzle...
<yegortimoshenko>
Yaniel: i'd guess it's just a missing package in steam chrootenv
<Nol>
haha. To be honest I don't like that it calls itself a puzzle game since the challenges have many, many solutions that vary in optimality
<Nol>
"puzzle" to me suggests a single solution
<yegortimoshenko>
oh, it's a strategy then
johnsonav has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
johnsonav has joined #nixos
<Nol>
A very programming-influenced one, like almost all Zachtronics games
<Nol>
Anyway, since I'm learning NixOS, how would I begin debugging this? The natural place to start seems to be to figure out what paths it's trying to invoke the game with and examine the chroot it creates, but I don't know how to do this
<radvendii>
Nol: Just taking a look at opus magnum now. I wish it had an interface where you could legit program things instead of drag-dropping :(
<Yaniel>
try Shenzhen I/O or Else Heart.Break()
<Nol>
radvendii, other games by Zachtronics do work that way - look at TIS-100 or SHENZEN I/O
<yegortimoshenko>
i also just finished downloading 64-bit Talos Principle and it's not specific to 64-bit games
<yegortimoshenko>
what i mean is, i can't reproduce it with 64-bit version of Talos Principle
<Nol>
hmmmm, admittently I assumed it was a 64-bit issue because the only distinction I could see between Opus Magnum and the other games I tried was 64 vs 32 bit
hellrazo1 has quit [(Ping timeout: 255 seconds)]
<yegortimoshenko>
so it might be some problem with the executable they ship
lambdamu_ has joined #nixos
<yegortimoshenko>
in my experience with debugging games on NixOS, problems tend to be googleable
phdoerfler has quit [(Quit: Leaving.)]
<Nol>
I did quite a bit of that. It might be that it was yesterday's steam update that broke it tbh
<Nol>
would explain why I couldn't find anyone else with this issue
<yegortimoshenko>
hm...
<yegortimoshenko>
i might have a theory
<yegortimoshenko>
do they have a wrapper script that rewrites LD_LIBRARY_PATH?
<yegortimoshenko>
(in ~/.local/share/Steam/steamapps/common/Opus Magnum)
drakonis has quit [(Read error: Connection reset by peer)]
<Nol>
# Restore the original library search path that was replaced by the Steam runtime.
<Nol>
yeah I can see why this would cause a problem on nix
<yegortimoshenko>
also i believe that all games by this developer have the same issue, when you mentioned SHENZEN I/O i remember some issue opened a while ago about that game and that it has the same problem
<crucialrhyme>
second dumb question, if i switch to the unstable channel right now (for my whole OS) and everything breaks, will I be able to roll back?
<yegortimoshenko>
crucialrhyme: yes
<crucialrhyme>
lol this is sick as hell. when do i have to watch out for the database/nix store format changing? does that happen often?
<Nol>
sorry about the 64-bit red herring
<yegortimoshenko>
Nol: it's ok! i should have identified the issue right away, having worked on it quite a while... overriding LD_LIBRARY_PATH typically results in libGL not being found
<yegortimoshenko>
crucialrhyme: if you're not running a server, you should not worry too much, in my opinion
nckx has quit [(Quit: Updating my GuixSD server — gnu.org/s/guix)]
justan0theruser has joined #nixos
justan0theruser has quit [(Client Quit)]
justan0theruser has joined #nixos
nckx has joined #nixos
<crucialrhyme>
yegortimoshenko: well, i'm gonna go for it, probably my system will crash in a few minutes
justanotheruser has quit [(Ping timeout: 272 seconds)]
boomshroom has joined #nixos
<boomshroom>
Hello!
<yegortimoshenko>
crucialrhyme: my family's laptops run on unstable, my laptop runs from master atm, don't worry
<boomshroom>
A while back, I was able to get renpy to work despite being marked as broken (it was updated relatively recently without the flag being removed), but now it doesn'
<boomshroom>
now it doesn't seem to want to build.
<boomshroom>
I get 2 errors from gen/pygame_sdl2.mixer.c about MIX_INIT_MODPLUG and MIX_INIT_FLUIDSYNTH being undeclared.
<yegortimoshenko>
boomshroom: can reproduce
<yegortimoshenko>
well, it's marked broken for a reason, typically broken packages don't build
<boomshroom>
yegortimoshenko: nice thing about nix is that `env NIXPKGS_ALLOW_BROKEN=1 nix-shell -p renpy` should give the same result for the same version. It worked the last time I tried, and it was still marked broken then.
crucialrhyme has quit [(Ping timeout: 264 seconds)]
orivej has quit [(Ping timeout: 264 seconds)]
<yegortimoshenko>
boomshroom: if you infer the nixpkgs commit where renpy used to build, you can build it again with `NIXPKGS_ALLOW_BROKEN=1 nix-shell -p renpy -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/<sha1>.tar.gz`
crucialrhyme has joined #nixos
<crucialrhyme>
yegortimoshenko: well, my keyboard stopped working, as did the KDE shutdown button, so i had to hard reset, but now everything seems very happy!
mizu_no_oto has joined #nixos
<crucialrhyme>
how do I turn on the nixos branding in the plasma theme?
<boomshroom>
Looking at my history, it didn't run with the nix renpy, but rather with python and a provided renpy sdk. The thing is that I don't remember deleting the sdk, but it seems to be gone anyway.
<boomshroom>
Now I remember why I deleted the sdk. Stupid executables pre-linked to /lib64/ld-linux-x86-64.so.2.
<yegortimoshenko>
Fare: i.e. `python.withPackages (ps: with ps; [ requests flask2 ... ])`
b has quit [(Ping timeout: 256 seconds)]
<yegortimoshenko>
boomshroom: just in case if you absolutely need to run such executables, you can use steam-run
<Fare>
I get this error when I try to put that in systemPackages: The option value `environment.systemPackages.[definition 4-entry 192]' in `/home/fare/nixos/common.nix' is not a package.
<boomshroom>
yegortimoshenko: I know I didn't need it last time.
gm152 has joined #nixos
tristan0x has quit [(Ping timeout: 240 seconds)]
b has joined #nixos
<yegortimoshenko>
Fare: if you want to put it directly into a list, wrap it with parantheses, e.g. `environment.systemPackages = [ (python.withPackages (ps: with ps; [ requests ])) ];`
<Fare>
oops, apparently it was the lack of parentheses around it, it seems
<Fare>
thanks!
<antonv>
Fare hello
<Fare>
I don't think I'd have found without help. I don't understand the nix syntax very well.
<Fare>
antonv, hi!
<antonv>
You once advised me to use nix, I'm trying now :)
<Fare>
antonv, welcome to the club! Michael Raskin has already done a lot of heavy lifting wrt to CL packages
crucialrhyme has quit [(Ping timeout: 256 seconds)]
<boomshroom>
Actually looking at the error messages, it's pygame that's broken, not renpy.
<boomshroom>
granted, nix's pygame is provided by renpy
<yegortimoshenko>
Fare: if Nix were to use s-expressions, i'd be so much happier
<antonv>
Fare: I'm not so much about CL packkages, just trying a convenient way for installing software. Had some difficulties with Nix on Ubuntoo, installed finally NixOS on an old, good laptop, 32 bit. Now I hope FireFox 57 will be available in 32 bit version on NixOS (currently it is not).
marsam has joined #nixos
<boomshroom>
antonv: do you the actual CPU model of your laptop?
<yegortimoshenko>
Fare: also, given the scope of quicklisp, it's probably possible to write a nix expression generator like what Nixpkgs currently has for emacsPackages
<antonv>
boomshroom: I can find out. Why do you ask?
<boomshroom>
most of the 32 bit systems running today are on 64 bit processors.
<boomshroom>
Many people install the 32 bit version either to be safe or because they don't know the difference. (Given this chat room, I very much hope it's the former.)
tristan0x has joined #nixos
<antonv>
boomshroom: my CPU is 32 bit
<antonv>
Intel Pentium M 730
crucialrhyme has joined #nixos
<Fare>
yegortimoshenko, probably, but MichaelRaskin is the person to ask about the current state of Lisp in Nix.
<crucialrhyme>
so on unstable everything seemed fine, except that clicking on the launcher icon (bottom left) would crash plasma
<boomshroom>
Wow, that's a 12 year old processor.
<antonv>
boomshroom: right
<crucialrhyme>
on the one hand, it was easy to revert, yay nixos, on the other hand, any thoughts on how to debug this?
<yegortimoshenko>
boomshroom: check out OpenBSD folks, they only recently stopped VAX support
<Fare>
antonv, I admit on my old 32-bit laptop I installed firefox in the debian part, after failing to recompile with nixpkgs
<antonv>
Fare: you use nix on Debian?
<antonv>
I tried to install Firefox on nix on Uboontoo - all fonts were nonospaced
<yegortimoshenko>
crucialrhyme: open an issue
<boomshroom>
I think I'm actually too young to remember a time when 32 bit x86 processors where being developed.
<Fare>
yegortimoshenko, well, ASDF still support Symbolics Ivory processors... although admittedly ASDF 3.3 currently fails to compile on Genera.
<antonv>
Fare: (menus, popups, web pages - everywhere)
<crucialrhyme>
yegortimoshenko: on nixpkgs github? is there a standard form for a bug report somewhere? any information i should include?
tristan0x has quit [(Ping timeout: 272 seconds)]
<Fare>
antonv, I use nixpkgs on top of debian. I believe the default nixos kernel wouldn't boot on that machine, and I really didn't want to recompile my own.
<yegortimoshenko>
crucialrhyme: when you open an issue, it starts with a form. there are steps, but it mostly involves running `nix-info -m` to get info for you to copy and paste
<Fare>
my gramps was programming FORTRAN on punch cards... yikes.
<antonv>
yegortimoshenko: stopping VAX support isn't an example to follow :)
<yegortimoshenko>
Fare: well, i'd guess it's more of a tribute, but very impressive.
<yegortimoshenko>
antonv: agree, they should have never stopped VAX support!
<yegortimoshenko>
and i had Zaurus with OpenBSD on it, and they dropped support for it too...
<Fare>
crucialrhyme, you can file bugs against nixpkgs on github.
<Fare>
ah, the Zaurus!
<Fare>
Reminds me I never managed to upgrade my Jornada 820 from kernel 2.4 to 2.6 and then all the distributions stopped supporting its ABI.
<Fare>
I also have an old ARM laptop somewhere that's essentially unupgradable.
<Fare>
Wondering whether to buy a new laptop now... Lenovo has been promising to repair mine for 2 months, but it's in limbo in the repair facility.
<yegortimoshenko>
Jornada 820 looks neat! i was a OpenBSD fan back then and bought Zaurus specifically because it had good OpenBSD support. it's very hard to use without prebuilt packages, however.
<yegortimoshenko>
if you want to buy a Lenovo laptop, then wait until they update their lineup, they usually do that at the end of the year
<yegortimoshenko>
they only recently (at 12th) ended their sale where they presumably try to sell as much of old stock as they can
<yegortimoshenko>
on 12th*
<yegortimoshenko>
i've recently upgraded to thinkpad x230. it's amazing.
crucialrhyme has quit [(Ping timeout: 265 seconds)]
<antonv>
yegortimoshenko: what OS do you use on tinkpad x230?
<boomshroom>
Well it doesn't look like I'm going to have much luck with renpy. I could have sworn I got it to work somehow before. Strangely enough, it doesn't seem like any of the packages used were updated since then.
<yegortimoshenko>
antonv: NixOS
<boomshroom>
Well, back to Xenoblade.
<Fare>
Jornada 820 was great; too bad I never got it all working properly. It was still a fine text terminal or very slow X client.
<yegortimoshenko>
boomshroom: i'm interested in getting renpy to run on NixOS, so i'd like to package it. could you open an issue on github?
<Nol>
I picked up an x1 carbon in that sale
<Nol>
very satisfied
<Fare>
yegortimoshenko, I don't want a next Lenovo, I want my old Yoga 900 back in working order. Sent it for repair twice, and they are incapable of telling more than it's "under testing" after repair --- since Nov 28.
<Fare>
so 2 weeks worth of "under testing".
<Fare>
My favorite machine ever was a 4-bit machine... the HP 28S.
<boomshroom>
DDLC is supposed to be a meta-game that messes with your computer, but I don't think it's meta enough to expect you to patch its dependencies manually. :P
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] yegortimoshenko opened pull request #32696: ccd2iso: init at 0.3 (master...ccd2iso/init) https://git.io/vbwX8
NixOS_GitHub has left #nixos []
crucialrhyme has joined #nixos
<crucialrhyme>
the kde launcher problem had something to do with something i had installed using nix-env bringing in an incompatible qt version
<yegortimoshenko>
boomshroom: it builds!!
<crucialrhyme>
so i just ran nix-env --upgrade and it worked
<yegortimoshenko>
crucialrhyme: ckb also probably had that problem, does the keyboard work after the upgrade?
<boomshroom>
yegortimoshenko: Sweet! Manually, or did you write a nix-expression for it? If the latter, I'm still working on that, so could you send it over?
astronavt has joined #nixos
<crucialrhyme>
yegortimoshenko: indeed it does! although it was sort of working before, just without the colors. it would only crash after running nixos-rebuild but would work after rebooting once
<crucialrhyme>
how do i just completely annihilate everything i've installed using nix-env?
<yegortimoshenko>
boomshroom: as a patch to nixpkgs, i want to open a pull request that unbreaks renpy
<yegortimoshenko>
i'll send a self-contained nix expression to you, hold on
<yegortimoshenko>
crucialrhyme: nix-env -Eir
<boomshroom>
yegortimoshenko: Before completely unflagging it, try running something with it. I remember that part still wasn't terribly simple.
<yegortimoshenko>
(my mnemonic is "air", poof, all gone)
<samueldr>
I saw DDLC in the previous logs: it runs with steam-run, using the itch.io windows+linux download
<boomshroom>
samueldr: I am in fact trying to play DDLC. I started a while back, SDL broke something, and now I'm trying to see if I can get to the good part.
<yegortimoshenko>
boomshroom: where do i get ddlc on renpy? i only could find prepackaged versions
<samueldr>
haven't yet had time to explore the game, but I was able to start the storyline without any issues using the pre-packaged version on itch.io, using steam-run ./DDLC.sh (IIRC)
crucialrhyme has quit [(Quit: WeeChat 1.9.1)]
<bhipple[m]>
Anyone have experience building Nix RPMs with the VM tooling? I'm giving it a try and it seems to be building with `-j1` even when I set `enableParallelBuilding = true;` and pass `-j4`
<boomshroom>
I think downloaded the windows version, but I somehow managed to find a ddlc.tar that I forget where I got it from.
pxc has quit [(Quit: WeeChat 1.9.1)]
<samueldr>
on itch.io, it's the same package for both windows and linux
<yegortimoshenko>
boomshroom: it reminds me of planescape: torment and torment: tides of numenera.
<boomshroom>
Trying to build the patched pygame seems to hang after `copying build/lib.linux-x86_64-2.7/pygame_sdl2/render.so -> src/pygame_sdl2`.
<yegortimoshenko>
i also have that for a while now. thought it's just my laptop is not beefy enough.
<crucialrhyme>
yegortimoshenko: by boot screen do you mean like somewhere in the bios somehow, or when kde starts up, or on login?
schoppenhauer has quit [(Ping timeout: 265 seconds)]
<boomshroom>
Well, my gaming desktop is probably more beefy than a laptop.
<crucialrhyme>
i just really enjoy that snowflake of lambdas
<yegortimoshenko>
crucialrhyme: after the bios screen, but before the desktop shows up
pikajude has left #nixos ["Leaving"]
<yegortimoshenko>
not on login
pikajude has joined #nixos
<boomshroom>
"snowflake of lambdas" How the bleep did I not notice that‽
schoppenhauer has joined #nixos
b has quit [(Quit: leaving)]
b has joined #nixos
rogue_koder has joined #nixos
<yegortimoshenko>
crucialrhyme: conceptually, if you want to do something similar later on, you can override the package, e.g. `boot.plymouth.themePackages = [(pkgs.breeze-plymouth.override { nixosBranding = true; })];`
<crucialrhyme>
ah ok! i was just reading the nix pill about that. thanks!
<boomshroom>
did you find anything about the pygame hang?
b has quit [(Client Quit)]
crucialrhyme has left #nixos ["WeeChat 1.9.1"]
<yegortimoshenko>
boomshroom: not yet, it's very weird that it would hang on render.so though, it wasn't even patched in that patch. also cpu doesn't go down when it happens, i.e. it certainly does something
<boomshroom>
Well it doesn't do very much. My CPU is only at 13%.
<bgamari>
How do people get by with the haskellPackages package sets?
<bgamari>
It seems they are far too old to be usable
b has joined #nixos
<neonfuz>
any downsides to running /tmp on a ramdisk on nixos?
notdaniel has quit [(Quit: Leaving)]
<yegortimoshenko>
neonfuz: as on any other linux distro, meaning you need enough RAM to do that
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<yegortimoshenko>
plus one more: boot.tmpOnTmpfs size is not configurable, it's always some % of your RAM, not remember how much exactly
<yegortimoshenko>
25% sounds right
gm152 has quit [(Ping timeout: 272 seconds)]
<jeaye>
You can configure it with systemd.mounts though.
crucialrhyme has joined #nixos
<crucialrhyme>
thanks to everyone for being so helpful and for helping to make a very good operating system!
astronavt has quit [(Remote host closed the connection)]
<Nol>
general question for veteran users: when you encounter software that isn't yet packaged, in general how much effort is it to get it running assuming fluency in nix?
<bgamari>
ahhh, I see, the default package set is held back at GHC 8.0.2 core packages
gm152 has joined #nixos
crucialrhyme has quit [(Quit: WeeChat 1.9.1)]
<yegortimoshenko>
Nol: range of several orders of magnitude depending on the software
<yegortimoshenko>
i've just packaged ccd2iso and it took 4-5 minutes to package it.
antonv has quit [(Ping timeout: 256 seconds)]
<yegortimoshenko>
however, i also have upcoming version of xfce in my overlay, with around 40 packages, it took at least 6 hours.
<Nol>
Mainly, I just have a worry as a new user about the lack of an "escape hatch" in nixos. On an FHS compliant system if something proves too annoying to package I can just fallback to directly compiling and installing it
<yegortimoshenko>
that also depends on package itself, some packages are much more presumptious than others.
<Nol>
I get the impression that doing so o nnixos tends to be just as much trouble as paking it would be
<yegortimoshenko>
i see, i don't think it takes more effort to package than to compile & install once you learn nix
<bhipple[m]>
Nix has some nice build-FHS-chroot hierarchy tools for the escape hatch
<bhipple[m]>
Of course, using them well requires some fluency in Nix ;)
<bhipple[m]>
but the basic idea is that it allows you to build a NixPkg that has a full chroot'd FHS compliant filesystem tree and build/run a binary within it
<hyper_ch>
hmmm, something is compiling but now idea what
leat has joined #nixos
vidbina has joined #nixos
rogue_koder has quit [(Ping timeout: 248 seconds)]
Synthetica has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jwiegley pushed 3 new commits to master: https://git.io/vbwFX
<NixOS_GitHub>
nixpkgs/master 0042a34 John Wiegley: coqPackages.coq-haskell: New expression
<NixOS_GitHub>
nixpkgs/master d1532bc John Wiegley: coqPackages.category-theory: New expression
<NixOS_GitHub>
nixpkgs/master 2e93f93 John Wiegley: z3_4_5_0: New expression, to access the release version's API
NixOS_GitHub has left #nixos []
<sphalerite>
hyper_ch: ls -d /tmp/nix-build*
<hyper_ch>
nix-build-gcj-6.4.0.drv-0
<hyper_ch>
nix-build.xSKpyG
<hyper_ch>
nixos-rebuild.sFwMeX
<sphalerite>
Huh
<sphalerite>
Does anyone even use gcj
<hyper_ch>
I think it's for pdftk
<hyper_ch>
sphalerite: btw, are you human?
<hyper_ch>
(you seem never to sleep)
<sphalerite>
I think I am
<sphalerite>
I do sleep
<sphalerite>
Slept from 10:30pm to 6:30am
<hyper_ch>
once every decade?
<sphalerite>
Nope, every night
<hyper_ch>
:)
<sphalerite>
I sleep more than most of the people I know
<sphalerite>
Although recently I've been slipping, sleeping too late almost regularly :/
<hyper_ch>
but you're always here
<sphalerite>
Nope
<sphalerite>
Only when you are :p
<hyper_ch>
yes you are always here.. I don't see any leave/join messages from you
<sphalerite>
(Actually no I think there are times when you're there and I'm not, and when I'm there and you're not)
<hyper_ch>
:)
<hyper_ch>
wtf is wrong with znc
<hyper_ch>
it keeps messing up my channel order
hyper_ch has left #nixos ["Konversation terminated!"]
hyper_ch has joined #nixos
pie_ has quit [(Remote host closed the connection)]
<hyper_ch>
I hope I didn't miss out on anything important while restarting konversation and znc
pie_ has joined #nixos
rogue_koder has joined #nixos
<sphalerite>
Oooooh right. I'm here through the matrix bridge, so I'm always in the channel (except in the not unheard-of case where the bridge goes down) even if I'm not present and reading the channel.
<sphalerite>
I waited for you to.come back to send that message :p
<vaibhavsagar>
there might be a clearer/simpler way but this works for me
<mbrock>
interpolating a whole Git repository into a string is very Nixish
dan_b has quit [(Ping timeout: 255 seconds)]
<vaibhavsagar>
I just tried `import ((fetchFromGitHub { ... }) + "/foo.nix") {}` and that seemed to work too
<Isorkin>
srhb: default kernel linux-4.9.68 work with zfs, 4.14 and hardened - not work
goibhniu has joined #nixos
sary has quit [(Remote host closed the connection)]
phdoerfler has quit [(Quit: Leaving.)]
stqism has quit [(Quit: Like 3 fire emojis lit rn 🔥🔥🔥)]
dywedir has joined #nixos
coup has joined #nixos
sary has joined #nixos
coup has quit [(Client Quit)]
coup has joined #nixos
coup has quit [(Changing host)]
coup has joined #nixos
coup is now known as stqism
dan_b has joined #nixos
marc__ has joined #nixos
periklis has joined #nixos
<srhb>
Isorkin: I have 4.14.5 with zfs.enableUnstable = true and that does work.
<srhb>
Isorkin: I can't try without unstable since my pool is encrypted.
<hyper_ch>
srhb: also using encryption :)
<mbrock>
vaibhavsagar: yeah that's a bit more easy to understand
radvendii has quit [(Ping timeout: 255 seconds)]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
marc__ has quit [(Remote host closed the connection)]
sary has quit [(Quit: Reconnecting)]
sary has joined #nixos
sary has quit [(Client Quit)]
sary has joined #nixos
thc202 has joined #nixos
marc__ has joined #nixos
<marc__>
I installed following packages from nixpkgs:
<marc__>
firefox thunderbird libreoffice
<marc__>
the versionnumber of these packages are stable
<marc__>
I am little bit confusig about the names nightly earlybird and devedition
<etu>
marc__: If you build firefox from source, you're not allowed to use the mozilla branding. So if you want the mozilla branding you can use firefox-bin.
<etu>
marc__: I know that you probably didn't build firefox from source, you probably got it pre-compiled by hydra.
<Isorkin>
srhb: thanks, with zfs.enableUnstable = true - works
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] cillianderoiste opened pull request #32705: Fix build of opensubdiv (master...opensubdiv-fix) https://git.io/vbwjR
NixOS_GitHub has left #nixos []
<etu>
marc__: firefox-bin gives you the binaries provided by mozilla. It's the same with thunderbird and thunderbird-bin :)
coot has joined #nixos
pie__ has joined #nixos
pie_ has quit [(Read error: Connection reset by peer)]
oida has quit [(Ping timeout: 248 seconds)]
pie__ has quit [(Read error: Connection reset by peer)]
<sphalerite>
hyper_ch: because people are always confused by nightly/earlybird
<hyper_ch>
In the unstable channel at the moment it is necessary to set boot.zfs.enableUnstable = true; to get zfs version based on master branch as zfsStable does not yet have this feature.
<roberth>
also, how do i find out in general who can and/or how can someone get merge access?
<joko>
Hello, guys, I was wondering for possible web app development on NixOS, e.g. you want nginx, php-fpm and postgres running, but not as system services all the time. Any thoughts on that?
<hyper_ch>
sphalerite: and why would debian's naming improve that?
bkchr has quit [(Ping timeout: 265 seconds)]
<Isorkin>
hyper_ch: I do not need encryption
<sphalerite>
hyper_ch: because it doesn't suggest "this isn't the stable version"
<hyper_ch>
sphalerite: I'm confused now
<mbrock>
joko: maybe the NixOS containers stuff would be useful
<hyper_ch>
sphalerite: could you elabrote a bit more?
periklis has quit [(Ping timeout: 256 seconds)]
<sphalerite>
hyper_ch: people often install Firefox and then think "why did I get nightly, I just want stable Firefox"
<hyper_ch>
ah
<sphalerite>
See marc___ just 20min ago
<hyper_ch>
never bothered me if it's stable or nightly
<etu>
hyper_ch: Well, if you install "firefox", many people would expect to get "firefox stable version xy.0" and not "firefox nightly" which can be read as "testing" or "unstable"
<hyper_ch>
otoh, many people using nixos aren't normal computer users and I tend to think most of them actually prefer a bit bleeding edge :)
yenzenz has joined #nixos
<etu>
But the "firefox" package isn't bleeding edge
<hyper_ch>
so where's the problem?
<etu>
Confusing for people
<hyper_ch>
you just say firefox isn't bleeding edge
<hyper_ch>
so they get stable
ma27 has joined #nixos
<etu>
But "Firefox Nightly" doesn't read as stable if you read the name
<hyper_ch>
I still fail to comprehend the problem
<hyper_ch>
nox firefox shows ff 57.0.2
<hyper_ch>
opening firefox and checking, it's 57.0.2
<hyper_ch>
I fail to understand the issue
<etu>
Yeah, but it's "Firefox Nightly", not "Firefox" as people expect
<goibhniu>
joko: indeed, nixos-containers are perfect for this (I use them extensively for web-dev)
<hyper_ch>
they get the version that's announced
<hyper_ch>
I fail to see what the perceived issue is
<etu>
hyper_ch: Yes, they get the stable version. But for example marc__ joined here and asked why he got "Firefox nightly" instead of firefox. Which proves that people get confused by the non-mozilla branding when they install a mozilla program.
<hyper_ch>
etu: 1 person != people
<goibhniu>
it comes up pretty regularly IMO
<etu>
yeah
<goibhniu>
it has the wrong name ... that's all
<etu>
I have encountered that myself
<goibhniu>
even if you deduce that the version looks right, it's confusing
<etu>
And I've seen it happen in other distros as well (gentoo) where they like building things from source.
<hyper_ch>
not really.. nox announces the same version that gets then installed
<marc__>
and gnome3 in setting says gnome 3.0
<etu>
And people asking about it
aarvar has quit [(Quit: Leaving.)]
<marc__>
but i think its 3.24
<etu>
marc__: You're on 17.09 I assume?
ThatDocsLady has joined #nixos
<etu>
marc__: It does that on 3.26 as well :p
<marc__>
if i seach package with 'nix-env -qaP | grep libreoffice'
aarvar has joined #nixos
<marc__>
there is sometimes a nixo. or nixos. prefix
<hyper_ch>
marc__: just look at it as a way to improve your english :) Don't worry, we won't bite your head off for making mistaked.... I think
<marc__>
ok thanks
<marc__>
now there are two versions
<etu>
marc__: You can probably do a "sudo nix-channel --remove nixo", if you don't use it you don't need it. I can't say if you use that or not though :)
mahalel_ has left #nixos []
<marc__>
ok only stable version in nix-channel
kini has quit [(Quit: No Ping reply in 180 seconds.)]
<etu>
Then you shouldn't see any nixo. things in your search results :)
<marc__>
thanks
athan has joined #nixos
<marc__>
there is no hunspell-dict-de-de
kini has joined #nixos
<marc__>
i like to use spelling correction in german language
rogue_koder has quit [(Ping timeout: 260 seconds)]
aarvar has quit [(Ping timeout: 272 seconds)]
iyzsong has joined #nixos
<srhb>
FossiWeb: Usually cherry-picking them onto the release-17.09 branch and fixing conflicts.
<srhb>
FossiWeb: (if all other dependencies are in place)
<FossiWeb>
it's basically only 3 lines in the chroot for steam, so i guess that should be easy enough
<FossiWeb>
so i do that and then just PR?
sigmundv__ has joined #nixos
Tucky has quit [(Remote host closed the connection)]
simukis has joined #nixos
<srhb>
FossiWeb: Yes. :)
<srhb>
FossiWeb: cherry-pick -x to get the (cherry-picked from...) line
Tucky has joined #nixos
<srhb>
FossiWeb: Testing it would be nice, too. :3
<FossiWeb>
well, i'll test it on my system to be sure :)
<srhb>
Nice! :-)
<FossiWeb>
well, it failed, so... XD
<FossiWeb>
maybe not that simple after all ;)
orivej has quit [(Ping timeout: 264 seconds)]
<srhb>
FossiWeb: I'll be able to help a little later when I get home. :)
<srhb>
(If I can figure it out)
sary has quit [(Ping timeout: 240 seconds)]
orivej has joined #nixos
Fare has joined #nixos
<FossiWeb>
i just merge
<FossiWeb>
*merged a bit too blindly ;)
<Fare>
Weird: after I last upgraded, now I get cannot connect to daemon at ‘/nix/var/nix/daemon-socket/socket’: Connection refused when I try to use nix-shell
<sphalerite>
Fare: systemctl status nix-daemon ?
<FossiWeb>
yay. i think it builds mono now :>
seanparsons has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
schoppenhauer has quit [(Ping timeout: 264 seconds)]
schoppenhauer has joined #nixos
reinzelmann has quit [(Ping timeout: 265 seconds)]
iyzsong has quit [(Read error: Connection reset by peer)]
zzamboni has joined #nixos
<aminechikhaoui>
Folks, I got an expression that generates about 122 *.drv, the weird thing is that the evaluation seems to abort while allocation some further mem with the error msg "Too many heap sections: Increase MAXHINCR or MAX_HEAP_SECTS"
<marc__>
i get an error 'Command "python setup.py egg_info" failed with error code 1 in /tmp/pip-build-wix7ffuy/sense-emu/'
<marc__>
'ModuleNotFoundError: No module named 'setuptools'
<marc__>
'
erictapen has quit [(Ping timeout: 260 seconds)]
pie_ has joined #nixos
<marc__>
'nix-env -iA nixos.python36Packages.setuptools' is no solution
<marc__>
:(
propumpkin is now known as contrapumpkin
<marc__>
:(
waleee has joined #nixos
drakonis has joined #nixos
<makefu>
marc__: pip3 may not work how you might expect it to work (it will not install the package globally). this is due to how nix' inner workings are set up
<hyper_ch>
marc__: what are you trying to install?
mizu_no_oto has joined #nixos
<marc__>
nix-env -iA python3
<marc__>
not with configuration.nix
Synthetica has quit [(Quit: Connection closed for inactivity)]
<marc__>
@hyper_ch do you install the python packages with configuration.nix?
<hyper_ch>
I just have those in my configuration.nix
<marc__>
ah ok
<marc__>
i try it
<hyper_ch>
because as user I needed to setup virtualenv to get a user installed pyrocore tools
<marc__>
thanks
<hyper_ch>
use 36 instead of 27
<hyper_ch>
and pyrocore I needed as enhancement for rtorrent-ps
marc__ has quit [(Remote host closed the connection)]
cmiles74 has joined #nixos
<lassulus>
is it possible to get internet in a nixos-test?
sary has quit [(Ping timeout: 265 seconds)]
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
LexLeoGryfon has joined #nixos
yegortimoshenko has joined #nixos
arximboldi has joined #nixos
babyflakes has quit [(Quit: Connection closed for inactivity)]
<Mic92>
lassulus: maybe without sandbox enabled?
<arximboldi>
Mmmm, `compass` from nixpkgs fails for me with: compass compile --output-style compressed
<arximboldi>
LoadError on line ["179"] of /nix/store/ck2mr9j2hmg5bhg5qyyndsc2afxb0r9v-compass-1.0.3/lib/ruby/gems/2.3.0/gems/compass-core-1.0.3/lib/compass/configuration/data.rb: cannot load such file -- compass/import-once/activate
<arximboldi>
seems to fail at: "require 'compass/import-once/activate'
<arximboldi>
"
<arximboldi>
which was added by the same command
<arximboldi>
I mean by "compass config"
<arximboldi>
:/
<arximboldi>
is there something special I need to do to use ruby based packages or something?
zzamboni has joined #nixos
johnsonav has quit [(Read error: Connection reset by peer)]
bennofs has quit [(Quit: Connection closed for inactivity)]
johnsonav has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gebner pushed 1 new commit to master: https://git.io/vbruH
<NixOS_GitHub>
nixpkgs/master e4d4b2e Gabriel Ebner: qutebrowser: 1.0.3 -> 1.0.4
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] domenkozar force-pushed haskell-outputs-bin-etc-lib from f73e524 to a0b51de: https://git.io/vbgMM
<NixOS_GitHub>
nixpkgs/haskell-outputs-bin-etc-lib ae47093 Domen Kožar: Revert "Merge pull request #32333 from NixOS/revert-32112-hask-split-out"...
<NixOS_GitHub>
nixpkgs/haskell-outputs-bin-etc-lib da183da Domen Kožar: hevm: disable bin output to prevent a cycle
<NixOS_GitHub>
nixpkgs/haskell-outputs-bin-etc-lib 2bb4a6a Domen Kožar: symlinkJoin: support multiple outputs...
NixOS_GitHub has left #nixos []
coot has quit [(Quit: coot)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vdemeester opened pull request #32712: notary: init at 0.4.3 (master...init-notary) https://git.io/vbrgU
NixOS_GitHub has left #nixos []
<lassulus>
Mic92: sadly the sandbox doesn't change anything
ssmike1 has joined #nixos
cement has joined #nixos
orivej_ has quit [(Ping timeout: 256 seconds)]
ssmike has quit [(Ping timeout: 255 seconds)]
ssmike1 is now known as ssmike
MercurialAlchemi has quit [(Ping timeout: 264 seconds)]
zzamboni has quit [(Quit: Leaving.)]
erasmas has joined #nixos
babyflakes has joined #nixos
<laggy_wifi>
does nix have any unique commands that accept integers as arguments
aneeshusa has joined #nixos
<sphalerite>
laggy_wifi: what do you mean?
yegortimoshenko has quit [(Ping timeout: 248 seconds)]
mental1ta has joined #nixos
dan_b has quit [(Ping timeout: 265 seconds)]
<laggy_wifi>
like does nix have any of its own commands that it itself created, that can accept numbers as arguments
<laggy_wifi>
for example "dd" does not count as nix did not make it
<laggy_wifi>
nix-env does but idk if it supports numbers as arguments
<Acou_Bass>
wondering how i can add that to my configuration.nix with my printer enabled? im adding it to services.printers.drivers = [ pkgs.cnijfilter_2_80 ]; but i need to modify it slightly to include my MG3200
<Guest67032>
laggy_wifi: Notably the --arg arguments to a bunch of commands take Nix expressions, of which integer literals are a case
leat has quit [(Quit: WeeChat 1.9.1)]
<Guest67032>
nix-build --max-silent-time
<Guest67032>
--cores, &c.
<laggy_wifi>
are u a nix dev? or just someone that only uses nix
<Guest67032>
I'm not sure what the distinction is
Guest67032 has quit [(Changing host)]
Guest67032 has joined #nixos
Guest67032 is now known as TweyII
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 1 new commit to master: https://git.io/vbrwo
<NixOS_GitHub>
nixpkgs/master 874dbb1 Peter Simons: git-annex: update hash for version 6.20171214
NixOS_GitHub has left #nixos []
<TweyII>
I have some nixpkgs contributions, if that's important to you for some reason… but it doesn't change the answer to your question :þ
<sphalerite>
laggy_wifi: what is the point of your question?
<laggy_wifi>
how would u reliably check is a user enters a integer in wich would overflow
<laggy_wifi>
assuming unsigned long can also overflow
<laggy_wifi>
say if a user enters something like 99999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999
<laggy_wifi>
as an argument
<laggy_wifi>
(without quoting it)
<sphalerite>
laggy_wifi: nix doesn't handle overflow, for example nix-instantiate --eval -E '4294967296 * 4294967296' gives you 0
<sphalerite>
Why do you care?
<laggy_wifi>
so a user could end up getting a completely different integer being processed then what he/she initially entered?
<sphalerite>
yes, but I don't know of any cases where you'd need such large numbers
<laggy_wifi>
ok
radvendii has joined #nixos
<sphalerite>
do you have one?
<sphalerite>
note that nix does not allow literal integers that big
<laggy_wifi>
what about in the event that an indirect numarical value causes an overflow?
<laggy_wifi>
like say the calculation of the total sum of packages about to be downloaded in bytes then converted to human readable form
<laggy_wifi>
total size of the sum of*
<sphalerite>
Then you have other problems than overflow, like storage
<laggy_wifi>
or in any case that an integer is given or required
<clever>
sphalerite: is the nix limit 32bit or 64bit?
<sphalerite>
2^64 bytes is 16384 petabytes
<sphalerite>
clever: seems to be 64 bit
<clever>
ah, yeah, your going to run out of storage first
<laggy_wifi>
is it possible to do mathamatical operations on a string directly without any conversion?
<sphalerite>
yes, but why do you want this
<laggy_wifi>
or in such a way that an overflow would never occur
<sphalerite>
any case where nix runs into numbers that big is likely to be pathological, i.e. you've got a bug somewhere
<sphalerite>
Or do you have a real-world usecase for extremely large numebrs?
<laggy_wifi>
i just want to figure out how to detect when a user enters a number (or if a calculation prefermed) would result in an overflow and if possible how to avoid it to allow for that calulation or user input to proceed
<alexteves>
I think nix isn't really intended to use for general purpose programming, only package management
<sphalerite>
yeah it's not a general-purpose programming language, and even for general-purpose programming you don't need numbers that big most of the time
<clever>
i just had an idea, one minute
<cransom>
user input, at least interactively, isn't something i've run into either
<clever>
testing...
<sphalerite>
yeah if you're trying to get interactive input you definitely shoudln't be using nix
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] primeos pushed 1 new commit to master: https://git.io/vbroG
<NixOS_GitHub>
nixpkgs/master 63596fa Michael Weiss: pfstools: 2.0.5 -> 2.1.0
NixOS_GitHub has left #nixos []
jacob_ has joined #nixos
<sphalerite>
laggy_wifi: and again, what is the real-world context of this? Why do you think you need to handle these cases?
<laggy_wifi>
fofor example say a user has 10 500tb harddrives connected and merged into a single 50,000 tb hard drive and decides to mass download every package possible to the hdd all at once
jacob_ is now known as Guest94605
<sphalerite>
Then 64 bits are still more than enough to handle that :)
<sphalerite>
You'd need 33 thousand 500TB arrays to reach that limit
<laggy_wifi>
wich would result in (assuming the user can fill up there 50k tb hdd entirely) in around 50000000000000000 bytes being processed at some stage
<sphalerite>
also, last I checked 10 times 500 is 5000, not 50000
<sphalerite>
Also, 500TB hard drives do not currently exist, and probably won't for another 10 years at least. So this does not qualify as a real-world use case.
<laggy_wifi>
what is the largest possible number a 64 bit achine can possibly handle untill an overflow would occur from being 1 number higher then the limit
<laggy_wifi>
machine*
<sphalerite>
If your next message isn't a real-world use case or a "kthxbai" I'm done with this discussion and will be ignoring you I'm 90% convinced you're trolling at this point
<sphalerite>
2^64
fendor has joined #nixos
<cransom>
if you are asking a group of people what 2^64+1 is... i... well anyway.
arximboldi has quit [(Quit: arximboldi)]
<laggy_wifi>
like does such a method for that detection exist
<catern>
laggy_wifi: sure, it's very possible to do overflow detection. though, it's not related to Nix
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
srdqty1 has joined #nixos
yegortimoshenko has joined #nixos
FossiWeb has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] matthiasbeyer opened pull request #32714: smenu: init at 0.9.10 (master...add-smenu) https://git.io/vbrim
NixOS_GitHub has left #nixos []
srdqty1 has quit [(Ping timeout: 265 seconds)]
mizu_no__ has joined #nixos
spietz has joined #nixos
radvendii has quit [(Ping timeout: 255 seconds)]
roberth has quit [(Ping timeout: 265 seconds)]
ertes has quit [(Ping timeout: 264 seconds)]
radvendii has joined #nixos
<fendor>
sphalerite, i think i am using i3
<sphalerite>
fendor: I thought you're using gnome
<fendor>
oh, no, i need a display manager, don't u?
<fendor>
*i
waleee has quit [(Quit: WeeChat 1.9.1)]
<fendor>
well, my setting is: desktopManager gnome3 enabled and windowMAnager i3 enabled
<sphalerite>
yeah and the things I named are display managers
<sphalerite>
right
<fendor>
oh, well, then it is obvious
<sphalerite>
if you don't have one set, you're probably using slim (light and minimal login screen?)
<fendor>
yeah, sounds like it
<fendor>
and also looks like it
<sphalerite>
I think lightdm, gdm and sddm all work
<fendor>
any preferences?
szicari_ has joined #nixos
<sphalerite>
I like sddm myself, but I don't know if/how well it works with GNOME, I've only used it with plasma
<samueldr>
when gnome-ing, it's generally recommended to go all-in
<fendor>
so, gdm?
zzamboni has joined #nixos
<samueldr>
probably the safer bet, for best integration
<fendor>
ok, thanks!
szicari has quit [(Ping timeout: 265 seconds)]
szicari_ is now known as szicari
<fendor>
another question, haskell builds fail because it cant find any zlib, altough i enabled in my global stack.yaml nixos and build with the --nix flag
radvendii has quit [(Ping timeout: 272 seconds)]
yegortimoshenko has quit [(Remote host closed the connection)]
FRidh has quit [(Quit: Konversation terminated!)]
tristan0x has quit [(Ping timeout: 240 seconds)]
Neo-- has quit [(Ping timeout: 265 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej closed pull request #32678: pqiv: add dependencies needed to enable all backends (master...feature/pqiv-backends) https://git.io/vbVjB
NixOS_GitHub has left #nixos []
marsam has joined #nixos
aneeshusa has quit [(Ping timeout: 265 seconds)]
marsam has quit [(Ping timeout: 240 seconds)]
tristan0x has joined #nixos
endformationage has joined #nixos
zzamboni has quit [(Quit: Leaving.)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vbr1i
<clever>
line 1 of that file brings everything into scope
<clever>
your file doesnt have that
<mitchty>
whoops, didn't copy the first line in the example
<mitchty>
its the same error with "with import <nixpkgs> {};"
<clever>
run "gist -p default.nix" and then it will upload the entire file for you
<clever>
Package ‘2048-in-terminal-2015-01-15’ in /Users/mitch/src/github.com/NixOS/nixpkgs/pkgs/games/2048-in-terminal/default.nix:25 is not supported on ‘x86_64-darwin’, refusing to evaluate.
alsoStevenXL has left #nixos []
<clever>
you somehow refers to the 2048 game, which is linux only
Tucky has quit [(Quit: WeeChat 2.0)]
Sonarpulse has joined #nixos
<mitchty>
updated the gist, the old default.nix contents are at the bottom, that used to work, was mainly trying to test out py3 support for a py2 application
<mitchty>
the commented out spiel used to work a while ago, thats mostly why I'm confused as to what changed that I need to change this derivation at all
<clever>
mitchty: what is the content of your config.nix file?
<mitchty>
i have none
<mitchty>
and trying the guidance gives me that other error message
<clever>
ok, try `nix-instantiate -v default.nix` into a log file, and gist the entire thing, as a file, not a comment
<plakband>
I'm a Haskell guy, I got interested in nixos after hearing how similar it was to stack. I have a few years of experience with development in the OS X shell, but I have never _really_ used linux. I want to start using linux on my macbook, is nixos a bad choice for a "my first linux"?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jwiegley pushed 1 new commit to master: https://git.io/vbrD6
<NixOS_GitHub>
nixpkgs/master 8123bc7 John Wiegley: coqPackages.coq-haskell: Update to add Hask.Ssr module
NixOS_GitHub has left #nixos []
<clever>
plakband: you can also use nix and haskell together on OSX if you want to
<mitchty>
clever: gist the derivation produced or just the instantiate output?
<clever>
mitchty: the output from instantiate
<catern>
plakband: before using NixOS, you should try using Nix on OS X
<clever>
mitchty: what is the content of this file?
<mitchty>
ah sorry thought you meant the ~/.config/nixpkgs/config.nix
<clever>
both paths have the exact same effect
<clever>
one just has a higher priority over the other
<plakband>
clever, catern: Forgot to mention, I have been using the package manager, with some success (some issues with PATH and python). My goal is mostly to move away from OS X
<clever>
mitchty: and the error doesnt exist in that run
<gchristensen>
Putti: barring errors (which happen, but are absolutely not intentional) unless you explicitly enable unfree software, you will not get unfree software.
<the-kenny>
As far as I know is stuff with license.unfree disabled by default.
<gchristensen>
(you have to set `allowUnfree = true` in your nixpkgs and nixos configs)
sbjorn has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 10 new commits to master: https://git.io/vbrFy
<Profpatsch>
garbas: Oh boy. What’s going on at Mozilla right now.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] markuskowa opened pull request #32719: airspy: fix installation of udev rules for USB access (master...airspy-udev) https://git.io/vbrxU
<sphalerite>
Can I copy paths ignoring signedness using nix copy as root?
<sphalerite>
What's the option I need for it?
<sphalerite>
Aah got it --signed-binary-caches ''
rogue_koder has joined #nixos
<pie_>
is it possible to have nix-shell expose librarieS? :/
<pie_>
im trying to build something but i get main.c:43:31: fatal error: readline/readline.h: No such file or directory , whereas i did nix-shell -p readline
mizu_no__ has joined #nixos
zzamboni has joined #nixos
cement has quit [(Ping timeout: 265 seconds)]
yegortimoshenko has joined #nixos
roberth has joined #nixos
drakonis has joined #nixos
goibhniu has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vbrjR
<NixOS_GitHub>
nixpkgs/master 1b555c5 Vladimír Čunát: texlive: more robust way to fetch...
NixOS_GitHub has left #nixos []
Judson1 has joined #nixos
<Judson1>
I'm trying to add postgresql support to liquibase in Nixpkgs. I'm finding that I need a more recent version of the JDBC drivers to do that.
<Judson1>
Is it okay to make that one PR?
jb55 has joined #nixos
<Judson1>
Or do I need to PR the JDBC update, and then build the Liquibase update on that?
<Ankhers>
Judson1: I believe you are allowed to do a single PR, just as long as the commits are in the "proper" order (i.e., update JDBC, then Liquibase).
<yegortimoshenko>
it has both fetcher for Maven and an example how to build a jar
erictapen has joined #nixos
<yegortimoshenko>
meaning, you might need to do this on top of this pull request to get the Maven fetcher
<pie_>
wow thats a pain in the ass installer
<yegortimoshenko>
pie_: buildFHSUserEnv
<pie_>
need extractors for that shit
aarvar has quit [(Ping timeout: 265 seconds)]
<kini>
I have some friends who are into machine learning and asked me if they could use nix but I had to tell them tensorflow is broken and MKL is not packaged yet, sadly
zzamboni has quit [(Quit: Leaving.)]
<kini>
tensorflow looks pretty darn complicated to build
<pie_>
SDE is just a tar archive, idk about getting the actual executables running, what i did with patchelf seems to have been insufficient
mizu_no__ has quit [(Quit: Computer has gone to sleep.)]
<pie_>
now that i think about it 256 bits might be more appropriate,...might not need avx512
ylwghst has quit [(Ping timeout: 248 seconds)]
ylwghst has joined #nixos
jb55 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 1 new commit to master: https://git.io/vbofk
<NixOS_GitHub>
nixpkgs/master 4859e70 Peter Simons: sane-backends-git: update to current HEAD of master branch
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vbofm
<NixOS_GitHub>
nixpkgs/master 0ea7ad1 Orivej Desh: scalafmt: change deps name to invalidate negative cache on Hydra [#32442]
NixOS_GitHub has left #nixos []
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<Judson1>
yegortimoshenko, I'm suprised that no-one has taken issue with fetchMaven yet.
<yegortimoshenko>
Judson1: what's wrong?
aarvar has joined #nixos
<Judson1>
If I understand correctly, you're shifting the resolution of deps to build time.
<Judson1>
So it's possible my clojure build and yours would be different.
<yegortimoshenko>
it's not me: Maven can't resolve deps any other way
<Judson1>
Even, my clojure build today and the next day
<yegortimoshenko>
no, it always resolves dependencies exactly the same way
<Judson1>
Not if they're SNAPSHOT...
<yegortimoshenko>
fetchMaven is a fixed-output derivation: if output doesn't match sha256, build fails
<Judson1>
Okay.
<Judson1>
I'm just saying I'm surprised
<orivej>
rust packaging works the same
<yegortimoshenko>
there is a npm fetcher that works the same way
goibhniu has quit [(Ping timeout: 272 seconds)]
aarvar has quit [(Ping timeout: 255 seconds)]
<catern>
Judson1: just to be clear, this doesn't resolve deps at build time, because the deps are still pinned by sha256
<yegortimoshenko>
it's impossible to resolve deps at nix runtime in nix
<yegortimoshenko>
given that you are building in a sandbox
aarvar has joined #nixos
ilyaigpetrov has quit [(Quit: Connection closed for inactivity)]
<catern>
yep
<gchristensen>
you resolve deps at .nix creation time :)
<orivej>
yegortimoshenko: Judson1 is right that the dependency are resolved during the build of the src derivation. Nix only guarantees that if they resolve differently, it fails the build.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] garbas pushed 3 new commits to master: https://git.io/vboUc
<NixOS_GitHub>
nixpkgs/master 390dc93 Peter Hoeg: libva: 1.7.3 -> 2.0.0
<NixOS_GitHub>
nixpkgs/master fddee84 Peter Hoeg: vaapiIntel: 1.8.2 -> 2.0.0
<NixOS_GitHub>
nixpkgs/master da864e5 Peter Hoeg: libva-utils: init at 2.0.0
NixOS_GitHub has left #nixos []
<jsgrant>
Can Systemd-boot be defined to add a custom boot option + location? I want to play around with Redox OS on my laptop's sdb drive-bay, and just trivially select it from bootscreen.
<jsgrant>
Thinking might have to messabout with Grub.
<jsgrant>
instead*
<yegortimoshenko>
orivej: true. however, Maven itself is deterministic, i.e. it guarantees that it will always resolve depednencies the same way.
<orivej>
AFAIK this is true (except for SNAPSHOT versions)
<orivej>
Judson1: Sorry, I did not mean that rust resolves dependencies while building src, only that it allows an arbitrary program to download dependencies. rust packaging requires lock files that list the concrete versions of dependencies.
marsam has quit [(Ping timeout: 264 seconds)]
<yegortimoshenko>
s/depednencies/dependencies/
amfl has quit [(Read error: Connection reset by peer)]
amfl has joined #nixos
<orivej>
jsgrant: not via configuration.nix, you'll have to investigate if it all
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] yegortimoshenko opened pull request #32721: Backport Steam fix to 17.09 (release-17.09...steam/13-dec) https://git.io/vboUb
NixOS_GitHub has left #nixos []
ertes has quit [(Ping timeout: 265 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vboTm
<Lisanna>
Does Nix hash files based purely on content rather than meta attributes like name, mode, etc.? For example, if I had a derivation which had a "path" type which was a /nix/store path, would it create a duplicate entry in the store for that with a different hash, or would it use that entry as-is?
<Lisanna>
s/derivation which had/derivation made up in part by
hamishmack has quit [(Ping timeout: 268 seconds)]
<vcunat>
there are two kinds of derivations
<yegortimoshenko>
Lisanna: nix hash is usually based on derivation itself rather than on contents that it produces
<vcunat>
fixed-output ones where the path is given by hash
qewzee has joined #nixos
<vcunat>
and the rest (noted above)
<yegortimoshenko>
examples of fixed-output derivations are fetchurl, fetchFromGitHub, etc.
<vcunat>
or most TeX Live
<Profpatsch>
garbas: A good maxim to live by. :)
<yegortimoshenko>
thanks for maintaining that, vcunat.
<vcunat>
I don't really maintain it, almost
<vcunat>
but it was fun to create, years ago
<Lisanna>
yegortimoshenko: so, I know that if you give a direct path in a nix expression (not a string), that path will get copied into the Nix store. How is the hash of that computed?
<yegortimoshenko>
every time i build tex on a new machine, and it downloads several hundred packages, i'm amazed that it works
<yegortimoshenko>
(full scheme)
<Lisanna>
is it just on the contents?
ma27 has joined #nixos
<yegortimoshenko>
Lisanna: i'll make an experiment, hold on.
<Lisanna>
Okay, I tried to think of a good one but couldn't think of anything trivial
mizu_no__ has joined #nixos
<yegortimoshenko>
i have a derivation where src points to a location in filesystem, and it will only rebuild if src has changed
<Lisanna>
Oh, duh
<Lisanna>
that's a good experiment :)
<yegortimoshenko>
so i will touch or chmod some file and see if it rebuilds
mizu_no__ has quit [(Read error: Connection reset by peer)]
goibhniu has joined #nixos
<yegortimoshenko>
Lisanna: touch doesn't trigger a rebuild
<Lisanna>
What about changing the name of the root of the path vs the name of one of the children?
<Lisanna>
I suspect that those might produce different results
<yegortimoshenko>
Lisanna: `chmod -w` also doesn't cause a rebuild. so i think it's just a recursive digest of the directory you point to.
<Lisanna>
I predict that changing the name of a child will trigger a rebuild, and that changing the name of the root of the path will not.
<yegortimoshenko>
Lisanna: changing root also triggers a rebuild
<yegortimoshenko>
because it's part of the artifact name in nix store
<Lisanna>
well, is that just because you also had to change your .nix file to have the different characters in it?
<Lisanna>
huh, so changing the name of ./. triggers a rebuild?
<yegortimoshenko>
if that's where your src is, yes. otherwise it won't.
<Lisanna>
that's unforuntate, that means if you have a path in an expression which is already a /nix/store path, it will create a new /nix/store path for that path when it's used
<Lisanna>
i.e. if you literally write "src = /nix/store/blah...;"
<yegortimoshenko>
only basename matters, but yes, you are right
<yegortimoshenko>
i.e. if i move the directory altogether without changing the name, it won't trigger a rebuild
<Profpatsch>
garbas: Oh lol, it really was a collab with MrRobot? Fascinating.
<Lisanna>
huh
Putti has joined #nixos
<Lisanna>
alright, and since the basnames of entries in /nix/store contain a hash, that will change every time
<yegortimoshenko>
if you point src to /nix/store, it won't change every time because /nix/store is read-only.
hamishmack has joined #nixos
crucialrhyme has joined #nixos
<Lisanna>
I mean, it will change for every "recursion" you do
<Lisanna>
sorry, didn't word that very well
<catern>
Lisanna: editing the .nix file won't always trigger a rebuild
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vboIb
<NixOS_GitHub>
nixpkgs/staging 24d81d6 Vladimír Čunát: Merge branch 'master' into staging
NixOS_GitHub has left #nixos []
<Lisanna>
catern: you're also right, yeah, I knew that x_x
<Lisanna>
only if it changes the derivation
<catern>
yes, it's pretty annoying that ./. changes if you change the location of the directory, I think there is some effort to fix tha
<catern>
t
<Lisanna>
catern: that's odd, since if they just used basename it would work
<Lisanna>
basename ./. => .
<catern>
oh sorry I meant the name of the directory
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 2 new commits to release-17.09: https://git.io/vboLc
<NixOS_GitHub>
nixpkgs/release-17.09 1b0dd72 Nikolay Amiantov: steam: add /steamrt/run.sh...
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
<betaboon>
i got aarch64 images running on several rpi3s. i guess that's why i just assumed ... assumptions ... very bad habit xD
<samueldr>
ah, then except for more building, you'll feel at home
orivej has quit [(Ping timeout: 272 seconds)]
<gchristensen>
unfortunately we don't make binaries for any ARM other than aarch64
ghostyy has quit [(Ping timeout: 264 seconds)]
<betaboon>
i went from: "home is where my unix is" to "home is where my linux is" to "home is where my shell is" to "home is where my nixos is" XD
zzamboni has joined #nixos
<samueldr>
as it should
<betaboon>
by now i realy can't understand why we havent done system deployment like this for the past 20 years ...
<samueldr>
inertia?
<betaboon>
someone should do plan9-nixos xD
<gchristensen>
betaboon: disk space wasn't cheap enough maybe
<betaboon>
notebook almost out of power, so i gotta say goodbye for today. samueldr thanks for the hint ;)
betaboon has quit [(Quit: WeeChat 1.9.1)]
Guest94605 has quit [(Quit: Lost terminal)]
<vcunat>
we go have plan9port package...
<vcunat>
s/go/do/
erictapen has quit [(Ping timeout: 256 seconds)]
crucialrhyme has quit [(Quit: WeeChat 1.9.1)]
yegortimoshenko has joined #nixos
drakonis_ has joined #nixos
drakonis has quit [(Ping timeout: 272 seconds)]
simpson has joined #nixos
<simpson>
Hi! I want to make a wrapper for a binary. The binary's named `metamath`, but I want to always run it behind `rlwrap`. Is there a correct way to set this up using makeWrapper?
<crucialrhyme>
yegortimoshenko: thanks, worked perfectly. is that the standard way of doing this? what exactly is going on?
<cmiles74>
Has anyone else had problems with automake? I'm getting an error when running aclocal: "No such file or directory at /nix/store/4ljzh46v2hp2vvk4817rk5gxixpm47c7-autoconf-2.69/share/autoconf/Autom4te/FileUtils.pm line 326"
<yegortimoshenko>
crucialrhyme: you need to call the function that produces derivation with arguments that it expects (stdenv, dependencies, so on)
<yegortimoshenko>
crucialrhyme: to do that, you should have something to call it with. so you import nixpkgs, and call package with callPackage to fill inputs
<crucialrhyme>
yegortimoshenko: and callPackage is a utlity function somewhere that just passes through... the usual set of inputs? everything in its scope? or what?
<yegortimoshenko>
crucialrhyme: everything in its scope, yes
<crucialrhyme>
yegortimoshenko: so then it must be relying on the weird introspection functions i remember skimming over
<yegortimoshenko>
last {} is overrides that you might want to set
<yegortimoshenko>
simpson: unfortunately, renaming logic is in wrapProgram, not in makeWrapper. i usually wrap such derivations in another derivation that only links executable, potentially wrapping it as in this case
<yegortimoshenko>
cmiles74: what are you trying to build?
<yegortimoshenko>
it looks for a perl module in an incorrect directory, probably some assumption made somewhere in their build system
drakonis_ has quit [(Read error: Connection reset by peer)]
babs__ has joined #nixos
mizu_no_oto has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] yegortimoshenko opened pull request #32723: Backport "steam: add /steamrt/run.sh" to 17.09 (second attempt) (release-17.09...steam/13-dec-2) https://git.io/vbo37
<Ankhers>
Can anyone tell me what the rebuild-*: 1-10 labels mean on GitHub?
<yegortimoshenko>
Ankhers: means accepting pull request will result in 1-10 packages rebuilt
<Ankhers>
yegortimoshenko: I thought so. THanks!
cement has quit [(Ping timeout: 240 seconds)]
ssmike has joined #nixos
<cmiles74>
yegortimoshenko: I'm trying to build snapd
<cmiles74>
yegortimoshenko: But that's a good point, it could be the project I'm building and not necesarily a problem with aclocal
szicari has quit [(Quit: szicari)]
simendsjo has quit [(Ping timeout: 265 seconds)]
<yegortimoshenko>
cmiles74: snapd? it doesn't seem to use autoconf, and it's written in go so it probably should be as easy to build as just running `go get ...`
<cmiles74>
yegortimoshenko: It's one of those projects that has uses autoconf (autoconf -i -f) to generate the configure script.
<yegortimoshenko>
cmiles74: is it cmd/configure.ac?
the-kenny has quit [(Ping timeout: 260 seconds)]
<cmiles74>
Yes, that's exactly right.
ghostyy has joined #nixos
<cmiles74>
yegortimoshenko: Yep, that's right. There's an aclocal.m4 file in that same directory.
sigmundv has quit [(Ping timeout: 248 seconds)]
the-kenny has joined #nixos
Putti has quit [(Ping timeout: 240 seconds)]
<yegortimoshenko>
the-kenny: new clojure-tools are in nixpkgs master
<yegortimoshenko>
cmiles74: interesting, at least their github releases don't have aclocal.m4... could you gist it?
crucialrhyme has quit [(Quit: WeeChat 1.9.1)]
<cmiles74>
yegortimoshenko: Maybe I am doing something wrong... I downloaded the source code for their latest release...
<yegortimoshenko>
it might have built aclocal.m4
<cmiles74>
yegortimoshenko: I couldn't find any build instructions in my (probably haphazard) web searching, so I'm modeling after the Arch AUR PKGBUILD file.
<cmiles74>
yegortimoshenko: yeah, it looks like calling "autoreconf -i -f" creates the aclocal.m4 file
<yegortimoshenko>
cmiles74: could you send the nix expression so far?
<iqubic>
I'm not sure why that is a thing that exists, or why I had it set.
simukis has quit [(Ping timeout: 240 seconds)]
<CrazedProgrammer>
anybody with zfs experience recommend upgrading to kernel 4.14 + zfsUnstable? currently running kernel 4.13 + zfs but there's no binary cache so i'm rebuilding the kernel on every update...
iqubic has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
ssmike has quit [(Remote host closed the connection)]