<pie_>
boomshroom, maybe clever can help if he's around
<boomshroom>
pie_: clever seems to be a master at all things Nix. :P
<gchristensen>
and other things
lopsided is now known as lopsided98
paraseba has quit [Ping timeout: 240 seconds]
maingo has joined #nixos
srdqty has quit [Ping timeout: 260 seconds]
lopsided98 has left #nixos [#nixos]
lopsided98 has joined #nixos
srdqty has joined #nixos
moth_ has quit [Ping timeout: 260 seconds]
thc202 has quit [Ping timeout: 256 seconds]
iyzsong has joined #nixos
jackdk has joined #nixos
kerrhau has quit [Ping timeout: 240 seconds]
paraseba has joined #nixos
srdqty has quit [Ping timeout: 264 seconds]
<boomshroom>
`SyntaxError: invalid syntax` On the plus side, that means my script at least tried to run. Yay!
srdqty has joined #nixos
<boomshroom>
What language should I write a toy PID1 in? Not C, because most production inits are in C.
random-nixor has joined #nixos
<boomshroom>
I've considered Rust because it's basically a better C, but I'm feeling a little adventurous, so I'm thinking of using Haskell.
<maurer>
boomshroom: Rust is well suited to the task. Haskell can do it too, but putting GCs or laziness into things that you want predictable time and memory bounds in is a bit perilous
jrolfs has quit [Ping timeout: 240 seconds]
detran` has joined #nixos
jrolfs_ has quit [Ping timeout: 240 seconds]
<achambe>
boomshroom: you should look at ziglang
<achambe>
bonus, the developer of ziglang uses nixos
<boomshroom>
achambe: This is the first time I've seen Zig outside of reddit. I've heard that it's safer than unsafe Rust, but that's the only positive thing I've heard so far.
<achambe>
if rust is like C++
<achambe>
zig is like C
<achambe>
I dunno, it has some nice stuff
detran has quit [Ping timeout: 265 seconds]
<achambe>
also, you can call C libraries directly
<achambe>
it doesn't have any fancy memory analysis like rust, just some tools to stop you from forgetting to do stuff
<maurer>
So, uh, what's the use case for Zig over Rust then?
<boomshroom>
Part of the reason I considered Haskell is because no one would expect and init in Haskell.
jrolfs has joined #nixos
<achambe>
rust is a disorganized clusterfuck but if you are willing to tolerate that
<achambe>
then rust is probably better
<maurer>
achambe: Uh, what do you mean by that?
<boomshroom>
achambe: Given that Rust is currently my favoourite programming language, I would say that I tolerate it.
<achambe>
well, for example - rust stdlib just assumed memory allocation couldn't fail, now they are trying to fix it post 1.0 by hacking on stuff
<achambe>
I can only see the complexity of rust increasing over time, but if you are willing to tolerate that
<achambe>
its a great language
<boomshroom>
What I want is a language that looks like Haskell and acts like Rust.
jrolfs has quit [Ping timeout: 256 seconds]
<achambe>
maurer: also, the fact that the developers shunted to the community on nearly every library
m0rphism1 has quit [Ping timeout: 276 seconds]
<achambe>
so there are lots of high quality libraries, but no way to know if it will be supported
<maurer>
achambe: So, I am not aware of any successful language which _didn't_ punt to the community on stuff beyond basic data structures
m0rphism1 has joined #nixos
<maurer>
there's too wide a variety of "basic functionality" and only so much horsepower in a core dev team
<achambe>
I dunno, Go, python and C# seem to be counter examples to me
<boomshroom>
Go's standard library is one of the fullest I've seen and people complain about it having things that shouldn't be there while omiting things it should have.
<achambe>
meh anyways, this might go way too far off topic :P
random-nixor has quit [Ping timeout: 260 seconds]
Fare has joined #nixos
<achambe>
does anyone use hydra as a general purpose CI system?
<elvishjerricco>
achambe: Hydra + Hail makes for pretty good continuous integration / deployment. It's a pain to setup though
<maurer>
achambe: I used to, the problem is that every time hydra needs to be updated, 123092183098 things break
<maurer>
and I ended up having to discard the database about once a month
<gchristensen>
yeah, I wonder if that should default to true
Mateon1 has quit [Ping timeout: 240 seconds]
Mateon2 is now known as Mateon1
Fare has joined #nixos
jrolfs has joined #nixos
<{^_^}>
[nixpkgs] @lopsided98 opened pull request #38818 → top-level: apply platform config to cross system → https://git.io/vxA1Y
jtojnar has quit [Quit: jtojnar]
jtojnar has joined #nixos
winsome has joined #nixos
sigmundv__ has quit [Ping timeout: 268 seconds]
m0rphism1 has quit [Ping timeout: 260 seconds]
m0rphism1 has joined #nixos
<boomshroom>
I just realized that it's completely possible to write an init in Java.
<Dezgeg>
can you handle SIGCHLD in java?
<Dezgeg>
(I also wonder if the JVM starts without /proc or /dev mounted, but I guess you could require the initrd to mount those)
jrolfs has quit [Ping timeout: 240 seconds]
<rotaerk>
hmm is it normal for a PR to nixpkgs to not get any attention?
<boomshroom>
Dezgeg: I guess the ability to write a signal handler would be the the biggest thing that would prevent a language from being able to work, and even then you could also implement that in a separate language that calls back to the main language.
<gchristensen>
handling sigchld is like "the thing" pid1 needs to do
<Dezgeg>
it probably isn't safe to hook SIGCHLD in FFI code in case the JVM has already hooked it
klntsky_ has quit [Ping timeout: 268 seconds]
<boomshroom>
It looks like the JVM hijacks SIGQUIT, SIGTERM, SIGINT, and SIGHUP, meaning SIGCHLD should be available for the init writer.
klntsky_ has joined #nixos
<boomshroom>
stack2nix is taking a long time to build, probably because it's building all of its transitive dependencies manually.
blankhart has joined #nixos
<boomshroom>
It seems that even python and bash are capable of signal handling.
c0ffee152 has quit [Quit: Lost terminal]
vidbina has joined #nixos
vidbina has quit [Ping timeout: 260 seconds]
<boomshroom>
Does stack2nix really have to include Win32 in the nix file and then not actually call it, and then have the audacity to prevent my build from working because of one package that isn't used?
<boomshroom>
Looks like it pulled in hfsevents as well and still doesn't use it. To make matters worse, I can't just fix it with NIXPKGS_ALLOW_BROKEN=1.
<boomshroom>
Opps, looks like I don't know how to use stack2nix
jperras has quit [Ping timeout: 268 seconds]
kerrhau has joined #nixos
jperras has joined #nixos
kerrhau has quit [Changing host]
kerrhau has joined #nixos
<boomshroom>
Nix isn't the fastest build system, but it can generate an entire linux system and build my init along with it.
<achambe>
maybe a stupid question, " sudo nixos-rebuild switch " doesn't take the curren't shells NIX_PATH into account?
<boomshroom>
Nix brands itself as a package manager, but it's really a build tool that just happens to build entire systems instead of just programs.
<achambe>
ah, need -E I guess
<elvishjerricco>
achambe: I usually use `-I` with nixos-rebuild
<elvishjerricco>
e.g. `-I nixpkgs=$(pwd)/nixpkgs -I nixos-config=$(pwd)/configuration.nix`
<achambe>
thanks
<achambe>
Another question, does nix inspect the contents of ./foo.patch files to determine if it should rebuild?
<achambe>
I'm trying to debug the gitlab runner
<elvishjerricco>
If any aspect of a derivation changes, it will be rebuilt. So if a patch file changes at all, the derivation will be rebuilt
<achambe>
neat, this is so handy.
<boomshroom>
I don't know of any other build tool capable of building an entire, fresh, linux system just to test out an init daemon.
<achambe>
checkout the same gitlab version, and am just using git diff -p to add another patch with my debug prints
<achambe>
boomshroom: yeah its awesome
<elvishjerricco>
I sometimes test my whole personal config with `nixos-rebuild build-vm` :P With home-manager, I can just deploy my config to a machine and have it boot up with everything I'm used to instantly. Completely unreasonable without nix.
<boomshroom>
Are all system service scripts handled by nixos itself rather than coming from upstream packages?
<zybell_>
to build a PID1:use C to write a proxy init;make 2 while()s;the inner while is wait()ing and writes pid+status to a pipe while a certain Process is running;the outer while fork()s that process and exec()s your real init with the other end of the pipe as stdin. The real init you can write in any language.
ardfard has joined #nixos
<boomshroom>
zybell_: Exactly why does that need to happen? Signal handlers can be writen in any language with a library for them, and SIGCHLD is the number one purpose of PID1.
jperras has joined #nixos
ardfard has quit [Client Quit]
seafood has joined #nixos
schoppenhauer has quit [Ping timeout: 264 seconds]
<zybell_>
That makes the proxy easy to debug (only a dozen lines)and therefore rocksolid. On the other hand the real handling can be augmented as needed,the process debugged and *restarted*. When did you last change a running init after a recompile;-) Short: it makes for a comfortable dev environment
<Myrl-saki>
boomshroom: Haskell is like Rust, and Rust is like Haskell. Therefore use Haskell and Rust.
pie_ has quit [Ping timeout: 264 seconds]
<boomshroom>
Myrl-saki: But I can't control memory in Haskell and I can't write `main = run >> forever (awaitSignals None)` in one line in Rust.
<zybell_>
Haskell *and* Rust
Myrl-saki has quit [Quit: WeeChat 1.6]
Myrl-saki has joined #nixos
<zybell_>
was the proposal
<boomshroom>
zybell_: But everything becomes easier if you're only using one language, and writing the Haskell runtime in Rust would match or exceed the size of the haskell code its supporting.
<Myrl-saki>
zybell_: Well, we can defined `likeenss` such that it is an equivalence class. Equivalence classes are reflexive, therefore `Haskell ~ Haskell ^ Haskell ~ Rust`.
klntsky_ has quit [Remote host closed the connection]
<usrshv>
Where are questions that, i.e. "my sistem wont start?". Are you all developers?
<zybell_>
Do you want to help?
<boomshroom>
usrshv: This is a good place to ask that. A lot of us are developers, but we mostly work with nix rather than on nix.
rosa has quit [Quit: rosa]
rosa has joined #nixos
rauno has quit [Ping timeout: 240 seconds]
<boomshroom>
What are your thaughts on basing an init system on Nix itself? I'm tempted to hook into the nix interpreter for handling the init config files.
<usrshv>
zybell_ boomshroom I will proudly help.. i great at breaking things, when i "repair" them! Seriously, i cannot help to anyone. Just curious.
<boomshroom>
Alternatively I can be boring and just use something like JSON.
<boomshroom>
usrshv: Nix is great for people who break stuff because of `nixos-rebuild switch --rollback` and `nix-env --rollback`
<boomshroom>
It would be kind of awkward if a package ended up built on bootup because a script requested a package that wasn't installed.
jtojnar has joined #nixos
hamishmack has quit [Quit: hamishmack]
<kuri0>
Mic92, i tried sandboxing bastet using apparmor and it works perfectly
ericsagnes has quit [Ping timeout: 240 seconds]
<kuri0>
it only needs access to it's dependencies and .bastetrc and .bastetscores
<kuri0>
I was going to try XBill but it doesn't seem to be packaged
<boomshroom>
And of course someone's already tried reimplementing Nix in Haskell... multiple times.
usrshv has quit [Ping timeout: 264 seconds]
<johnw>
boomshroom: hnix is almost passing the Nix language tests
rauno has joined #nixos
<boomshroom>
johnw: Alternate implementation are always nice.
<johnw>
it sure does teach you about odd corner cases
<johnw>
like that "builtins.deepSeq (let x = x; in x) true" diverges, but "builtins.deepSeq (let y = { x = y; }; in y) true" does not
<zybell_>
I thought that nix is developed from haskell`?
<johnw>
well, the current implementation is in C++
rosa has quit [Quit: rosa]
<zybell_>
I meant nix the lang from haskell the lang whatever the impl
rosa has joined #nixos
astiles has joined #nixos
<fearlessKim[m]>
I am packaging A who works with a patched program B. The patch lives in A source folder. I thus wrote my derivation B = mkDerivation { patches ] [ A + "/B.patch"] but A has B as buildInput which generates an infinite recursion. Should I extract B.patch from A source and put it in nixpkgs ? or is there another way ? like A.src
<boomshroom>
fearlessKim[m]: Is the patch in A's source tree? Then A.src would be your best bet.
Judson1 has quit [Ping timeout: 264 seconds]
xy2_ has joined #nixos
<astiles>
Hi, I am trying to package R (a particular version). During the "configure" step of installation, I receive an error because the readline headers (readline-devel in centos) aren't available. That makes sense, but how would I express "this package requires building readline-devel"?
<boomshroom>
Fuck it: init scripts in Nix! :D
<astiles>
lol
<astiles>
haven't learned about those yet
<fearlessKim[m]>
boomshroom: yep it worked
<symphorien>
astiles: add readline to buildInputs
Ariakenom has joined #nixos
kuri0 has left #nixos ["Leaving"]
rosa_ has joined #nixos
rosa has quit [Read error: Connection reset by peer]
<astiles>
symphorien: that seemed to work, thanks! my understanding is that because I used "nixpkgs.readline", it'll fetch the "readline" package from the nixpkgs repository. How do I know that "nixpkgs.readline" means the same thing tomorrow as it does today?
<{^_^}>
→ 30bff422 by @roberth: linux module handling: support kernels without modules
<boomshroom>
I remember there was a function for generating an executable script within Nix, but I forget what it's called.
<{^_^}>
→ 38e04bbf by @7c6f434c: Merge pull request #38470 from roberth/linux-without-modules
astiles has quit [Ping timeout: 260 seconds]
asuryawanshi has quit [Remote host closed the connection]
hamishmack has joined #nixos
seafood_ has joined #nixos
<seafood_>
Does Eelco Dolstra hang out here on this channel? If so, what is his IRC handle?
<symphorien>
WriteScriptBin ?
rosa_ has joined #nixos
mog has quit [Ping timeout: 240 seconds]
mog has joined #nixos
ardfard has quit [Remote host closed the connection]
<boomshroom>
symphorien: That's it. I was also wondering where it was and it's in <nixpkgs>
<MichaelRaskin>
boomshrom: do you want to write an init or a daemon manager?
<boomshroom>
MichaelRaskin: The difference?
rosa_ has quit [Quit: rosa]
<MichaelRaskin>
Well, my lang-os runs sinit as init, and has only rudimentary service management that is done by the Lisp daemon. You could easily write a config that removes the Lisp/Scheme daemons and run something like runit under sinit.
ericsagnes has joined #nixos
<MichaelRaskin>
Basically, PID1 has a very limited list of irreplacable obligations: reap orphans, not die, exec the shutdown script (or perform the shutdown)… and that's all, I think.
<boomshroom>
MichaelRaskin: Link? I'd like to see your work.
<seafood_>
Question: is the “closure” of a store path dependent on the keep-derivation and keep-outputs flags?
<MichaelRaskin>
closure of the output path is the runtime dependency closure
<seafood_>
i.e. is the closure smaller if the keep-outputs flag is false?
<boomshroom>
I guess I'm kind of going the route of putting both together. PID1 needs something to execute, so I figured I would tell it what by way of the service scripts.
<seafood_>
MichaelRaskin: I’m reading through the Nix Manual right now and reading the glossary, the definition of “closure” does not make this distinction.
rosa_ has joined #nixos
<seafood_>
MichaelRaskin: Would you agree?
<MichaelRaskin>
seafood: yes and no
<MichaelRaskin>
The closure goes by hard dependencies
<MichaelRaskin>
Hard dependencies of an output path are exactly its runtime dependencies (as found by Nix)
<seafood_>
MichaelRaskin: I’m promise I’m not trying to be difficult here but the term “hard dependency” isn’t defined in the manual :-)
<MichaelRaskin>
Manual just calls them dependencies
<seafood_>
Okay.
asuryawanshi has joined #nixos
<MichaelRaskin>
Basically, keep-derivations means that GC (and only GC) keeps both dependencies and (if available) the derivations of output paths
<{^_^}>
[nixpkgs] @dotlambda pushed commit from lassulus to release-18.03 « bepasty: make it a package again »: https://git.io/vxANW
reinzelmann has joined #nixos
<seafood_>
MichaelRaskin: Okay, this is helpful. I am thinking about proposing some wording changes to the Nix Manual in a PR.
jensens has quit [Ping timeout: 256 seconds]
<seafood_>
But I want to have some of the dialog here on IRC, in real time, so I can get through most of the objections/misunderstandings faster.
rosa_ has quit [Client Quit]
<fearlessKim[m]>
I believe the doc for cleanSourceWith is wrong ? it says it to call it ( filter: src: ...) when it expects in fact {filter=; src= ;};
<MichaelRaskin>
Might be a good idea, as long as people writing the manual are people who have already read the thesis, the wording might miss the target audience.
<MichaelRaskin>
I am leaving soon.
asuryawanshi has quit [Remote host closed the connection]
asuryawanshi has joined #nixos
Fare has quit [Ping timeout: 260 seconds]
<seafood_>
MichaelRaskin: I was wondering if Eelco hangs out here?
<MichaelRaskin>
boomshroom: note that lang-os just ignores the idea of good initramfs design and goes the «Nix on <4GiB is painful, so I can have a 500MiB initramfs if I clean up afterwards» way
<boomshroom>
MichaelRaskin: It's all (or mostly) writen in Lisp? Cool.
<{^_^}>
→ 26019419 by @dtzWill: git: 2.16.2 -> 2.16.3
<{^_^}>
→ ad076ff7 by @layus: git: 2.16.3 -> 2.17.0
<{^_^}>
→ 2ddba49f by @layus: git: fix perl libs path
<MichaelRaskin>
seafood_: not sure if he is still in this channel, but getting enough of his attention is hard even for questions where his opinion is actually needed. You don't look like someone who will ask a question today that can only be answered by him, though — a lot of time there are some people who know Nix well.
<seafood_>
MichaelRaskin: You’re right. Perhaps that was too forthright of me. I’ll keep asking the channel as a whole.
* boomshroom
tells himself "I will not stay up until 3:30 debugging Haskell tonight!"
MercurialAlchemi has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @peti pushed commit from @dtzWill to master « git: 2.16.2 -> 2.16.3 »: https://git.io/vxAAU
<niksnut>
yes, I'm still here :-)
<MichaelRaskin>
boomshroom: bash for bootscripts, then it spawns sinit as PID1 and spawns a Common Lisp daemon to implement the actual work
ardfard has quit [Ping timeout: 268 seconds]
seafood has quit [Quit: seafood]
seafood_ is now known as seafood
<MichaelRaskin>
boomshrom: and then it is a collaboration of a user's Lisp session, the system daemon, user's StumpWM process…
<MichaelRaskin>
niksnut: nice. Didn't want to highlight your nick without a clear question.
MercurialAlchemi has joined #nixos
ardfard has joined #nixos
rosa_ has joined #nixos
Guanin has quit [Read error: Connection reset by peer]
zarkone has quit [Ping timeout: 256 seconds]
Guanin_ has joined #nixos
rosa_ has quit [Client Quit]
ardfard_ has joined #nixos
ardfard has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @peti pushed commit from @mdorman to master « taffybar: fix build »: https://git.io/vxAA6
ldlework has quit [Quit: co'o ro pendo be mi]
ldlework has joined #nixos
ldlework has quit [Changing host]
ldlework has joined #nixos
ardfard has joined #nixos
<boomshroom>
MichaelRaskin: I see an init in fat-initramfs.nix that does a "switch_root" before calling out to $targetInit. Those are...?
<MichaelRaskin>
What do you mean?
<MichaelRaskin>
There is a stage 1 init in shell
<MichaelRaskin>
switch_root is implemented in busybox, for example.
ardfard_ has quit [Ping timeout: 264 seconds]
<MichaelRaskin>
(which I think I use)
reinhardt has joined #nixos
<boomshroom>
I'm trying to find the stage 1 init.
reinhardt has quit [Remote host closed the connection]
<MichaelRaskin>
Well, it is combined together from a few strings
jw_p has joined #nixos
rosa_ has joined #nixos
<seafood>
MichaelRaskin: I’ve now read through more of the manual and, in many places, the distinction between the closure of a derivation and the closure of an output is made, just not in the “closure” definition in the glossary. I’m thinking about doing some clarifications there.
<MichaelRaskin>
Technically, applying the definition of closure to the path and to the derivation file works; but better wording would be helpful
<boomshroom>
MichaelRaskin: Really? Because it looks like that script writes a buch of scripts to the root and that one is called "init" which from what I can tell, is called implicitly by the kernel as PID1.
dvim has joined #nixos
hakujin has quit [Ping timeout: 260 seconds]
<boomshroom>
Now that I've cloned your repo, I can now git grep on it.
rosa_ has quit [Quit: rosa]
rosa_ has joined #nixos
<MichaelRaskin>
It runs sinit from /run/
<MichaelRaskin>
The expression writes to $out that gets linked as /run/current-system/bin
<MichaelRaskin>
local/ contains top-level calls
MichaelRaskin has quit [Quit: MichaelRaskin]
jperras has joined #nixos
<boomshroom>
MichaelRaskin: It looks like fat-initramfs is the stage 1.
<boomshroom>
targetInit would be /run/currentSystem/bin/init... Shouldn't it be /run/currentSystem/sw/bin/init?
spear2 has quit [Ping timeout: 264 seconds]
chisui has joined #nixos
jperras has quit [Ping timeout: 264 seconds]
__Sander__ has joined #nixos
ardfard has quit [Remote host closed the connection]
coot has joined #nixos
ardfard has joined #nixos
reinzelmann has quit [Remote host closed the connection]
reinzelmann has joined #nixos
selaux has joined #nixos
vaninwagen has joined #nixos
selaux has quit [Client Quit]
selaux has joined #nixos
<boomshroom>
This has been very interesting, and I already I have better understanding of the startup process than before. (I'll have an even better understanding after more time and work.) You really can break everything down into micro-services, or have one binary handle everything. I will need to see kernels other than Linux in nixpkgs so I can test them out.
<boomshroom>
Goodnight everyone! It's almost one and I have my last exam tomorrow.
boomshroom has quit [Quit: WeeChat 2.0]
thblt has quit [Remote host closed the connection]
thblt has joined #nixos
Jetien has joined #nixos
reinzelmann has quit [Quit: Leaving]
hexagoxel has quit [Ping timeout: 265 seconds]
reinzelmann has joined #nixos
<{^_^}>
[nixpkgs] @vbgl opened pull request #38826 → yosys: fix build on darwin → https://git.io/vxAjq
lo_mlatu has quit [Quit: Connection closed for inactivity]
<fearlessKim[m]>
I made a mistake and rebuild without X, the process is kinda violent: after the rebuild nixos killed X and I faced a terminal. what If I had unsaved files in background ???
clefru has joined #nixos
<joko>
fearlessKim[m]: the files could get corrupted / not properly saved, but this is to be expected since you did such a big system change.
<fearlessKim[m]>
joko: except that wansn't on purpose :'(
<fearlessKim[m]>
couldn't there be some kind of warning "gonna kill X, press enter" before actually killing everything
nschoe has quit [Read error: Connection reset by peer]
<clefru>
is there a way for nixops to update the machine definition? for instance, after running deploy, I changed the number of vcpus my libvirtd VM gets. I can't seem to find a way to update the machine without destroying it.
<fearlessKim[m]>
clefru: I opened an issue because of the same pb, no answer though :/
ditadi has joined #nixos
rosa_ has quit [Quit: rosa]
<clefru>
fearlessKim[m]: thanks, at least I know that it doesn't work
<clefru>
(and I can stop reading the man page a 3rd time)
humanoyd has joined #nixos
rosa_ has joined #nixos
sigmundv__ has joined #nixos
<{^_^}>
[nix] @sseefried opened pull request #2071 → Add to glossary and clarify garbage collection → https://git.io/vxxtG
<nschoe>
Hi everyone, can someone explains why I have this error "collision between xxx/firefox-bin-58.0.2/bin/firefox and firefox-59.0.2/bin/firefox ; use ‘nix-env --set-flag priority NUMBER PKGNAME’ to change the priority of one of the conflicting packages" when trying to upgrade / install firefox?
<srhb>
cx405: Your question makes little sense. There are no actions or instructions(*), just values.
<srhb>
cx405: I think you're thinking of nix in an imperative way, which will not work :)
<cx405>
srhb: no, actually, I want to call two groups sequentially, which should unpack during evaluation.
<srhb>
cx405: There is no "sequentially" either. Perhaps you should pose the actual problem you have. :)
romildo has joined #nixos
<cx405>
srhb: this is actually the actual problem, I am learning nix, carefully, turtle tempo and made it to loops in "nix by example", where the suggested recursion just outputs one of the arguments - and I am trying to figure how to do a "_call_back" "function" equivalent.
<srhb>
cx405: But notice how this recursion is just producing a value. That's what the questions were about before.
<cx405>
however, I am trying to expand whats possible. I know how nix evaluates tree nodes, so theoretically this should be possible.
<cx405>
srhb: well, its fine, the functions can also produce values.
<srhb>
Sure, the question is "how do you want to combine the values"
<cx405>
srhb: I mean nix-type of "function", when I say "function" ofc.
<cx405>
srhb: hmmmmm, interesting!
jrolfs has quit [Ping timeout: 264 seconds]
<srhb>
So assuming you have two functions f and g that you want to apply to x and y, and combine with g, your answer is h (f x) (g y)
<srhb>
But it's not clear that that's what you're asking, because you seem to be mixing in a "do this, *then* that"
<srhb>
And there is no "then"
<cx405>
Well, "then" can be concatenation of results ofc, but I get the idea. :)
<srhb>
OK! Then yes, my answer stands. :)
<srhb>
(Although I meant combine with *h* not g)
<cx405>
The build of last-current [18.03] update, which is few minutes ago, breaks on my system with "flash_player_npapi_linux.x86_64.tar.gz.drv" failing.
<cx405>
srhb: I can figure out from here, I guess - I got what you mean. Thank you! :)
<srhb>
cx405: Great! :)
<srhb>
cx405: About that error, am I right to guess that it's complaining about a wrong hash?
<srhb>
cx405: Adobe has a tendency to mutate the file in place, so our hashes are no longer true when they update the version (more or less silently...)
<srhb>
Hm, no, guess in this case they just remove the old url entirely. Fun.
<cx405>
srhb: Yes, I had same fun experience with teamviewer, when I was writing my first "configuration.nix". Yet, people behind teamviewer use nix - but don't want to take responsibility :/
<{^_^}>
→ 280b421c by Timo Kaufmann: m4ri: init at 20140914
<{^_^}>
→ 077bf8de by @dotlambda: Merge pull request #38762 from timokau/m4ri-init
shikiryogy has joined #nixos
mojjo has joined #nixos
<srhb>
mkaito: Explicitly, this method: find $(nix-build '<nixpkgs>' -A hello --no-out-link)
<mkaito>
I kinda miss `pacman -Ql` :P
<srhb>
What does that do?
<mkaito>
list files contained in an installed package
<srhb>
Well.. What's wrong with find :P
<mkaito>
nothing, just more convenient to type lol
<srhb>
Make an alias :)
<mkaito>
yep already on it lol
<srhb>
\o/
<srhb>
The nice thing here is how little magic is involved
telent has joined #nixos
dan_b has joined #nixos
<srhb>
Every package has its own subtree in the store, so finding all the files really is as simple as listing them.
<mkaito>
I already wrote myself a shell function that wraps `nix-env -qaP | grep -i <stuff>` with a cache file.
<Dezgeg>
maybe one could add such a feature to nix-locate, then you wouldn't even need to download the package
<mkaito>
yeah, the hard part is finding which path in the store it is sometimes
<symphorien>
mkaito: nix-index / nix-locate might make up for some use cases
<srhb>
Dezgeg: Yes, that would be handy
<symphorien>
Ah too late
<shikiryogy>
Hi guys, after upgrading to 18.03, I don't have sound. I read the release notes and added "sound.enable", I enabled pulseadio, but still no sound whatsoever, is there something else I should enable? Sound was working fine on 17.09
<srhb>
mkaito: Right, that's where nix-build helps :)
<mojjo>
hi! I had good experiences with running NixOs on a macbook. A question: Is there a good support for a "DELL XPS 13 9360R" laptop? I guess so, but just to make sure...
Fare has joined #nixos
ThatDocsLady has joined #nixos
<goibhniu>
hi shikiryogy, did you check if the volume is turned up in alsamixer?
<shikiryogy>
goibhniu checked, it is
rosa has joined #nixos
<goibhniu>
shikiryogy: hrm ... in pavucontrol ... are applications using the right device? (they might be defaulting to HDMI or something)
<srhb>
d0nkey: Yeah, though that's just a package. Not sure what the difference is
rosa has quit [Read error: Connection reset by peer]
<cx405>
srhb: hmm, I remember there was a certain case, where PA would disable "monitors" of output,- but the only way to see their status or to enable them was via pavucontrol. This results in inability to record any sound, whats played on the system. They are listed on "input devices". Thats something to remember and check if there is pavucontrol alternative. I use what KDE offers plus pavucontrol (just for this case).
<srhb>
cx405: I think part of that message was aimed at shikiryogy? But thanks! :)
rosa_ has quit [Remote host closed the connection]
zzamboni has joined #nixos
rosa has joined #nixos
seafood has joined #nixos
rauno has quit [Ping timeout: 260 seconds]
maxventu has joined #nixos
pkill9 has joined #nixos
Fare has quit [Ping timeout: 264 seconds]
rosa has quit [Quit: rosa]
hodapp has quit [Ping timeout: 265 seconds]
bbsl has joined #nixos
rosa has joined #nixos
<bbsl>
can anoyne tell me how nixos links its packages in the store? say if I install package A that depends on lib B then I install package C that allso depends on lib B. Given that they both depend on the same version will package C be able to use the same lib first installed by package A (dinamicly linked) or is it all staticaly linked in package A?
<bbsl>
gchristensen: ok that makes sense I guess but then how are the packages also isolated from eachother?
<benny>
I thought I had a handle on nixpkgs but apparently I do not. I want to nix-env -iA something that is only in my copy of nixpkgs so I exported NIX_PATH=nixpkgs=$HOME/nixpkgs but it's not using the latest version in my nixpkg repo
Fare has joined #nixos
<gchristensen>
bbsl: notice the very interesting URL by which they reference their libs :)
<gchristensen>
s/URL/file path/ d'oh
<gchristensen>
benny: nix-env breaks rules :(
<TweyII>
benny: nix-env supports -I, so you can nix-env -I$HOME/nixpkgs -iA hello
<gchristensen>
or you may need nix-env -f $HOME/nixpkgs? I'm not sure ... I don't use nix-env because it is confusin to me.
<TweyII>
Oh, yeah, -f is probably better
hodapp has joined #nixos
<TweyII>
benny: nix-env knows about channels, so it does channel magic & probably doesn't support NIX_PATH in the same way as other things
<gchristensen>
fwiw I either put packages in systemPackages globally, or use nix-shell. I don't use nix-env.
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @timokau to master « pythonPackages.cysignals: init at 1.6.9 (#38781) »: https://git.io/vxxu8
<benny>
that is very confusing and also doesn't work for me, hehe. nix-env -I $HOME/nixpkgs -iA blabla picks up the old blabla from before my modifications
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<TweyII>
I think I used to do nix-env -f~/nixpkgs -iA hello
<benny>
pointed at my local nixpkgs repo "error: attribute 'nixos' in selection path 'nixos.libsForQt5.qtstyleplugin-kvantum' not found"
<TweyII>
benny: That's to be expected — the attribute will be relative to pkgs
<benny>
yeah without nixos it works, thanks!
<TweyII>
NP
<TweyII>
Mmmm
<TweyII>
benny: -I and NIX_PATH will only affect <angle-bracket paths> anyway
<TweyII>
benny: So ‘nixos’ probably refers to your env nixos channel
<TweyII>
benny: But if you evaluate <nixpkgs> you'd probably get your ~/nixpkgs
<benny>
my nix-build used <nixpkgs> to get it to build as it requires an argument, that's why I was confused between the tools
<steveeJ>
hey, I'm experiencing very distorted font rendering in umlet and plantuml, which are both java based. seems like it's an issue with java font rendering but I wouldn't know how to fix it. I'm on 18.03
MercurialAlchemi has quit [Ping timeout: 256 seconds]
<TweyII>
benny: If you're doing something like nix-build -E '(import <nixpkgs> { }).callPackage ./package.nix { }' you can also just do nix-build -E '(import /home/benny/nixpkgs { }).callPackage ./package.nix { }'
<TweyII>
And bypass NIX_PATH entirely
MercurialAlchemi has joined #nixos
<benny>
good to know, because when you're new you don't know how much magic is behind the various things
xcmw has joined #nixos
<TweyII>
benny: nix-env has more magic than most things :þ
tertle||eltret has quit [Quit: Connection closed for inactivity]
seville has joined #nixos
<fadenb>
benny: if you intend to use it for some time and keep it up to date most people add themselves as maintainer
shikiryogy has quit [Ping timeout: 276 seconds]
<benny>
for this package I decided against it, I just wanted to try it and found out it was broken and it was a good excuse to learn more nix
<benny>
thanks :-)
m0rphism1 has joined #nixos
FareTower has quit [Ping timeout: 240 seconds]
<thor>
I made the changes that dotlambda suggested in https://github.com/NixOS/nixpkgs/pull/38842, when I now run nix-build, it started pulling in rust, didn't do that with the original PR, why would it do that for a python package?
<nschoe>
I have a small question regarding channels: when I run `nixos-rebuild build/switch/test`, it will rebuild against the channels that were added as root, i.e. those I see with `sudo nix-channels --list`, right? But when I install packages as standard user (with `nix-env -i`), the versions which are installed ar ebased on the user's channel, right?
<nschoe>
And *if* I don't have any user channels (output of `nix-channels --list run as non-root user being empty), *then* it takes the root channels, is that correct?
nocnoc has joined #nixos
<tilpner>
All of that behaviour is customizable and depends on the values of NIX_PATH and the contents of ~/.nix-defexpr, for each user
<nocnoc>
0/
nocnoc has left #nixos [#nixos]
<Myrl-saki>
I have no idea how the hell sudo works now.
<nschoe>
tilpner, hum okay, I see nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs so I guess my regular user uses root's channels.
robstr has joined #nixos
<tilpner>
nschoe - All tools that use <nixpkgs> will use root's channels like that, yes. Unfortunately, nix-env doesn't care for your NIX_PATH
m0rphism1 has quit [Ping timeout: 245 seconds]
<crooksey>
I am trying to create a nix configuration for my desktop, so I can simply run a nixos-install, then pull in a user profile, install that, then pull in any configuration files etc, would this be the most logical approach?
<nschoe>
tilpner, okay, thanks. Thing is: I have had 'nixos' and 'unstable' added as root's channel and no user channels since I installed nixOS. And it has caused me some troubles. So my thought was to have only 'nixos' in root's channel and add 'unstable' only in the user's channels.
romildo has quit [Quit: Leaving]
<nschoe>
But with what you've jsut told me, it means that nix-env won't care about that, and always use root's channels?
<nschoe>
So I can't effectively do what I wanted.
<tilpner>
nschoe - No, nix-env will look at the contents of ~/.nix-defexpr
fendor has quit [Ping timeout: 260 seconds]
<tilpner>
nschoe - If you want to make nix-env honor NIX_PATH, make sure that ~/.nix-defexpr contains a single file "default.nix" with the contents "import <nixpkgs>"
<nschoe>
tilpner, hum this is very confusing :/ in my .nix-deferxpr I have two symbolic links: 'channels' and 'channels_root'.
<Myrl-saki>
I have no idea anymore
<nschoe>
Based on this, it should take both the root and the user's channel (I checked: 'channels' point to my user's channels). And since I want to enable 'unstable' as my user's channels, this should work, right?
<zybell_>
Myrl-saki:Do you really want to preserve PATH on nixos? As root not having the root profile seems to me wrong, sudo nix install cmd works, but sudo cmd after that not?
<Myrl-saki>
zybell_: I wanted it on Debian. NixOS does that by default, I believe.
rosa has quit [Quit: rosa]
d0nkey has quit [Ping timeout: 260 seconds]
rosa has joined #nixos
<zybell_>
On nixos it feells wrong. Check that it really preserves path (and not accidentally set to the same value)
<{^_^}>
→ a60e1743 by Russell O'Connor: trezord: 1.2.1 -> 2.0.12
<{^_^}>
→ fbcbac67 by @adisbladis: Merge pull request #38704 from roconnor-blockstream/trezor
<zybell_>
Thats bug prone
<petersjt014[m]>
has anyone been able to run wicd?
<electrocat>
hey, are there some docs available for building things with musl instead of glibc?
<petersjt014[m]>
I found an obnoxious error that keeps it from starting, and the solution I usually see online does nothing
<gchristensen>
what is the problem and typical solution
<gchristensen>
I think most people use networkmanager?
<maurer>
I just use wpa_supplicant directly
<maurer>
it works fairly well
zzamboni has quit [Quit: Leaving.]
<petersjt014[m]>
trying to start the daemon gives me "error backing up resolv.conf ". the suggested solution involves doing something with symlinks, but nix's not following the FHS seems to make that not work
ZeDestructor has joined #nixos
spear2 has quit [Ping timeout: 240 seconds]
vidbina has quit [Ping timeout: 268 seconds]
<tilpner>
petersjt014[m] - Are you starting it manually or via networking.wicd.enable?
seville has quit [Quit: Leaving]
<petersjt014[m]>
it's a service? I didn't know that. one sec while I try that
rosa has quit [Quit: rosa]
<petersjt014[m]>
.....
zzamboni has joined #nixos
<petersjt014[m]>
one metric crapton of symlinks later...
<{^_^}>
→ 8a41729c by R. RyanTM: ncmpcpp: 0.8.1 -> 0.8.2
<{^_^}>
→ 21042229 by @Mic92: Merge pull request #38841 from r-ryantm/auto-update/ncmpcpp
xAFFE has left #nixos ["Error from remote client"]
<{^_^}>
[nixpkgs] @Mic92 pushed commit from R. RyanTM to release-18.03 « ncmpcpp: 0.8.1 -> 0.8.2 »: https://git.io/vxxNg
rosa has joined #nixos
ottidmes has joined #nixos
mojjo has quit [Ping timeout: 240 seconds]
nschoe_ has joined #nixos
rosa has quit [Client Quit]
<ottidmes>
I have a submodule, but and when I was defining an additional option for it and forgetting to use mkDefault for the default config value, I got an error that the value could not be merged (it had type int). However I have no clue why it would be assigned multiple times, since the only assignment to it is the config block of the submodule where I forgot to wrap it in a mkDefault
knupfer has quit [Ping timeout: 268 seconds]
<{^_^}>
[nixpkgs] @pbogdan opened pull request #38855 → eq10q: fix build with glibc 2.27 → https://git.io/vxxAH
<{^_^}>
→ 69017e46 by @expipiplus1: ruby bundlerApp: Allow passing gemConfig through to bundlerEnv
<{^_^}>
→ cb3f5a2d by @expipiplus1: asciidoctor: Use an explicit path to patchelf
<{^_^}>
→ 16bc4ffc by @expipiplus1: asciidoctor: 1.5.6.1 -> 1.5.6.2, also add mathematical and coderay
raynold has joined #nixos
crooksey has quit [Ping timeout: 264 seconds]
bpa has joined #nixos
vaninwagen has quit [Ping timeout: 240 seconds]
<ekleog>
dotlambda: about the issue with dovecot 2.3.1 I was mentioning, after investigating during the switch from 2.3.0.1 to 2.3.1 dovecot stopped supporting Return-Path: foo@bar and wants the Return-Path: <foo@bar> (as mandated by the RFC), which breaks as opensmtpd forgets the <>
vidbina has joined #nixos
dan_b has quit [Ping timeout: 256 seconds]
telent has quit [Ping timeout: 260 seconds]
<ekleog>
so I think it should be backed out at least from release-18.03 for now, waiting for the 2.3.1.1, which should fix this issue? (I'm also contacting opensmtpd to fix this RFC-non-compliance)
socksy_ has quit [Remote host closed the connection]
xAFFE has joined #nixos
<TweyII>
If I have some binary that's linked in a FHS kind of way, can I relink it into my Nix environment using patchelf? I know I can set the full rpath and then shrink it, but I'm not sure how to get the rpath of all the inputs
astiles has joined #nixos
astiles has quit [Remote host closed the connection]
<boomshroom>
TweyII: You should just need to provide each dependency to patchelf individually.
astiles has joined #nixos
vidbina has quit [Ping timeout: 256 seconds]
rosa has quit [Quit: rosa]
rosa has joined #nixos
drakonis has quit [Read error: Connection reset by peer]
krey has quit [Remote host closed the connection]
rosa has quit [Client Quit]
astiles has quit [Remote host closed the connection]
astiles has joined #nixos
astiles has quit [Remote host closed the connection]
<TweyII>
boomshroom: Oh yeah! I guess that works — thanks
astiles has joined #nixos
hakujin has quit [Ping timeout: 240 seconds]
winem_ has quit [Ping timeout: 256 seconds]
Ariakenom has joined #nixos
rosa has joined #nixos
astiles has quit [Remote host closed the connection]
astiles has joined #nixos
toby has joined #nixos
toby is now known as Guest5112
rosa has quit [Client Quit]
toby1851 has quit [Ping timeout: 260 seconds]
lonokhov has quit [Quit: Lost terminal]
sigmundv__ has quit [Ping timeout: 256 seconds]
astiles has quit [Ping timeout: 255 seconds]
rauno has joined #nixos
rosa has joined #nixos
nschoe_ has joined #nixos
<Mic92>
tilpner: done
<Mic92>
I have not worked a lot after then with ocaml
rosa has quit [Client Quit]
<tilpner>
Mic92 - Aww, I was hoping you'd have found a way to use topfind without an ocaml wrapper :/
<Ralith>
how do I make `nix build` show me the full output of a failed build?
<tilpner>
Ralith - Try nix log
<Mic92>
tilpner: maybe vbgl has better insights
<boomshroom>
Ralith: nix log is the command.
<{^_^}>
[nixpkgs] @pbogdan opened pull request #38863 → caps: fix build with glibc 2.27 → https://git.io/vxpIz
<tilpner>
Ralith - Or stty ocrnl (didn't try, pasted from irc log)
rauno has quit [Ping timeout: 260 seconds]
<Ralith>
huh, neat
nschoe_ has quit [Ping timeout: 256 seconds]
<tilpner>
Mic92 - I'll ask #ocaml, and then mail vbgl if I can't get anything else to work. A wrapper would have to go into nixpkgs, and you weren't sure about that either (with utop)
<TweyII>
What should I do if my package generates only static binaries? Usually I'd call the shared binaries out and the static binaries static, but this one never produces shared binaries
Synthetica has quit [Quit: Connection closed for inactivity]
endformationage has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @jtojnar pushed commit from @r-ryantm to master « babl: 0.1.44 -> 0.1.46 (#38862) »: https://git.io/vxpsc
rosa has joined #nixos
ryanartecona has joined #nixos
<coconnor>
Hi all. I've noticed that the rustc build is ignoring NIX_BUILD_CORES and is building using max number of cores
<coconnor>
which runs into resource limits and fails
<coconnor>
anyone familiar with this build? Seems like it should be inheriting the cmake limits.
oahong has quit [Ping timeout: 265 seconds]
<coconnor>
interesting. rustc derivation includes "dontUseCmakeConfigure = true;" which would disable the cmake defaults
alex`` has joined #nixos
<ThatPako>
How does one write an expression like `buildPythonPackage` which "knows" how to add the dependencies to that path of the compiler?
alex`` is now known as alexherbo2
<boomshroom>
Does anyone know how to pass the ffi flag to stack2nix? GHC keeps failing when parsing the foreign import line.
alexherbo2 is now known as alex``
oahong has joined #nixos
Jetien has quit [Ping timeout: 256 seconds]
<coconnor>
haha ouch: ""-DLLVM_PARALLEL_COMPILE_JOBS=32" I don't have enough RAM for that
<lejonet>
clever: I'm thinking about giving up on this compilation and just order a pine64 LTS version... sure it takes forever to get stuff from them, but it'll probably arrive before the compilation is done anyway
<ThatPako>
lejonet: next shipment is on 13th of april
<ThatPako>
so you better be quick :^)
<lejonet>
ThatPako: oh lol :P
Turion has joined #nixos
<Turion>
Why does nixos-unstable take so long to update?
ndrei has joined #nixos
<coconnor>
lots to compile
<boomshroom>
NVM, I found it. `{-# LANGUAGE ForeignFunctionInterface #-}`
<lejonet>
ThatPako: awh, paying in their store seems to not work atm :(
sehqlr has joined #nixos
rosa has quit [Quit: rosa]
<ThatPako>
So... Anything regarding my "setting the env var automatically when building" thing?
<boomshroom>
I suppose I can ask: what should I name my init system? currently it's just "myinit".
<lejonet>
boomshroom: "herp derp, I'm a init"? :P
<TweyII>
init8
<ThatPako>
init^2
<ThatPako>
*init²
<lejonet>
init xor win
<boomshroom>
lejonet: Too many spaces and punctuation marks. "derpinit" could work, but it doesn't really roll off the tongue.
romildo has joined #nixos
<lejonet>
boomshroom: haha :P
<lejonet>
hey, I have herpderpvpn so :P
<ThatPako>
> doesn't roll off the tongue
<lejonet>
tonginit :D
<ThatPako>
as if people have to pronounce it unless when cursing it for breaking (again)
<ThatPako>
initongue
<joepie91>
boomshroom: there's always yainits
<joepie91>
Yet Another INIT System
<lejonet>
Any word can become a curse word in the right context :)
<romildo>
I want to do a shell variable substitution in the following nix string literal. How should it be escaped?
<TweyII>
Turion: I think they're extra careful with nixos in case of breakage
<lejonet>
romildo: ''$ escapes the $
<TweyII>
romildo: \${…}
rosa has joined #nixos
<boomshroom>
How about "ainobfih" pronounced a-nob-fee "An Init NoBody Asked For In Haskell"
<romildo>
lejonet, that works, thanks.
<TweyII>
Oh, double-singles — yeah, ''
<lejonet>
boomshroom: and you said my suggestion didn't roll of the tongue... ;)
<boomshroom>
lejonet: This is naming. There are no rules except the ones the namer chooses to enforce. :P
<lejonet>
boomshroom: :P
<boomshroom>
That said, the p-i transition seemed a bit hard.
<lejonet>
pinit :D
<lejonet>
pin-it
<Turion>
Tweyll, well, that's great ;) I'm just wondering which single bug takes 7 days to resolve, or rather whether I can do anything useful like testing and reporting
apostolis has quit [Read error: Connection reset by peer]
<boomshroom>
lejonet: That works, but I did want a reference to Haskell in there.
<ThatPako>
Unfortunately that's not what I'm trying to achieve :/
bara has joined #nixos
<ThatPako>
I'm trying to make a nix function creating a derivation which builds Emojicode packages and programs. For that I have to add dependencies to the search path of the compiler
seville has joined #nixos
romildo has quit [Quit: Leaving]
<ottidmes>
aha, figured out my problem of multiple definitions being applied, I converted a `listOf (submodule someModule)` to `attrsOf (submodule someModule)` and my expectation was that the config block (of the submodule) would only be applied once, but its applied every time its being type checked, causing those merges to happen when I did not use mkDefault
vil has quit [Ping timeout: 256 seconds]
* lejonet
shuns away from the mention of emoji and code in the same word
<TweyII>
ThatPako: The way most things do it is that the compilers are wrapped in a script that adds a bunch of flags calculated from buildInputs
<boomshroom>
lejonet: I'd like to mention that I seriously use emoji in my i3 status bar instead of names. My UPS charge displays as "🔋 100.0"
<silver_hook>
benny: Cool. So, what changes now in 18.03 with the introduction of the borg service?
<lejonet>
boomshroom: my font refuses to show that xD
<silver_hook>
benny: So far I’ve simply installed it as a package and wrote a systemd timer and unit for it.
<yrashk>
whom can I ping to take a look at https://github.com/NixOS/nixpkgs/pull/36957 to see if it can be merged? All other PRs I had before got merged fairly quickly but I received no comments or action on this one. It's been sitting there for a month now.
<silver_hook>
benny: …but what now?
<lejonet>
boomshroom: and sure, I can see the value of using it on a statusbar, kindof, but to _code_ with?!
<ThatPako>
TweyII: is the `NIX_` prefix necessary or just something used for cc?
<boomshroom>
: To troll people like you. :P
<lejonet>
xD
<ThatPako>
I'm not trolling
<ThatPako>
I actually wrote a Discord client with it
<ThatPako>
and a web server
<lejonet>
I like emojis, I just... don't want to code with emojis xD
<lejonet>
Tho it would be cool to use the poop emoji for a trash variable or so :P
<TweyII>
ThatPako: NIX_CFLAGS_COMPILE is just the name used by the cc-wrapper
<ThatPako>
Where would I write my custom "makeEmojicodePackage" function? How do I make it accessible?
jrolfs has joined #nixos
<ThatPako>
(and how would I differentiate between a lib and an executable)?
rosa has joined #nixos
mmercier has joined #nixos
odites has left #nixos ["Konversation terminated!"]
mmercier has quit [Client Quit]
<boomshroom>
ThatPako: The source would either go in build-support, or in the emojicode compiler's directory. The attribute would either go in all-packages or as an attribute of the compiler.
mmercier has joined #nixos
mmercier has quit [Client Quit]
vidbina has quit [Ping timeout: 240 seconds]
mmercier has joined #nixos
<ThatPako>
boomshroom: so I *have* to clone pkgs? I can't do it without?
<woffs>
nix-channel --update fails if my additional binary cache is down.
rosa has quit [Client Quit]
<electrocat>
elvishjerricco: i put it everywhere, but it's still not working
<boomshroom>
ThatPako: If you're not working in nixpkgs, then ~/.config/nixpkgs/overlays/ or the config.nix.
<{^_^}>
[nixpkgs] @LnL7 pushed commit from @knedlsepp to release-18.03 « darwin.opencflite: fix build »: https://git.io/vxpBc
Myrl-saki has joined #nixos
dbe has joined #nixos
<ThatPako>
boomshroom: guess I'll have to learn how to use overlays then
<ThatPako>
thanks
<{^_^}>
[nixpkgs] @matthewbauer pushed 4 commits to master: https://git.io/vxpBb
<{^_^}>
→ 831cb416 by @matthewbauer: Revert "arm-frc-linux-gnueabi-binutils: init at 2.28"
<{^_^}>
→ 9ed96e76 by @matthewbauer: Revert "arm-frc-linux-gnueabi-gcc: init at 4.9.4"
<{^_^}>
→ ea67004b by @matthewbauer: Revert "arm-frc-linux-gnueabi-eglibc: init at 2.21-r0.83"
<ThatPako>
boomshroom: um... How do I add a function using an overlay? Judging by the way the manual explains it I can only override packages with overlays?
<clever>
ThatPako: the overlays can just insert anything they want into the pkgs set
<ThatPako>
ohhh
<boomshroom>
ThatPako: an overlay takes self and super arguments and returns a set of packages to append to nixpkgs.
<{^_^}>
→ 34191242 by R. RyanTM: knot-dns: 2.6.5 -> 2.6.6
<clever>
so i could just make an overlay that says "4" = 5;
<ThatPako>
clever: Any idea how to merge the "object/record" (or whatever nix calls them) that my function would receive with my preset setup and build phases?
ndrei has quit [Remote host closed the connection]
dan_b has joined #nixos
telent has joined #nixos
odi has quit [Ping timeout: 240 seconds]
Sonarpulse has quit [Ping timeout: 260 seconds]
dbe has quit [Ping timeout: 240 seconds]
djahandarie has joined #nixos
vaninwagen has joined #nixos
medvid has joined #nixos
<djahandarie>
Hmm... I have a very simple stdenv.mkDerivation which I'm using to have a nix-shell with R and ggplot2 in it. I want to provide R access to some fonts in that nix-shell, but I'm not sure how to do this. Simply adding them to buildInputs doesn't seem to work.
<ThatPako>
damn, the cc-wrapper setup-hook.sh is making no sense to me
<{^_^}>
→ 59b75e06 by R. RyanTM: brotli: 1.0.3 -> 1.0.4
<{^_^}>
→ 7e7f4701 by @Mic92: Merge pull request #38861 from r-ryantm/auto-update/brotli
jrolfs_ has joined #nixos
<astiles>
Hi, I'm trying to create a package for an older version of R. My build script is pretty simple and just calls configure and then make. During the configure step, I receive the following error: "configure: error: C++ preprocessor "/lib/cpp" fails sanity check"
<astiles>
I don't receive this error when running configure inside nix-shell
<astiles>
Do I need to somehow point at nix's C++ preprocessor during the configure step?
olto has joined #nixos
ndrei has joined #nixos
orivej has joined #nixos
Jetien has joined #nixos
jrolfs_ has quit [Ping timeout: 240 seconds]
Guest5112 has joined #nixos
<lejonet>
Hmm, I'm getting an odd error related to shared memory when trying to start postgresql, and I don't seem to have shmmax, shmmin and those in /proc/sys/kernel either
<{^_^}>
→ b30497c8 by @pbogdan: caps: fix build with glibc 2.27
<{^_^}>
→ 6bb609d6 by @Mic92: Merge pull request #38863 from pbogdan/caps-glibc-2.27
Sonarpulse has joined #nixos
aarvar has quit [Ping timeout: 240 seconds]
<anonymousemice>
I do bring apologies, I found file located in .local/share/tor-browser/TorBrowser/Data/Tor/torrc , and could be easily edited per user as required. My question is solved.
<tilpner>
Nix 2.0 is complaining about old settings from your Nix 1.* config file
<arianvp2>
Also installation failed halfway because /boot is too small
<astiles>
pstn: there's this in the manual https://nixos.org/nixpkgs/manual/#r-packages. I'm trying to package an R package using Nix (not NixOS or nixpkgs) "from first principles". I think I'm close to getting the R binary to compile. I plan to then list that as a dependency in nixfiles that compile R packages.
<arianvp2>
How can I remove old generations from /boot?
<{^_^}>
→ 367db080 by @dtzWill: radare2: 2.4.0 -> 2.5.0
<{^_^}>
→ b21283e9 by @dtzWill: Merge pull request #38873 from dtzWill/updatea/radare2-2.5.0
taktoa has joined #nixos
<{^_^}>
[nixpkgs] @yegortimoshenko merged pull request #38871 → alot: Put included themes where alot will find them → https://git.io/vxpiy
<{^_^}>
[nixpkgs] @yegortimoshenko pushed 2 commits to master: https://git.io/vxpM5
<{^_^}>
→ 44a7be94 by @mkaito: alot: Put included themes where alot will find them
hiratara has quit [Ping timeout: 276 seconds]
<{^_^}>
→ 47848756 by @yegortimoshenko: Merge pull request #38871 from mkaito/fix-alot-themes
hiratara has joined #nixos
<kini>
It seems the avr-gcc package in nixpkgs is unrelated to the gcc package -- why is that? Couldn't avr-gcc be built by just overriding the configure flags for the usual gcc?
<mkaito>
now someone just needs to fix the nixos-unstable channel so I can stop building off nixpkgs HEAD :D
<astiles>
how can I set my library search path in my build script? I need the equivalent of "libreadline-dev" and I've listed readline as a dependency in the nix file. I tried export LD_LIBRARY_PATH="$readline/lib" but that didn't seem to work
jrolfs_ has joined #nixos
jrolfs has quit [Quit: WeeChat 2.0]
<mpickering>
woop! I fixed the problem I was having with QGIS and GRASS. Feels amazing to package something.
<mpickering>
astiles: Can you be more specific?
<mpickering>
listing it as a dependency should usually be enough
<mpickering>
Why are you using a custom builder :o ?
Jetien has quit [Ping timeout: 276 seconds]
<astiles>
oh I'm trying to understand what's going on with as little magic as possible
<astiles>
I can probably write it in a more canonical form once I get this working
<mpickering>
I would scrap this way if you want it to work at all.
bpa has quit [Remote host closed the connection]
<mpickering>
What are you trying to understand?
Giovani_ has joined #nixos
<Giovani_>
When installing KDE5 Kontact/Korganizer, Kontact is empty,
<Giovani_>
Onlh has welcome screen,and FILE context menu omly has CLOSE
<Giovani_>
Are additional xonfiguratiin steps required to allow it to work,kde forums statethey cannot assist as they believe it to be nixos specific
<{^_^}>
[nixpkgs] @LnL7 opened pull request #38876 → pinentry_mac: enable xcbuild on darwin → https://git.io/vxpyK
<astiles>
mpickering: well, I read about the language constructs and can kind of reason through why the code is doing what it's doing. I tried to follow this: https://nixos.org/nixos/nix-pills/why-you-should-give-it-a-try.html I'm sure the community has some niceties (e.g. for autotools-based things), but doing it this way just seemed simpler. With some of the generic things, I couldn't convince myself that I understood how things became in scope (e.g. w
<{^_^}>
[nixpkgs] @yegortimoshenko pushed 2 commits to master: https://git.io/vxpSh
<{^_^}>
→ ba337a2d by R. RyanTM: riot-web: 0.13.5 -> 0.14.0
<{^_^}>
→ 6207cb93 by @yegortimoshenko: Merge pull request #38827 from r-ryantm/auto-update/riot-web
asuryawanshi has quit [Remote host closed the connection]
jperras has quit [Ping timeout: 260 seconds]
<benny>
astiles: as someone that is himself new, I think of it as preparing environmental variables for me. Which I guess is also what is happening. ;-)
<benny>
because I know this best in nixos: "nix-shell -p python" will change PYTHONPATH for you, compare it with "nix-shell -p pythonPackages.scrapy" and you'll see how much more it contains now due to the dependencies
<{^_^}>
[nixpkgs] @tadfisher opened pull request #38878 → steamcmd: init at 20180104 → https://git.io/vxpHC
hakujin has quit [Ping timeout: 268 seconds]
<zybell_>
I have read them. You dont set buildInputs. configure searches for gfortran48 by that name. That name is not imported. Maybe you should do what line 9 of your error proposes.
ottidmes has quit [Ping timeout: 264 seconds]
spietz has quit [Quit: WeeChat 1.9.1]
redfish65 is now known as redfish64
kquick has joined #nixos
vidbina has quit [Ping timeout: 264 seconds]
jtojnar_ has joined #nixos
d0nkey has quit [Read error: No route to host]
<{^_^}>
[nixpkgs] @dezgeg pushed to release-18.03 « nixos/make-ext4-fs: Use closureInfo »: https://git.io/vxpQw
<astiles>
zybell_: thanks. I just learned about nix-build's "-K" option.
jtojnar has quit [Ping timeout: 268 seconds]
jtojnar_ is now known as jtojnar
freeman42x[nix] has joined #nixos
jperras has joined #nixos
xeji has quit [Quit: WeeChat 2.0]
freeman42x[nix] has quit [Ping timeout: 256 seconds]
<mpickering>
Why are you doing the unpacking in the install phase?
<astiles>
mpickering: no worries! I did make it past the gfortran error and am now on to an x11 error. Thanks for your offer to help another day if I'm still stuck.
<day|flip>
so do that in postFetch?
<mpickering>
Do that in unpackPhase
<mpickering>
but you don't need to do anything as the default unpackPhase will untar it for you and set sourceRoot appropiately
<mpickering>
so infact, just delete that line
<day|flip>
so 'unpackPhase = ture;' ?
<mpickering>
no just nothing
<mpickering>
Then how is the font packaged, I assume there is no Makefile?
<day|flip>
so I don't need to define tar to unpack
<mpickering>
It looks like you actually want sourceRoot = "."; though, it depends on the structure of the tarball
<clever>
this problem only happens with nasty tar files that lack a common directory, and barf files all over your downloads folder
<clever>
the timestamps of such files are generally scattered all over the place, to make sure you cant just find those files in one area of `ls -ltrh` :P
jperras has joined #nixos
<day|flip>
unpacker produced multiple directories
ndrei has quit [Ping timeout: 268 seconds]
<day|flip>
im still getting
<clever>
can you gist the current expression?
sigmundv has quit [Ping timeout: 264 seconds]
jackdk has quit [Ping timeout: 260 seconds]
<day|flip>
do i have to make a acc to used gist github?
<clever>
day|flip: *.otf on line 17 matched zero files
<day|flip>
only thing I change was 'preUnpack = "sourceRoot=.";'
<clever>
so it parsed as "cp $out/share/fonts/opentype"
<astiles>
mpickering: do you know if there's some way to include X headers? I'm seeing this error: "configure: error: --with-x=yes (default) and X11 headers/libs are not available" but am listing libX11, libXt, and xorgserver as build inputs
qvvqj has joined #nixos
Mateon3 has joined #nixos
<clever>
day|flip: insert an ls -l on line 16-ish
<day|flip>
so it have a DIR with OTF/
<clever>
day|flip: then you want cp OTF/*.otf
<qvvqj>
What is the final args object supposed to look like in `callCabal2Nix`? I'm trying to override the version of a hackage dependency but I keep getting "called with unexpected argument 'bifunctors'"
<day|flip>
thank you so much clever. It now working
<clever>
qvvqj: i think it was callCabal2Nix "packagename" ./source {}
<qvvqj>
I see some examples on github where it looks like I should just be able to do `{ bifunctors = pkgs.haskellPackages.bifunctors_5_4_2; }`
<qvvqj>
Yeah. I'm just confused about what are permissible values in place of that final `{}`
Mateon1 has quit [Ping timeout: 264 seconds]
Mateon3 is now known as Mateon1
<day|flip>
ya. I can Now enjoy my fav fonts without the lowcase k loop on it
<clever>
qvvqj: the {} at the end is the same as the {} in callPackage
<clever>
qvvqj: it passes arguments into the generated default.nix, the same ones you can set with normal .override
ericsagnes has quit [Ping timeout: 240 seconds]
<clever>
qvvqj: so you need to look at the file that cabal2nix generates when ran on that source
orivej has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #38881 → cc-wrapper: More intelligent sierra hack → https://git.io/vxpFg