<clever>
sphalerite: you could use a modified form of my netboot config to boot them on a ramdisk, and just ignore the hdd, or use the hdd as swap only
<sphalerite>
They have 240GB hard disks
<sphalerite>
oooh using them purely as swap and having everything as tmpfs could be nice
<sphalerite>
although that might involve a lot of downloading stuff from the netboot server
<petersjt014[m]>
I posted about a thing earlier and then had to (soz).
<petersjt014[m]>
I was wondering if anyone had any working configs for minimal iso on virtualbox?
<samueldr>
clever: the issue needing a build machine for armv6 and one for armv7 stems from binfmt not discriminating between binaries, right?
ssmike has quit [(Ping timeout: 260 seconds)]
<clever>
sphalerite: its something like 200mb download to boot, and then nix has to push the whole closure over
<petersjt014[m]>
I get that one issue with grub+blocklists
<clever>
samueldr: yeah, there is no way to tell them apart at binfmt-misc
jb55 has quit [(Ping timeout: 248 seconds)]
<clever>
petersjt014[m]: what did you set the grub.rootDevice to?
davidak has quit [(Quit: Leaving.)]
<samueldr>
looking at readelf, there's Tag_CPU_arch, but I'm guessing it's not at a good location for that... but could a binfmt wrapper read that and redirect to the proper qemu?
<clever>
samueldr: its pretty deeply nested into the elf structure, a few pointers down
<sphalerite>
clever: yeah… would be nice to have the downloaded stuff cached on the disk rather than redownloading it on every boot
digitus has quit [(Quit: digitus)]
<clever>
samueldr: binfmt-misc only works on a fixed offset mask and value
<petersjt014[m]>
so is that not taken from grub.device?
<sphalerite>
clever: samueldr : a wrapper could do it though, no?
<clever>
petersjt014[m]: you can get some blocklist related errors if you set the device wrong
<petersjt014[m]>
also Is is meant to be on a a separate partition?
<sphalerite>
clever: since you already have a wrapper in the process especially…
<clever>
sphalerite: ah, the argv[0] thing could inspect every elf file, and pass the right args
<samueldr>
register hacky-wrapper to binfmt, haky-wrapper looks at that elf structure, redirects then
ssmike has joined #nixos
<sphalerite>
yeah that
<samueldr>
I'll eat and look at that
<clever>
samueldr: you know where the existing wrapper is?
<petersjt014[m]>
yeah, that one. I have it as /dev/sda, which is not working rn
<clever>
petersjt014[m]: what partition scheme is on sda?
<petersjt014[m]>
ext4
<petersjt014[m]>
which i do have included in the extra modules list too
<clever>
paritition table, not filesystem
<clever>
"blkid /dev/sda"
slack1256 has quit [(Remote host closed the connection)]
pxc2 has joined #nixos
<petersjt014[m]>
If It's supposed to tell me if it's gpt or mbr, I dont see that
<petersjt014[m]>
name, label, uuid, type
<petersjt014[m]>
is what It shows me
<clever>
petersjt014[m]: what does "fdisk -l /dev/sda" say?
DerGuteMoritz has quit [(Ping timeout: 276 seconds)]
<clever>
petersjt014[m]: and also "mount | grep sda"
Biappi has quit [(Ping timeout: 268 seconds)]
lassulus has quit [(Ping timeout: 268 seconds)]
DerGuteMoritz has joined #nixos
lassulus has joined #nixos
Biappi has joined #nixos
griff_ has quit [(Quit: griff_)]
<petersjt014[m]>
the first one is gonna get super mangled by by ocr screen copier (need to edit it first), the second says /dev/sda on /mnt type ext4 (rw,relatime,data=ordered)
<clever>
petersjt014[m]: yep, theres your problem
<clever>
petersjt014[m]: you dont have any partition tables!
<clever>
grub cant be installed on such a setup
akfp has quit [(Read error: Connection reset by peer)]
<petersjt014[m]>
wat
jensens has quit [(Ping timeout: 248 seconds)]
akfp has joined #nixos
pxc2 has quit [(Ping timeout: 240 seconds)]
<clever>
you formated sda as ext4, so there are no partition tables
<petersjt014[m]>
I can have a filesystem w/o one of those?
<clever>
you can, but grub wont install to it
SOO7 has quit [(Ping timeout: 255 seconds)]
<clever>
so it wont boot
nalc has joined #nixos
<petersjt014[m]>
so do I need a different filesystem? that seems like weird
<petersjt014[m]>
a weird option
<clever>
petersjt014[m]: you need to wipe the disk, make a partition table, and put the ext4 at sda1
jb55 has joined #nixos
<tobiasBora>
clever: samueldr Ok thanks!
<tobiasBora>
sphalerite: Interesting post! Just, when you say "Since Nix doesn’t provide binaries for ARMv7, I built it from source", you are talking about "nix", not "nixos" right?
<petersjt014[m]>
where would I do that? I used grep and cant find anything that looks right in mkfs.ext4
drakonis has joined #nixos
<sphalerite>
tobiasBora: well to be properly accurate I'm talking about cache.nixos.org/the official Hydra
<clever>
petersjt014[m]: you would create the partition tables with "fdisk /dev/sda" then format it with "mkfs.ext4 /dev/sda1"
<tobiasBora>
clever: And by the way, is it normal that it takes soooooo much time to compile? Is has not yet finished the first "./configure" !
<petersjt014[m]>
ah
<clever>
tobiasBora: fairly normal
<petersjt014[m]>
that is way simpler than i thought it would be
<petersjt014[m]>
thanks!
<clever>
petersjt014[m]: you will also need to umount the drive before you repartition it, and any data on it will be lost
<petersjt014[m]>
gotcha
<sphalerite>
tobiasBora: yeah, I think a typical bootstrap up to building hello would take a couple of hours natively, and emulation makes it a lot slower
<tobiasBora>
clever: and even with that speed, I can use it also in real life? I'll compile 30x slower than the raspberry pi...
oida has quit [(Ping timeout: 255 seconds)]
erictapen has joined #nixos
<tobiasBora>
(and actually, I guess that 30x may be even a good thing)
<gchristensen>
"copying 19 missing paths (774.90 MiB) to ..." it might be time to turn off debug builds when I deploy ofborg :D
<gchristensen>
$
<clever>
gchristensen: seperateDebugInfo = true; i think
<gchristensen>
for rust stuff?
oida has joined #nixos
<clever>
not sure how that interacts with rust
nalc has quit [(Read error: Connection reset by peer)]
erasmas has quit [(Quit: leaving)]
ylwghst has joined #nixos
<sphalerite>
tobiasBora: as I mentioned in the blog post, I'd really like to get my old phone running as a build slave
<sphalerite>
tobiasBora: I think a better solutino than throwing lots of emulated hardware at it is throwing cheap real hardware at it :p
<tobiasBora>
sphalerite: hum, it's a pretty good idea...
pareidolia has quit [(Ping timeout: 240 seconds)]
<tobiasBora>
sphalerite: and it shouldn't be that hard to do
cybrian has joined #nixos
<LnL>
the packet host could build armv7 stuff no?
<sphalerite>
LnL: I'd love one of those but that definitely doesn't qualify as cheap :D
<tobiasBora>
you can use "Linux deploy" to get a debian chroot, and then install nix on it.
<clever>
LnL: it might be aarch64 only
<sphalerite>
tobiasBora: I want build sandboxing though, which takes some kernel features which aren't enabled in android iirc
<LnL>
the host aarch64 but can't that build armv7 just like x86_64 can build i686
<clever>
LnL: yeah, but that would be cross-compiling
<sphalerite>
tobiasBora: plus I can't operate my old phone using the normal UI because the screen is shot
<clever>
LnL: so when you do "nix-env -iA nixpkgs.hello" it asks for the natively built glibc, and has to rebuild everything all over
<tobiasBora>
sphalerite: why do you mind sandboxing if it's an old phone?
<sphalerite>
tobiasBora: for build purity
<LnL>
clever: oh, then it's not the same
<sphalerite>
it's not to stop the stuff inside from getting out, it's to stop the stuff outside from getting in ;)
<clever>
LnL: oh, and also, when you do 32bit builds under nixpkgs, it runs a 32bit build of gcc, so the cpu needs 32bit support
<gchristensen>
LnL: aarch64 can't natively run armv7 instructions
<clever>
LnL: arm can be made without 32bit
<gchristensen>
in the way an x86_64 bit cpu can run i686
qmm has quit [(Ping timeout: 260 seconds)]
<clever>
gchristensen: depends on the chip
<petersjt014[m]>
for partitioning I'd wanna use gdisk, right? any reason to give mbr the time of day?
<gchristensen>
this one*
<clever>
the raspberry pi can do v6, v7, and aarch64
qmm has joined #nixos
<clever>
petersjt014[m]: fdisk can also make gpt tables
<clever>
petersjt014[m]: also, if you want to use legacy booting on GPT, you must create a bios boot partition, about 1mb in size, no fs, dont mount it
jb55 has quit [(Ping timeout: 260 seconds)]
lewo`` has quit [(Read error: Connection reset by peer)]
<tobiasBora>
By the way, I'm wondering by it's so hard to do efficient cross compiling. For me there is no link between the host language, and the compilation output. For example, why can't we imagine a program that is like gcc, armv6 version, except that all instructions are translated into 64 bits ? A program like "armv6_to_x86_64", that just do elf conversion
<clever>
tobiasBora: nix doesnt understand that part though
<clever>
tobiasBora: and there are impurities that would leak in if you tried to force it
<clever>
ive even caused such impurities with qemu-user
<sphalerite>
tobiasBora: normal cross-compilation is much faster, it's just that cross-compiled and natively-compiled stuff won't mix well
<clever>
if you use a dynamicaly linked qemu-user, then the arm ldd will spit out x86 libraries
<clever>
resulting in x86 libraries in your arm initrd!
<sphalerite>
on the nix side, that is
<clever>
yeah
<tobiasBora>
so the "hard part" comes from the libraries?
<tobiasBora>
And why cross-compiled and natively-compiled stuff don't mix well?
<clever>
more that nix doesnt understand 2 different compilers being capable of producing the same output
<clever>
so nix treats each compiler as a seperate task, producing different outputs
<tobiasBora>
and it's not possible to let nix think that the compiler is the original one, but in fact we just replaced the exec file?
ssmike has quit [(Ping timeout: 260 seconds)]
<andi->
argh... RPi3 with nixos wont boot anymore (so no music before bed time), spent another evening trying to cleanup some mess (looking at you unix.h), emptied a bottle of local wine… still loving NixOS anyway <3
<clever>
tobiasBora: basically, in nix, you have a set of environment variables, 2 of them being special, $builder and $args
<sphalerite>
something something intensional store
<clever>
tobiasBora: nix will then hash that entire set of builders, and use it to compute $out's path
<clever>
tobiasBora: then to make $out, it runs $builder with $args
<clever>
tobiasBora: if you change the gcc from a native to a cross-compiler, that affects the hashes in strings, and changes the $out
<tobiasBora>
clever: the hash is on the name, not the real file no ?
<clever>
the hash is in the storepath
<clever>
so /nix/store/hash1-gcc and /nix/store/hash2-gcc
<tobiasBora>
ok. And now, let's do a ugly hack. I manually put the content of /nix/store/hash2-gcc into /nix/store/hash1-gcc. Then nix will compile with the cross-compiler, and think that he is compiling with the native one no?
<samueldr>
tobiasBora: on my armv7 board, a kernel build takes around 17h, on my E5-1660 (6 cores 12 HT) kernel takes around 4h to build using qemu-user
<clever>
tobiasBora: all paths in the store are immutable, and must never be modified once built
<samueldr>
pretty sure that your builds will be faster than on the raspberry pi
<clever>
tobiasBora: and the whole design of nix relies on paths having been made by the script that was hashed to make the path
ilja_kuklic has quit [(Ping timeout: 248 seconds)]
<tobiasBora>
samueldr: interesting... The thing is that in 1h, the ./configure of the first tool (bootstrap) has not even been finished... And it's the ./configure, not even the make! It takes like 6 seconds for each line like "checking whether printf supports infinite 'double' arguments... yes"
<LnL>
that would break caching
yegortimoshenko has joined #nixos
<tobiasBora>
clever: Ok I understand better now. But if as an experimentation I just try to do that, it could *in theory* work? And does that mean that cross-compiling will never be usable in nix? :'( It's so sad to have such a powerful build system, and not being able to do cross compiling.
pxc2 has joined #nixos
<clever>
tobiasBora: you would also need to swap in all the libraries the gcc would be linking against
heath has joined #nixos
<tobiasBora>
clever: ok thanks.
<clever>
tobiasBora: oh, and swapping those libraries out will also break anything refering to them
ssmike has joined #nixos
<sphalerite>
tobiasBora: cross-compilation can still be used very effectively with nix in principle if you don't need to natively compile stuff as well.
jb55 has joined #nixos
<sphalerite>
tobiasBora: you could use nix with cross compilation to build system images for embedded devices very well for instance
<sphalerite>
tobiasBora: this is in principle, I think nixpkgs's support for cross-compilation is still very much WIP, but Sonarpulse has been doing some awesome work on that
cybrian has quit [(Read error: Connection reset by peer)]
<Sonarpulse>
sphalerite: summoned
<Sonarpulse>
tobiasBora: what wants to be done?
<sphalerite>
There's also the pie-in-the-sky intensional store stuff which may magically fix the issue if it ever happens, idk
<Sonarpulse>
converting architecture post compilation?
<Sonarpulse>
looked like first question
<tobiasBora>
Sonarpulse: I'm trying to understand if it may be possible, one day, to generate a binary cache for nix using cross compiling.
<Sonarpulse>
oh, sure
<clever>
2017-11-29 12:27:08 * dtzWill is using allexe-built terminal/shell/ssh/etc. to chat presently, many of which are JIT'd just because I can xD
<clever>
this is about using llvm IR for all applications
<clever>
and then specializing it to the cpu at runtime
<sphalerite>
tobiasBora: that's already possible. It just won't be compatible with natively compiled stuff
<Sonarpulse>
yeah
<sphalerite>
oh crap it's almost 1am
<Sonarpulse>
:D
<Sonarpulse>
I just headed west
<tobiasBora>
sphalerite: so you mean that either your raspberry pi will use *only* stuff compiled by a cross compiler, or *only* stuff compiled by a native armv6?
<clever>
sphalerite: ive said that about 7am a few times this week :P
<LnL>
exactly, nix makes the distinction which is arguably correct since one of the variants might not work, etc.
<Sonarpulse>
so now I feel very mature; it's *earlier* than it feels for once!
<Sonarpulse>
e.g. headers encode wrong bash
<Sonarpulse>
the build time bash
<Sonarpulse>
could sneak in dev output
<Sonarpulse>
(not actually in header files)
<sphalerite>
tobiasBora: it can use both, but you'll essentially have two completely different dependency trees, and it will take up twice the space correspondingly
<Sonarpulse>
we could do a trust thing
<tobiasBora>
sphalerite: ok yes, that what I thought after asking the question
mrkgnao has joined #nixos
<Sonarpulse>
"beleive me, the output may not be at all fixed, but these two very different derivations produce the same thing"
<Sonarpulse>
intensional store does make that nicer to implement
<Sonarpulse>
but it could be hacked today with nars/substitutors
srdqty has quit [(Quit: WeeChat 1.9.1)]
jb55 has quit [(Ping timeout: 248 seconds)]
<sphalerite>
what confused me a great deal is having two different derivations that produced the same output
<Sonarpulse>
we technically do have that already with fixed output
ssmike has quit [(Ping timeout: 260 seconds)]
<tobiasBora>
Do you have any reference on how to use/build a tool that first try to use the already available binary caches, then if not available try to ask to a cross-compiler in the network to build an equivalent program, and if no other cross-compiler is present, then it tries to compile it by itself?
<sphalerite>
(fixed-output ones that downloaded sources; I copied an ARM one to my main laptop)
<sphalerite>
yeah
<Sonarpulse>
the "I changed URL but forgot to update hash and nix didn't do anything oh no!!"
<LnL>
yeah that and different nix versions might generate different drv files
<sphalerite>
tobiasBora: yes, that tool is called nix
<sphalerite>
:D
<sphalerite>
wait, except for the cross-compiler bit
dtzWill has quit [(Ping timeout: 252 seconds)]
<sphalerite>
right I really need to sleep though. Gnight all!
<Sonarpulse>
good night!
<Sonarpulse>
tobiasBora: look up nix substitutors
<tobiasBora>
good night, thanks for your help!
<Sonarpulse>
/ ask shlevy about the new cache C++ caches
<Sonarpulse>
it isn't hard, just requires deep knowledge of the guts of Nix's *implementation* (as opposed to concept/ interface) to do today
<Sonarpulse>
*new Cache/store C++ class
<Sonarpulse>
I recall some disparate concepts were combined together
<gchristensen>
is there as easy way to see closure size of a given path?
<clever>
gchristensen: du -hc $(nix-store -qR path)
<clever>
maybe |sort -h at the end
ssmike has joined #nixos
<gchristensen>
671M :(
<clever>
gchristensen: the above will also show what the fat deps
<clever>
are
<gchristensen>
heh
<gchristensen>
mkRustCrate never sheds dependencies
<tobiasBora>
Sonarpulse: so if you need deep knowledge of the Nix's implementation, it may be considered as hard ^^ And I don't see lot's of references to nix substitutors, I just feel like people use substitutors instead of "binary cache" in the ML ^^'
<Sonarpulse>
tobiasBora: sure it's hard, just not a lot of code to be changed
<Sonarpulse>
that's what i meant
dan_b has quit [(Ping timeout: 260 seconds)]
<Sonarpulse>
"substitutors" i think at one point was it's own concept
<Sonarpulse>
then binary cache was the main type of substitutor
<Sonarpulse>
then people decided binary cache was the only good notion of "substitutor"
<Sonarpulse>
and the binary cache just became "remote" store, as opposed to local store
<Sonarpulse>
in other words, a bunch of implementation rejiggering
<tobiasBora>
by the way, any program should "cross-buildable", or some of them can fail?
<tobiasBora>
s/should /should be /
thc202 has quit [(Ping timeout: 248 seconds)]
griff_ has joined #nixos
<dhess>
Sonarpulse: is the plan for Nixpkgs cross-compiling support to do something crazy for TemplateHaskell like running it in a QEMU VM for the host architecture? Or is TH just not going to work?
<samueldr>
hmm, anyone has pointers (eh) for a documented-enough elf library that includes the arm attributes?
<Sonarpulse>
dhess: angerman has out of process th working
<samueldr>
elfutils/libelf seems to implement everything... but I can't seem to find any documentation
<clever>
dhess: in theory, qemu-user can be used as an interperter for the TH daemon, the same way nodejs handles ghcjs TH
<Sonarpulse>
you can use real external server!
<clever>
dhess: and then you can do cross-compiled TH without binfmt-misc
<Sonarpulse>
but no idea how that will work with nix's sandboxing
<Sonarpulse>
the permenant solution though
<Sonarpulse>
is to just....stop making template haskell need the target platform
<Sonarpulse>
the semantics of the host platform are more correct anyways
<clever>
Sonarpulse: ghcjs uses the out of process th under nodejs to deal with this issue
<clever>
Sonarpulse: half the problem, is that TH can access modules you import, and what platform did ghc compile those modules for?
<Sonarpulse>
clever: yeah TH is wrong :p
<Sonarpulse>
$(import This.Is.How.It.Should.Work)
<Sonarpulse>
(in practice $import is probably better)
<dhess>
Sonarpulse: sure, that's a disastrous decision in retrospect, but I don't expect it to change anytime soon.
<Sonarpulse>
dhess: I don't think anyone is actually against that
<clever>
Sonarpulse: you still need to have a host build of that module available
<dhess>
anyway if Nixpkgs can handle it through herculean efforts, it would be awesome
<Sonarpulse>
clever: absolutely
<clever>
Sonarpulse: which means building all of the deps twice
<Sonarpulse>
sure
<Sonarpulse>
well not *all* the deps
<dhess>
Sonarpulse: who knows? So many people using it and it would be a pretty major breaking change.
<clever>
at that point, its simpler to just build every module twice, and transparently pick the right variant
<johnw>
I just started getting "cycle detected" when building the documentation of Haskell packages
<Sonarpulse>
because you will have th-specific deps in cabal file
<clever>
Sonarpulse: which is reminding me of how apple handles binaries
<johnw>
I'll have to git bisect to find the culprit
<clever>
Sonarpulse: apple can have multiple arches in the same binary, and the runtime linker picks the right one
<Sonarpulse>
indeed that is similar
<Sonarpulse>
basically my approach to nixpkgs was "fuck it; do the right thing at any cost"
<Sonarpulse>
....next up GHC!
<clever>
lol
<Sonarpulse>
but i need to really finish nixpkgs first
<Sonarpulse>
which is close
<joepie91>
I might be slightly crazy, but I'm experimentally designing a language that's loosely based off Nix...
<joepie91>
an embeddable one :P
<joepie91>
(also steals some concepts from Rust and JS)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vbGTq
<NixOS_GitHub>
nixpkgs/master 5490e58 Tuomas Tynkkynen: v8: Fix aarch64 build...
NixOS_GitHub has left #nixos []
cybrian has joined #nixos
yegortimoshenko has quit [(Remote host closed the connection)]
nalc has joined #nixos
<Sonarpulse>
joepie91: you had me until JS :p
pxc2 has quit [(Ping timeout: 248 seconds)]
Alchemical has quit [(Ping timeout: 268 seconds)]
<joepie91>
Sonarpulse: what I'm stealing from JS is actually a feature that isn't *in* JS yet :P
<joepie91>
(bind operator)
<joepie91>
it's still a stage-not-enough proposal
m0rphism has quit [(Ping timeout: 268 seconds)]
srk has quit [(Ping timeout: 268 seconds)]
<joepie91>
no idea when or if it will make it into the language
flyx|znc has joined #nixos
<joepie91>
I don't think I'm using anything else that's really obviously part of JS
nwspk has quit [(Ping timeout: 276 seconds)]
flyx has quit [(Ping timeout: 258 seconds)]
flyx|znc is now known as flyx
m0rphism has joined #nixos
<joepie91>
not yet, at least
Alchemical has joined #nixos
m0rphism is now known as Guest47064
Wizek_ has quit [(Quit: Leaving)]
<clever>
joepie91: i always use .bind to deal with that, yet ive delt with libraries before, that always took 2 arguments for every callback, the func, and the 'this'
nwspk has joined #nixos
<joepie91>
clever: mind that 'bind operator' in this context means something like `foo::bar(baz)` translating to `bar(foo, baz)` :P
<joepie91>
which is a handy way to specify object-ish behaviour out of band
srk has joined #nixos
<joepie91>
without messing about with prototype mutations (in the case of JS)
erictapen has quit [(Remote host closed the connection)]
<joepie91>
hm
<joepie91>
I think that in the case it actually sets the thing before :: as the context (`this`) but I'm not 100% sure on that
<clever>
joepie91: i was thinking more along the lines of bar.bind(foo) causing 'this' to point to foo when the returned function is ran
<joepie91>
been a while since I actually read the proposal
<joepie91>
clever: ah yeah, no, that's not what the bind operator does
<clever>
joepie91: and it can also prefix arguments into the function call
<joepie91>
well it might be what it does behind the scenes, but that's not its purpose
<joepie91>
sec
<clever>
which roughly gives you partial function application
<joepie91>
clever: yeah, for partial function application I just use .bind where necessary
<joepie91>
anyway, that repo readme has some examples of the bind operator
<joepie91>
as it is currently proposed
<joepie91>
in my hypothetical language I don't have a concept of a special `this` variable so it's just passed as the first argument
<clever>
behind the scenes, thats pretty much what c++ did to c
<jeaye>
So we have foo.bind(f) (aka std::bind), monadic bind, and now some new other bind?
<clever>
the compiler just invisibly treats that special object syntax as an extra arg
<joepie91>
clever: anyway, I'm still experimenting with it, but in case you're curious what I've come up with so far: https://gist.github.com/joepie91/7d462acf74fd552103f48e8cca28f10a (context: imagine a HTTPd like nginx or Apache that allows specifying 'request hooks' using this embedded language, and that gist is an example of a 'require password for a certain directory and hostname' implementation)
<clever>
joepie91: at that point your basically just rewriting php, lol
<joepie91>
clever: no, this is just one example scenario - you might also think of typical cases where Lua is currently used like game scripting
<joepie91>
it's meant as a general-purpose embeddable language where the builtins are provided by the embedding application, similar to how Lua is used
<angerman>
clever, dhess, Sonarpulse: multi-target ghc will eventaully happen.
<joepie91>
clever: a HTTPd request hook is just one of the simpler cases to try out that anybody can understand (unlike game scripting, which is often incredibly specific)
<joepie91>
so that's what I started with :P
<clever>
joepie91: when i got into nix, i was considering a game engine as an extension of nix, lol
<clever>
joepie91: using the nix store to cache pure operations
<joepie91>
heh
<joepie91>
I feel like that may not end well :P
<joepie91>
this embeddable language idea is actually a result of me liking the general ideas behind Nix (as a language), but being irritated with some of the quirks and realizing that it's really very specifically designed for NixOS-like scenarios
<joepie91>
so I guess this is my closest thing to "using Nix for things it wasn't made for", except it'd be a variant
<joepie91>
will probably take forever to actually get it to a usable point, but hey, fun project to have
mizu_no_oto has quit [(Quit: ["Textual IRC Client: www.textualapp.com"])]
pxc2 has joined #nixos
cement has quit [(Ping timeout: 260 seconds)]
<dhess>
angerman: of course, for some value of "eventually" :)
<angerman>
dhess: well. if you trust me persuing this :-)
<dhess>
angerman: Are you zw3rk on Medium?
pxc2 has quit [(Ping timeout: 276 seconds)]
simpson has left #nixos ["WeeChat 1.0.1"]
<angerman>
Yes, that’s me
<dhess>
angerman: then I trust you more than anybody else, at least! :D
fiddlerwoaroof has quit [(Read error: Connection reset by peer)]
<angerman>
Ha ;-)
<angerman>
multi target ghc will not automatically give us peoper TH support though. And as TH stands the out of process approach is the only one that can give the correct result in every case.
<angerman>
However TH is like a sledgehammer used for a lot of stitching.
fiddlerwoaroof has joined #nixos
<dhess>
sure, I agree with all the dont-use-TH/TH needs to change/etc sentiment. I just don't think it's very realistic.
<clever>
angerman: ive had projects fail to build because the TH cant find the git executable at compile time
<angerman>
Most of the TH use cases I came across are just AST expansions.
<dhess>
but the Haskell community is probably as agile as any so maybe it can happen sooner than later.
<clever>
angerman: yeah, those are better ones, things like acid state and lenses
<angerman>
dhess: it’s not realistically. If it was, I wouldn’t try to get a “workable” solution.
<dhess>
clever: wow that's a bad one.
<angerman>
clever: yes. For process logic we need the open PR to be integrated into ghc.
Sonarpulse has quit [(Ping timeout: 240 seconds)]
<dhess>
I can't even imagine trying to use TH for something like that. What a bastardization of the intent
<angerman>
So the gradual progress will be to be more explicit about build/host IO in TH.
<dhess>
angerman: anyway that is not to denigrate your work. Thank you so much for doing all of this.
{^_^} has joined #nixos
{^_^} has quit [(Changing host)]
{^_^} has joined #nixos
<dhess>
I look forward to being able to cross-compile for my ARM targets :D
<dhess>
because even on a decent aarch64 box my Haskell builds take sooooooo loooooong.
<angerman>
The next step might be trying to figure out how to restrict TH so that we know a splice is architecture/is independent. (Pure ast expansion)
<dhess>
god I have to bump nixpkgs again soon and I am so not looking forward to it.
mbrgm has quit [(Ping timeout: 260 seconds)]
ma27 has quit [(Ping timeout: 255 seconds)]
<dhess>
angerman: yeah I would imagine that will cover like 99.99% of all cases!
<angerman>
dhess: i might have a fresh set of prebuilt relocatable ghc cross compiler binary distributions ready by the weekend
<dhess>
angerman: wow, cool! With a nix expression?
<dhess>
oh "prebuilt".
<angerman>
No
<dhess>
Yeah my whole build system is built around Nix so that won't help me for now.
<angerman>
But they basically built with my hadrian fork and my ghc fork.
<angerman>
dhess: nix can use binary blobs no?
<dhess>
I guess I could try to package it. I did that for the Debian debs to get a GHC for ARMv7/v8
<dhess>
angerman: sure, just a bit of work. Probably worth it for something like this. But then, it's not just ghc that needs work to get cross-compiling going on Nixpkgs
mbrgm has joined #nixos
<dhess>
anyway it seems like that might be a lot better soon-ish.
laggy_wifi has joined #nixos
<angerman>
Well at least you could use the binary dists to see what issues with you TH come up ;-)
<clever>
Wizek: i think it translates to the cabal flag --enable-library-profiling
<Wizek>
clever: so this flag is specific to haskell from what I see in that file, right?
<clever>
yeah
pxc2 has joined #nixos
lambdamu_ has joined #nixos
nuncanada has quit [(Read error: Connection reset by peer)]
nuncanada has joined #nixos
lambdamu has quit [(Ping timeout: 248 seconds)]
dieggsy has quit [(Quit: ERC (IRC client for Emacs 27.0.50))]
pxc2 has quit [(Ping timeout: 248 seconds)]
akfp has quit [(Read error: Connection reset by peer)]
nuncanada2 has joined #nixos
akfp has joined #nixos
nuncanada has quit [(Ping timeout: 255 seconds)]
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
<Wizek>
clever: okay, thanks.
jsgrant has joined #nixos
vidbina has joined #nixos
jereem1 has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
vidbina has quit [(Ping timeout: 248 seconds)]
vidbina has joined #nixos
jb55 has joined #nixos
nuncanada2 has quit [(Read error: Connection reset by peer)]
nuncanada2 has joined #nixos
vidbina has left #nixos []
griff_ has quit [(Quit: griff_)]
vidbina has joined #nixos
laggy_wifi has quit [(Remote host closed the connection)]
pxc2 has joined #nixos
laggy_wifi has joined #nixos
vidbina has quit [(Ping timeout: 260 seconds)]
vidbina has joined #nixos
pxc2 has quit [(Ping timeout: 248 seconds)]
griff_ has joined #nixos
nuncanada2 has quit [(Quit: Leaving)]
ylwghst has quit [(Ping timeout: 260 seconds)]
<kini>
in https://github.com/NixOS/nixpkgs/commit/0094ca20cc I changed the package "acl2-v6-5" to the package "acl2-7.4", but since the format of the version number has changed, users who had installed the former won't be able to upgrade to the latter, right?
<kini>
if I'm not mistaken, in "acl2-v6-5", the package name is considered to be "acl2-v6" and the version number is considered to be "5"
<kini>
what's the normal procedure for renaming a package in such a way that users will be able to upgrade across the name change?
danl1240 has quit [(Quit: My iMac has gone to sleep. ZZZzzz…)]
dieggsy has joined #nixos
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
nalc has joined #nixos
dieggsy has quit [(Quit: ERC (IRC client for Emacs 27.0.50))]
mizu_no_oto has joined #nixos
pie_ has quit [(Ping timeout: 255 seconds)]
babyflakes has quit [(Quit: Connection closed for inactivity)]
dieggsy has joined #nixos
dieggsy has quit [(Remote host closed the connection)]
<Havvy>
I'm specifically working on the rust compiler.
<clever>
Havvy: what about "nix-shell '<nixpkgs>' -A rustc" ?
ylwghst has joined #nixos
<Havvy>
Gonna pull in the wrong source and bootstrap, but I can try.
<Havvy>
And that seems to be working though.
endformationage has quit [(Quit: WeeChat 1.9.1)]
<Havvy>
Nope, it also failed.
ylwghst has quit [(Ping timeout: 276 seconds)]
<clever>
same or different error?
<Havvy>
Same error.
<clever>
cant think of anything else then
<Havvy>
It's not even showing the "simple program" it's failing on anywhere. >_>
Harpalus has joined #nixos
fiddlerwoaroof has quit [(Ping timeout: 260 seconds)]
adamt has joined #nixos
adamt has quit [(Changing host)]
adamt has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
athan has joined #nixos
ma27 has joined #nixos
adamt has quit [(Ping timeout: 268 seconds)]
ma27 has quit [(Ping timeout: 255 seconds)]
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
Harpalus has quit [(Quit: Leaving)]
leat has quit [(Quit: WeeChat 1.9.1)]
kreetx has joined #nixos
reinzelmann has joined #nixos
asuryawanshi has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
griff_ has quit [(Quit: griff_)]
<fearlessKim[m]>
when does nixops create the users ? I can't diagnostic why I don't have network access in Qemu if I can't login, and I am not sure why my users are not created :'(
pie_ has quit [(Ping timeout: 260 seconds)]
dieggsy has quit [(Ping timeout: 255 seconds)]
<lejonet>
adisbladis: nope, not yet
cybrian has joined #nixos
orivej has joined #nixos
asuryawanshi has quit [(Remote host closed the connection)]
laggy_wifi has quit [(Ping timeout: 260 seconds)]
pie_ has joined #nixos
leat has joined #nixos
jb55 has quit [(Ping timeout: 248 seconds)]
cybrian has quit [(Read error: Connection reset by peer)]
lewo``` has quit [(Read error: Connection reset by peer)]
lewo has joined #nixos
oahong has quit [(Ping timeout: 252 seconds)]
asuryawanshi has joined #nixos
<wmertens[m]>
I'm not sure how to proceed with https://github.com/NixOS/nixpkgs/issues/31611 - if you use the sudo-via-sshagent setting, sudo authentication checking can be entirely circumvented. The easiest fix is a breaking change, and nobody is responding to the issue. Do I just merge the breaking change?
oahong has joined #nixos
jasom has quit [(Ping timeout: 240 seconds)]
georges-duperon has joined #nixos
asuryawanshi has quit [(Ping timeout: 260 seconds)]
jmeredith has quit [(Quit: Connection closed for inactivity)]
kreetx has quit [(Ping timeout: 240 seconds)]
digitalmentat has quit [(Quit: Leaving)]
reinhardt has joined #nixos
reinhardt has quit [(Remote host closed the connection)]
freusque has joined #nixos
ssmike has joined #nixos
ylwghst has joined #nixos
asuryawanshi has joined #nixos
justbeingglad has joined #nixos
justbeingglad has left #nixos []
ylwghst has quit [(Ping timeout: 248 seconds)]
civodul has joined #nixos
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
das-g[m] has left #nixos ["User left"]
das-g[m] has joined #nixos
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
asuryawanshi has quit [(Remote host closed the connection)]
<srhb>
Can anyone test this out on a recent unstable that has a fqdn? Look up the fqdn with hostname -f. Does the domain appear? Then, try disabling services.nscd.enable and rerunning hostname -f. Does the domain name appear now?
<srhb>
"has a fqdn" => networking.domain is set
jsgrant has quit [(Remote host closed the connection)]
Isorkin_ has joined #nixos
jsgrant has joined #nixos
georges-duperon has quit [(Ping timeout: 260 seconds)]
Isorkin has quit [(Ping timeout: 255 seconds)]
johann__ has joined #nixos
<sphalerite>
clever gchristensen: I usually add -d0 to du when measuring closure size to keep the output a manageable size
i-am-the-slime has quit [(Ping timeout: 260 seconds)]
zzamboni has joined #nixos
tusj has quit [(Remote host closed the connection)]
dan_b has joined #nixos
<sphalerite>
srhb: is there a difference between setting networking.hostname to a FQDN and setting networking.domain?
ssmike has quit [(Ping timeout: 255 seconds)]
notdaniel has joined #nixos
cybrian has joined #nixos
Stammon has joined #nixos
<sphalerite>
because I've done the former on a couple of machines and could test it on there
zzamboni has quit [(Quit: Leaving.)]
freusque has quit [(Ping timeout: 255 seconds)]
<Stammon>
My Polybar, which is started with home-manager, complains, that when executing a userscript,it can't find any interpreter. E.g: /usr/bin/env sh file not found. I assume, that sh is not in it's path variable. How can I change that?
<srhb>
sphalerite: Debian-likes tend to say setting hostname to fqdn is wrong, while red hat-likes tend to say that's the right way.
<srhb>
sphalerite: That's as far as my Google-fu has brought me.
<sphalerite>
srhb: huh ok… I had no idea
<srhb>
sphalerite: There's also some services that break on NixOS if you use fqdn for your hostname. avahi comes to mind.
<srhb>
sphalerite: On the other hand, as far as I can tell, networking.domain is broken whenever nscd is enabled (which it is by default)
<sphalerite>
that is, I didn't even know there was a distinct setting for domain until you mentioned it earlier
hotfuzz_ has joined #nixos
asuryawanshi has joined #nixos
<srhb>
sphalerite: Setting networking.domain will (at least) set the sysctl kernel.domainname correctly.
<srhb>
(I also discovered that that's impure, but meh)
<srhb>
That is, if you go from domain unset to domain and back to unset, it will persist to the set version.
<srhb>
At least until you reboot.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti opened pull request #32333: Revert "Revive multiple outputs for Haskell packages." (master...revert-32112-hask-split-out) https://git.io/vbGBL
NixOS_GitHub has left #nixos []
<srhb>
I can't find heads or tails in this. I need a networking course. :-P
<sphalerite>
Stammon: how are you enabling polybar?
asuryawanshi has quit [(Remote host closed the connection)]
<sphalerite>
eeeh it uses a systemd user service? something something systemd shouldn't manage X applications
<sphalerite>
anyway
asuryawanshi has joined #nixos
<sphalerite>
you can probably get it to work by setting systemd.user.services.polybar.Service.Environment
<MichaelRaskin>
sphalerite: if you accept systemd, systemd will manage everything.
<sphalerite>
MichaelRaskin: no this is systemd's own position iirc
zzamboni has quit [(Ping timeout: 260 seconds)]
zzamboni has joined #nixos
<MichaelRaskin>
I guess the question is not whether you remember it correctly, but whether it has been updated.
<sphalerite>
because the systemd user daemon has a lifetime independent of (longer than) the X session's, and so it can't manage stuff within the X session well. Or something.
cybrian has quit [(Read error: Connection reset by peer)]
<Stammon>
maybe I should start the polybar somewhere in my X config instead?
<sphalerite>
Stammon: yeah I'd put it in my .xsession script. Not sure how home-manager handles that stuff though
<MichaelRaskin>
Well, given that systemd does track the lifetime of X session as it will start a new one when the current one ends…
<Stammon>
There are other issues as well like it isn't able to list all my i3 workspaces. I'm not sure whether the right environment is enough to fix this.
<Stammon>
there is xsession.initExtra which executes arbitrary commands
<sphalerite>
Stammon: yeah I'd put it there, or maybe in your i3 config
<sphalerite>
I'd leave that until you've fixed the actual issue though
<Stammon>
I had it there previously, but then I wanted to manage it with home-manager.
cybrian has joined #nixos
vidbina has joined #nixos
<Stammon>
The actual issue?
<sphalerite>
the PATH one
<sphalerite>
oh actually it's a systemd-specific solution… idk
zzamboni has quit [(Ping timeout: 258 seconds)]
zzamboni has joined #nixos
<Stammon>
so the i3 solution works as expected, it's somewhat unclean though. Maybe I should fix the package instead
<Stammon>
I have no experience in writing nixpackages. I'm not sure whether this is a big start or not
<sphalerite>
in this case it's a module you want to fix, not a package
<tpanum>
How do I get foomatic printing drivers on NixOS? Adding foomatic-filters to services.printing.drivers does not seem to work :/
akfp has quit [(Read error: Connection reset by peer)]
Dr8128 has joined #nixos
akfp has joined #nixos
tpanum has quit [(Remote host closed the connection)]
davidak has joined #nixos
zzamboni has quit [(Ping timeout: 240 seconds)]
ilja_kuklic has joined #nixos
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
zzamboni has joined #nixos
townsend has joined #nixos
<srhb>
Is there an unlines-analogue somewhere? ([String] -> String)
<srhb>
Ah, concatMapStringsSep will do.
<tilpner>
srhb - concatStringsSep "\n"
<srhb>
tilpner: Thanks.
zzamboni has quit [(Client Quit)]
i-am-the-slime has quit [(Ping timeout: 248 seconds)]
FRidh has quit [(Remote host closed the connection)]
vidbina has quit [(Ping timeout: 255 seconds)]
zzamboni has joined #nixos
zzamboni has quit [(Client Quit)]
zzamboni has joined #nixos
tpanum has joined #nixos
i-am-the-slime has joined #nixos
ylwghst has joined #nixos
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
FRidh has joined #nixos
<magnetophon>
how can I tell if a pkg needs fftw, fftwFloat, or one of the others?
laggy_wifi has joined #nixos
ylwghst has quit [(Ping timeout: 240 seconds)]
freusque has joined #nixos
<tpanum>
I simply cannot get foomatic drivers to show up in CUPS, when I add it to services.printing.drivers. Anybody has a clue why that is? :/
joshuaks has joined #nixos
vidbina has joined #nixos
<sphalerite>
Li: with nixUnstable you can also use nix-build --hash
thetet has joined #nixos
i-am-the-slime has quit [(Ping timeout: 240 seconds)]
joshuaks has quit [(Ping timeout: 260 seconds)]
ssmike has joined #nixos
phreedom has quit [(Ping timeout: 255 seconds)]
zzamboni has quit [(Quit: Leaving.)]
zzamboni has joined #nixos
phreedom has joined #nixos
zzamboni has quit [(Client Quit)]
zzamboni has joined #nixos
i-am-the-slime has joined #nixos
<Li[m]>
Linus: thx
vidbina has quit [(Ping timeout: 260 seconds)]
vidbina has joined #nixos
Cheng has joined #nixos
Avaq has quit [(Quit: Leaving.)]
<Cheng>
Hi, how do I get a <nixpkgs> with my /etc/nixos/configuration loaded as the config?
<Cheng>
I know 'with import <nixpkgs> {}; something', but how do I get /etc/nixos/configuration.nix loaded as the config?
<Cheng>
I tried 'with import <nixpkgs> { config = import /etc/nixos/configuration.nix; } something', but that doesn't work
cybrian has joined #nixos
oahong has quit [(Ping timeout: 255 seconds)]
ma27 has joined #nixos
<tpanum>
b
tpanum has left #nixos ["ERC (IRC client for Emacs 25.3.1)"]
oahong has joined #nixos
johann__ has quit [(Quit: Leaving.)]
johann__ has joined #nixos
zzamboni has quit [(Quit: Leaving.)]
phdoerfler has joined #nixos
sigmundv__ has joined #nixos
zzamboni has joined #nixos
vidbina has quit [(Ping timeout: 260 seconds)]
vidbina has joined #nixos
jensens has joined #nixos
tmaekawa has joined #nixos
ssmike has quit [(Ping timeout: 248 seconds)]
vidbina has quit [(Ping timeout: 248 seconds)]
vidbina has joined #nixos
vidbina has quit [(Client Quit)]
vidbina has joined #nixos
vidbina has quit [(Client Quit)]
vidbina has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] makefu opened pull request #32335: init devpi-server at 4.3.1, bump devpi-client (master...pkgs/devpi-common/update) https://git.io/vbGwV
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 opened pull request #32336: Vim codefmt (master...vim-codefmt) https://git.io/vbGwK
NixOS_GitHub has left #nixos []
<hedning[m]>
Cheng: I might be mistaken here, but I'm pretty sure that the nixos-config only gets picked up by nixos-rebuild, in which case it happens automatically
<b73_>
Hi, I'm using Nix on macOS. I'm attempting to use ruby bundler (`bundle install`) - but I get a `Could not verify the SSL certificate for https://rails-assets.org/.` error. Suggesting that I'm missing some OpenSSL certificate files on my system.
<b73_>
I have installed the OpenSSL package from Nix
zzamboni has joined #nixos
ilja_kuklic has quit [(Remote host closed the connection)]
<thblt>
b73_: it's probably worth filing a bug
<thblt>
b73_: I can try reproducing on NixOS first if you can give me exact steps.
townsend has quit [(Ping timeout: 240 seconds)]
ilyaigpetrov has joined #nixos
<Cheng>
hedning[m]: I digged into nixpkgs source codes, and found a way: with import <nixpkgs/nixos/lib/eval-config.nix> { modules = [ <nixos-config> ]; }; pkgs.SomePackage
<b73_>
ruby_2_1, bundler and openssl in a nix-shell.
<b73_>
When I `bundle install` in a ruby project (with a Gemfile) I get the error.
johann__ has quit [(Quit: Leaving.)]
<b73_>
The issue is I need to update my OpenSSL certs
orivej has quit [(Ping timeout: 255 seconds)]
<b73_>
In the past I have done this using tools like RVM (Ruby Version Manager)
notdaniel has quit [(Quit: Leaving)]
<b73_>
but I'm not using RVM on this machine - trying to just do it all with Nix
zzamboni has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] P-E-Meunier opened pull request #32338: Add optional MathJax support (master...patch-1) https://git.io/vbG6L
NixOS_GitHub has left #nixos []
cybrian has quit [(Read error: Connection reset by peer)]
zzamboni has joined #nixos
orivej has joined #nixos
tmaekawa has quit [(Quit: tmaekawa)]
ilja_kuklic has joined #nixos
zzamboni has quit [(Ping timeout: 248 seconds)]
Taneb has joined #nixos
erictapen has quit [(Remote host closed the connection)]
<Taneb>
Hi! I'm using a multiuser nix installation on Ubuntu and I've somehow messed it up
<Taneb>
When I try to run nix-env -i hello, I get an error message (nix-env -iA nixpkgs.hello works as expected)
<Taneb>
The error message is error: syntax error, unexpected $undefined, expecting IND_STR or DOLLAR_CURLY or IND_STRING_CLOSE, at /nix/store/gg431ln3rc9jhdc2x1vbjifpmch1g0fz-nixpkgs-18.03pre120540.b8f7027360/nixpkgs/pkgs/applications/misc/hubstaff/default.nix:30:77
ilja_kuklic has quit [(Ping timeout: 248 seconds)]
<sphalerite>
Taneb: generally, don't use nix-env -i without -A anyway
<sphalerite>
But what does nix-info say?
<Taneb>
Command not found?
pie_ has joined #nixos
stphrolland has joined #nixos
<sphalerite>
Oh, maybe it only comes preinstalled on nixos
<sphalerite>
What about nix-shell -p nix-info --run nix-info
<stphrolland>
Hi. Do you do { system.autoUpgrade = true; } in your configurations ? DO you think it's good practice ?
<sphalerite>
Hm OK. Running nix-channel --update both as root and as nathan may fix it. Not sure though
<sphalerite>
But yeah, as I said before — you shouldn't generally use nix-env -i without -A anyway
<Taneb>
What's the reasoning for that?
<Taneb>
(it seems to be fixed btw)
darlan has joined #nixos
<sphalerite>
-i is less predictable and slower
<Taneb>
OK, thanks a lot for your help!
<sphalerite>
It searches through all packages by name, so it has to evaluate all of nixpkgs
Cheng has quit [(Quit: Leaving)]
<sphalerite>
With -A, it selects by attribute path, which means it's very precise, and will only need to evaluate the package you actually want
darlan has quit [(Client Quit)]
<sphalerite>
One day we will have `nix install` and everything will be wonderful
zzamboni has joined #nixos
ssmike has joined #nixos
<akfp>
I'm interested in getting the ROCm code (OpenCL for AMDGPU) into nixpkgs working with the 4.15rc2 kernel. Anyone willing to do that for $$$?
cybrian has joined #nixos
<sphalerite>
Ugh why do the imagemagick source tarballs keep disappearing
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
orivej has quit [(Ping timeout: 260 seconds)]
vidbina has quit [(Ping timeout: 260 seconds)]
vidbina has joined #nixos
<pstn>
akfp: Pinging Corngood
pie_ has quit [(Remote host closed the connection)]
ylwghst has joined #nixos
pie_ has joined #nixos
zzamboni has quit [(Quit: Leaving.)]
<corngood>
philipp: hello, what's up?
<akfp>
corngood: see my message above
<pstn>
See the message from akfp I thought you might be interested, if anyone.
zzamboni has joined #nixos
<corngood>
Sorry, I see... On my phone. I was thinking of taking a look at the recent amdgpu-pro problems
zzamboni has quit [(Client Quit)]
zzamboni has joined #nixos
<corngood>
I could look at that as well.
ylwghst has quit [(Ping timeout: 240 seconds)]
drakonis has joined #nixos
<akfp>
corngood: sounds good!
simukis has joined #nixos
<corngood>
akfp: maybe make an issue on my nixpkgs fork with any info you have (corngood/nixpkgs)?
<gchristensen>
is there a way to tell nix I really don't want a daemon restarted?
johbo_ has quit [(Quit: Bye)]
johbo has joined #nixos
<akfp>
corngood: enable issues on your fork
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 2 new commits to master: https://git.io/vbG1T
<NixOS_GitHub>
nixpkgs/master 7a9784c Benjamin Staffin: dropbox: fix version sort order...
<NixOS_GitHub>
nixpkgs/master f3cf114 Thomas Tuegel: Merge pull request #32328 from benley/dropbox-wrapper...
NixOS_GitHub has left #nixos []
adamt has joined #nixos
adamt has quit [(Changing host)]
adamt has joined #nixos
akfp has quit [(Remote host closed the connection)]
akfp has joined #nixos
<stphrolland>
not sure anyone saw my question. Do you think it's good practice to use {system.autoUpgrade = true;} ? Is it something that is needed for security updates or things like taht ?
adamt_ has joined #nixos
<gchristensen>
stphrolland: I regularly run nixos-rebuild --switch --upgrade... once a week or so by hand
<gchristensen>
stphrolland: but it should be safe to autoUpgrade
<corngood>
akfp: done
adamt has quit [(Ping timeout: 240 seconds)]
<stphrolland>
gchristensen: but the role of the --upgrade switch... it cannot change the configuration.nix file ?
<stphrolland>
i'm a bit puzzled
<gchristensen>
what do you mean?
<stphrolland>
gchristensen: i don't see what the --upgrade does upgrade indeed...
<gchristensen>
oh --upgrade updates the channels
<stphrolland>
related to system.stateVersion ?
<pstn>
Corngood: I'll have a look at at on thursday. If I can achieve anything, I'll send a PR on your fork but no promises.
<srhb>
stphrolland: Have you used a Debian-like before? It's like apt-get update -- it fetches the newest nixpkgs for your channel (eg. the newest nixos-17.09 nixpkgs set)
vidbina has quit [(Ping timeout: 248 seconds)]
zzamboni has quit [(Quit: Leaving.)]
<stphrolland>
yep I come from Ubuntu and Debian, so unless I call upgrade, it still relies on the last version on my system. got it. thanks
<srhb>
stphrolland: It has nothing to do with system.stateVersion whatsoever, I should add. :)
vidbina has joined #nixos
<gchristensen>
stphrolland: no, don't change stateVersion
pie_ has quit [(Read error: Connection reset by peer)]
pie__ has joined #nixos
<stphrolland>
gchristensen: it's written in the comment :-) I would not have. But I thought it was the variable that made my system be on the channel 17.09 nixos
<gchristensen>
it isn't.... I should have clarified!
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vbGMf
<NixOS_GitHub>
nixpkgs/staging 689d630 Vladimír Čunát: Merge branch 'master' into staging...
NixOS_GitHub has left #nixos []
pie__ has quit [(Remote host closed the connection)]
pie_ has joined #nixos
TweyII has joined #nixos
zzamboni has joined #nixos
jluttine has joined #nixos
TweyII is now known as Guest37228
gm152 has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
zzamboni has quit [(Ping timeout: 248 seconds)]
pie_ has quit [(Remote host closed the connection)]
pie_ has joined #nixos
zzamboni has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] oxij opened pull request #32339: uim: more options (master...pkg/uim) https://git.io/vbGM7
<NixOS_GitHub>
[nixpkgs] oxij opened pull request #32342: pkgs: fix some metas (master...pkg/metas) https://git.io/vbGDs
NixOS_GitHub has left #nixos []
<corngood>
philipp: sounds good. I'll do a bit of reading, but it'll probably be the weekend before I get to put any time into it.
zzamboni has joined #nixos
{^_^} has quit [(Remote host closed the connection)]
<AlanPearce[m]>
How can I change the firefox package configuration when firefox is installed as a system package? I tried `nixpkgs.config.firefox.enableOfficialBranding = true;`, as indicated in the manual but it's still "Nightly"
{^_^} has joined #nixos
<AlanPearce[m]>
I'm using the `firefox` attribute in my `environment.systemPackages`
<makefu>
FRidh: regarding python-packages ordering, i really do not know where to put the new packages, ordering by name would be the most logical. however there are multiple beginnings and endings right now...
babyflakes has joined #nixos
gfixler has quit [(Quit: WeeChat 1.4)]
<etu>
AlanPearce[m]: I use firefox-bin, then I get the official branding :)
<gchristensen>
orivej I fixed it, and re-triggered evals for PRs from the last 8hrs which were incorrectly labeled. thanks for the heads up!
<gchristensen>
oh orivej isn't here either... :)
<sphalerite>
Alan Pearce: put (firefox.override {enableOfficialBranding = true;}) in systemPackages I think
<AlanPearce[m]>
etu: I want some other options though :(
<gchristensen>
as of today, if a PR causes an evaluation error which prevents enumerating outputs, it'll provide a gist of that error output as well. previously, it would just say it failed to enumerate outputs.
<AlanPearce[m]>
Hmm, now I get "attribute `override` is missing", which is weird, because I'm using the same syntax elsewhere
<sphalerite>
I suppose the neater way would be to have an overlay that just modifies firefox-unwwrapped
<AlanPearce[m]>
Even more confusing is that I want to use `pkgs.unstable`, so I had to specify unstable twice
<thblt>
Any idea where I should start troubleshooting "builder for ‘/nix/store/7viw4zcvnnsr2k3g52w0yjc4z2nyxxwk-nvidia-x11-384.90-4.14.3.drv’ failed with exit code 2" on nixos-rebuild switch --upgrade?
<sphalerite>
heh, I've now set up my controller to control the mouse cursor in general
<sphalerite>
it's pretty nice to use actually
<srhb>
I've never managed to get the hang of controllers. I do need it for my in-bed-battlestation, of course. :-P
<srhb>
(No, I don't have an in-bed-battlestation >_>)
<joepie91>
... yet
<srhb>
yet :-P
<sphalerite>
:D
<sphalerite>
I set up the left joystick to move the cursor and the right one to scroll
<srhb>
Oh, that's actually pretty nifty.
davidak has quit [(Ping timeout: 260 seconds)]
<thblt>
Stupid question, because I'm confused. I've been fighting with configuring a stupid cheap samsung printer on my NixOS box, only to find there were a few samsung-linux-unified-driver packages waiting to be installed, which didn't appear in the online NixOS packages list <https://nixos.org/nixos/packages.html>
<thblt>
Should I just distrust this list?
<srhb>
thblt: Are they marked as unfree?
<srhb>
thblt: The list doesn't show unfree drivers.
<gchristensen>
thblt: where is that drivee
<gchristensen>
oops
<srhb>
(Or unfree packages in general)
<thblt>
srhb: I haven't thought of that, thanks.
zzamboni has quit [(Quit: Leaving.)]
vidbina has quit [(Ping timeout: 260 seconds)]
ylwghst has joined #nixos
<thblt>
I should just use nix-env -qa I guess
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 3 new commits to master: https://git.io/vbGHt
<NixOS_GitHub>
nixpkgs/master 06edc9c Tim Steinbach: linux: 4.4.103 -> 4.4.104
<NixOS_GitHub>
nixpkgs/master c0b09e9 Tim Steinbach: linux: 4.9.66 -> 4.9.67
<NixOS_GitHub>
nixpkgs/master 93e68fd Tim Steinbach: linux: 4.14.3 -> 4.14.4
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.09: https://git.io/vbGHm
<NixOS_GitHub>
nixpkgs/release-17.09 007a867 Tim Steinbach: linux: 4.4.103 -> 4.4.104...
NixOS_GitHub has left #nixos []
<thblt>
Lesson learned, though: never use cheap printers.
vidbina has joined #nixos
<srhb>
thblt: That will also only show them if you have unfree packages enabled though.
CrazedProgrammer has quit [(Quit: Ping timeout (120 seconds))]
CrazedProgrammer has joined #nixos
<thblt>
So according to the internet, if you want something that's easy to configure under Linux, you should buy HP printers.
erictapen has joined #nixos
zzamboni has joined #nixos
<sphalerite>
the controller is actually nicer to use than the touchpad on my laptop since I can bind buttons to back, forward, next tab, previous tab, etc
<sphalerite>
the only thing that's more of a pain is switching to the keyboard to type
<joepie91>
thblt: wellll sort of kind of
<joepie91>
thblt: HP has a Linux version of their bloatware, but for *native, OSS* support iirc Brother(?) is the usual recommendation
<joepie91>
if I'm not misremembering
<srhb>
I've heard that as well and had the same experience, most of the time.
jaeckel has quit [(Ping timeout: 240 seconds)]
<joepie91>
the problem with Brother is mostly that it's really difficult to find an affordable Brother printer
<joepie91>
:P
<joepie91>
at least around here
<thblt>
joepie91: my Brother printer is usually easy to setup indeed, let me check if the drivers are free.
zzamboni has quit [(Ping timeout: 258 seconds)]
<thblt>
yes they are
a6a3uh has quit [(Quit: Bye)]
zzamboni has joined #nixos
<magnetophon>
I'mupdating a rust pkg, and get "error: the lock file needs to be updated but --locked was passed to prevent this" Google is not helpful. Ideas?
tomster has joined #nixos
<srhb>
Wow, it seems like the Logstash people just didn't bother to update the docs for the 6.0 release..
<tomster>
hi, i'm trying to use vagrant with nixos as a host system and using lxc as container but i cannot get vagrant-lxc installed
<tomster>
i tried lxc.enable = true; and lxd.enable = true; in my configuration.nix
<tomster>
but i still get the error "The provider 'lxc' could not be found, but was requested to
<tomster>
back the machine 'default'. Please use a provider that exists.
<tomster>
"
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] grahamc pushed 1 new commit to master: https://git.io/vbGQy
<NixOS_GitHub>
nixpkgs/master 5e19122 Graham Christensen: gucharmap: fix meta
NixOS_GitHub has left #nixos []
<tomster>
the same setup works fine with virtualbox as driver, but i want to use lxc because i need to work with a 10-node cluster and firing up 10 vbox instances froze my machine... o.O
vidbina has quit [(Ping timeout: 240 seconds)]
vidbina has joined #nixos
<tomster>
oh, my bad. i should have read the README... "vagrant plugin install vagrant-lxc" does the trick :)
a6a3uh has joined #nixos
<tomster>
ok, so how could i add that to my default.nix?
ryantrinkle has joined #nixos
b73_ has quit [(Quit: Connection closed for inactivity)]
cement has joined #nixos
dywedir has joined #nixos
Stammon- has joined #nixos
zzamboni has quit [(Quit: Leaving.)]
Harpalus has joined #nixos
CrazedProgrammer has quit [(Quit: Ping timeout (120 seconds))]
zzamboni has joined #nixos
alsoStevenXL has joined #nixos
CrazedProgrammer has joined #nixos
<tomster>
hm, still not working, now i get cannot create 'lxc/<redacted>: no such pool 'lxc'
zzamboni has quit [(Client Quit)]
<alsoStevenXL>
Hi everyone. I am trying to get a Haskell web app to build. Unfortunately, I am getting an error related to a library called `hlibsass`. I opened a PR in the project itself, which includes logs of the errors that I'm getting, as well as "things" that I've tried (https://git.snowdrift.coop/sd/snowdrift/issues/40)
zzamboni has joined #nixos
<etu>
tomster: If you get vagrant lxc to work well, tell me how. I can't wait to stop using vbox.
vidbina has quit [(Ping timeout: 240 seconds)]
<alsoStevenXL>
One of the active users on that project suggested teh following: Without looking at it too hard, here is what I think is the problem. hlibsass is bundled with its own copy of libsass, which it builds via a custom Setup.hs/cabal build process. That process is fragile and does not work on OS X, as it uses compiler flags unknown to OS X's C compiler. Now, hlibsass should have a Cabal flag that specifies
<alsoStevenXL>
to use the system install of libsass, rather than relying on the custom bundled copy. Using Nix, it should be quite possible to build hlibsass with a system-level install of libsass that is guaranteed to be compatible.
<alsoStevenXL>
(Sorry for the wall of text)
mrkgnao has quit [(Ping timeout: 260 seconds)]
vidbina has joined #nixos
jaeckel has joined #nixos
townsend has quit [(Remote host closed the connection)]
mrkgnao has joined #nixos
iyzsong has joined #nixos
<srhb>
alsoStevenXL: "Agreed." :-)
<srhb>
alsoStevenXL: (As in, that sounds doable!)
<alsoStevenXL>
srhb: cool!
<alsoStevenXL>
I'm not sure how I would do it but I'm game to give it a try. I have about an hour before work.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vbG5A
<NixOS_GitHub>
nixpkgs/master f38560a John Doe: oraclejdk8: support aarch64...
NixOS_GitHub has left #nixos []
orivej has joined #nixos
<srhb>
alsoStevenXL: I'm at work right now, so I can't easily guide you right now. Perhaps someone else is available.
<alsoStevenXL>
srhb: No problem. I'll give this a go on my own, but if someone else wants to help just ping me. ;-)
<stphrolland>
Another noob question. I hope to understand and try with local dev environment this afternoon. As far as I have understood, I need to understand Nix Stores. I'm into reading the NixOs manual, I have only skimmed the beginning with installing and nix languages primitive. Do you think the manual will be sufficient for "mastering" nix stores, or are there any complementary resources you would advise ?
<srhb>
stphrolland: It's a bit confusing to hear about "Nix Stores" in the plural at all.
<srhb>
stphrolland: I would say you'll gradually gain an understanding of the Nix store (/nix/store) as you go on.
mrkgnao has joined #nixos
<stphrolland>
srng: ok so there is one main nix store.
<srhb>
stphrolland: the NixOS manual is probably not sufficient, though. The Nix and Nixpkgs manuals are also good resources.
<srhb>
stphrolland: Usually, yes.
<tomster>
etu: i'm giving up for now :( i'll focus on subsets of the cluster instead...
ma27 has joined #nixos
vidbina has quit [(Ping timeout: 248 seconds)]
vidbina has joined #nixos
a6a3uh has quit [(Quit: Bye)]
MP2E has quit [(Remote host closed the connection)]
roberth has quit [(Ping timeout: 248 seconds)]
i-am-the-slime has quit [(Ping timeout: 240 seconds)]
<tomster>
damn, i just did a sudo nix-channel --update and now suddenly rebuilding fails with "error: attribute ‘linuxPackages_4_13’ missing,"
i-am-the-slime has joined #nixos
<tomster>
has the 4.13 kernel been dropped for unstable?
<etu>
yeah
<etu>
4.14 is in there now
<tomster>
i tried building with 4_14 but that fails with a different error
<etu>
4.14 will be the new stable since it is an LTS kernel
<etu>
tomster: yeah, you can't build vbox 5.1.x on 4.14, you need 5.2.1
<tomster>
etu: how do i up to 5.2.1?
<etu>
tomster: Which should be in nixos-unstable, I use it on 4.14 from there :)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dtzWill opened pull request #32345: r2-cutter: init at 1.0(.0), radare2 qt gui (master...feature/r2-cutter) https://git.io/vbGbl
NixOS_GitHub has left #nixos []
<tomster>
ah right
<tomster>
i will try that
<srhb>
The virtualbox modules should be 5.2.2 by default on unstable currently.
vidbina has quit [(Ping timeout: 276 seconds)]
<tomster>
hm, i switched virtualbox to unstable.virtualbox but it's still trying to build 5.1.26?
<tomster>
and fails of course
<etu>
tomster: yeah, 5.2.x also broke vagrant for me, because vagrant needed updating, and that vagrant update broke for me but that is fixed in unstable :)
<tomster>
i switched both vagrant and virtualbox to unstable but it's still building the old versions
<tomster>
invalid channel URL ‘nixos-stable’ at /run/current-system/sw/bin/nix-channel line 68.
<tomster>
i want to add the stable channel with a new name before removing it
<srhb>
tomster: Looks like you flipped the arguments
<srhb>
Notice that it's complaining that nixos-stable is not a valid channel URL
<tomster>
indeed!
<tomster>
thx
<tomster>
now i get ok.error: Package ‘spl-kernel-0.7.3-4.14.3’ in /nix/store/1f7ryx474g7yygabbvnr42gs5b0m45r0-nixos-18.03pre121925.3eccd0b11d1/nixos/pkgs/os-specific/linux/spl/default.nix:49 is marked as broken, refusing to evaluate.
<tomster>
should i allow broken or simply give up until this has been fixed?
<tomster>
what the hell, i will go for it. all in, might as well
erictapen has quit [(Remote host closed the connection)]
asuryawanshi has quit [(Ping timeout: 264 seconds)]
a6a3uh has joined #nixos
<sphalerite>
tomster: the nix pills at https://nixos.org/nixos/nix-pills/ are great for gaining a fairly in-depth understanding of how nix and nixpkgs work with very little prior knowledge
pie_ has quit [(Ping timeout: 255 seconds)]
<sphalerite>
oops I meant stphrolland ^
<thblt>
In case anyone interested, I drafted a small Emacs package to browse Nix package list. It's really, really basic right now, but it can be kinda useful (faster than nix-env -qa, at least). Here it is: https://github.com/thblt/nixpkgs-emacs
<stphrolland>
sphalerite: thx
tomster has quit [(Remote host closed the connection)]
dieggsy has quit [(Remote host closed the connection)]
<thblt>
civodul: thanks! this seems *much* more sophisticated than my small toy. I just wrote this to be able to to quickly look up packages, since the list on the website is incomplete and nix-env -qa is extremely slow and is virtually unusable for searching.
<gchristensen>
thblt: you might also be interested in `nox`, which does some caching.
dieggsy has joined #nixos
<thblt>
gchristensen: thanks!
Harpalus has quit [(Remote host closed the connection)]
stphrolland has quit [(Quit: leaving)]
Taneb has left #nixos []
reinzelmann has quit [(Quit: Leaving)]
dieggsy has quit [(Quit: ERC (IRC client for Emacs 27.0.50))]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vbGpq
<NixOS_GitHub>
nixpkgs/master de8f941 Peter Hoeg: syncthing: 0.14.40 -> 0.14.41
NixOS_GitHub has left #nixos []
roberth has joined #nixos
cnu- has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vbGp8
<NixOS_GitHub>
nixpkgs/master f649b04 Peter Hoeg: freerdp: 2.0.0-rc0 -> 2.0.0-rc1
NixOS_GitHub has left #nixos []
cnu- has joined #nixos
cybrian has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] rycee pushed 1 new commit to master: https://git.io/vbGph
<NixOS_GitHub>
nixpkgs/master 601c895 Robert Schütz: abcMIDI: 2017.06.10-> 2017.11.27
NixOS_GitHub has left #nixos []
thblt has quit [(Changing host)]
thblt has joined #nixos
<manveru>
why would you use `nox` if you can use `nix search` ^^;
<tilpner>
magnetophon - This channel is more appropriate :)
<manveru>
thblt: btw, spacemacs has a nixos layer
adamt has quit [(Remote host closed the connection)]
<thblt>
manveru: I don't use spacemacs, so I'd rather not contribute code to them directly. But if the package becomes something more useful than a tabulated list of packages, I'll publish it somehow, so the Spacemacs people will be able to grab it :)
<thblt>
(Also, IIUC, Spacemacs layers are just sets of preconfigured packages, there's no actual code to contribute)
<manveru>
keybinds mostly :)
<manveru>
but yeah, nix-emacs would be the better place
<NixOS_GitHub>
[nix] shlevy closed pull request #1698: fetchGit: Ignore tarballTtl if rev is set and not in the repo. (master...fetchGit-fast-revision-update) https://git.io/vFxxQ
NixOS_GitHub has left #nixos []
<srhb>
gchristensen: Ugh, I disconnceted. Wouldn't the new nix-diff have made that easier? :)
<gchristensen>
it would have! A)
nuncanada has joined #nixos
vidbina has joined #nixos
<srhb>
:D
mrkgnao has quit [(Ping timeout: 276 seconds)]
slyfox_ has joined #nixos
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<tilpner>
hyper_ch - Because nix-prefetch-git has higher data usage for not-very-recent revs, and the JSON output wasn't directly compatible with fetchgit (and there was some weird hash mismatch)
<hyper_ch>
ok :)
<clever>
tilpner: why did it need ls-remote at all?
jluttine has quit [(Ping timeout: 248 seconds)]
<tilpner>
clever - I needed to know the current rev to insert into the github.com archive URL
<tilpner>
(branch -> rev)
<hyper_ch>
nixos without needing to compile vbox is so much better :)
<clever>
tilpner: ah, you mean for pre-fetching the tip of a branch and making it pure
jb55 has joined #nixos
<tilpner>
clever - Yes. Previously I maintained a .json by hand, copying around hashes, now it's faster
goibhniu has quit [(Ping timeout: 250 seconds)]
<clever>
nice
<tilpner>
(The .json contained multiple repos, like nixpkgs-mozilla or riot-electron-nix)
<clever>
have you played with jq before?
<tilpner>
I'm aware of it, but I don't remember much of the syntax
<clever>
its pretty powerful
<adamt>
jq is neat.
<hyper_ch>
I used jq a while back, don't recall what for but it's a nice tool for working with json
jtojnar has joined #nixos
<tilpner>
How do I check why a package has no binary cache? Building wine gets annoying after a few hours :/
<clever>
tilpner: are you building it with nixos-rebuild, nix-env, or nix-build?
cybrian has joined #nixos
<hyper_ch>
how long does it take to build wine? it usually took me like 2h to compile vbox with extension.... I can't fathom wine taking longer
<tilpner>
clever - nix-build --no-out-link '<nixpkgs/nixos>' -A system --keep-going
<tilpner>
hyper_ch - You're right, one build doesn't take forever, but it gets rebuilt every time I update
<hyper_ch>
oh, every update? that sounds weird
<clever>
tilpner: what do you have in nixpkgs.config?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] adisbladis pushed 2 new commits to master: https://git.io/vbZlD
<NixOS_GitHub>
nixpkgs/master cb5ac6a Frank Doepper: extremetuxracer: 0.6.0 -> 0.7.4
<NixOS_GitHub>
nixpkgs/master d8787a8 adisbladis: Merge pull request #32350 from woffs/etr-0.7.4...
NixOS_GitHub has left #nixos []
<tilpner>
clever - I don't have nixpkgs.config, but <nixpkgs> is imported with { config = { allowUnfree = true; useSandbox = true; }; }
townsend has joined #nixos
<tilpner>
Rather, <nixpkgs> is a .nix file that imports nixpkgs with that attrset
Isorkin_ has quit [(Ping timeout: 248 seconds)]
griff_ has quit [(Quit: griff_)]
<tilpner>
The packages in question are wineFull and wineUnstable, both trigger a rebuild
Isorkin has joined #nixos
<clever>
tilpner: can you gist all of the expressions involved?
erictapen has joined #nixos
jluttine has joined #nixos
sigmundv__ has quit [(Ping timeout: 260 seconds)]
<tilpner>
It's not important enough for someone to read all that, it's quite a few files in a non-standard arrangement. I'm just making sure there's no easy checklist :)
<clever>
tilpner: things like the state of the pulseaudio flag, and the arch
nevermind has joined #nixos
chaker has quit [(Quit: WeeChat 1.9.1)]
nevermind is now known as Guest29357
pxc2 has quit [(Ping timeout: 240 seconds)]
Shados has quit [(Remote host closed the connection)]
pxc2 has joined #nixos
townsend has quit [(Ping timeout: 260 seconds)]
<tilpner>
system is "x86_64-linux", I don't know how to check the pulseaudio flag. I vaguely remember it being a nixpkgs thing, instead of nixos, right?
Shados has joined #nixos
<tilpner>
So if my nixpkgs.config doesn't have anything pulse-related, what's the default?
<tilpner>
Is there a list of used config options, like allowUnfree or firefox.enableAdobeFlash?
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
orivej has quit [(Quit: No Ping reply in 180 seconds.)]
pxc2 has quit [(Ping timeout: 255 seconds)]
orivej has joined #nixos
Sonarpulse has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
phdoerfler has joined #nixos
leat has joined #nixos
berce has joined #nixos
townsend has joined #nixos
<hyper_ch>
clever: sind you use zfs and nixos, have you ever pondered to do an auto zfs snapshot when running any nix-sensible command?
ma27 has quit [(Ping timeout: 250 seconds)]
marzzzello has joined #nixos
roberth has joined #nixos
georges-duperon has joined #nixos
ma27 has joined #nixos
<adamt>
When using systemd.services.foo.script, how can i run something other than bash?
<adamt>
(a shebang line pointing to bash is automatically prepended to the script)
<hyper_ch>
what's wrong with bash? :)
<adamt>
(i'll just pretend to ignore that question, despite not really ignoring it at all :p)
<hyper_ch>
well, one simple way would be to point it to a script that then calls the actual script with your desired shell
<hyper_ch>
and bash is one of the best inventions since sliced bread
tmaekawa has joined #nixos
<adamt>
Yeah, but it's kinda neat that i can just add a one-line script directly in the unit, so i don't have to go create extra files for tiny scripts. I just really can't be bothered with bash.
<adamt>
Besides, it's an artificial limitation if it can't be changed, it's just the shebang that needs changing anyways (as long as i add the desired interpreted to the path). :)
<hyper_ch>
dunno
magnetophon has joined #nixos
<hiroshi>
adamt: which shell do you have
cybrian has joined #nixos
<adamt>
hiroshi: Not understood? I just want it to generate another shebang line, pointing to e.g. Python or Ruby.
<symphorien>
adamt: use execstart= instead of script= in the unit declaration
<tilpner>
adisbladis - Thank you! :)
<adisbladis>
No problem :)
<kini>
repeating my question from yesterday:
<kini>
<kini> in https://github.com/NixOS/nixpkgs/commit/0094ca20cc I changed the package "acl2-v6-5" to the package "acl2-7.4", but since the format of the version number has changed, users who had installed the former won't be able to upgrade to the latter, right?
<kini>
<kini> if I'm not mistaken, in "acl2-v6-5", the package name is considered to be "acl2-v6" and the version number is considered to be "5"
<kini>
<kini> what's the normal procedure for renaming a package in such a way that users will be able to upgrade across the name change?
<adamt>
symphorien: yeah, but then i'll have to writeText to a file and exec that, instead of having it all neat inside the unit declaration.
<adamt>
hiroshi: That doesn't change the fact that nix generates a systemd unit that has a bash shebang line in the way :-)
drakonis has quit [(Read error: Connection reset by peer)]
Guest29357 has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<adamt>
Meh, i'll just use writeText and ExecStart.
<hiroshi>
adamt: that is because it is the standar. you can install python and hack it
<clever>
adamt: if you use serviceConfig.ExecStart, then the .service file points directly to the thing you gave it
<clever>
adamt: but if you use .script, it will writeScriptBin it for you, and put that into ExecStart
<hiroshi>
I don't understand why nixOS has nano in the base and not only vim -_- adamt but as I said . standars xD
<clever>
adamt: so you need to remove the .script option
<adamt>
clever: But there's really no reason that you couldn't change the interpreter from bash to something else. Somebody just has to hack the code together. I just wondered whether this was already possible.
<adamt>
(like scriptInterpreter = ${pkgs.ruby}/bin/ruby; # you get the idea)
<clever>
adamt: but at this point, its probably simpler to just serviceConfig.ExecStart = pkgs.writeScriptBin "foo-start" "#!{pkgs.ruby}/bin/ruby ... ";
<adamt>
clever: 1) nice, the code looks quite hackable, if need be. 2) thanks, that's a neat alternative. :-)
<clever>
adamt: oh, writeScript is probably better then writeScriptBin
<clever>
adamt: the issue, is that ExecStart needs an absolute path to an executable, but writeScriptBin returns the path to a directory containing "bin/executable"
<sphalerite>
I want to do some IFD stuff — specifically, use builtins.readFile on a derivation. How can I do this? Just doing `builtins.readFile foo` says "cannot read /nix/store/...-foo, since path /nix/store/...-foo.drv is not valid"
cybrian has joined #nixos
<Stammon|>
Is there someone successfully running uim on nixos? I've got fcitx running, but now I want to replace it with uim. All my applications completely ignore uim.
<elvishjerricco>
sphalerite: I don't know if you can `readFile` a derivation. I know you can `import` from one
JosW has quit [(Quit: Konversation terminated!)]
vcunat has joined #nixos
<sphalerite>
elvishjerricco: so I have to convert the file to a nix expression containing the string of its contents? What a pain
<ryantrinkle>
sphalerite: reading the acutal .drv isn't right
<ryantrinkle>
it should probably be reading derivation.outPath
<ryantrinkle>
though i'm a bit surprised that didn't happen automatically
<sphalerite>
ryantrinkle: it's trying to read the out path correctly
akfp has quit [(Read error: Connection reset by peer)]
<sphalerite>
clever: same error
<elvishjerricco>
That exact command line ran for me
<elvishjerricco>
oh wait you were responding to clever. my bad
akfp has joined #nixos
<sphalerite>
elvishjerricco: I don't think toFile involves a derivation though
<elvishjerricco>
I think it does
<elvishjerricco>
Oh you may be right
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 5 new commits to master: https://git.io/vbZKK
<NixOS_GitHub>
nixpkgs/master d7d8b76 Peter Simons: LTS Haskell 9.16
<NixOS_GitHub>
nixpkgs/master 1e027f1 Peter Simons: hackage-packages.nix: automatic Haskell package set update...
<NixOS_GitHub>
nixpkgs/master 8737c96 Peter Simons: haskell-Cabal: update overrides for new 2.0.1.1 release
NixOS_GitHub has left #nixos []
<elvishjerricco>
urgh
<sphalerite>
huh, it works substituting writeText for builtins.toFile though
<sphalerite>
maybe it's failing to build the derivation?
<clever>
sphalerite: i think its that readFile doesnt force the derivation to build
<sphalerite>
ooooh it just refuses to build stuff from nix-instantiate --eval
<sphalerite>
I suppose that makes sense
<clever>
sphalerite: readFile doesnt trigger the building, but import from derivation does
<sphalerite>
clever: yes it does trigger it, it just won't happen with nix-instantiate --eval
<clever>
amd writeText is just another derivation that properly depends on it
TweyII has quit [(Ping timeout: 248 seconds)]
nevermind has joined #nixos
bennofs has quit [(Quit: WeeChat 1.9.1)]
nevermind is now known as Guest99960
jgeerds has joined #nixos
roberth has quit [(Ping timeout: 260 seconds)]
<sphalerite>
so nix-instantiate --eval fails with that error message but nix-build doesn't (although the thing I'm trying to build isn't a derivation so it ends up failing later on)
dywedir has quit [(Remote host closed the connection)]
tempeh has joined #nixos
Mic92 has joined #nixos
roberth has joined #nixos
<sphalerite>
With a Haskell program, I'm getting `<stdout>: commitBuffer: invalid argument (invalid character)` in a nix build and in a pure nix shell, and it works fine in an impure nix shell. Which env vars might cause this difference?
<Stammon|>
So there is nobody around using uim?
civodul has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to release-17.09: https://git.io/vbZiR
<NixOS_GitHub>
nixpkgs/release-17.09 561425b Vladimír Čunát: nvidia-x11: 384.90 -> 384.98...
<sphalerite>
Stammon|: ask your specific question rather than asking if anybody uses xyz. You're just increasing latency for solving your problem, and potentially excluding people who can help you in spite of never having touched uim.
<Havvy>
sphalerite: I've my own coding hands tied with getting the Rust reference into a usable state.
<Stammon|>
So more specifically: I'm starting uim-xim and and uim-toolbar-gtk-systray in my .xsession. Also Im exporting the GTK_IM_MODULE, the XMODIFIERS and QT_IM_MODULE Environment variables(they are set in the environment). However no application seems to accept uim as the input method, as no application shows the uim input window.
cybrian has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vbZXY
<NixOS_GitHub>
nixpkgs/master 94dbb3a Tuomas Tynkkynen: linuxPackages.phc-intel: Only works on x86...
NixOS_GitHub has left #nixos []
georges-duperon has quit [(Ping timeout: 260 seconds)]
<Stammon|>
Someone told me that on other distributions it is sometimes necessary to run "gtk-query-immodules-3.0 --update-cache" to make uim work. This changes the systems global state which ,I think, is against the design of nixos. Also it isn't available on nixos
ilyaigpetrov has quit [(Quit: Connection closed for inactivity)]
<woffs>
sphalerite: LANG=en_US.UTF-8
griff_ has quit [(Quit: griff_)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
<sphalerite>
I've finally tried X on my chromebook after 6 months of owning it and 1 month of having nixos on it. It works!
griff_ has joined #nixos
<sphalerite>
woffs: nope, doesn't help
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
<woffs>
sphalerite: maybe LC_something too
<sphalerite>
LC_ALL=en_US.UTF-8 also doesn't help
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
orivej has quit [(Ping timeout: 240 seconds)]
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
Isorkin has quit [(Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)]
griff_ has quit [(Client Quit)]
Synthetica has quit [(Quit: Connection closed for inactivity)]
griff_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 5 new commits to release-17.09: https://git.io/vbZ1o
<NixOS_GitHub>
nixpkgs/release-17.09 a8b76c4 Vladimír Čunát: nvidia_x11_legacy340: patch to support kernel 4.14...
<NixOS_GitHub>
nixpkgs/release-17.09 422f786 Vladimír Čunát: nvidia_x11_legacy304: 304.135 -> 304.137...
<NixOS_GitHub>
nixpkgs/release-17.09 120ebc0 Vladimír Čunát: nvidia_x11_legacy304: fix build with kernel 4.14...
NixOS_GitHub has left #nixos []
griff_ has quit [(Client Quit)]
ilja_kuklic has joined #nixos
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
<woffs>
sphalerite: and LANG=de_DE.UTF-8 ? just in case ...
<NixOS_GitHub>
nixpkgs/master b04040c Michael Raskin: Merge pull request #32329 from fare-patches/gerbil...
NixOS_GitHub has left #nixos []
cybrian has quit [(Read error: Connection reset by peer)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vbZ9o
<NixOS_GitHub>
nixpkgs/master 53bf3cc Tuomas Tynkkynen: pipelight: Only works on x86...
NixOS_GitHub has left #nixos []
<sphalerite>
woffs: it was LOCALE_ARCHIVE that was missing
jensens has quit [(Ping timeout: 250 seconds)]
phreedom has quit [(Remote host closed the connection)]
mkoenig has quit [(Ping timeout: 255 seconds)]
ssmike has joined #nixos
jb55 has joined #nixos
<DIzFer[m]>
anyone awake that can help me debug what's probably a silly mistake on a package? I'm trying to substituteInPlace in an installPhase, but I'm consistently getting an Operation not Permitted with it
<puckipedia>
egh. I've been struggling for multiple days to get my multi-GPU multi-screen setup to work and I just can't ...
<puckipedia>
I have two screens, HDMI1 and DP1, on an intel GPU, and they work fine for all intents and purposes, even when using the nixos monitor configuration thing ... but then I try to add a third screen, HDMI-1-1, on the other GPU
mkoenig has joined #nixos
pSub_ has quit [(Remote host closed the connection)]
ati-radeon has joined #nixos
pSub has joined #nixos
pSub has quit [(Changing host)]
pSub has joined #nixos
orivej has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar opened pull request #32358: gnome3.gnome_online_accounts: enable more backends (master...goa-last.fm) https://git.io/vbZQl
NixOS_GitHub has left #nixos []
<ati-radeon>
I just bought a second-hand video card so I could use my new 4k display, it is an ATI Radeon HD7870 XT Boost, which apparently does not work with the "ati" driver like my previous ATI Radeon HD5850 was able to do
mkoenig has quit [(Remote host closed the connection)]
jacob has quit [(Quit: Lost terminal)]
<ati-radeon>
So I tried "amdgpu_pro" instead, which failed to build due to xfont missing
<ati-radeon>
Then I tried "amd_unfree", which too failed to compile
hiroshi has quit [(Ping timeout: 248 seconds)]
mkoenig has joined #nixos
<ati-radeon>
Does anybody else has a similar video card?
<DIzFer[m]>
shouldn't there be a "radeon" driver?
<vcunat>
"ati" is the free one
<vcunat>
that might be preferable, or without any specified (probably ending up the same)
ris has joined #nixos
<DIzFer[m]>
oh
<DIzFer[m]>
why doesn't it work then? it should AFAIK
<vcunat>
and there's free "amdgpu" - actually the best free choice probably (I'm not sure)
<vcunat>
"ati" is probably just for the older cards now
<akfp>
how should the ROCK kernel (modified 4.9) which is used by the ROCm project (AMD hetregeneous computing thing) be packaged?
<woffs>
sphalerite: never heard LOCALE_ARCHIVE before
<ati-radeon>
So, is "radeon", one of the non-free versions, or does it simply not exist? I will try "amdgpu" (without the pro), once the updates are done
hiratara has joined #nixos
Dr8128 has quit [(Ping timeout: 248 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar pushed 3 new commits to master: https://git.io/vbZ5u
<NixOS_GitHub>
nixpkgs/master 3514af8 Jan Tojnar: gnome3.gnome_online_accounts: clean up
<NixOS_GitHub>
nixpkgs/master 29121b6 Jan Tojnar: gnome3.gnome_online_accounts: build devdocs
<NixOS_GitHub>
nixpkgs/master 3f45966 Jan Tojnar: gnome3.gnome_online_accounts: enable more backends
<ati-radeon>
joepie91: I will try, but first the upgrade has to finish (first not enough disk space, then the manual JDK thing) and damnit, still not enough space on disk even after nixos GC of 5G
babyflakes has quit [(Quit: Connection closed for inactivity)]
<ati-radeon>
joepie91: radeon failed to sync rings (-22)
<joepie91>
ati-radeon: is that the *only* error in the log?
<sphalerite>
I'm probably more proud of myself than I should be for having set up a bunch of servers which are named after pokemon to display pictures of themselves in the terminal when you log in to them
<joepie91>
ati-radeon: try to rebuild with that setting, and see if it solves the problem; I've had some really bizarre issues with DPM on AMD cards
<joepie91>
(dynamic power management)
<ati-radeon>
joepie91: I will try that
<tokudan[m]>
Linus: hmm... then i guess I don't understand the difference between what nix-repl shows me and what nix-env -qaP display
<joepie91>
ati-radeon: if you have an AMD CPU, you may also need to disable the CPU equivalent in your BIOS/UEFI
<joepie91>
ati-radeon: it'll be called something-PM
<joepie91>
DPM, APM
<joepie91>
something along those lines
<ati-radeon>
joepie91: I have an i7 920 (overclocked)
<joepie91>
I had to disable both to get things to run stable on my system
<joepie91>
okay, then you're probably fine on that (hopefully)
<clever>
joepie91: what are the symptons of those PM problems? i happen to have amd all over my box
<sphalerite>
tokudan: the names available in nix-repl '<nixpkgs>' are the same as those from `nix-env -qaP` without the nix{pkgs,os}. at the beginning
<joepie91>
clever: memory corruption primarily leading to all kinds of bizarre crashes as well as GPU lockups, the symptoms changed depending on desktop env, kernel version, Xorg version, exact drivers...
<joepie91>
clever: I'd be talking for an hour if I dug up all the symptoms
<joepie91>
it was a proper mess
<joepie91>
been debugging it for months :P
<sphalerite>
tokudan: to work with the same stuff as you get with nix-repl '<nixpkgs>' when using nix-env, you can pass -f '<nixpkgs>' to nix-env. So nix-env -qaPf '<nixpkgs>' should get you the same info as nix-repl '<nixpkgs>'
<joepie91>
clever: the one consistent thing was that it was somehow corrupting data in working memory - the exact affected region varied by driver
<tokudan[m]>
Linus: then shouldn't "nix-repl> <nixpkgs>" + hitting tab show me "pkgs" somewhere?
<clever>
!ping
<joepie91>
I can tell you, debugging memory corruption issues is the worst thing :D
<clever>
joepie91: that sounds a lot like some of my things
<sphalerite>
clever: !pong
jtojnar has quit [(Remote host closed the connection)]
asuryawanshi has quit [(Ping timeout: 240 seconds)]
<clever>
joepie91: ive been seeing memory corruption in the framebuffer that runs deep enough that i can capture it in a screenshot
<joepie91>
clever: so yeah, if you get these issues, disable APM/DPM on the CPU in the BIOS/UEFI, disable DPM in the radeon driver (fglrx does not allow you to do this), and it will likely magically go away
<joepie91>
I still need to complain at AMD about this
<sphalerite>
tokudan: putting pkg and pressing tab in nix-repl suggests pkgs.
<clever>
joepie91: ive seen heavy signs of the gpu rendering tooltips before the buffer is filled, causing it to render undefined data
<joepie91>
clever: you're running fglrx I suspect?
<clever>
joepie91: amd-gpu now
cybrian has joined #nixos
<joepie91>
clever: also with the visual artifacts?
jtojnar has joined #nixos
<joepie91>
because I found that fglrx was the only driver for me that kept the memory corruption contained to the framebuffer, mostly leading to weird flickering in chrome and agmes
<joepie91>
games*
<tokudan[m]>
Linus: when i have "nix-repl> pkg" and hit tab nothing happens
<joepie91>
all the other drivers just wrote random shit into application and kernel memory
<clever>
tokudan[m]: nix-repl '<nixpkgs>'
<joepie91>
crashing random processes in the process
<clever>
joepie91: amd-gpu has the random corruption in one area of the screen after a few months of uptime, and the machine hard-locks if i unplug a monitor
griff_ has quit [(Quit: griff_)]
<clever>
joepie91: the closed-source driver hard-locks if i enable any extra monitors
<joepie91>
hm, might be a different issue, although DPM is somewhat notorious for being buggy
<tokudan[m]>
clever: that did the trick, thanks!
<joepie91>
in more than one way
alexteves has joined #nixos
<joepie91>
no idea if amdgpu has a flag to disable it
<clever>
joepie91: the "ati" driver mostly works, but hard-locks if i unplug a monitor before disabling it
<sphalerite>
tokudan: did I write nix-repl without '<nixpkgs>' anywhere?
<ati-radeon>
joepie91: it crashed again, I will try amdgpu now
<clever>
joepie91: also, the default driver from a few months ago, it would vsync after every character in xterm, leading to horid draw rates
<joepie91>
ati-radeon: hmm, hold on
ylwghst has quit [(Quit: Lost terminal)]
<alexteves>
Is there a way to check which binary caches the nix-daemon is using? Have a script that adds a cache and am trying to figure out when I need to restart the daemon (at least os x, it seems)
<tokudan[m]>
Linus: i probably overlooked it. your tips helped me a lot. thanks :)
<clever>
-r--r--r-- 1 root root 4.0K Dec 5 18:29 /sys/module/amdgpu/parameters/dpm
radivarig has joined #nixos
<clever>
joepie91: and it lacks the write bit
<clever>
so i would have to rmmod to change it
ylwghst has joined #nixos
<clever>
joepie91: a quick skim thru the source confirms, the option is only read in several init functions
<joepie91>
:(
<clever>
it may also leave it in the current state at rmmod, and fbcon makes rmmod difficult
<joepie91>
also, wtf is up with missing binary builds for WINE in nixpkgs
cybrian has quit [(Quit: cya)]
<ati-radeon>
joepie91: I just tried, amdgpu is worse, it keeps resolution at 800x600, so I will read the link you gave me to see if that fixes the ati driver
hiratara has joined #nixos
<clever>
joepie91: what is the value of the hydraPlatforms option on it?
<joepie91>
ati-radeon: that might just be a config setting
<clever>
absent
ylwghst has quit [(Client Quit)]
<clever>
nix-repl> wineFull.meta.hydraPlatforms
<clever>
[ ]
<clever>
joepie91: aha, this package is specially flag for hydra to ignore
<akfp>
can nix resume failed downloads of large files? I'm having trouble downloading 150M files on a bad line.
<clever>
akfp: not on its own, youll have to manually download it with something else, then nix-store --add-fixed sha256 ./foo.tar.gz
<clever>
and that only works for fixed-output things, not nar's
<gchristensen>
I have two bugs in {^_^}: one is that since the gateway and the factoid daemons are in the same cargo project, nixops wants to restart the gateway more often than it really needs to, since it rebuilds the gateway a lot. (2) the amqp library is kind of slow, causing multi-second delays in replies :(
<joepie91>
it's the happiest bot around, though
<gchristensen>
yeah it is!
<clever>
"D
<clever>
:D
<gchristensen>
poor [0__0], always so shocked
<akfp>
clever: does it work for github downloads?
gm152 has joined #nixos
<clever>
akfp: depends on which fetch function nix is using
<sphalerite>
clever: for nars it should be doable by downloading them manually and using nix-store --restore though, right?
<clever>
sphalerite: you will need to know the original name, and have root
<sphalerite>
Related: does nix not preserve signatures when fetching paths from a binary cache?
<clever>
sphalerite: --restore is like `tar -xf` and lacks the <hash>-<name> part of the paths
<clever>
all signatures are lost after download
<sphalerite>
Oh right
<clever>
and all hydra's re-sign everything they share
<sphalerite>
Ugh that's a pain
<sphalerite>
Why would it work like that?
<clever>
the signatures look like a tacked-on extra
<clever>
and are only checked in one or 2 spots
zraexy has quit [(Ping timeout: 276 seconds)]
<clever>
and once verified, they go into the trash
<sphalerite>
That's silly. It makes nix-serve so much less practical
<sphalerite>
Idk if this is only in unstable, but the nix store can store signatures
<clever>
ive also had another idea on improving it further
<clever>
if you store the signatures after downloading, you can verify that a given storepath hasnt been tamperd with
<clever>
given root, i could trojan every binary in your store, then fix db.sqlite to claim they are perfectly fine
<clever>
but if you keep the signatures, you can verify those hashes at a later date, and detect that
<tobiasBora>
Hello,
<sphalerite>
clever: yeah the store already allows storing sigs, at least in unstable
<sphalerite>
So that's already feasible
martinklepsch has quit [(Ping timeout: 246 seconds)]
<sphalerite>
Just that imports from bianry caches don't seem to copy the sigs in
pie_ has joined #nixos
<clever>
ah
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] caugner opened pull request #32362: phpPackages.mailparse: init at 3.0.2 (master...php-pecl-mailparse) https://git.io/vbZjK
NixOS_GitHub has left #nixos []
<alexteves>
Is there a way to check which binary caches the nix-daemon is using? Have a script that adds a cache and am trying to figure out when is it needed to restart the daemon (at least on os x, it seems)
<sphalerite>
clever: there is nix copy-sigs though which can retrieve signatures after the fact
pie_ has quit [(Read error: Connection reset by peer)]
pie__ has joined #nixos
Dr8128 has quit [(Ping timeout: 276 seconds)]
<sphalerite>
Oh silly me. I did nix copy-sigs --all -s https://cache.nixos.org on my Chromebook. The Chromebook that everything builds locally on because cache.nixos.org doesn't have ARMv7 binaries.
spietz has quit [(Ping timeout: 268 seconds)]
<ati-radeon>
joepie91: I will try Ubuntu tomorrow, to see if it does work and what driver it is using to make it work, if it does, which then hopefully helps me figure out what I need to do in NixOS to get it to work. Thanks for all the help you have given me! I am going to sleep now, goodnight
<sphalerite>
alexteves: very hacky, but maybe looking into /nix/var/nix/binary-cache-v3.sqlite could be helpful. Not sure though.
<sphalerite>
You may also need to remove that file when actually restarting the daemon because it's a bit finicky with that stuff (the file is safe to remove)
<NixOS_GitHub>
nixpkgs/release-17.09 66f2d6a Michael Raskin: Merge pull request #32363 from tokudan/ff5701...
NixOS_GitHub has left #nixos []
schoppenhauer has joined #nixos
ryantrinkle has quit [(Ping timeout: 248 seconds)]
joshuaks has joined #nixos
jb55 has quit [(Ping timeout: 255 seconds)]
zraexy has joined #nixos
<tobiasBora>
sphalerite: for the nixos, I don't see pre: "18.03.git.66b63d2f5a (Impala)
ryantrinkle has joined #nixos
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
Olgierd has quit [(Ping timeout: 255 seconds)]
drakonis has joined #nixos
<gchristensen>
do we really need like 11 years of dutch tax return software?
dan_b has quit [(Ping timeout: 248 seconds)]
<joepie91>
gchristensen: mind that Dutch tax return software is a bit weird in that each year has its own version that can't interact with tax returns from any of the previous years
<joepie91>
gchristensen: so if you want to be able to access your past tax returns, you *need* the old software
<gchristensen>
lol
<joepie91>
(at least it used to be that way)
jb55 has joined #nixos
<joepie91>
the Dutch tax return software predates the ability to file taxes online, even
<joepie91>
you used to get it on a floppy disk, and you'd fill it in
<joepie91>
and ship back the floppy
<gchristensen>
that is so cool
<joepie91>
it is :D but the idea of each year having its own non-interoperable software version lived on for a bit too long, heh
<Ralith>
anyone have a home theater config to share?
<Ralith>
currently trying to work out how kodi plugins are managed, but that is just for starters