kampfschlaefer has quit [(Ping timeout: 258 seconds)]
<dash>
simpson: No :-[
<dash>
simpson: i asked this once here.
<dash>
the most I remember is that Hydra Is Different.
<gchristensen>
simpson: what is POLA?
<simpson>
Yeah, I'm just wondering how different. I mean, we now use the same kind of knot that Python and Haskell use.
<simpson>
gchristensen: The Principle of Least Authority.
<simpson>
gchristensen: A common USA phrase in security is "need to know", as in "You don't need to know that information." POLA says that agents should only have as much authority and knowledge as required to do their appointed tasks, and no more.
alx741 has joined #nixos
<gchristensen>
gotcha
filterfish_ has joined #nixos
<simpson>
What's interesting is how close we get to cap-safe semantics with Nix.
markus1189 has joined #nixos
<gchristensen>
shanemikel: I don't know what NixOS-as-mainstream looks like, but I think it is making progress. I've seen two job postings this week for nixos skills.
<simpson>
Let's pretend that we have Ideal Nix. In Ideal Nix, there's only three ways to reference the Nix store: Raw preshared paths of /nix/store/xxxx...-package, being passed `package` in a function input, or letting `package = stdenv.mkDerivation {...}`.
<simpson>
This looks almost *exactly* like a theorem in object-capability theory!
markus1199 has quit [(Ping timeout: 252 seconds)]
<gchristensen>
shanemikel: honestly, I think it makes too much sense to ignore, as a business. and not in a wishy-washy sense
<gchristensen>
shanemikel: rollbacks are _powerful_ and basically impossible with other tooling.
<gchristensen>
auditability is incredibly powerful, too
<shanemikel>
that's why I'm here!
Shou has joined #nixos
<gchristensen>
oh?
<dash>
gchristensen: a few weeks ago i discovered that nixos contributors worked at my very own employer!@
<gchristensen>
dash: where??
<dash>
mozilla
<gchristensen>
oh, yep!
<gchristensen>
I applied for a job there but never heard back :( maybe next time around ;)
deepfire has quit [(Ping timeout: 246 seconds)]
deepfire` has quit [(Ping timeout: 246 seconds)]
eacamero_ has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
systemfault has quit [(Quit: Bye!)]
digitus has quit [(Quit: digitus)]
mudri has quit [(Quit: WeeChat 1.6)]
shanemikel has quit [(Quit: My MacBook Air has gone to sleep. ZZZzzz…)]
AllanEsp1nosa has quit [(Ping timeout: 248 seconds)]
thc202 has quit [(Read error: Connection reset by peer)]
msd has quit [(Remote host closed the connection)]
shanemikel has joined #nixos
<shanemikel>
I'd love to work for Mozilla one day. Though I was a little turned off by some PNACL discourse on HN...
<shanemikel>
those FF guys are pretty unhappy about it afaict
Shou has quit [(Ping timeout: 258 seconds)]
fare__ has quit [(Ping timeout: 260 seconds)]
<simpson>
shanemikel: Well, Native Client was always a bit silly.
<simpson>
And more than a bit Googley IMO.
fare__ has joined #nixos
hugoduncan has quit [(Ping timeout: 252 seconds)]
digitalmentat has quit [(Ping timeout: 245 seconds)]
<shanemikel>
I've been pretty impressed.. from a PL point of view, it feels like the most language and browser agnostic approach, not to mention lightweight. I won't hesitate to concede I'm no expert.. ASM calls for special backend support, where for langs (without much dynamic codegen) with LLVM backends, getting pnacl going is a minor retrofit
AllanEsp1nosa has joined #nixos
<shanemikel>
the main question remaining for me: how does ASM work for high-level languages. I'm a bit skeptical about layered runtimes (or do I have the wrong idea?)
ambro718 has quit [(Ping timeout: 252 seconds)]
<dash>
shanemikel: webasm? it's basically a "get c++ game engines working in browsers" thing
<dash>
shanemikel: have you seen the conference talk "the birth and death of javascript"?
<simpson>
shanemikel: But it's woefully insecure, which should have made it a nonstarter.
<shanemikel>
that's definitely a work in progress.. I wholeheartedly disagree with Google for shipping it so early.. I think you need to treat the nacl implementation as a serious and dangerous civil engineering problem, but with the right tech and discipline... who knows. have you seen this? https://github.com/pilki/FPdNaCl
<shanemikel>
proof of concept in Coq.. I haven't had time to peruse it yet, but soon
msd has joined #nixos
AllanEsp1nosa has quit [(Ping timeout: 248 seconds)]
<simpson>
The correct discipline is object capabilities, not three sandboxes.
<LnL>
gary's talks are great!
<simpson>
shanemikel: A very nice fourth sandbox.
acertain has joined #nixos
<shanemikel>
they shouldn't be so liberal with their use of "sandbox".. they're not so much sandboxes as OS abstractions, virtual memory and so forth, right?
<nekroze>
how can I get the index of an element in a list of say, strings, given a string as input?
<shanemikel>
sure, the trampolines in the API.. but most of it is static afaict
<simpson>
Hm. It is a correct-by-construction sandbox, which is definitely great.
<NixOS_GitHub>
[nixpkgs] dermetfan opened pull request #21538: rename sound.enableMediaKeys to sound.mediaKeys.enable (master...media-keys) https://git.io/vMLjw
<shanemikel>
I'll have to read up on this cap-safe stuff you keep mentioning
<shanemikel>
and dash I'll watch that video tonight
<simpson>
I was actually in the audience in that filming. It's a great talk.
mizu_no_oto has joined #nixos
<fare__>
shanemikel, pretty cool
Wizek_ has quit [(Ping timeout: 246 seconds)]
Wizek has quit [(Ping timeout: 252 seconds)]
ryantrinkle has joined #nixos
shanemikel has quit [(Quit: My MacBook Air has gone to sleep. ZZZzzz…)]
Cutu_Chiqueno has quit [(Ping timeout: 265 seconds)]
Cutu_Chiqueno has joined #nixos
roxma has quit [(Quit: WeeChat 1.4)]
shanemikel has joined #nixos
cjhowe has joined #nixos
danharaj has joined #nixos
sigmundv has quit [(Ping timeout: 252 seconds)]
shanemikel has quit [(Quit: My MacBook Air has gone to sleep. ZZZzzz…)]
<martijn923>
Anyone familiar with the rustc nixpkgs expressions? I'm trying to build rustc with musl (goal is to compile static binaries), but the following expression ends in a failing curl invocation: http://sprunge.us/KdWE
<martijn923>
This is the last part of the output I get instantiating it: http://sprunge.us/SHEj
cjhowe has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<martijn923>
rustc compiles fine (which of course takes quite a while) but then it looks like curl is not in the path, but I don't actually expect it to be called anyway
<martijn923>
I'm now trying again with curl added to buildInputs
eacameron has joined #nixos
<LnL>
it shouldn't be trying ro download stuff during the build
eacamero_ has quit [(Read error: Connection reset by peer)]
AllanEsp1nosa has joined #nixos
shanemikel has joined #nixos
<martijn923>
Exactly...
filterfish_ has quit [(Ping timeout: 246 seconds)]
<martijn923>
Also, I have nix.useSandbox=true which iirc prohibits network access during build (but I'll find out in a few minutes I guess)
<NixOS_GitHub>
[nixpkgs] mdaiter opened pull request #21539: wxGTK30: added Darwin support (master...wxGTK30Mac) https://git.io/vMtfL
shanemikel has quit [(Quit: My MacBook Air has gone to sleep. ZZZzzz…)]
<msd>
LnL: pushed wxGTK30 Darwin patches, if you'd like to check them out
<LnL>
I'll look at them tomorrow, it's sort of late :)
<msd>
LnL: pretty much the same as wxGTK29
<msd>
LnL: sounds good! Thanks so much!
msd has quit [()]
<ToxicFrog>
Oh dear. wife has new laptop, was hoping to install Nix on it to make things easier to manage.
<ToxicFrog>
Problem: Nix doesn't support Cinnamon.
fare__ has quit [(Remote host closed the connection)]
<martijn923>
gnome3 is pretty good actually :)
<ToxicFrog>
Yeah, but she's used to Cinnamon and doesn't particularly enjoy changing DEs.
<ToxicFrog>
On the other hand, new laptop has a touchscreen, so she might want something more touch-friendly to replace it, in which case gnome3 or plasma5 are the leading contenders.
<martijn923>
indeed, change will always take some effort
xadi has quit [(Quit: Leaving.)]
cjhowe has joined #nixos
<martijn923>
well, retrying my rustc build didn't work out, it still fails on its get-stage0.py script which i think shouldn't run in nix
marusich has joined #nixos
<martijn923>
but that's for another day, bedtime for me
shanemikel has joined #nixos
cjhowe has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
shanemikel has quit [(Read error: Connection reset by peer)]
PragCyphr has quit [(Read error: Connection reset by peer)]
mizu_no_oto has quit [(Quit: ["Textual IRC Client: www.textualapp.com"])]
PragCyphr has joined #nixos
<NixOS_GitHub>
[nixpkgs] veprbl opened pull request #21541: libc++3.7: fix to use with clang 3.9 (master...libcpp37) https://git.io/vMtU8
aarvar has joined #nixos
aarvar has quit [(Quit: aarvar)]
aarvar has joined #nixos
hexagoxel has quit [(Ping timeout: 264 seconds)]
aarvar has quit [(Client Quit)]
hexagoxel has joined #nixos
<NixOS_GitHub>
[nixpkgs] grahamc pushed 6 new commits to master: https://git.io/vMtU5
<m`>
hi! is there a shorthand for builder bash scripts, which basically just contain: ''source $stdenv/setup; cp -R $src $out''
Heffalump has left #nixos []
proteusguy has joined #nixos
Shou has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
aarvar has quit [(Ping timeout: 245 seconds)]
lhart has joined #nixos
indi_ has quit [(Remote host closed the connection)]
indi_ has joined #nixos
lhart has quit [(Ping timeout: 260 seconds)]
<sphalerite>
m`: as a nix expression? Probably copyPathToStore
tvon has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vMtGo
<NixOS_GitHub>
nixpkgs/master 2dcbb8e Peter Hoeg: kirigami: add pre-release v2...
<NixOS_GitHub>
nixpkgs/master b4cab20 Michael Raskin: Merge pull request #21511 from peterhoeg/u/kirigami...
indi_ has quit [(Ping timeout: 265 seconds)]
tvon has quit [(Ping timeout: 252 seconds)]
ebzzry has quit [(Ping timeout: 248 seconds)]
aarvar has joined #nixos
thc202 has joined #nixos
m` has quit [(Ping timeout: 268 seconds)]
ertesx has joined #nixos
<gchristensen>
well just referencing a path in nix copies it to the store
ertes has quit [(Ping timeout: 265 seconds)]
ertesx is now known as ertes
ebzzry has joined #nixos
latk has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
indi_ has joined #nixos
aarvar has left #nixos []
<ToxicFrog>
Say I want to 'nixos-install' into a mounted system using additional channels. How do I do that? It doesn't find channels added to the host system with --nix-channel
<ToxicFrog>
Er, with nix-channel --add
<NixOS_GitHub>
[nixpkgs] bjornfor pushed 1 new commit to master: https://git.io/vMtZQ
<kmicu>
Then with nixos-install you should an empty system. What are those ‘host system packages’?
<ToxicFrog>
To be clear, by "host system" I mean "the currently booted LiveUSB I'm using to run nixos-install" and by "target system" I mean "the hardware it's booted on that I'm trying to install nixOS to"
Fare has joined #nixos
<ToxicFrog>
And looking at it again, no, the liveUSB isn't using unstable either; I was thinking of something else.
<LnL>
you'll want to change the version and the sha256, I usually change the sha to 000... first and build then nix will tell you what it's supposed to be
hotfuzz_ has joined #nixos
mudri has quit [(Ping timeout: 245 seconds)]
hotfuzz has quit [(Ping timeout: 268 seconds)]
uralbash has joined #nixos
uralbash has quit [(Ping timeout: 265 seconds)]
<sphalerite>
why does nix-prefetch-git output a different hash from the one that fetchFromGitHub expect?
<sphalerite>
expects*
<LnL>
fetchFromGitHub uses fetchurl not fetchgit
simendsjo has joined #nixos
<sphalerite>
yeah, but it should still fetch the same tree and that tree should still have the same hash
<hodapp>
how can I intentionally make a derivation's hash different to avoid it being confused with existing stuff in the store?
<kmicu>
Change something meaningful for hash in a derivation e.g. buildInputs order.
<MichaelRaskin>
Give it an additional unused string attribute, it will get passed in the environment without any effect, and the hash will change
<hodapp>
oh, buildInputs order matters? didn't know that
<kmicu>
Could you elaborate on ‘confused with existing stuff in the store’?
<LnL>
hodapp: you can also force a rebuild with nix-build --check
<hodapp>
to clarify: I'm working on the build for darktable 2.2.0, but also wanting to leave my working installation of darktable 2.2.0 alone, and from what I've seen, changes in postFixup and the like are not changing the hash
<MichaelRaskin>
Changes in postFixup do change the hash
<hodapp>
hm, for some reason that's not what I was seeing
<kmicu>
That does not sound right. postFixup must change hash otherwise there is not much of determinism here.
<kmicu>
Probably there is an another explanation, but w/o a minimal reproducible example it is difficult to find it ( ͡~ ͜ʖ ͡°)
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 2 new commits to master: https://git.io/vMtlk
<NixOS_GitHub>
nixpkgs/master 6305b85 Michael Raskin: Merge pull request #21549 from mimadrid/update/rofi-1.3.0...
<MichaelRaskin>
Or maybe not a minimal, but reproducible
<LnL>
urgh, remote builders are so hard to debug
<kmicu>
Minimal in a sense that we should not need to download GBs of data to reproduce it :)
<hodapp>
yes, I know what a minimal example is :P
mudri has joined #nixos
<kmicu>
sphalerite: GitHub’s release tarballs do not always correspond to git tree. Iirc we should not expect fetchurl and fetchgit to always have the same hash for the same tag/revision.
cjhowe has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
tvon has quit [(Ping timeout: 265 seconds)]
latk has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
indi_ has quit [(Remote host closed the connection)]
zagy has joined #nixos
tokudan has joined #nixos
indi_ has joined #nixos
latk has joined #nixos
* ToxicFrog
bonks head against keyboard again
<ToxicFrog>
Setting up new laptop for wife
<ToxicFrog>
Install gnome3 because it's a convertible touchscreen dealy and g3 supposedly has good touchscreen support
<ToxicFrog>
-> gnome and gdm crash whenever the touchscreen is touched
<gchristensen>
faaancy....
<ToxicFrog>
I used sddm+plasma on the liveUSB, but I figured she'd like gnome3 more and it seems to have a more touchscreen-friendly layout in any case
<ToxicFrog>
Or would if not for all the crashing
<ToxicFrog>
(this is at least not a Nix-specific bug; it was reported in ubuntu nearly a year ago)
<dmj`>
have a GTX1080 which needs v367, but nixos is providing v375
<MichaelRaskin>
Should be compatible
<MichaelRaskin>
8 versions is not that large a gap
<LnL>
GiGa: you're missing a semicolon at line 20, but apart from that you should be able to run nix-build now
<LnL>
GiGa: nix-build -A gramps
<GiGa>
LnL, How do I re-edit? I'm stuck in compare mode
<dmj`>
MichaelRaskin: just upgraded, tried to start display manager, it failed with “no screens found” and the machine immediately rebooted. nvidia-smi says it cannot communication w/ the driver
<LnL>
GiGa: not sure what you mean
<MichaelRaskin>
dmj`: have you also upgraded the kernel version?
<GiGa>
I want to go back and add that semicolon
<GiGa>
but I can't work out how to do that - sorry, I probably sound very lame!
<kmicu>
dmj`: I’m sorry, but I only wanted to help with Nixpkgs related issue. I’m not interested in proprietary drivers, but I’m also sure other folks will help you. Also consider reporting the problem on Nixpkgs repo.
* jack[m]
is using nvidia-x11-375.26-4.4.39/lib/modules/4.4.39/misc/nvidia.ko
<MichaelRaskin>
Obviously, because the interesting question is what the nvidia modules says when loaded, not what happens after you start nvidia-smi
<MichaelRaskin>
I mean, nvidia is not like nouveau which is a part of a kernel and plays well with the kernel
<MichaelRaskin>
It is a fragile thing
<MichaelRaskin>
The first thing to try (if you don't want to switch to nouveau) is using linux_4_4
simendsjo has quit [(Remote host closed the connection)]
<LnL>
GiGa: maybe it's not actually broken, otherwise that will have to be fixed first
<mpickering>
gchristensen: Perfect thanks
roconnor_ is now known as roconnor
<GiGa>
LnL, I'll be honest, I'm not entirely sure what that link is showing me. I'm assuming it's saying "if is py3k then don't get the bsddb bindings"?
tokudan has left #nixos ["Leaving"]
<dmj`>
MichaelRaskin: switching...
<LnL>
GiGa: indeed, you can try to remove that and add doCheck = (!isPy3k);
ixxie has quit [(Remote host closed the connection)]
joepie91_to_go has quit [(Ping timeout: 245 seconds)]
Fare has quit [(Ping timeout: 268 seconds)]
<stepcut>
this package has a bunch of hardcoded references to /opt/vc in the build system and to the even the .c files. And other libraries like SDL2 also have hardcoded references to /opt/vc in their ./configure scripts to deal with this. What is the right way to package this in NixOS? If the package was installed via nix-env then I could fake things with a symlink -- but that doesn't seem like it would work for nix-shell
<NixOS_GitHub>
[nixpkgs] gebner opened pull request #21554: emscriptenfastcomp: build using cmake and use cc-wrapper (master...emscripten-wrapped) https://git.io/vMt2a
hugod has quit [(Ping timeout: 260 seconds)]
[0x4A6F] has joined #nixos
sigmundv has quit [(Ping timeout: 256 seconds)]
tokyo_jesus has joined #nixos
<flxw>
a question about the expression language: the difference between "inherit e", and "inherit (e)". Is it, in the first case, the _attribute_ e is added to the lexical scope (of the inherit statement), and in the latter, that all attributes of the _attribute set_ e are added to the same scope?
<LnL>
inherit e -> e = e; and inherit (e) a b -> a = e.a; b = e.b;
<flxw>
thank you.
<flxw>
Could you point me to a place where such details are documented? I read the nixpkgs documentation and the nix manual, but I couldn't find it in either of them.
<LnL>
apparently inherit (e); is valid syntax just doesn't do anything
copumpkin has quit [(Quit: My MacBook Pro has gone to sleep. ZZZzzz…)]
<kmicu>
gchristensen: Well, first of all become a programmer and discover a marvelous fix for something, and then, when the software engineering profession really starts to take notice of you, you can jolly well tell them what to do and make sure they get everything right so there'll never be any bugs ever again. ― https://www.youtube.com/watch?v=tNfGyIW7aHM (parody inspired by your marvelous exchange yesterday with
<kmicu>
a Nix community troll ( ͡~ ͜ʖ
<kmicu>
͡°)) Happy New Year ʕノ•ᴥ•ʔノ ︵ ♥❤❣💞
joepie91_to_go has joined #nixos
<gchristensen>
kmicu: hah! thank you :D
<flxw>
LnL: Yeah, this is the manual I was refering to. I just grepped through it again, and except for a release note, the inherit (e) a b syntax gets gets not mentioned.
<flxw>
s/gets//
hotfuzz_ is now known as hotfuzz
<gchristensen>
kmicu: if only all our software was provably correct and written in Coq.
<kmicu>
gchristensen: ‘why don’t *you* fix everything? It’s so easy to create a distro w/o bugs. You should be happy that I share with you my precious un–constructive criticism. I hope this helps. Next time I expect all bugs fixed. I have expectations.’
<gchristensen>
kmicu: anything less than fully automated updates isn't worth my time. it is easy!
<gchristensen>
you should do it!
<gchristensen>
:)
<kmicu>
I’m a little bit surprised that above is tolerated. Un–constructive criticism has zero benefits and many drawbacks. But :) they call me fascist here with zero response from community so maybe that’s why ( ͡~ ͜ʖ ͡°)
<gchristensen>
I banned them before, but was asked to undo it
<gchristensen>
so I just be frank with them now
ryanartecona has quit [(Quit: ryanartecona)]
<hodapp>
huh, what's it mean when my pull request is approved but not merged?
<gchristensen>
hodapp: I think it means someone says it is good but aren't necessarily feeling ready to merge. which one?
<aszlig>
LnL: anyway, hope that helps, need to run for work now, c'ya :-)
<LnL>
yep it's working now :)
<aszlig>
cool :-)
<LnL>
hodapp: clever has a bunch of raspberries connected to a hydra
Cutu_Chiqueno has joined #nixos
<LnL>
oh --check doesn't work with remote builders?
<gchristensen>
hodapp: more cost effective I think is getting an ARM server which has like 90 cores in it
<gchristensen>
that requires up-front investment though, instead of being able to grow it slowly
Shou has joined #nixos
<MichaelRaskin>
Well, from up-front investment point of view the optimum is to improve reproducibility, improve ARM portability (so Raspberry and Cubie and a smartphone can build the same stuff) and then make a survey of unused ARM hardware…
<MichaelRaskin>
If you have bit-perfect reproducibility, you don't need to actually centralize the stuff.
<gchristensen>
I'm not understanding
cjhowe has joined #nixos
Ivanych has quit [(Quit: Leaving.)]
<simpson>
gchristensen: If builds were sufficiently reproducible and correct, then it doesn't matter *who* does the building, which means that it doesn't have to be a central authority doing the building.
<MichaelRaskin>
And then you don't need to buy any hardware, because who doesn't have a rooted smartphone that is lying unused somewhere.
<gchristensen>
sure, whatever, but you need to know what to expect / trust that output. anyway, whataver, point being with what we have -- it is more cost effective to get 1x96 core server than N raspberry pis.
<MichaelRaskin>
trust is simpler than you have 4 committers do the build and independently get the same result.
<MichaelRaskin>
But we are not there, and if you do want 32+ cores, a single server is probably more cost-effective
<gchristensen>
yes
<gchristensen>
esp. when considering space / power / all the accessories you need to support a slew of pi's
<simpson>
OTOH those multicore servers for ARM are going to sometimes only be cost-effective when it's secretly a box full of Beowulf'd Pis.
<simpson>
Or whatever we do these days instead of Beowulf.
<maurer>
Sadly denver got canned, or that might have been a good way to convince someone to donate a server :(
<gchristensen>
denver?
cjhowe has quit [(Read error: Connection reset by peer)]
<maurer>
gchristensen: Nvidia was working on ARM64 servers
<gchristensen>
ah
<maurer>
but since it was impacting their partnerships with other businesses, who saw them edging in on their turf
<maurer>
they backed off
Ivanych has joined #nixos
cjhowe has joined #nixos
<gchristensen>
anywho
NeQuissimus has quit [(Ping timeout: 248 seconds)]
<flxw>
LnL: Thank you for the other url also. I was away, but poked a little in the nix sources now. Gladly I found the definitive reference documentation of the expression language in all glory detail: src/libexpr/parser.y ;)
<toogley>
how can i disable the simultaneous download of two packages?
<gchristensen>
MichaelRaskin, simpson: I think if people independently set this up, it'd be helpful, but I think any sort of distributed package building by normal people is suuuper far off for nixos official
<Dezgeg>
maybe we could do distributed pull request testing, though
<gchristensen>
sure
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<hodapp>
gchristensen: where does one even find such an ARM server?
Shou has joined #nixos
<maurer>
hodapp: My understanding is that X-Gene is the current standard, but they basically sell the cpu + chipset on a single chip by default, and if you want a server from them you're actually buying a devkit
<maurer>
since they're mostly focusing on customers like google/facebook/msft
<hodapp>
I guess there's Scaleway too
<gchristensen>
hodapp: gigabyte sell arm servers
<gchristensen>
~6kUSD for 96 cores and a few TB storage
<maurer>
Those arem ARM32 though, yeah?
<maurer>
*are
<gchristensen>
64bit ARMv8
<maurer>
Oh, OK, sounds good
<maurer>
I was worried about getting ARM32 given that a lot of new phones are ARM64, which suggests ARM32 will probably be gone outside of embedded in the next few years
<maurer>
ertes: Yes, but you probably want to use a commit or tag for anything you want to release, since it'll confuse things if the branch updates
<hodapp>
interesting, haven't heard of packet.net
<ertes>
maurer: it's only for some local stuff
<maurer>
Yeah, then it's fine, just commenting
<ertes>
=)
<gchristensen>
hodapp: nathan7 turned me on to them. they seem good
<hodapp>
neat!
<gchristensen>
I like the hourly dedicated hardware, makes nixos testing faster.
pikajude has quit [(Quit: Quit)]
mizu_no_oto has quit [(Ping timeout: 264 seconds)]
cjhowe has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<nathan7>
note that the packet.net servers are Cavium ThunderX boxes
<nathan7>
which are 64-bit *only*
<nathan7>
(the ARM ones, that is)
pikajude has joined #nixos
<gchristensen>
speak of the devil.
<gchristensen>
yeah, same as the arm servers I was looking at.
cjhowe has joined #nixos
pikajude has quit [(Client Quit)]
pikajude has joined #nixos
cjhowe has quit [(Read error: Connection reset by peer)]
cjhowe has joined #nixos
<alphor>
how do you configure a ipv4 gateway with networking.interfaces?
mizu_no_oto has joined #nixos
<alphor>
ah, nvm. strange that you can't do gateways per interface.
<gchristensen>
I don't think gateways work that way
systemfault has quit [(Quit: Bye!)]
jazzencat has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
jazzencat has quit [(Client Quit)]
eacameron has joined #nixos
jazzencat has joined #nixos
andi- has joined #nixos
eacameron has quit [(Ping timeout: 256 seconds)]
<alphor>
is setting openssh.enable true the only thing needed to start sshd? it doesn't seem to be running under a pgrep, and all requests are being refused.
<alphor>
do I have to start that every time I rebuild my configuration?
<pikajude>
no
<pikajude>
shouldn't have to
<MichaelRaskin>
Maybe it is just socket-activated?
hiratara has joined #nixos
eacameron has quit [(Ping timeout: 245 seconds)]
eacameron has joined #nixos
<alphor>
idts, if I know what you mean by socket-activated. I tried multiple times to connect through a wireless network and ad hoc (when I was able to send packets in both scenarios). Strange that I don't have to put that in my configuration, though. Should I anyway?
<gchristensen>
alphor: whatversion? nixos-version
<alphor>
17.03.git.c311871 (Gorilla)
<LnL>
services.openssh.startWhenNeeded is not on by default
eacameron has quit [(Ping timeout: 265 seconds)]
eacameron has joined #nixos
<alphor>
^ that would be my idea of socket-activation.
eacameron has quit [(Ping timeout: 260 seconds)]
<gchristensen>
alphor: anything in journalctl -u sshd?
eacameron has joined #nixos
cjhowe has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
eacameron has quit [(Ping timeout: 256 seconds)]
eacameron has joined #nixos
tvon_ has joined #nixos
<alphor>
gchristensen: can't ssh in as root atm. I'll let you know after I eat.
<alphor>
(sudo isn't installed)
<MichaelRaskin>
ssh root login is normally prohibited
eacameron has quit [(Ping timeout: 260 seconds)]
<gchristensen>
not by default, no
<gchristensen>
fpletz: ping! :D :D :D
<jazzencat>
Okay, I'm not finding where enlightenement stores the system themes in NixOS. They are supposed to be in share/elementary/themes, but there is no elementary directory in /nix/store/*enlightenment-0.21.2/share.
<jazzencat>
How would I find the location of *enlightenment*/share/elementary/themes in the console?
<gchristensen>
pikajude: oh, man, I totally had the same question, solved it, and forgot. :(
<pikajude>
nice
hiratara has joined #nixos
<LnL>
mount -t vboxsf foo /mnt
<pikajude>
oh, you have to mount them
<pikajude>
nice
<LnL>
so you might be able to use the fileSystem options
<gchristensen>
and `deployment.virtualbox.sharedFolders`
eacameron has quit [(Ping timeout: 246 seconds)]
<pikajude>
so you use sharedFolders and pass the same name to mount as you put in sharedFolders
<pikajude>
surprisingly straightforward
<jazzencat>
c74d: That just echoes the input. The two locations I found for enlightenment, neither has the same directory structure as the one listed in the Themes setting panel when I go up a directory.
<jazzencat>
It would be nice to install a theme globally into system rather than having to put it in each individual user.
[0x4A6F] has quit [(Ping timeout: 258 seconds)]
eacameron has joined #nixos
<jazzencat>
It appears as though the path that the Enlightenement settings panel is using doesn't exist in the /nix/store/*enlightenement* directory, which is odd. It has to be somewhere or NixOS wouldn't be showing the structure.
<jazzencat>
Oh, seems that it's located in the EFL directory, not the *enlightenment* directory.
eacameron has quit [(Ping timeout: 256 seconds)]
jgertm_ has joined #nixos
mizu_no_oto has joined #nixos
eacameron has joined #nixos
<jazzencat>
And apparently read-only, so it seems that I have to add each theme I want to each user individually.
eacameron has quit [(Ping timeout: 246 seconds)]
nil has quit [(Ping timeout: 264 seconds)]
jgertm_ has quit [(Ping timeout: 258 seconds)]
<jazzencat>
Would there be any risks to changing my efl directory in /nix/store to read-write?
mindB has joined #nixos
systemfault has joined #nixos
<mindB>
Is there a quick way to tell what depends on an installed package?
nil has joined #nixos
JagaJaga has quit [(Quit: WeeChat 1.6)]
JagaJaga has joined #nixos
thc202 has quit [(Ping timeout: 245 seconds)]
jazzencat has quit [(Quit: WeeChat 1.6)]
latk has quit [(Ping timeout: 260 seconds)]
katyucha has quit [(Ping timeout: 245 seconds)]
<FRidh>
mindB: $ nix-store -q --referrers $(nix-build -A glibc)
<NixOS_GitHub>
[nixpkgs] bdimcheff opened pull request #21558: astroid: init at 0.6 (master...add-package-astroid) https://git.io/vMtXZ
<bdimcheff>
truth
takle has quit [(Ping timeout: 265 seconds)]
takle has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]