<{^_^}>
[nixpkgs] @xeji pushed to master « nixos/tests/flatpak: explicitly disable gdm to fix eval »: https://git.io/fAlZQ
dhess has joined #nixos
* gchristensen
does his customary switch-to-nixos-unstable-right-before-branch-off
<colemickens>
Why do some Go applications have `goDeps = ./deps.nix` and others don't?
<gchristensen>
if they're 100% vendored I don't think they need to?
<gchristensen>
^guess
<colemickens>
For example, @gchristensen, you packaged 'packet' a while back and I'd like to update it. I'm unsure what I'd need to do to update the Nix definition totally.
<colemickens>
looks like ebsarr/packet does not vendor deps, so that's probably why.
<gchristensen>
hmm
<gchristensen>
let's see!
<gchristensen>
oh I used go2nix to build that ... hmm!
<colemickens>
okay, should I do the same, or have best practices changed since the time of packaging?
<colemickens>
I'm not familiar with Go + ARM, but I'm guessing the error about GOARM isn't specific to the `packet` tool? (RE: https://logs.nix.ci/?key=nixos/nixpkgs.45923&attempt_id=3ced5555-129a-40f2-8c45-f465f45b6979 )
<tnks>
before I get lost in override code, does anyone know for sure if it's safe to call overrideAttrs on Haskell or Python packages, which have their respective overrideCabal and overridePythonAttrs functions?
<p_l>
managed to get RubyMine working as it should
<p_l>
took "just" 9 hours of searching through snippets etc. :/
<p_l>
"yay"
silver has quit [Read error: Connection reset by peer]
<samueldr>
hopefully your learned through the whole 9 hours
<p_l>
the answer might contain metric tons of sarcasm that might be a bit off putting
<samueldr>
oh, sorry :/
lassulus_ has joined #nixos
<p_l>
(tl;dr I'm currently seriously considering bootstrapping gentoo while having something to use while it sets up)
lassulus has quit [Ping timeout: 252 seconds]
lassulus_ is now known as lassulus
<p_l>
(the final solution was to use buildFHSUserEnv with ruby-mine, ruby and necessary deps)
Mateon1 has quit [Quit: Mateon1]
Mateon1 has joined #nixos
<infinisil>
Okay I'll give up on the module system for today
<infinisil>
Turns out option declarations like `foo = mkOption { options = ...; }` exist, and I didn't know until now what they do
<infinisil>
Spoiler: They add options to submodules
<infinisil>
So when you have `foo = mkOption { type = types.loaOf (types.submodule { ... }); }` somewhere, the above code will add additional options to the submodule
<clever>
infinisil: you can also have 2 different modules doing foo = mkOption { options = ...; } i believe
<clever>
infinisil: and the options they defined will be merged
<infinisil>
Indeed
<infinisil>
Any number of modules can add submodule options like that
<clever>
infinisil: it looks like you can even mix loaOf and options
<infinisil>
Yup, I just stumbled over the implementation of that
<infinisil>
Options remember what submodules they have, and the options attribute works with that
<infinisil>
Having done some changes, these instances of options now give me trouble though, they don't seem to be properly merged anymore :(
<infinisil>
And that's why I'm giving up for now, can't find the error
<infinisil>
It's so so so friggin hard to debug nix..
<clever>
infinisil: what i tend to do is expose the thing via an attribute
<clever>
then i can eval them seperately in the repl
<p_l>
hmm, why is `assert args ? localSystem -> !(args ? system || args ? platform);` triggered when mixing stable and unstable? The only thing I did was override a module according to wiki
<clever>
p_l: can you gist the files you changed?
<infinisil>
clever: Yeah, won't work quiet so well in some module system computation 4 layers deep in lambdas :P
<clever>
infinisil: yeah, you sort of need to modify the eval-modules function to return its internal state, rather then the result of merging
<{^_^}>
[nixpkgs] @romildo opened pull request #45926 → deepin-gettext-tools: init at 1.0.8 → https://git.io/fAllR
<MarkRBM_>
when trying to add pkgs.1password to environment.systemPackages I get 'undefined variable password' error on that line, do I need to escape the '1' somehow?
shreyansh_k has left #nixos ["Leaving"]
shreyansh_k has joined #nixos
<infinisil>
> <nixpkgs>/pkgs
<{^_^}>
attempt to call something which is not a function but a path, at (string):171:1
<infinisil>
> pkgs."1password"
<{^_^}>
attribute '1password' missing, at (string):171:1
<MarkRBM_>
so I have basically all the packages I want installed, do I use dotfiles the same way I would on another distro or on mac in nixos? It seems wierd to go to all the trouble to define the system dependencies and then just dump a bunch of dotfiles in the home folder. I have things like zshrc, emacs conf, tmux conf etc. Any good blog posts or similar on the topic of user specific configs?
jonreeve has quit [Ping timeout: 252 seconds]
mayhewluke has quit [Ping timeout: 246 seconds]
mayhewluke has joined #nixos
<worldofpeace>
I've been really surprised by how crazy nixpkgs gets before the branch off. Is it always like this?
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
endformationage has quit [Quit: WeeChat 1.9.1]
work_ has quit [Quit: Connection closed for inactivity]
<c15ade4>
MarkRBM_: there is something called home manager
<c15ade4>
I haven't tried it
<{^_^}>
[nixpkgs] @xeji pushed commit from @ckauhaus to staging « poppler 0.61: patch against CVE-2018-13988 (#45916) »: https://git.io/fAl80
<{^_^}>
[nixpkgs] @xeji pushed commit from @romildo to master « deepin-gettext-tools: init at 1.0.8 (#45926) »: https://git.io/fAl8w
worldofpeace has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @xeji merged pull request #45924 → bird: set reloadIfChanged to true → https://git.io/fAlnM
<{^_^}>
[nixpkgs] @xeji pushed commit from @flokli to master « bird: set reloadIfChanged to true (#45924) »: https://git.io/fAl8o
<drakonis>
c15ade4, its nice
logzet has joined #nixos
semilattice has joined #nixos
jD91mZM2 has joined #nixos
MarkOtaris has quit [Quit: issued !quit command]
<MarkRBM_>
hmm I am lookng at it now and I guess I dont really understand its unique selling point. In my head how it would work was I would tell nix about my dotfiles and where they need to be symlinked to and maybe get some rllback niceness if I mess them up but maybe thats out of scope for nix
<MarkRBM_>
with emacs for example I can see how to install it and do some config for it through nix and even install emacs packages. but I have a bunch of elisp files with various settings that I dont see how to manage with nix
drakonis has quit [Remote host closed the connection]
<MarkRBM_>
actually this repo makes me think there is a patch to what I would like to do with home manager github.com/yrashk/nix-home/blob/master/home.nix
<{^_^}>
[nixpkgs] @xeji pushed commit from @danieldk to master « backblaze-b2: 1.1.0 -> 1.3.6 (#45929) »: https://git.io/fAlBC
simukis has joined #nixos
xeji has joined #nixos
Fare has quit [Ping timeout: 245 seconds]
tg has quit [Excess Flood]
tg has joined #nixos
Izorkin has joined #nixos
gspia has quit [Quit: Leaving]
slack1256 has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @xeji opened pull request #45931 → nixos/release.nix: disable blivet test → https://git.io/fAlBQ
<{^_^}>
[nixpkgs] @etu opened pull request #45932 → nixos/switch-to-configuration: Fix reload of user units for users of fish shell → https://git.io/fAlB7
gspia has joined #nixos
<{^_^}>
[nixpkgs] @utdemir opened pull request #45934 → bluejeans-gui: init at 1.6.39 → https://git.io/fAlBd
<rfold>
Does the gnatmake package from nixpkgs some sort of custom version of gnatmake? It does not support the --version flag, causing dependent packages to fail configuring.
<clever>
> gnatmake.src.urls
<{^_^}>
undefined variable 'gnatmake' at (string):171:1
<clever>
rfold: i dont see gnatmake in nixpkgs
Thra11 has quit [Quit: WeeChat 1.4]
Ariakenom has quit [Ping timeout: 244 seconds]
<rfold>
clever: sorry, the package is called gnatboot and it has an executable gnatmake
<clever>
> gnatboot.src.urls
<{^_^}>
undefined variable 'gnatboot' at (string):171:1
<clever>
rfold: not sure where the source is, the wrapper is complicating things
<rfold>
Hmm thanks, seems to be a wrapper script for Gentoo.
bahamas has quit [Ping timeout: 272 seconds]
<rfold>
I'll see if I can patch it or work around its limitations.
<rfold>
Oh I think gnatboot is a package for bootstrapping GNAT, and so it has different tools.
<Mic92_>
can nixops already deploy aarch64 from x86?
<Mic92_>
or in other words, can I set the target system architecture somewhere?
<gchristensen>
sure
<Mic92_>
what do I need to set?
<gchristensen>
just like you would in any configuration.nix ... iirc it is nixpkgs.system = "aarch64-linux", but maybe Cross stuff changed that
booglewoogle has joined #nixos
<infinisil>
There's nixpkgs.localSystem and nixpkgs.crossSystem
<booglewoogle>
hey, anyone here using a chromium-based browser? it seems the pepperFlash and widevine plugins aren't working for vivaldi. maybe this is somehow connected to chromium being broken? https://pastebin.com/7VvCi4DZ
iyzsong-x has quit [Ping timeout: 252 seconds]
Ridout has joined #nixos
<Mic92_>
infinisil: thanks, I had to set nixpkgs.localSystem.system, not nixpkgs.system
fendor has joined #nixos
fenedor has quit [Read error: Connection reset by peer]
booglewoogle has quit [Read error: Connection reset by peer]
<Thra11>
Running "nix -u -eq" can reduce the number of dependencies floating around the system, but is there a command which can just list the 'duplicates' in a profile? e.g. tell me how many different glibcs I have on my system.
<{^_^}>
anonymous function at /var/lib/nixbot/state/nixpkgs/pkgs/build-support/fetchurl/boot.nix:5:1 called with unexpected argument 'recursiveHash', at /var/lib/nixbot/state/nixpkgs/pkgs/build-support/fetchz...
<infinisil>
There are so many eval errors hidden in nixpkgs..
<infinisil>
I have a branch of nixpkgs where I get rid of all such errors which prevent package evaluation
<Dezgeg>
ok the 'recursiveHash' wasn't something I expected :P
<infinisil>
Same
endformationage has joined #nixos
MinceR has quit [Ping timeout: 246 seconds]
drakonis has quit [Remote host closed the connection]
MinceR has joined #nixos
jdnavarro has quit [Quit: WeeChat 2.1]
v0latil3 has joined #nixos
<pie__>
symphorien, where does it output the vm image?
* pie__
looks for docs meanwhile
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jdnavarro has joined #nixos
hotfuzz has joined #nixos
<symphorien>
pie__: it prints the path
<symphorien>
I guess a result symlink
<pie__>
ok
kenshinC1 has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @jtojnar pushed 74 commits to gnome-3.30: https://git.io/fAlPO
<pie__>
symphorien, so, no grub, that means i should remove the enable grub stuff in configuration.nix right? will it still be able to boot?
<{^_^}>
[nixpkgs] @etu closed pull request #45932 → nixos/switch-to-configuration: Fix reload of user units for users of fish shell → https://git.io/fAlB7
<symphorien>
the grub stuff will just be ignored
<symphorien>
and yes it will boot like a charm
<pie__>
awesome
<{^_^}>
[nixpkgs] @Ma27 closed pull request #45945 → nixos/switch-to-configuration: support shells which aren't POSIX compliant → https://git.io/fAlzJ
baimafeima has quit [Quit: Leaving]
baimafeima has joined #nixos
MarkRBM_ has quit [Ping timeout: 246 seconds]
<Thra11>
Nix aborted an operation due to lack of disk space, and now I'm getting errors like "error (ignored): aborting transaction: SQL logic error (in '/nix/var/nix/db/db.sqlite')" Is there a way to magically fix the database?
<pie__>
thats odd. id have figured nixos would be doing everything sufficiently atomically at this point ¯\_(ツ)_/¯
baimafeima has quit [Remote host closed the connection]
<Thra11>
Also, that error happened during nix-collect-garbage, and seems to have resulted in it using *more* space than before.
<Thra11>
Maybe I need a certain amount of free disk space to successfully collect the garbage?
<pie__>
symphorien, so how can i control how much disk space it gets and whatnot?
<symphorien>
no idea
<pie__>
"The VM mounts the Nix store of the host through the 9P file system." uh so the vm image may not actually be transferrable
<symphorien>
I know there is a mksquashfs function in nixpkgs
<yegortimoshenko>
is there some way to do something like `nix-store --query --requisites`, or libstore's `computeFSClosure` from within a nix expression?
<Orbstheorem>
Hello, why does builtins.toPath requires absolute paths? where as paths can be relative to the nix file?
<{^_^}>
[nixpkgs] @grahamc merged pull request #44526 → nixos/activation: Identifies the snippet that failed → https://git.io/fNXQg
<infinisil>
It would be nice to have https://github.com/NixOS/nixpkgs/pull/42793 in 18.09, to reduce everybody's closure size, but I really don't feel like cleaning this PR up right now
<gchristensen>
wayyyyy out of my wheelhouse but I know who to ping@
<gchristensen>
!
<johanot>
please ping away :)
aarvar has quit [Ping timeout: 240 seconds]
<infinisil>
johanot: Um, how many options does this add?
<infinisil>
Because I'd prefer not adding too many, since they add eval time for everybody, even though 99.99% won't use these options
<clever>
infinisil: i just had an idea on how to further improve performance
<gchristensen>
^ this is a challenge we need to solve: capturing these hugely, hugely valuable services while also protecting a commons
<clever>
if the type of options could somehow depend on the config tree?
<infinisil>
clever: I think they can
<clever>
if services.nginx.enable = false, then services.nginx is just an untyped attrset, and essentially ignored
<clever>
NOP out every option, dont even put them in the options tree
<clever>
it would need an off switch, because it would even vanish from the docs :P
<infinisil>
Not possible unfortunately, how the module system works currently at least
<johanot>
infinisil: haven't counted to be honest. added a handful I think, but also removed some. srhb and I actually talked about maybe moving the k8s module to a separate repo. which has.. both pros and cons I guess.
<infinisil>
clever: It first needs to go through all modules to find valid options, then evaluate the config of each module and check if the values you assigned are valid options
v0latil3 has quit [Ping timeout: 250 seconds]
<clever>
infinisil: yeah, it would be a drastic change in how things currently work
v0latil3 has joined #nixos
<gchristensen>
johanot: I _love_ those new docs!
<infinisil>
Oh and actually, I think submodules can be lazy, but they aren't effectively because the man pages include every options
<infinisil>
option*
<johanot>
clever infinisil how about "rm module-list.nix" ? :P
<samueldr>
there was a proposal to let the end-user fill in their modules list
<infinisil>
johanot: If you mean to include modules manually, that would work, but it would remove one of the selling points of nix, namely that you can just `services.foo.enable = true;` and it works
<samueldr>
but IIRC it's tricky since some modules will reference options from other modules
<infinisil>
Oh, but the module can be enabled by default then
<gchristensen>
*nixos
<clever>
johanot: oh, thats an idea, just omit modules, and require the user to do `imports = [ <nixpkgs/nixos/modules/services/web-servers/nginx/default.nix> ];`
<clever>
johanot: i did try doing something similar, with the flag to disable modules, but it ran into dependency problems
* johanot
reading
<johanot>
clever: I think it's a nice idea in theory, but it would probably require a major refactor of the entire module tree.
periklis has joined #nixos
<samueldr>
it may very well need one!
<johanot>
infinisil: I especially like the "Migration path: TODO" in that gist :D
<samueldr>
let's hope it's as transparent as possible to the end-users
<infinisil>
Heh yeah
<Orbstheorem>
found lib.strings.fileContents ^^
<johanot>
gchristensen: Thanks. Credit to srhb again who always advocates for more docs. I got nudged a couple of times there. :)
<gchristensen>
yay!
<gchristensen>
I should send her a sticker or something as a thank you
<samueldr>
oof, that's a lot of changes johanot :)
<johanot>
samueldr: sorry :( I mean... Especially sorry for squashing new features and just moval of existing code together. 2000 additions and 2000 deletions is a bit misleading because of that.
<samueldr>
ah, then it's probably not so bad
<johanot>
you live you learn
<pie__>
aszlig, hey, you're listed as a maintainer for gajim, any chance for updating it to a newer version?
asslick has joined #nixos
<infinisil>
What the hell: fixed-output derivation produced path '/nix/store/ingwqhaav953m4m4yhzjhf7sna87xlha-cabal-helper-0.8.1.0.tar.gz' with sha256 hash '098jjdjzkzpr4lgb78fhy9k80r5m96shyj39km69m5x4ls16cj4g' instead of the expected hash '098jjdjzkzpr4lgb78fhy9k80r5m96shyj39km69m5x4ls16cj3z'
periklis has quit [Ping timeout: 250 seconds]
<infinisil>
Those hashes differ by 2 letters..
<samueldr>
that's eerily close
<emily>
infinisil: are you sure that's just not someone typoing
<{^_^}>
[nixpkgs] @cjpbirkbeck opened pull request #45956 → Added an example for environment.variable. → https://git.io/fAl9k
<emily>
when copying it
<emily>
sounds more plausible than sha256 suddenly breaking
<infinisil>
Definitely, but I have to first find where this hash is coming from
<clever>
infinisil: its generally done to fix build errors because of mistakes in the version constraints
<clever>
infinisil: the odd part, is that the hashes in your error (which match the 2 cabal files perfectly), arent the hashes that had a single-bit error
<clever>
i was hoping/worrying that the cabal files where a near hash collision :P
periklis has quit [Ping timeout: 240 seconds]
<infinisil>
Ah yeah, it was not the hash of the cabal file, but of the package itself
<clever>
yeah, but cabal revisions dont effect the hash of the source
<infinisil>
Yup, just learned that
<{^_^}>
[nixpkgs] @xeji merged pull request #45956 → Added an example for environment.variable. → https://git.io/fAl9k
<{^_^}>
[nixpkgs] @xeji pushed commit from @cjpbirkbeck to master « Added an example for environment.variable. (#45956) »: https://git.io/fAlHC
<infinisil>
Phew, at last, I have succeeded in updating haskell-ide-engine, and it works!
<philippD>
clever: I would rather fix the underlying issue that the cert file is missing.
<{^_^}>
[nixpkgs] @samueldr pushed 0 commits to release-18.09: https://git.io/fAldl
selfsymmetric-mu has quit [Remote host closed the connection]
<gchristensen>
vcunat++
<{^_^}>
vcunat's karma got increased to 1
<gchristensen>
samueldr++
<{^_^}>
samueldr's karma got increased to 18
cransom has quit [Ping timeout: 252 seconds]
<p_l>
clever: btw, yesterday's attempt at getting LXD from unstable didn't work ultimately - it required overriding package list globally to include unstable lxd in place of nixpkgs.lxd
<clever>
p_l: thats usually required with services
<{^_^}>
input-output-hk/stack2nix#126 (by domenkozar, 6 hours ago, open): Pin revision according to stackage plan
cransom has joined #nixos
Dedalo has quit [Client Quit]
<cransom>
and thank you kexec_nixos... finally got around to zfs'ing my linode.
<clever>
infinisil: aha!
Denommus` has joined #nixos
Denommus has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @samueldr opened pull request #45958 → nixos: Sets default channel for 18.09 → https://git.io/fAlFq
IRCsum has joined #nixos
selfsymmetric-mu has joined #nixos
<selfsymmetric-mu>
I'm having trouble with `kdeconnect`. I've installed the package and opened some ports. However (I'm on XFCE), there does not seem to be a working GUI. My phone can see my desktop client, but not vice versa. Anybody have success with it?
<selfsymmetric-mu>
More generally I'm looking for a way to get Android notifications through an API. Another option for that would be PushBullet's websocket interface.
<selfsymmetric-mu>
But maybe there are other solutions that I haven't seen yet?
<dhess>
probably a stupid question but I'm jetlagged and not exactly thinking straight. Is there a way to set a NIX_PATH component in a .nix file? i.e., do what -I does ?
<gchristensen>
..yes... but usually doing that is a bad path to walk down
* dhess
cries
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
semilattice has joined #nixos
<gchristensen>
...why?:)
<aszlig>
pie__: it is already up to date, the current version is only a release canidate
<semilattice>
anyway to get the root paths for the current nix-shell?
<aszlig>
pie__: unless i missed something?
Thra11 has quit [Ping timeout: 244 seconds]
<semilattice>
so if I run 'nix run nixpkgs.hello', I want it to include /nix/store/nic2bl8ry6vfyxr9717983d5b2l4sn1c-hello-2.10
<dhess>
gchristensen: there's probably a better way to do what I'm doing, but with my NixOps deployments, I've started using overlays to provide configuration state, so that I can easily swap out production for testing or whatever.
<dhess>
and right now I do that via -I state=path/to/state
<aszlig>
pie__: and last time i checked a few plugins don't yet work with the rc
theunknownxy has quit [Quit: theunknownxy]
<dhess>
and I'd like to provide a default path if i don't provide one
<semilattice>
the end goal in this case is to be able to use nix-shell/nix run for vim plugins (or other non-bin portions)
<symphorien>
semilattice: nix-build "<nixpkgs>" -A hello
<symphorien>
and --no-out-link
<semilattice>
@symphorien, I'm not asking for building specific packages. I'm saying, if I run an arbitrary nix-shell or nix run command
<dhess>
I can probably just switch to args rather than path and then provide a default argument
<semilattice>
I want to know the root(s) associated with that command, while inside of it
camsbury has joined #nixos
sigmundv has quit [Ping timeout: 246 seconds]
<symphorien>
I don't understand
camsbury has quit [Client Quit]
<symphorien>
ah maybe: inside nix-shell -p hello, the path to hello is in $buildInputs
<symphorien>
no idea for nix run
<semilattice>
symphorien: that actually works for nix-shell. That's exactly what I wanted
kisik21 has joined #nixos
mayhewluke has quit [Ping timeout: 252 seconds]
semilattice has quit [Ping timeout: 272 seconds]
mayhewluke has joined #nixos
glasserc has quit [Read error: Connection reset by peer]
worldofpeace has quit [Quit: worldofpeace]
georges-duperon has quit [Ping timeout: 252 seconds]
selfsymmetric-mu has quit [Remote host closed the connection]
<philippD>
I have a strange issue that I can build on travis-ci and my lokal NixOs machine but on my arch machine I get this error:"collision between `/nix/store/fw9d7b7mr0nws6ddvxrkflg4f22kbndb-texlive-tetex-3.0/scripts/texlive/texconfig.sh' and `/nix/store/7xlkmw909saqvk8m3c0mpjv95na1xbw3-texconfig-2017/scripts/texlive/texconfig.sh'"
<wpcarro>
does anyone have a configuration.nix file available that is known to work on a Raspberry Pi 2? Having a lot of trouble getting this to work
<wpcarro>
I have some prompt that say root@nixos, at which point I'm supposed to edit the existing configuration.nix file. I put what was on the wiki in, but that is yielding me errors
<samueldr>
wpcarro: what kind of errors?
<wpcarro>
error: while evaluating the attribute 'config.system.build.toplevel' at /nix/var/nix/profiles/per-user/root/channels/nixos/nixos/lib/eval-config.nix:53:5:
<samueldr>
also note, you should use `boot.kernelPackages = pkgs.linuxPackages_latest` and not the _rpi packages
<wpcarro>
samueldr: I'm using pkgs.linuxPackages_latest
<samueldr>
good
<samueldr>
(just a common gotcha, since the other one says _rpi)
<wpcarro>
samueldr: I'm not connected to internet but I was expecting that to be a problem down the line
<wpcarro>
I'm also getting a Bad initial deap size 100000 warning
<wpcarro>
and then `error: invalid integer '4294967295'`
<wpcarro>
I'm surprised there isn't an existing configuration.nix file that is known to work for RPi2. Feels like a common use case
<{^_^}>
[nixpkgs] @tadfisher opened pull request #45963 → uefitool: init at 0.25.1 → https://git.io/fAlpy
justanotheruser has joined #nixos
<aszlig>
gchristensen: well, docbook isn't bad per-se, but i think it has its own learning curve and i probably wouldn't call it beginner-friendly if you want to do something more substantial things than simple formatting
<slabity>
Has anyone here got Plasma/KDE working with i3 in NixOS?
<aszlig>
gchristensen: i actually prefer docbook to something like eg. markdown
<gchristensen>
me too!!
MarkRBM_ has joined #nixos
<aszlig>
because it allows to properly cross-reference stuff and it's easier to transform
<gchristensen>
and has real semantics which you can predict :)
<aszlig>
i mean, if it's just some comment, something like markdown is okay, but for technical documentation i wouldn't recommend it =)
drakonis_ has quit [Remote host closed the connection]
<aszlig>
also there is restructured text, which while it's kinda similar to markdown in readability is better suited for technical documentation and you can extend it using roles
<aszlig>
s/d text/dtext/ ;-)
<aszlig>
but still, it's not as powerful as docbook
<infinisil>
aszlig: xml will always be a big hurdle for getting lots of people to write docs though :/
rprije has joined #nixos
astrofog has joined #nixos
<aszlig>
infinisil: yeah, unfortunately... however just writing docs is probably as easy as writing TeX
<infinisil>
I admit that I almost never wrote docs for nixpkgs/nixos myself for my whole time in the nix community (1.5 years), while I wrote 2 wiki entries already since it's rebirth (half a year ago?)
<aszlig>
infinisil: but you did write option descriptions, right?
<infinisil>
Ah yeah, good point, I did
<aszlig>
because if yes then you did write docs for nixos ;-)
jluttine has quit [Ping timeout: 272 seconds]
<infinisil>
Indeed :)
jluttine has joined #nixos
<infinisil>
(Would've been *much* quicker with markdown though..)
<infinisil>
Also: I often see option descriptions having malformed xml, because people assumed it's markdown
<aszlig>
infinisil: yeah, but how would you cross-reference something from an option to another page in the manual?
drakonis has joined #nixos
<aszlig>
(btw. one thing that's also on my todo list is cross-referencing <option/>)
<infinisil>
I personally haven't ever needed this, but it might arise when I write more docs I guess
<infinisil>
Ah yeah, option cross references would be nice tbh
<infinisil>
I guess that's a point for xml
<aszlig>
hm, lemme implement that real quick to see how many bad references we have there :-D
hl` has joined #nixos
<infinisil>
But still, I think the disadvantages of docbook xml outweigh the ones of markdown, will take a bit more than that to convince me :P
phreedom_ has joined #nixos
phreedom has quit [Ping timeout: 250 seconds]
ilikeheaps has quit [Ping timeout: 246 seconds]
edef_ has joined #nixos
edef has quit [Killed (hitchcock.freenode.net (Nickname regained by services))]
edef_ is now known as edef
q3k1 has joined #nixos
hl has quit [Ping timeout: 240 seconds]
q3k has quit [Ping timeout: 240 seconds]
nbp has quit [Ping timeout: 240 seconds]
hl` is now known as hl
<aszlig>
infinisil: well, then maybe take a look at restructuredtext, as it's better suited than markdown, as mentioned before
* emily
likes ReST too
shreyansh_k has quit [Ping timeout: 260 seconds]
davidcl has quit [Ping timeout: 260 seconds]
<emily>
sphinx is a nice manual generator
<infinisil>
aszlig: Rst works well with [cross-]references?