barrucadu has quit [(Quit: Rebooting for kernel upgrade...)]
jb55 has quit [(Ping timeout: 276 seconds)]
akaWolf has quit [(Ping timeout: 250 seconds)]
jb55 has joined #nixos
barrucadu has joined #nixos
<catern>
hmmmmmmmmmmm
rosa has quit [(Quit: rosa)]
rosa has joined #nixos
<rosa>
does nix os use ptrace to help redirect?
<catern>
rosa: no
<rosa>
Ok
rosa has quit [(Quit: rosa)]
rosa has joined #nixos
erasmas has quit [(Quit: leaving)]
rosa has quit [(Client Quit)]
rosa has joined #nixos
rosa has quit [(Client Quit)]
fendor has quit [(Quit: Leaving)]
rosa has joined #nixos
rosa has quit [(Client Quit)]
rosa has joined #nixos
rosa has quit [(Client Quit)]
<Lisanna>
Hey, if I have 50 nix-build processes launched from e.g. bash, but I specify that I can only have 10 builds running at a time, then will 40 of those block until they get a free "slot" to run in?
<Lisanna>
e.g. for i in `seq 1 50`; do nix-build -A "package-$i"; done
markus1199 has joined #nixos
mariav has joined #nixos
gerschtli has quit [(Quit: WeeChat 1.9.1)]
markus1189 has quit [(Ping timeout: 248 seconds)]
<catern>
hey #nixos
<manveru>
Lisanna: that will still do one build at a time, since nix-build waits until the build is done
<Lisanna>
what do you mean "still do one build at a time"? I'm assuming that I have build-max-jobs set to 10
erictapen has quit [(Ping timeout: 264 seconds)]
<manveru>
yes, but nix-build is synchronous
<catern>
I have a package manager "mypkg" full of packages to import into Nix, and I have some options of which approach to take. I can either expose the package locations on a colon-separated path like MYPKGPATH, or I can list the packages in a file and identify that file with MYPKG_INDEX, or I can list the packages in a file and have that file be looked up relative to the executing file or something like that, or...
<manveru>
replace the `nix-build -A` part with `sleep 1` and it's the same
<catern>
what do you think? which is better, a MYPKGPATH or a database full of package locations?
<Lisanna>
...you're telling me that build-max-jobs is always per-invocation of nix-build, and that if two users e.g. invoke nix-build at the same time on the same system, one of them will be completely blocked?
<catern>
database/index
<Lisanna>
regardless of what the other is building?
<manveru>
Lisanna: no
<Lisanna>
I'm confused then
<manveru>
for i in `seq 1 50`; do nix-build -A "package-$i" &; done
<manveru>
this would work
<Lisanna>
that would have 10 builds running?
<manveru>
but without doing the nix-build in the background, your shell will block :)
<Lisanna>
and 40 waiting?
<manveru>
yes
<Lisanna>
oh, yeah, sorry, forgot to include &
<Lisanna>
I know that:)
<manveru>
:D
<Lisanna>
okay, good
<Lisanna>
had me worried for a second
<manveru>
catern: generally i like env vars better...
<manveru>
catern: or you could make a directory with links to each package
<catern>
manveru: a directory with links to each package seems a lot worse?
<manveru>
don't think an actual "database" brings much
<catern>
sure sure, it would be a plain-text file
<manveru>
well, that's how we do it with ruby and yarn packaging
<catern>
but I guess links would be equivalent-ish
<catern>
oh, I guess ruby and yarn need some kind of specific directory structure, and you recreate that structure using symlinks?
<manveru>
we build each package separately and then symlink them together
<manveru>
yeah
<catern>
I have no constraints on directory structure myself
<manveru>
any more details about that "mypkg" thing? :)
<catern>
it's a proprietary package manager at my job that I want to replace with Nix :)
<manveru>
for python?
mariav has quit [(Quit: mariav)]
<catern>
it's for Python, C/C++, and Java
<catern>
but it also has some crazy random databases inside various packages
<catern>
which are currently loaded through a bespoke mechanism, which I guess ideally I would keep that bespoke mechanism working inside Nix
<catern>
but it's a really sucky mechanism, and I could just make a new one, and add a backwards compatible wrapper :)
<sphalerite>
Why do I get "you are not privileged to add signatures" when I try to use nix sign-paths? Surely there shouldn't be a problem with allowing all permitted users to add signatures to paths, provided they can read the key?
<catern>
so that's why I have a decent amount of freedom: I only need to figure out how to reference these random pieces of data inside packages, and I can invent whatever mechanism I want to do so
<manveru>
catern: i'm not sure anyone can help you much in figuring out your design...
<catern>
well, I effectively have complete freedom, though :)
<manveru>
yeah, but that's the problem :)
<catern>
surely people have some idea about what the ideal package manager to integrate with Nix would be :)
<catern>
or, maybe in other terms: the ideal programming language? hmm
<manveru>
well, have you heard about nix? it's my ideal package manager ;)
<catern>
:)
<manveru>
just put stuff into buildInputs and you're fine
<catern>
unfortunately I guess the issue is that I need to provide support for loading these pieces of data, from many different languages
<sphalerite>
builtins.toJSON?
<catern>
sphalerite: it's a different kind of problem than can be solved by that I'm afraid :) it's big binary blobs, not anything that the Nix language is touching, just shuffling them around
<catern>
but I need to load them from, say, Python, which doesn't support baking-in paths, so it absolutely needs to be an environment variable... I guess?
<Ralith>
catern: or you could write a list of paths to a file, or do any number of other things
<Ralith>
pass it on the command line via a wrapper script; substitute it into the source
thc202 has quit [(Ping timeout: 255 seconds)]
<catern>
Ralith: if I write a list of paths to some database file, how do I find that database file, from an arbitrary executing Python file?
<catern>
passing it on the command line is no better than an environment variable, since it still requires a wrapper script
<catern>
substituting into the source requires either a specially prepared file for substitution, or a file that I create, which is difficult to locate.
<catern>
I guess a specially prepared file could be okay though
<catern>
hmmm...
<Ralith>
catern: build the python thing with nix
<catern>
sure, of course, I am doing that already?
<Ralith>
you can generate the file adjacent to the source, or substitute the paths directly into the source
_ris has quit [(Ping timeout: 264 seconds)]
<Ralith>
python must have *some* way to access the prefix you installed to or at least the physical location of the current file
<gchristensen>
is vbgl on IRC?
<Ralith>
even if not a static one, and even if you don't want to substitute in the static path via nix
drakonis has joined #nixos
<catern>
I don't think Python has a way to access the prefix I installed to
<catern>
it can certainly access the location of the current file
<Ralith>
the latter is a predictable function of the former, so
<catern>
but that function is unique for each individual file
<Ralith>
so?
<Ralith>
you only need it once
<catern>
if I have /nix/store/hash-pkg/foo/bar/a.py and hash-pkg/foo/b.py and hash-pkg/.index_file, then how do I generically find index file from both a and b?
jb55 has quit [(Ping timeout: 240 seconds)]
ilja_kuklic has quit [(Ping timeout: 268 seconds)]
<Ralith>
generate c.py at a well-known location relative to the prefix, import it as needed
<catern>
also, more concretely, how would this kind of index file work for programs run under nix-shell, before building? I don't think it would work
<catern>
for the same reason
<catern>
(as that it doesn't work when running under nix-shell)
<catern>
also, again, then how do I find c.py??
<gchristensen>
github is down right?
<Ralith>
the same way you find any file?
<Ralith>
o_O
<catern>
if c.py is at a well-known location, that's no better than .index_file being at a well known location...
<Ralith>
gchristensen: works for me
<gchristensen>
huh
<catern>
works for me too
<gchristensen>
double huh...
<Ralith>
catern: python allows you to access code written in other files, I am given to understand
<Ralith>
something to do with 'import' statements
<catern>
(for what it's worth, there is a convenient site "downforeveryoneorjustme.com" which you can use to check these kind of things)
<Ralith>
there is nothing magically difficult to import about a generated file
<Ralith>
as far as the rest of the code is concerned it is no different than any other part of the source
<catern>
so how do I import the c.py you mentioned, from both a.py and b.py from my example above?
<Ralith>
the same way you would if it wasn't generated
<Ralith>
pretend there is a c.py in your source distribution
mumux has quit [(Remote host closed the connection)]
<Ralith>
it sounds like you also need to make up your mind as to whether you want this data to be compiled-in or fetched dynamically from the environment; both are feasible but they're semantically different
<catern>
but even in normal Python, that is a pain, because you use different paths to import c.py in each file {a,b}.py
<Ralith>
gchristensen: yep
<gchristensen>
???
<catern>
same
<Ralith>
gchristensen: I saw some reports of issues elsewhere, perhaps they're having issues on just a few of their servers
Mateon1 has quit [(Quit: Mateon1)]
<Ralith>
catern: nontrivial projects in any programming language involve spreading code between many files and doing whatever import-analogous thing
orivej has joined #nixos
<Ralith>
this should not be a barrier
<Lisanna>
heya, if I have some test binary that depends on a library via dynamic linking, both built with Nix, and if I have 100 different builds of that library, won't I need 100 different builds of that test?
Mateon1 has joined #nixos
<Ralith>
I'm sure you can re-export it from some utility module or whatever if you want to minimize impact
<catern>
alright alright... hmmmm
<Lisanna>
even if the test itself never changed?
<Ralith>
wait
<Ralith>
why are we even talking about this as if it needs to be generated anyway
<catern>
regardless, that is still really inconvenient in Python
<Ralith>
just go to whatever utility or root module you put other simple helper functions in and jam it in there
<catern>
to do "import c" in one file and "import ../c" in another
<Ralith>
this is literally no less convenient than any other routine programming task
<manveru>
Lisanna: you could use patchelf to change the link location before running the test, but usually yes
<Ralith>
like, IANA pythonista, but if using code from other files is unbearably onerous then python is fatally broken and would never have reached meaningful adoption
<Lisanna>
manveru: or LD_LIBRARY_PATH?
<manveru>
or that :)
<manveru>
depends on how you run the test
<Lisanna>
because if I use patchelf then I'd still end up with 100 different binaries
<Lisanna>
each the output of patchelf
<manveru>
true
<manveru>
if your test is huge... you could optimize that way, but if you want to package the test + the lib in one, then you need 100 different builds
<catern>
Ralith: so there is a variable list of dependencies at build time, the file that is being generated contains that list of dependencies
hyphon81 has joined #nixos
<catern>
the API I want is from resource_lib import get_resource; get_resource(name_of_resource)
<Ralith>
and the generated file is at a fixed location in the installed tree, and can hence be trivially referred to by whatever helper function parses it, which will only be in one file anyway
<catern>
that is possible to do with an environment variable and not possible if there are any generated files at some location
<catern>
er
<catern>
not possible if it's done with some generated file at some location*
<Ralith>
I can't tell you how to write pretty python APIs, but the information is there
<catern>
hmm
<Ralith>
perhaps resource_lib is c.py
<Lisanna>
okay, but say I want to go and run the test against some version of the library. Since that dependency isn't being tracked in nix anymore, is there some pretty way I could package it so that I could have 100 small wrappers or whatever for the test that link agaisnt the correct thing, but all use the same base test binary?
<Ralith>
if need be, nix likely trivializes the task of literally generating a different instance of an external python package for every dependency set
<Ralith>
but if that's truly necessary it's only because python is dumb
<Ralith>
which would not be my initial assumption
<catern>
well, generating an external python package which contains the dependency set, would solve the problem because the location of that python package would be put on PYTHONPATH
<manveru>
Lisanna: sure
<Lisanna>
so if I ask Nix for a list of the dependencies of that wrapper, it would give me the base test binary and the correct library, and then if I install that somewhere and then run that wrapper, it links against the correct thing?
<manveru>
Lisanna: basically, you write a derivation that outputs a shell file like `LD_LIBRARY_PATH=${pathToTheLib} ${pathToTheTest}`
<catern>
i dunno, let me think about it, but I think that Python is indeed dumb, or at least Haskell and Idris are way smarter :)
<Lisanna>
...oh yeah, I guess that could work like that
<Lisanna>
I was thinking to myself that it could be a bash script, but bash scripts don't have rpaths :)
<manveru>
you'd have 100 scripts, but still way smaller than 100 binaries
<Lisanna>
but just doing ${library} in the bash script in LD_LIBRARY_PATH would work
Arcaelyx_ has joined #nixos
<manveru>
out of curiousity, why do you have so many versions of that lib? :)
<manveru>
all built with different flags?
<Ralith>
catern: presumably you could shortcut through whatever mechanism PYTHONPATH is being communicated through directly and reduce the faffery
<catern>
Ralith: well, PYTHONPATH is communicated through a wrapper script
<catern>
I could set another environment variable in that wrapper script, pointing to the generated file :)
<Ralith>
given that the wrapper already exists, that seems pretty natural
<catern>
Ralith: are you maybe familiar with bash enough that I could use it as an example of how doing this purely with a generated file is tricky?
<Ralith>
I could understand wanting to avoid it if there wasn't one already
<Lisanna>
manveru: different versions, different flags
<catern>
yeah I mean
<catern>
I am probably going to use an env var I guess
<catern>
now I just need to decide how to format that env var... :)
<Ralith>
colon-separated is traditional
<catern>
key-value pairs like NIXPATH, or just a straight list like PATH?
<Ralith>
do you need keys?
<catern>
certainly colon-separated either way :)
Arcaelyx has quit [(Ping timeout: 276 seconds)]
<Lisanna>
I'm writing a CI / build system basically
<catern>
well, the keys are the names of the resources I'm exporting...
<clever>
catern: with NIX_PATH, if you search for <nixpkgs>, it will check each bare path to see if it contains a nixpkgs
<catern>
you could scan the directories to find out the names, since the name is in a well-known file under each directory
<manveru>
Lisanna: oh, then you could even do fancy things like avoid making the shell scripts...
<catern>
but it might be a nice optimization to have the names in the env var itself
<clever>
catern: and anything in the form of nixpkgs=path, will strip the nixpkgs off the search, and directly use that path
<clever>
catern: so an entry of /tmp/foo and <nixpkgs> will check for /tmp/foo/nixpkgs
<catern>
clever: check each bare path to see if it contains a nixpkgs directory? oh, interesting
<clever>
catern: but nixpkgs=/tmp/foo will map <nixpkgs> directly to /tmp/foo
<catern>
that's pretty flexible
<clever>
it also supports sub-directories
<clever>
so <nixpkgs/foo.nix> will only return the nixpkgs that contains a foo.nix
<catern>
woah
<clever>
so you could have several nixpkgs within NIX_PATH
<Lisanna>
manveru: what do you mean?
<catern>
I had no idea about that :)
<Ralith>
I think there's a finite size budget for environment variables, bear in mind
<clever>
and it will return the full path to the foo.nix it finds
<clever>
catern: and -I will prepend things to the search path
<clever>
catern: which has a nasty side-effect, if you -I nixpkgs=/does-not-exist, it will fail to find it, and silently check the original $NIX_PATH
<manveru>
Lisanna: if you have a json like `{flags:[["-a"], ["-b", "-c"]],versions:["1","2"]}` to configure your builds
<clever>
catern: if you want to ensure purity, you want to set NIX_PATH, rather then use -I to prepend
<catern>
clever: hah!
<Ralith>
bouncing off disk may become prudent
<catern>
that goes to show, fancy features, you don't need them :)
<Lisanna>
manveru: I have a datastructure like that already actually, you pass it into the builder and the builder will build the package with that exact configuration
<Lisanna>
version is one of the options
<catern>
Ralith: prudent or imprudent? thinking about it, I am certain that I won't run into env var size limits
<Ralith>
okay then
<manveru>
Lisanna: then just iterate through them and run `LD_LIBRARY_PATH=$(nix-build --no-out-link build-my-lib.nix --arg flags "$flags" --arg version "$version") $(nix-build build-test-file.nix)`
<Ralith>
presumably easy to change in the future if you need to
<Lisanna>
manveru: yeah, basically my plan
<manveru>
so you don't write the 100 bash files :)
<Lisanna>
except I'll do the iteration in Nix itself
<Lisanna>
most of the logic for this is written in Nix
<Lisanna>
including a constraint solver :)
<manveru>
nice
<catern>
hmm, maybe I should put JSON in the environment variable?
<catern>
no that's dumb, I'll stick to key=value:colon=separated
<clever>
catern: also, you can do something like myconfig=/path/to/foo.json in the path
<clever>
catern: then just builtins.fromJSON (builtins.readFile <myconfig>)
<clever>
which will return a nix attrset
<catern>
clever: these are nice tricks... though I think I could abuse that quite horribly :)
<catern>
(in unrelated ways to my current dilemma)
<catern>
Ralith: yeah I think I have figured it out, thanks for talking me through it
Neo-- has quit [(Remote host closed the connection)]
Neo-- has joined #nixos
<Ralith>
\o/
sigmundv__ has joined #nixos
Neo-- has quit [(Ping timeout: 240 seconds)]
lambdamu_ has joined #nixos
sibi_ has quit [(Quit: Connection closed for inactivity)]
lambdamu has quit [(Ping timeout: 268 seconds)]
drakonis has quit [(Quit: Leaving)]
taktoa has joined #nixos
dj_goku has quit [(Ping timeout: 240 seconds)]
dan_b has quit [(Ping timeout: 250 seconds)]
dj_goku has joined #nixos
<clever>
catern: another thing thats neat
<clever>
catern: <nixpkgs> is translated into a function call at parse time
<clever>
catern: this generates a nixos image, then boots that image under qemu
<clever>
the image contains the xen hypervisor
<clever>
so that runs nixos, under xen, under qemu
<clever>
you can now test xen unikernels without having xen installed, without using root
<samueldr>
dunno if you saw, but sandboxed builds are fine just by including the qemu-user, the file corruption (?) only happened when adding the arm bash to the sandboxed paths
<samueldr>
now that's a bit spooky, I don't know how I could prepare a standalone test case
<>
changed the topic of #nixos to: Topic set by gchristensen!~gchristen@unaffiliated/grahamc on 2017-10-18 23:09:41 UTC
<catern>
it's basically the only way I install or upgrade Linux distros at this point
<>
changed the topic of #nixos to: Channel #nixos created on 2008-04-25 12:32:07 UTC
<clever>
and with the rescue boot option, you can auto-generate the whole ramdisk
mbrgm has quit [(Ping timeout: 276 seconds)]
<catern>
because I don't have to screw around with thumb drivers or much less CD drives
<samueldr>
the void of ping timeouts ate my messages!
mbrgm has joined #nixos
<samueldr>
[21:27:32] <samueldr> dunno if you saw, but sandboxed builds are fine just by including the qemu-user, the file corruption (?) only happened when adding the arm bash to the sandboxed paths
<samueldr>
[21:28:01] <samueldr> now that's a bit spooky, I don't know how I could prepare a standalone test case
<clever>
samueldr: does the empty file persist after the build has ran?
<samueldr>
the build didn't run because the file became empty, and yes, the empty file persisted
dj_goku has quit [(Ping timeout: 240 seconds)]
dj_goku has joined #nixos
dj_goku has quit [(Changing host)]
dj_goku has joined #nixos
srdqty has joined #nixos
srdqty has quit [(Client Quit)]
<clever>
samueldr: was the file empty before the build started?
srdqty has joined #nixos
<samueldr>
no
<clever>
that is strange
<clever>
and does nix-store --verify --check-contents complain about it being corrupt?
<samueldr>
this was verified, during the build I tried "file" on it and it emptied after the fetches to the cache
<clever>
samueldr: yeah, thats not right, its corrupting the nix store!
<samueldr>
exactly
<samueldr>
I'm having a thought
pxc has joined #nixos
<pxc>
I don't really know/understand how /nix/store gets mounted as read-only/write-once, but I'd like to do a NixOS install where /nix/store is separate from the root partition. Is there anything I should know about mount options or will Nix just figure it out if during the install I mount it as a normal R/W partition?
<samueldr>
happy to report that only the bash that will be used in the build process gets corrupted, not the gnu hello I added to the sandboxed paths
<samueldr>
since that bash can be used without issue when adding qemu-user-arm to the sandbox, without putting that bash in, could it be corrupting itself when "mounting" it again over the already included bash?
dj_goku has joined #nixos
dj_goku has quit [(Changing host)]
dj_goku has joined #nixos
marusich has joined #nixos
<samueldr>
I really don't know much about the sandboxing though
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] adisbladis pushed 2 new commits to master: https://git.io/vbfPe
<NixOS_GitHub>
nixpkgs/master 05c00a8 Benjamin Hipple: haskell docs: fix typo
<NixOS_GitHub>
nixpkgs/master c8a60e3 adisbladis: Merge pull request #32132 from bhipple/docs-typo...
NixOS_GitHub has left #nixos []
dieggsy has quit [(Quit: ERC (IRC client for Emacs 27.0.50))]
remexre has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
dj_goku has quit [(Ping timeout: 276 seconds)]
dj_goku has joined #nixos
remexre has joined #nixos
<pxc>
can /nix/store be on a separate partition while /nix is on the root partition?
hamishmack has quit [(Quit: hamishmack)]
hamishmack has joined #nixos
hamishmack has quit [(Client Quit)]
proteusguy has quit [(Remote host closed the connection)]
dj_goku has quit [(Ping timeout: 248 seconds)]
pie__ has quit [(Ping timeout: 268 seconds)]
dj_goku has joined #nixos
dj_goku has quit [(Changing host)]
dj_goku has joined #nixos
dieggsy has joined #nixos
<pxc>
I guess I'll just try it
dj_goku has quit [(Ping timeout: 260 seconds)]
dj_goku has joined #nixos
dj_goku has quit [(Changing host)]
dj_goku has joined #nixos
<eacameron>
I just update my 17.09 server to newer commit of nixos, and now my haskell app has no logs in journal and it isn't recording any...
<eacameron>
Do I need to configure something differently?
<eacameron>
It's running via systemd
<eacameron>
<facepalm> I misspelled the service.
<eacameron>
Please ignore me.
henrytill has quit [(Ping timeout: 248 seconds)]
henrytill has joined #nixos
dieggsy has quit [(Remote host closed the connection)]
dieggsy has joined #nixos
dieggsy has quit [(Client Quit)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] yrashk opened pull request #32133: tlwg: init at 0.6.4 (master...tlwg) https://git.io/vbf13
NixOS_GitHub has left #nixos []
Gohla has quit [(Ping timeout: 240 seconds)]
pxc has quit [(Ping timeout: 268 seconds)]
uwap has quit [(Ping timeout: 248 seconds)]
Gohla has joined #nixos
hamishmack has joined #nixos
uwap has joined #nixos
dj_goku has quit [(Quit: leaving)]
dj_goku has joined #nixos
noobineer has joined #nixos
noobineer has quit [(Max SendQ exceeded)]
mizu_no_oto has joined #nixos
schoppenhauer has quit [(Ping timeout: 240 seconds)]
schoppenhauer has joined #nixos
CustosLimen has quit [(Ping timeout: 252 seconds)]
asuryawanshi has quit [(Remote host closed the connection)]
ckauhaus has joined #nixos
asuryawanshi has joined #nixos
zzamboni has joined #nixos
Capprentice has joined #nixos
reinzelmann has quit [(Ping timeout: 250 seconds)]
periklis has quit [(Ping timeout: 248 seconds)]
<Lisanna>
Is there a better way to make the Nix garbage collector not delete certain builds than to manually add a symlink to /nix/var/nix/gcroots/bar?
asuryawanshi has quit [(Remote host closed the connection)]
<Lisanna>
err, /nix/varr/nix/gcroots/
<srhb>
Lisanna: Well, you don't have to manually create the roots, you can use nix-build to do it.
asuryawanshi has joined #nixos
<Lisanna>
oh?
<srhb>
Lisanna: It's what happens by default if you don't specify --no-out-link
dhess has joined #nixos
<ckauhaus>
Lisanna: I've written a small tool which registers GC roots for _unmanaged_ builds (e.g., manual configure && make)
<ckauhaus>
no need in case the build is managed via nix-build though
<Lisanna>
wait, that makes a symlink in the current directory, and makes a gcroot to that "result" symlink, so it'll break if I remove "result" from the current directory
<ckauhaus>
correct
<srhb>
Lisanna: I tend to place those symlinks in ~/gcroots or something :)
<Lisanna>
srhb: and the you have a permanent gcroot to ~/gcroots?
<ckauhaus>
if you need something permanently, put it into your user-env
<srhb>
Lisanna: No, the /nix/var/nix/gcroots just point to those.
<ckauhaus>
like nix-env -i -f ...
<Lisanna>
..but the thing I said would also work, wouldn't it? since it traverses directories
Tucky has joined #nixos
<srhb>
Lisanna: I suppose so, I never thought of that :)
<Lisanna>
ckauhaus: for certain things, I'm not sure what it would even mean to add it to the user-env. like, a derivation which just produces a text file. would nix-env'ing that keep it from being GCd?
zzamboni has quit [(Quit: Leaving.)]
<Lisanna>
since it doesn't fall under the regular categories of /bin, /share, /lib, etc.
dhess has quit [(Quit: ERC (IRC client for Emacs 25.3.2))]
<ckauhaus>
if it is referenced from ~/.nix-profile, yes
reinzelmann has joined #nixos
dhess has joined #nixos
<Lisanna>
okay, that seems a littttle bit hacky, since the user that builds it is now also responsible for maintaing a nix profile that keeps a reference to it...
townsend has joined #nixos
<ckauhaus>
...or just keep the symlink
<ckauhaus>
you have to keep anything at least
<Lisanna>
I'm liking the symlinked gcroots directory idea though, only issue is that you'd have to specify that with --out-link, which means you have to be careful to always choose a unique name
<Lisanna>
it works if it was the thing that made that symlink
zzamboni has joined #nixos
tomster has quit [(Ping timeout: 248 seconds)]
tomster has joined #nixos
<Lisanna>
interesting... if it's a symlink not made by nix-build, it fails, but if it's a symlink made by nix-build, even if for a different derivation, it will overwrite it
thc202 has joined #nixos
<Lisanna>
that's unfortunate, that means there's no safe way to do what I'm trying to do
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vbfpy
MP2E has quit [(Remote host closed the connection)]
ssmike has joined #nixos
<aw>
Hello, everybody. Anyone here who might be able to help me to set-up uwsgi with PHP?
<tomster>
gchristensen: mornin! it's me again :)
<tomster>
gchristensen: i was wondering what would be the best approach to use your upgraded expression for chefdk permanently?
<aw>
Actually, uwsgi is already working with php, but I'd like to enable APCu (which is available from phpPackages) but I'm too dumb to understand how to get uwsgi/php-embed/<some arbitrary phpPackages package> married… Oo
davidak has quit [(Ping timeout: 260 seconds)]
<ertes-w>
is there a way to make 'nix-env -u' default to '--always' without using an alias/wrapper?
<sphalerite>
Can I use a chroot/bind mount nix store, as supported by nix 1.12, as a binary cache? That is, put it in the nix.binaryCaches option
<ertes-w>
BTW, the current default is still --lt, and that's very unfortunate… i *strongly* suggest that the default becomes at least --leq
<ertes-w>
imagine that there is an OpenSSL upgrade for yet another security hole, and that you have a program installed that depends on OpenSSL
<ertes-w>
--lt will *not* update that program, so it will keep using the vulnerable version
<sphalerite>
ertes-w: I'd say just avoid nix-env -u in general, just like nix-env -i without -A
Thra11 has joined #nixos
<ertes-w>
sphalerite: why? nix-env -u is fine with --always
<ertes-w>
sphalerite: but if we're gonna have an upgrade feature, we might as well do it correctly
<sphalerite>
ertes-w: it will still often do stuff you don't actually want it to because it searches by name
<ertes-w>
sphalerite: in my case all the names start with "ertes-"
<sphalerite>
Like upgrading mumble to its unstable version. It's better to have an expression that you just install again (with -r) to upgrade
<ertes-w>
i have a bunch of package groups in my ~/.nixpkgs/config.nix
pxc has quit [(Ping timeout: 276 seconds)]
<sphalerite>
Is there a particular way nix will prioritise binary caches if more than one has the same path available?
qqq has joined #nixos
antonv has joined #nixos
<antonv>
hello, if I use nix package manager on Ubuntu, how can I install window managers / desktop environments?
<antonv>
I did nix-env -i i3
<antonv>
but the login screen doesn't suggest i3 as an desktop environment for the session
<adisbladis>
antonv: Thats expected. Your display manager does not know of i3.
<adisbladis>
antonv: You could use xsession
darlan has joined #nixos
<antonv>
adisbladis: I understand it's expected, but not sure what is the best way to solve that
darlan has quit [(Client Quit)]
<antonv>
googling about sxession
ThatDocsLady has joined #nixos
chaker has joined #nixos
ckauhaus has quit [(Quit: Leaving.)]
ilja_kuklic has joined #nixos
dan_b has quit [(Ping timeout: 260 seconds)]
ssmike has quit [(Ping timeout: 240 seconds)]
ssmike has joined #nixos
ssmike has quit [(Read error: Connection reset by peer)]
antonv has quit [(Ping timeout: 258 seconds)]
dhess has left #nixos ["ERC (IRC client for Emacs 25.3.2)"]
<sphalerite>
anton__: if you have a script called .xsession in your home directory it will supersede any choice made at login time in the display manager
<sphalerite>
anton__: you could symlink that to i3, or if you want to do more, make it a shell script that starts other stuff in the background and ends with exec i3
ilja_kuklic has quit [(Remote host closed the connection)]
ShalokShalom has quit [(Remote host closed the connection)]
sigmundv__ has joined #nixos
<sphalerite>
clever: any idea about using qemu-system-arm for a similar purpose to your qemu-user magic? I want to have a virtual ARM builder that's isolated from its host machine rather than sharing everything
<hyper_ch>
hmmm, I tried to create udev rules to let me assign usb devices to a qemu vm: https://paste.simplylinux.ch/view/65ebb9d6 however it still complains about readonly... the only thing that works this far is chmod -R 0777 /dev/bus/usb
<kuznero>
nixops is using ssh to connect to provisioned servers. Is there any way I can pass `-o "ProxyCommand=..."` to ssh that nixops is using? I am working behind the proxy and there is no other way to make ssh work other than to pass custom ProxyCommand to it...
Khetzal has quit [(Remote host closed the connection)]
patrl has joined #nixos
<patrl>
@hyper_ch thanks, I'll try this
<patrl>
I guess the "persistentKeepalive = 25;" line might make a difference
<ocharles>
Is there anyway to change the type of a configuration option without forking nixpkgs? services.grafana.database.type won't let me use `postgres`
<ocharles>
obviously I'm fixing that and pushing upstream
<ocharles>
but in the meantime it would be nice to override it
<niksnut>
don't think so
<ocharles>
I don't think so either (
<ocharles>
:(
ssmike has quit [(Remote host closed the connection)]
<ocharles>
probably for the best, it's such a niche use case that would require a fair bit of engineering
Khetzal has joined #nixos
ssmike has joined #nixos
simukis has joined #nixos
Thra11 has joined #nixos
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
tomster has joined #nixos
<hyper_ch>
patrl: that shouldn't make a difference... just check journalctl
<patrl>
guh, still getting "device not found" errors for wireguard
<patrl>
does anyone have a confirmed working wireguard config on 17.09?
<patrl>
looking at the log for the generated unit wireguard-wg_home, this is the relevant error i'm getting "ip link del dev wg_home (code=exited, status=1/FAILURE)"
amfl has quit [(Read error: Connection reset by peer)]
<gchristensen>
silver_hook: I don't know, but you might want to look in to steam-run
Neo-- has joined #nixos
amfl has joined #nixos
<silver_hook>
gchristensen: As in a) using it directly to run Lutris; or b) to base a nix expression/package for Lutris on the steam-run one?
<gchristensen>
maybe both! I don't know, but I'd start with (a) and see how it goes :)
<silver_hook>
gchristensen: Hmmm, I may take a stab at it. Lutris could actually be a pretty good solution for gaming on NixOS.
patrl has quit [(Ping timeout: 268 seconds)]
patrl has joined #nixos
ertes-w has quit [(Ping timeout: 240 seconds)]
bennofs has joined #nixos
<gchristensen>
"output path ‘/nix/store/qv6mdc3hilbxsicjgn3sivhx3w8fl0qd-firefox-59.0a1.en-US.linux-x86_64.tar.bz2’ has sha512 hash ‘0d17ys3z2wvn72r24va5k8py4grrd6z23rszisf78j8srh8rfrli0j10cq6ppypzamcpffygdkvpp236vvqz883m4dbfyz5gacd6abs’ when ‘2kh572qsw9hpxjhaqyxw9vh8rw0620rpjh9p9wi0rc9cj4xc4pzybc3yiy8lv0i7h3zss7x9xasbl7dn6bbrrxdngdwsjf805x92fy9’ was expected" :( firefox still hasn't fixed the cache coherency on
<gchristensen>
their CDN
asuryawanshi has quit [(Ping timeout: 240 seconds)]
paraseba has quit [(Remote host closed the connection)]
ertes-w has joined #nixos
bennofs has quit [(Quit: WeeChat 1.9.1)]
ylwghst has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vbJ84
<NixOS_GitHub>
nixpkgs/master 3314e18 Yurii Rashkovskii: tlwg: init at 0.6.4 (#32133)...
NixOS_GitHub has left #nixos []
Thra11 has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] garbas pushed 1 new commit to master: https://git.io/vbJ8K
<NixOS_GitHub>
nixpkgs/master 652842d Léo Gaspard: clamav module: make services.clamav.daemon.enable actually work
NixOS_GitHub has left #nixos []
sg2002 has joined #nixos
patrl has quit [(Ping timeout: 264 seconds)]
asuryawanshi has joined #nixos
ylwghst has quit [(Ping timeout: 252 seconds)]
patrl has joined #nixos
ilyaigpetrov has joined #nixos
ylwghst has joined #nixos
patrl has quit [(Client Quit)]
stphrolland has joined #nixos
asuryawanshi has quit [(Ping timeout: 248 seconds)]
<sg2002>
Hello. Can someone point me at what lib.optional does?
<stphrolland>
Hi. I'm a little lost. I want to use fluxbox, which is considered a windowManager (like awesome) and not a desktopManager (like plasma5 or gnome3). However although I have enabled fluxbox, the fluxbox entry is not listed as available wm inside the lightdm choices. I have only listed plasma5 and gnome3. I can make fluxbox run, as if it was becoming a default, when I disable plasma5 and gnome3. Any idea what config marameter should
Khetzal has quit [(Remote host closed the connection)]
oida has quit [(Ping timeout: 252 seconds)]
Oida has joined #nixos
<sg2002>
Another question from the tutorial - when I do "nix-build -A libfoo" in which folder should I do it? In the root, the one that contains README.md?
ylwghst has quit [(Ping timeout: 260 seconds)]
<mg_>
is there any way to install a nix overlay globally for the whole nixos system? All guides I see tell you to put them in $HOME/.config/nixpkgs/overlays, but I guess that only works for a single user
townsend has quit [(Remote host closed the connection)]
Khetzal has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 2 new commits to master: https://git.io/vbJ0Z
<NixOS_GitHub>
nixpkgs/master acbaefa Tuomas Tynkkynen: Revert "linux-testing: 4.14-rc8 -> 4.15-rc1"...
<NixOS_GitHub>
nixpkgs/master e8fcced Philipp Steinpass: linux-testing: 4.14-rc8 -> 4.15-rc1...
NixOS_GitHub has left #nixos []
patrl has joined #nixos
<gchristensen>
Dezgeg: ^ that is a funny push
<Dezgeg>
heh
<Dezgeg>
yay for standardized commit message formats
<gchristensen>
:)
<hyper_ch>
why is it a revert?
<sphalerite>
mg_: nixpkgs.overlays option
<sg2002>
So I've added my own variable. Now I get "undefined variable". What should I do to define it?
<Dezgeg>
I revert a non-working version and push the working version
pxc has joined #nixos
<sphalerite>
sg2002: use a let binding, a function parameter binding or a with. We'll need some more context to be able to help you any more than that
<sg2002>
sphalerite: Ok, my bad there - missed a comma.
phreedom has quit [(Quit: No Ping reply in 180 seconds.)]
nallar has joined #nixos
<ylwghst>
sg2002: I see
<ylwghst>
sg2002: There is better way to do it.
<ylwghst>
sg2002: You should use override within configuration.nix
Ross has quit [(Ping timeout: 276 seconds)]
nallar is now known as Ross
<sg2002>
ylwghst: I was thinking of, but how to install those to?
<WilliamHamilton>
hi, I'm trying to install a haskell program in my user environment via `nix-env -iA nixos.haskellPackages.brittany`. I'm on nixos-unstable, and the build fails in the test phase, so I'd like to disable the test phase, with something like dontCheck, but I don't know how to do that on the command line
<ylwghst>
sg2002: but I actually forgot how it should be done exactly. I haven't used it for long time. I can just remember I once built polybar with i3 support somehow this way.
<pcarrier>
anybody configured qt and gtk to use breeze icons+theme with a window manager like i3?
<pcarrier>
currently breeze-icons, breeze-qt5, breeze-gtk are in my env, but dolphin doesn't find icons for example
zzamboni has joined #nixos
<ylwghst>
pcarrier: how you did installed it?
<pcarrier>
logs a lot of "QPixmap::scaled: Pixmap is a null pixmap"
<sg2002>
ylwghst: But would this work with an option that was just added in the local repo. Wouldn't this look into the nixpkgs channel and fail, not finding this option?
<pcarrier>
ylwghst: all of them are in the buildEnv for my user that I nix-env -riA
<sg2002>
ylwghst: That's why I need some kind of nix-build solution.
<pcarrier>
ylwghst: not sure what's needed here.
jtojnar has quit [(Quit: jtojnar)]
<ylwghst>
sg2002: this way you can completely override the original derivation so you can make it build it exactly how you want
bitchecker has quit [(Quit: bye!)]
<ylwghst>
pcarrier: try build them system wide first by defining them in environment.systemPackages = with pkgs; []; in configuration.nix
<pcarrier>
why do they need to be system-wide?
bitchecker has joined #nixos
<pcarrier>
FWIW breeze-qt5 and breeze-icons are already system-wide for ssdm
<pcarrier>
sddm sorry
<ylwghst>
pcarrier: adn you can't use them?
<pcarrier>
IDK, how would I use them?
<ylwghst>
pcarrier: try lxapperance
<pcarrier>
I can choose them in lxappearance, and apply, and dolphin still doesn't look themed and still doesn't have icons
cybrian has quit [(Read error: Connection reset by peer)]
<Neo-->
hey all, how can you set a package as an alias? I have firefox-esr and pkgs-master.firefox in my conf (https://gist.github.com/uskudnik/996f2214f2282656914444e1def5a596), but for some reason firefox points to firefox-esr and I can't seem to find how to even access the master's firefox
<ylwghst>
pcarrier: the problem is that they aren't in usual location and dolphin problem cant reach them
<ylwghst>
pcarrier: so
<ylwghst>
pcarrier: try this install them with nix-env
<ylwghst>
pcarrier: and then
<pcarrier>
how am I supposed to tell Qt which theme to use BTW? I don't think lxappearance wrote anything but gtk config
<ylwghst>
just try if it will work, but propably it should work without it if they are already in system wide profile
<ylwghst>
pcarrier: so there shold another env variable for qt icons?
<tobiasBora>
I'd like to install cups to manage some printers on a newly created nix server, and as a nix newbie, I'm not sure how to do that. I saw that the github contains this a "cupsd.nix" package, but I'm not sure how I should use it. Should I copy paste it and modify it to fit my needs ? Or am I supposed to import it, and then, somehow, happened things to the configuration file ? For example, how could I change the
<tobiasBora>
line "Listen localhost:631" if I don't do a big copy/paste ?
<sg2002>
ylwghst: When I try to get nix-shell in my local git repo it crashes with "error: out of memory".
<Wizek>
TweyII: that sounds like a great idea! checking
Thra11 has quit [(Ping timeout: 240 seconds)]
<TweyII>
Wizek: There's probably something useful in there, though I don't know the structure of cabal2nix packages
zzamboni has joined #nixos
<TweyII>
Or whatever the cool kids use these days :)
Thra11 has joined #nixos
<Wizek>
TweyII: this is great! how have I not heard of or tried nix-shell before...
zzamboni has quit [(Ping timeout: 258 seconds)]
asuryawanshi has joined #nixos
<Wizek>
actually, that gave me the idea to: `cabal list --installed | grep -4 directory`, TweyII
<Wizek>
so it would seem 1.3 is used in there
<Wizek>
which is odd
zzamboni has joined #nixos
<TweyII>
Why's that odd?
himmAllRight has quit [(Remote host closed the connection)]
himmAllRight has joined #nixos
zagy has quit [(Quit: Leaving.)]
<sg2002>
So, after I've build my version using "nix-build -E 'with (import /path/to/local/nixpkgs{}); lm_sensors.override { sensord = true; }'" I still get the basic version whenever I try to install the build package. Is there some place where I should also register my build version?
fusion809 has joined #nixos
mkoenig has quit [(Ping timeout: 260 seconds)]
zzamboni has quit [(Ping timeout: 260 seconds)]
mkoenig has joined #nixos
<sg2002>
Is there some way to double check that the version that I install using nix-env -f . -iA lm_sensors is the same as my build?
<Wizek>
TweyII: because when I try to use ghcid with that project, I get `canonicalizePath: does not exist (No such file or directory)` and a rapid, endless reload cycle
<Wizek>
TweyII: and that seems to be a bug in directory, fixed in 1.2.3.0
<Wizek>
attempting to upgrade ghcid
drakonis has joined #nixos
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
Sonarpulse has quit [(Ping timeout: 248 seconds)]
<fusion809>
Is there an equivalent to autoremove and autoclean features of apt for nix? I've searched the nixos.org website for the works "autoremove" and "autoclean" but no results.
<fusion809>
no relevant results I should say in the case of autoclean search^
Thra11 has quit [(Ping timeout: 248 seconds)]
<maurer>
fusion809: Those features don't make sense in nix - if you install a package, its dependencies are only visible to it, not the system
<sg2002>
fusion809: nix-collect-garbage is probably what you're looking for.
<maurer>
Yeah, if you're just trying to reclaim space, nix-collect-garbage
<fusion809>
Ah thanks. I just want to clear up disk space taken up by packages I've uninstalled as du -sh /nix has shown no significant change in disk space usage.
apeyroux has quit [(Read error: Connection reset by peer)]
<Wizek>
okay, upgraded ghcid, now I'm getting `/home/wizek/sandbox/reflex-project-skeleton/frontend/frontend/src/: can't watch what isn't there!: does not exist` Which is weird, but at least it seems to be less related to nix
<tobiasBora>
Is there any stackexchange-like website for nix?
<fusion809>
Not that I'm aware of, Unix & Linux StackExchange is probably the best you're going to get
sg2002 has quit [(Read error: Connection reset by peer)]
<fusion809>
That's if you want just StackExchange / AskBot sites
sg2002 has joined #nixos
<tobiasBora>
fusion809: ok thanks. And do you know any active forum ?
asuryawanshi has quit [(Ping timeout: 260 seconds)]
<Wizek>
tobiasBora: I know of reddit.com/r/nixos, but it's not very active
asuryawanshi has joined #nixos
<TweyII>
fusion809: Are you using NixOS?
<tobiasBora>
ok thank you!
<fusion809>
Not at the moment, I'm running Arch, but I was using Nix on it. I've decided to uninstall it. I used it as a way of running VLC when Arch had a major bug in it.
<TweyII>
fusion809: You might need to ‘nix-env --delete-generations old’ to get rid of old roots
<srhb>
tobiasBora: You probably wanted "lan" in networking.domain or something.
<tobiasBora>
makes sens
<tobiasBora>
ok, I'll try that
<tobiasBora>
thanks
chaker has joined #nixos
<srhb>
tobiasBora: I guess it's *technically* legal, but I think it will cause all sorts of dns absurdities, possibly affecting avahi as well.
asuryawanshi has joined #nixos
<tobiasBora>
ok, I'll try removing it, thanks
ylwghst has quit [(Ping timeout: 276 seconds)]
<srhb>
Weird, I find totally conflicting information on whether to use FQDN or hostname there...
<srhb>
But since yours isn't FQ anyway...
ylwghst has joined #nixos
erasmas has joined #nixos
<ij>
I want to try to change something in nixpkgs. How do I test it out with by building some package? (I want to add an argument to pkgs/development/haskell-modules/generic-builder.nix.)
ylwghst has quit [(Client Quit)]
<srhb>
ij: nix-build path/to/your/clone -A someAttribute ?
<ij>
Duh! Thanks.
jb55 has quit [(Ping timeout: 276 seconds)]
<tobiasBora>
srhb: Woa, that's just great, I removed the ".lan", and no problem anymore! Thanks!
szicari has quit [(Ping timeout: 248 seconds)]
<srhb>
tobiasBora: Great! :)
patrl has joined #nixos
szicari has joined #nixos
pxc has joined #nixos
cybrian has joined #nixos
patrl has quit [(Client Quit)]
patrl has joined #nixos
pxc has quit [(Ping timeout: 260 seconds)]
ylwghst has joined #nixos
Tucky has quit [(Quit: WeeChat 1.9.1)]
Neo-- has quit [(Ping timeout: 258 seconds)]
seanparsons has quit [(Ping timeout: 248 seconds)]
<Wizek>
New question: If I `nix-env -i gitg` and launch it on elementaryOS, how come I can't resize the window? And as it happens, many GUI apps I install with nix have this limitation
<Wizek>
any ideas why and-or how I could overcome it?
<tobiasBora>
And by the way, is it possible to have two drivers in cups, the Gutenberg and the Gutenprint and gutenprintBin ? I tried to put both of them and if told me that an assert failed in gutenprintBin.
seanparsons has joined #nixos
ylwghst has quit [(Ping timeout: 240 seconds)]
<tobiasBora>
hum I meant Gutenprint and gutenprintBin.
<tobiasBora>
(By the way, what is the difference between these two drivers ?)
reinzelmann has quit [(Quit: Leaving)]
cybrian has quit [(Read error: Connection reset by peer)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vbUeV
<NixOS_GitHub>
nixpkgs/master 2e55aec Orivej Desh: altcoins.zcash: disable parallel building
NixOS_GitHub has left #nixos []
<tobiasBora>
waou gutenprint is downloading so much stuf ! O_o
<tobiasBora>
(and it compiles lot's of things also)
Lisanna2 has quit [(Quit: Page closed)]
<tommyangelo>
nixops has a pubkey/privkey pair for a host, and the host has the pubkey in the /etc/ssh/authorized_keys.d/root file, but sshd log on the host says `failed publickey for root` and nixops cannot connect. Anybody seen this before?
orivej has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 2 new commits to release-17.09: https://git.io/vbUfk
<tommyangelo>
this was first set up of a host: I set an initial nixos on the VM with `services.openssh.permitRootLogin = "yes";` but not in the nixops config, so that it would be deactivated by the nixops deploy
dywedir has joined #nixos
akaWolf has joined #nixos
erictapen has quit [(Ping timeout: 255 seconds)]
<barrucadu>
I'm having a bit of difficulty with containers. One of my containers almost always fails to start, saying it has timed out. According to the journal, it's making progress (printing out messages about services starting and suchlike), but I guess it just doesn't finish in time. Any idea how I can go about debugging this (or increasing the timeout?)
<barrucadu>
fwiw I don't think the container should take a long time to start, it's just running a few servers
DerGuteMoritz has joined #nixos
cybrian has joined #nixos
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
bennofs has quit [(Ping timeout: 240 seconds)]
cement has joined #nixos
dan_b has joined #nixos
<barrucadu>
It may just wishful thinking, but I think it tends to work better if I want a bit before trying to start it; rather than trying to stop and immediately start (or restart) the container.
fragamus has joined #nixos
civodul has quit [(Remote host closed the connection)]
sigmundv__ has quit [(Ping timeout: 276 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vbUTe
<NixOS_GitHub>
nixpkgs/staging addd20e Vladimír Čunát: Merge branch 'master' into staging
NixOS_GitHub has left #nixos []
hdaugherty has quit [(Ping timeout: 240 seconds)]
pxc has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 2 new commits to staging: https://git.io/vbUTO
<NixOS_GitHub>
nixpkgs/staging 93294fc adisbladis: pcre2: Patch for CVE-2017-8786
<NixOS_GitHub>
nixpkgs/staging e6a7a2e Vladimír Čunát: Merge #32134: pcre2: Patch for CVE-2017-8786
NixOS_GitHub has left #nixos []
tommyangelo has quit [(Ping timeout: 248 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] strout opened pull request #32144: cnijfilter2: specify --datadir for cnijlgmon3 (master...cnijfilter2-fix) https://git.io/vbUTo
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] wizeman closed pull request #30790: kodi: downgrades kodiPlugin.joystick to compatible version 1.3.2 (master...master) https://git.io/vFe97
NixOS_GitHub has left #nixos []
Neo-- has joined #nixos
JosW has joined #nixos
patrl has quit [(Ping timeout: 258 seconds)]
dywedir has quit [(Remote host closed the connection)]
goibhniu has quit [(Ping timeout: 276 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] wizeman pushed 1 new commit to release-17.09: https://git.io/vbUkk
<NixOS_GitHub>
nixpkgs/release-17.09 29a4d94 Stefan Huchler: kodi: downgrades kodiPlugin.joystick to compatible version 1.3.2...
NixOS_GitHub has left #nixos []
patrl has joined #nixos
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
aloiscochard has quit [(Quit: Connection closed for inactivity)]
<WilliamHamilton>
hi, I'm trying to install a haskell program in my user environment via nix-env -iA nixos.haskellPackages.brittany. I'm on nixos-unstable, and the build fails in the test phase, so I'd like to disable the test phase, with something like dontCheck, but I don't know how to do that on the command line
<ij>
I have a haskell package, whose "src = ./.;" and whose nix files are in the src directory, so I'll have a cache miss even when I wouldn't have one beacuse of haskell files. How could I avoid that while keeping the nix files in the repo?
<ij>
It also copies the .stack-work to the nix store, which is horrible.
<srhb>
ij: src = ./somedir ?
<ij>
I would really like to avoid that if I could. I'd basically have only two root repo directories at that point — nix and the_rest.
ambro718 has joined #nixos
jtojnar has joined #nixos
sg2002 has quit [(Ping timeout: 240 seconds)]
<tobiasBora>
Hum, it's really strange, "wget ftp://ftp.gnutls.org/gcrypt/gnutls/v3.5/gnutls-3.5.15.tar.xz " works great, but
dan_b has quit [(Ping timeout: 264 seconds)]
<tobiasBora>
"curl ftp://ftp.gnutls.org/gcrypt/gnutls/v3.5/gnutls-3.5.15.tar.xz " does not work...
<tobiasBora>
and because of that I can't download the software... Does it work for you? And if it does not, do you know if nix can use wget in place of curl ?
jgt has joined #nixos
<jgt>
on MacOS, how should I store my list of installed packages?
<jgt>
e.g., I can do nix-env -q to list packages I've installed
<jgt>
and I'd like to version control that list
<srhb>
jgt: That exact list? It may be less useful than you think, since those are just more or less arbitrary names and not the attributes used for installing packages.
sg2002 has joined #nixos
<srhb>
WilliamHamilton: Why not just override it via an overlay or similar?
<WilliamHamilton>
srhb: I was reading about overlays for a user in ~/.nix/overlays.nix or something; if that's the idea you're proposing, do you have an example of that?
<jgt>
srhb: not necessarily exactly that list
<jgt>
srhb: but I should be able to install all my stuff on a new machine in one go
cybrian has quit [(Read error: Connection reset by peer)]
Ivanych has joined #nixos
<srhb>
jgt: How about creating an overlay with the packages you want instead? Ie. one metapackage that contains them all
<WilliamHamilton>
wow, that works; you're always so helpful srhb , thanks :)
<jgt>
srhb: I don't know what that means. Is there any documentation on that?
<srhb>
WilliamHamilton: You're welcome!
<srhb>
jgt: Overlays are documented in the nixpkgs manual, but essentially, one file in ~/.config/nixpkgs/overlays/mybundle.nix containing: self: super: { mybundle = with super; [ vim git firefox ]; }
<srhb>
Actually, that's probably less than useful...
<srhb>
Scratch that
<srhb>
Actually, hmm.
<srhb>
It should work
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vbULh
<NixOS_GitHub>
nixpkgs/master c06c2cd Orivej Desh: git-up: mark as broken
NixOS_GitHub has left #nixos []
<jgt>
srhb: your example at least looks like the kind of thing I want
<srhb>
I'm just unsure if it will remove things you take off the list.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to release-17.09: https://git.io/vbUtT
<NixOS_GitHub>
nixpkgs/release-17.09 ecfcd0b Orivej Desh: git-up: mark as broken...
NixOS_GitHub has left #nixos []
<srhb>
(ie when you upgrade it)
<jgt>
ah, I see. I guess the real answer is "Use NixOS"
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 3 new commits to master: https://git.io/vbUtm
<NixOS_GitHub>
nixpkgs/master 2492f45 Andreas Rammhold: ffmpeg-3.4: apply fix CVE CVE-2017-16840...
<NixOS_GitHub>
nixpkgs/master 64d8cc7 Andreas Rammhold: ffmpeg-full-3.4: apply patch for CVE-2017-16840
<antonv>
strange thing with nix, I deleted a file and now even if I reinstall package the file does't come back
vincent_vdk has quit [(Read error: Connection reset by peer)]
<antonv>
(when manipulating with xsession files I deleted ~/.nix-profile/share/xsessions/i3.desktop, and now I uninstall and reinstall i3, the file doesn't come back)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to release-17.09: https://git.io/vbUs1
<tobiasBora>
so it means that for now, all my system will be built ?
<samueldr>
don't worry, arm platform names are confusing
<samueldr>
everything depending on anything needing a rebuild, yes
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 2 new commits to staging: https://git.io/vbUBy
<NixOS_GitHub>
nixpkgs/staging 98e3af0 Vladimír Čunát: xorg.libXcursor: security 1.1.14 -> 1.1.15...
<NixOS_GitHub>
nixpkgs/staging db6adec Vladimír Čunát: xorg.libXfont*: security update for CVE-2017-16611
NixOS_GitHub has left #nixos []
danimal has joined #nixos
<tobiasBora>
samueldr: whooa. No way to avoid that? By the way, it's possible to help to build the packages for armv6 ?
<samueldr>
I talk like I know what I'm talking about, but in reality I've only dipped my toes in arm stuff
sigmundv__ has joined #nixos
<samueldr>
and to help build, Dezgeg would know, in fact, I'd also like to provide anything I can with the arm6l/arm7l efforts
dan_b has joined #nixos
MP2E has joined #nixos
berce has joined #nixos
erictapen has quit [(Ping timeout: 276 seconds)]
<tobiasBora>
So all my system will need to compile from scratch every single package... The rasp won't like that (it's so slow to compile...), it's not possible to have several "trees", like "if you can install package A with prebuild binaries, even with different options, do that" ?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 2 new commits to release-17.09: https://git.io/vbUR2
<NixOS_GitHub>
nixpkgs/release-17.09 4f12cdf Vladimír Čunát: xorg.libXcursor: security 1.1.14 -> 1.1.15...
<NixOS_GitHub>
nixpkgs/release-17.09 efe5a97 Vladimír Čunát: xorg.libXfont*: security update for CVE-2017-16611...
NixOS_GitHub has left #nixos []
lfam has joined #nixos
erictapen has joined #nixos
ravloony has quit [(Ping timeout: 248 seconds)]
<samueldr>
tobiasBora: once built on one, it should be possible to copy the built result (e.g. with nix-copy-closure) thus needing a one-time build effort
<samueldr>
but yet, I'm not even close to an expert with all that right now
<samueldr>
(oh, didn't even open the link, only looked at the end of the url, sorry)
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
Isorkin_ has joined #nixos
<berce>
Thanks all for correcting and improving my answer :-)
sbjorn has joined #nixos
<tobiasBora>
samueldr: no, but I wrote the image maybe 2-3 weeks ago, it can cause the trouble?
ylwghst has joined #nixos
Sonarpulse has joined #nixos
Isorkin has quit [(Ping timeout: 276 seconds)]
<samueldr>
Dezgeg's cache is long-lived, not sure when the cut-off is, but pretty sure it would still be cached
<tobiasBora>
what do you mean ?
<Sonarpulse>
thoughtpolice: pong
<Ralith>
there doesn't even appear to be a terminus related conf file in /etc/fonts/conf.d/
<samueldr>
asking since channels might advance before a new build is complete, iirc it takes about 2-3 days for a complete armv6l build
<samueldr>
I mean that if you haven't updated the channel since you wrote the image, it should still pick up everything it can from the cache
<tobiasBora>
samueldr: oh, you sayed that it may be because my raspberry pi just updated the channel todya ?
<tobiasBora>
interesting
<samueldr>
sorry, no, but if you manually did update the channel, it could cause even more rebuilds
_ris has joined #nixos
<samueldr>
pretty sure you're simply falling outside of the subset
<samueldr>
I'm now curious and wonder *what* the subset is
<tobiasBora>
ok. But why did he recompile perl for example? I guess that perl is in the subset no?
fragamus has joined #nixos
joshuaks has joined #nixos
<samueldr>
now this falls outside of my knowledge, for many reasons
<samueldr>
1) I don't know how perl is packaged
<samueldr>
2) don't know why it could need a rebuilt (if build already exists)
<joshuaks>
Hey guys. I'm new to Nix and have a few questions.
<joshuaks>
If I'm asking really broad/stupid things, just send me back to the documentation.
<samueldr>
don't worry, it's better not to assume and ask broad/stupid things :)
<joshuaks>
I'm using Nix on macOS 10.12.6. Moving away from homebrew. I really have a thing for clean systems :).
<joshuaks>
Anyways. Everything has gone well except for oh-my-zsh and weechat (plugins specifically).
ylwghst has quit [(Quit: Lost terminal)]
<joshuaks>
Both install. In the case of Weechat (I find this one more interesting), it's not reading my config files to install plugins.
<joshuaks>
For OMZ, it seems like some required files (for example "
<tobiasBora>
samueldr: ok thank you. Just, does that mean that all the next packages that would rely on perl will need to be rebuilt, even if there exists a binary version with "the old perl", or both perl version will co-exist on the system?
<joshuaks>
(sorry) (for example "oh-my"zsh.sh") isn't where I'd expect it. It's in the root /nix dir under the hash dir, etc. Which is fine. But it feels wrong to point my zshrc to that "random" directory. Or is that ok?
<samueldr>
tobiasBora: I can't answer since I don't know *why* there's a rebuild, but the way nix works, two concurrent thing can exist in the system and generally isn't an issue
bennofs has joined #nixos
<tobiasBora>
samueldr: ok thank you. Hope I won't need to rebuilt everything from now ;) By the way, if you or Dezgeg have any idea how I could help the binary channel of armv6, please tell me!
Ivanych has quit [(Quit: Leaving.)]
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<samueldr>
there's one detail I don't know much enough, and may be relevant to your issues
Ivanych has joined #nixos
<samueldr>
I don't know if your log is showing complete failure from the binary cache or transient issues that were resolved
<samueldr>
I often got those "still waiting" messages when playing around with arm[67]l stuff, but never was sure why, or how to fix them
<FenTiger>
Is this the right place for Nix language questions?
<makefu>
samueldr: it might be the problem that either the proxy is too slow or your arm device' IO is too slow
ravloony has quit [(Ping timeout: 240 seconds)]
<samueldr>
in this case, I'm pretty sure that canada<->finland link + slower server is causing the issue
<samueldr>
I'm juste unsure how to confirm that the "still waiting" resolved or they timed out
<Ralith>
fontconfig bitmap font support seems totally broken in latest 17.09
<Ralith>
trying to fix it by hacking localConf but that doesn't seem to actually be written anywhere :|
<berce>
samualdr: that would be good, it's the first place I looked :-)
<tobiasBora>
samueldr: ok thanks. Just, do you know if it's possible to disable the build of the tests and example ? I really don't mind, and because it will already take 2 days to compile, don't think it's relevant to include examples ^^'
drakonis has quit [(Read error: Connection reset by peer)]
<Isorkin_>
How to need to merged pr ? 32075
miko__ has quit [(Ping timeout: 276 seconds)]
plakband has joined #nixos
<joshuaks>
Looks like I'll be doing some research. Little in over my head. That documenation helped a ton though. Thanks berce
<plakband>
I've just installed the nix package manager, and I'm reading through the manual. The manual has an example where they install firefox, but when I try to do the same, it says that there's no firefox derivation. Did I misconfigure something? I could install irssi and git just fine. I'm on OS X 10.13.
<berce>
joshuaks: about the files in /nix: when you do things right, nix does it's magic and makes all files available to the programs that need them. Just like with real magic, you don't have to understand it to make it work.
<symphorien>
plakband: how did you try to install it ?
<plakband>
symphorien: nix-env -i firefox
<joshuaks>
Yeah that's what I figured. It looks like (based off the other docs) I need to work on my config files a bit
<symphorien>
try nix-env -iA nixpkgs.firefox
<joshuaks>
"magic" has been the case for my other binaries so far. although, I can't help but look under the hood :)
stanibanani has joined #nixos
<plakband>
symphorien: ah, that gives the error that firefox is not supported for darwin, is that the reason why nix-env -i did not work?
szicari has quit [(Quit: szicari)]
<symphorien>
that is quite possible, yes. The difference is that nix-env -i looks for "derivations whose name is matching" whereas nix-env -iA is when you already know the exact attribute name.
davidak has joined #nixos
stanibanani has quit [(Ping timeout: 240 seconds)]
chaker1 has joined #nixos
erictapen has quit [(Ping timeout: 240 seconds)]
chaker has quit [(Ping timeout: 240 seconds)]
stanibanani has joined #nixos
ambro718 has joined #nixos
szicari has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 4 new commits to staging: https://git.io/vbUac
<NixOS_GitHub>
nixpkgs/staging 994cb76 John Ericson: cc-wrapper: Don't leave CMD defined after setup hook...
<NixOS_GitHub>
nixpkgs/staging f4cb1e2 John Ericson: cc-wrapper: Export env vars for objdump and readelf in setup-hook...
<NixOS_GitHub>
nixpkgs/staging 43e1137 John Ericson: cc-wrapper: Define new- and old-style cross env vars...
NixOS_GitHub has left #nixos []
fragamus has joined #nixos
<pierrebeaucamp>
hey, I hope I'm not disrupting anyones discussion, but I have a quick question: I'm experimenting with NixOS and I'm using EXWM as my wm. That means emacs gets installed as a system package. Is it possible that a user can run a different emacs with the same window manager?
<pierrebeaucamp>
I mean Emacs seems to be initialized after login, but X server starts earlier (for the login prompt)
<pierrebeaucamp>
This might also be too specific to EXWM, but I though I'd ask here
<jeaye>
I don't know anything about EXWM, but my understanding is that your display manager, which shows your login, is not running X. Only once you log in does X start and your session is bound to X being alive.
<pierrebeaucamp>
hm, interesting. In that case I could try to overwrite the exwm package from the user environment
sg2002 has quit [(Ping timeout: 276 seconds)]
<pierrebeaucamp>
I suppose it is not possible to overwrite configuration.nix with something similar from the user env?
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<sphalerite>
heh. sshfs with the reconnect option will happily try over and over again if it can't authenticate to the server… which will easily get it blacklisted by fail2ban >_>
gsmith has joined #nixos
<Ralith>
okay, wildly flailed at my fontconfig setup and how it's somehow working without any apparent changes
<gsmith>
allo, does anyone have any examples of a bridged bond configuration? for some reason the configuration that I have causes bonding to break badly to the point I sysfs is not functioning properly
* Ralith
gives up
sg2002 has joined #nixos
<pierrebeaucamp>
Ah, what I was trying to do should be able using ~/.config/nixpkgs. Sorry for asking so many beginner questions
Neo-- has quit [(Remote host closed the connection)]
Wizek_ has joined #nixos
Neo-- has joined #nixos
pierrebeaucamp has quit [(Quit: Leaving)]
acarrico has joined #nixos
<berce>
Ralith: look at public dotfiles from other users: search for "configuration.nix fontconfig" and you 'll get several working examples.
acarrico has quit [(Max SendQ exceeded)]
<Ralith>
berce: I have a working example; it does not cause me to be any less mystified as to why it broke itself temporarily.
<berce>
Ralith: you might need to make the system reload the fonts, by logging out or even rebooting
<Ralith>
it is currently working
<Ralith>
hopefully it does not stop working when I next reboot, I guess
acarrico has joined #nixos
rihards has joined #nixos
asuryawanshi has quit [(Ping timeout: 252 seconds)]
hariel has joined #nixos
ssmike has joined #nixos
stanibanani has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] zimbatm pushed 1 new commit to master: https://git.io/vbUr9
<NixOS_GitHub>
nixpkgs/master 55709e4 zimbatm: python3Packages.bash_kernel: allow different python interpreters
NixOS_GitHub has left #nixos []
wolfcub has quit [(Ping timeout: 276 seconds)]
TweyII has quit [(Ping timeout: 240 seconds)]
plakband has quit [(Quit: leaving)]
ma27 has quit [(Ping timeout: 255 seconds)]
<FenTiger>
Any Nix language experts on here? Is there a way to find the dependencies of a package, or ideally compute its closure, from within the Nix language itself? A look at the source suggests not, because dependencies seem to be represented as Nix store paths rather than as references to the original derivation. Am I missing anything?
<clever>
FenTiger: the runtime dependencies are not known at eval time
<clever>
FenTiger: after the eval and build have finished, nix will basicaly grep the output, to see what build-time deps it still refers to, and those become the runtime deps
<joshuaks>
Look into OMZ, I have another question
<joshuaks>
On what is correct
<joshuaks>
I see the install symlink at $HOME/.nix-profile/share/oh-my-zsh -- should my zshrc reference that or is that incorrect?
<FenTiger>
OK. So let's say I use the exportReferencesGraph feature to get hold of the closure from within a derivation, and then read it into a file using builtins.readFile (playing around in the REPL suggests this'll work). Is it then possible to translate the resulting store paths into references to the original definition in nixpkgs or wherever?
<gsmith>
I got the bridged bond working, thanks anyway!
gsmith has quit [(Quit: leaving)]
<joshuaks>
that symlink seems like a reliable path
<clever>
FenTiger: exportReferencesGraph creates a variable/file at build time, for the derivation its in, which i believe contains the runtime closure of the referenced thing
<clever>
FenTiger: but the eval cant access it
<FenTiger>
clever: Maybe I should try harder to prototype it, but from looking at the source and trying it in the REPL, doing "readFile (mkDerivation {...})" seems to be sufficient to "force" the derivation and leave its results in a file that can then be read back into the outer evaluation.
<clever>
FenTiger: thats doing import from derivation, which causes nix to build some things at eval time, which generally ruins performance
bennofs has quit [(Ping timeout: 276 seconds)]
hiratara has quit [(Ping timeout: 276 seconds)]
<FenTiger>
OK, but if I can do what I want with poor performance, that beats not being able to do it at all.
hamishmack has quit [(Quit: hamishmack)]
hiratara has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] zimbatm pushed 1 new commit to master: https://git.io/vbUKI
<FenTiger>
Maybe I should take a step back. I have a derivation that operates on an existing package in the Nix store and transforms it into a different format. I would like to apply this derivation to all the packages in the closure of a given package list. Any suggestions for how I should do this?
zraexy has quit [(Ping timeout: 240 seconds)]
<clever>
FenTiger: do you want to generate a list of derivations, that each transform one dependency?
<FenTiger>
clever: yes, that's right.
<clever>
FenTiger: most things like the tarball generator just put the entire closure into a single output
<clever>
so when you rebuild, the derivations that havent changed can reuse the layers they made before
<FenTiger>
I can fall back on doing it like that, but it seems a bit unclean in a functional language ;)
<clever>
and at runtime, docker puts all the layers back together
<FenTiger>
Damn, that is exactly what I was trying to build! I'll take a look at what they've done.
ma27 has joined #nixos
vcunat has quit [(Ping timeout: 264 seconds)]
sigmundv__ has quit [(Ping timeout: 248 seconds)]
chaker1 has quit [(Ping timeout: 240 seconds)]
bennofs has joined #nixos
<FenTiger>
Actually, on a closer look, I don't think hocker is what I had in mind at all. Oh well. I think you've answered my question, though: I need to do all the work in the main derivation - I can't realistically split it into smaller pieces. Thanks!
dbmikus has quit [(Quit: WeeChat 1.9.1)]
<catern>
hey #nixos
<catern>
so, I am a little confused about the proper usage of nix-support
<catern>
I can build a derivation that has a nix-support/setup-hook
<catern>
and any derivation B which has derivation A in its buildInputs, sources nix-support/setup-hook for each buildInput
<catern>
right?
<catern>
and if C is in propagatedBuildInputs somewhere in the dependency, then C/nix-support/setup-hook is sourced for *any* package above C, right?
<catern>
all at derivation build time, specifically
<catern>
right????
<clever>
catern: yeah, i believe thats all right
<clever>
catern: you could try adding echo's to the setup hook and confirm it all
sigmundv__ has joined #nixos
<catern>
oh wait
<clever>
catern: one anoying thing ive noticed, nixops has itself and python in the propagatedBuildInputs
<clever>
catern: python's setup hook adds all inputs to PYTHONPATH
<catern>
right right
<catern>
the setup-hook does whatever it wants
<catern>
it doesn't *have* to look at buildInputs
<clever>
catern: so if you enter a nix-shell with nixops in the buildInputs, nixops gets added to PYTHONPATH as well
<clever>
catern: if you then nix-build nixops, and run a new build via ./result/bin/nixops
<catern>
it can look at propagatedBuildInputs instead or as well, and that's what Python does, right?
<clever>
it obeys PYTHONPATH, and uses the old version
<clever>
and silently ignores the rebuilds
<catern>
it puts both propagatedBuildInputs and buildInputs on PYTHONPATH
acarrico has quit [(Ping timeout: 276 seconds)]
szicari has quit [(Quit: szicari)]
zzamboni has quit [(Quit: Leaving.)]
<danimal>
Hi all. Question, what could cause Nixos to prefer an older version of a derivation, when installing the environment.systemPackages? `nix-channel --list` is empty, `system.stateVersion = 17.09`.
spietz has joined #nixos
berce has quit [(Quit: leaving)]
<danimal>
The derivation in question is firefox, for which `nix-env -qaP | grep firefox` shows 22 offers. nixos.firefox is offered at 56.0.2, and this is what I got. What I was trying to install was nixpkgs.firefox at 57.
<danimal>
I'm not actually sure :p I haven't cloned nixpkgs or anything, I'm on a recent nixos install.
<digitalmentat>
do we have documentation in the manual for updating hackage-modules.nix??
mizu_no_oto has joined #nixos
<digitalmentat>
I know it's updated automatically on a schedule, but I was curious
<danimal>
I may have accidentally deleted a channel, while exploring that section of the manual. I think I tried to add a channel and then remove it, a week ago...
Ivanych has joined #nixos
<sphalerite>
danimal: there should be a nixos channel
<catern>
also, how do wrapper scripts for binaries get made??!
<sphalerite>
catern: often setupHooks live in their own derivation. But for example cmake has one, which replaces the configurePhase
<sphalerite>
catern: wrapper scripts are typically made using the (appropriately named 😉) makeWrapper function
<sphalerite>
or wrapProgram which uses makeWrapper. You could say wrapProgram is a wrapper maker wrapper.
cybrian has joined #nixos
<catern>
I see, I see
* catern
investigates
jgt has quit [(Ping timeout: 276 seconds)]
fragamus has joined #nixos
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
hamishmack has joined #nixos
ylwghst has joined #nixos
dhess` has joined #nixos
simukis has quit [(Ping timeout: 248 seconds)]
ylwghst has quit [(Quit: Lost terminal)]
ssmike has quit [(Ping timeout: 255 seconds)]
cybrian has quit [(Read error: Connection reset by peer)]
<danimal>
sphalerite: Thanks for the tips. In the end, I found that while my own user had no channels set, root had both nixos and nixpkgs channels set. With all user and configuration channels sychronized to nixos-unstable, the derivation query is much cleaner.
danimal has quit [(Quit: leaving)]
the-kenny has quit [(Ping timeout: 252 seconds)]
rihards has quit [(Quit: rihards)]
the-kenny has joined #nixos
kriztw has quit [(Remote host closed the connection)]