gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
ekleog has joined #nixos-chat
Synthetica has quit [Quit: Connection closed for inactivity]
jasongrossman has joined #nixos-chat
drakonis1 has joined #nixos-chat
drakonis has joined #nixos-chat
drakonis1 has quit [Ping timeout: 258 seconds]
jackdk has joined #nixos-chat
jasongrossman has quit [Ping timeout: 268 seconds]
jtojnar has quit [Read error: Connection reset by peer]
jtojnar_ has joined #nixos-chat
jtojnar_ has quit [Quit: jtojnar_]
jasongrossman has joined #nixos-chat
Drakonis__ has joined #nixos-chat
Drakonis__ has quit [Ping timeout: 244 seconds]
jasongrossman has quit [Ping timeout: 268 seconds]
drakonis has quit [Ping timeout: 258 seconds]
drakonis has joined #nixos-chat
cjpbirkbeck has quit [Quit: Quitting now.]
endformationage has quit [Ping timeout: 244 seconds]
jasongrossman has joined #nixos-chat
drakonis_ has joined #nixos-chat
drakonis has quit [Ping timeout: 246 seconds]
jasongrossman has quit [Ping timeout: 246 seconds]
__monty__ has joined #nixos-chat
jasongrossman has joined #nixos-chat
<eyJhb> ,locate touch
<{^_^}> Found in packages: atom, z88dk, gwt240, radarr, sonarr, toybox, busybox, brackets, atom-beta, coreutils, plan9port, coreutils-full, gnome3.caribou, nodePackages.bower, plasma-workspace.bin, gnome3.gnome-devel-docs
<infinisil> ,locate bin touch
<{^_^}> Found in packages: toybox, busybox, coreutils, coreutils-full
ashkitten has quit [Quit: WeeChat 2.4]
ashkitten has joined #nixos-chat
<eyJhb> infinisil: you are always after me with this :D
<eyJhb> I thought of it right after
<infinisil> :P
<ashkitten> modern build systems are so wild because if you configure them wrong you can accidentally start building giant runtimes and stuff
<sphalerite> MichaelRaskin: woo!
<MichaelRaskin> Still not using xpra for client isolation in X session, though
<pie_> i cant remember who the fancy people who i as talking to about dynamic linking stuff was last time but maybe some of you fabulous people can help
<pie_> im trying to use dlopen in chicken scheme to load C APIs at runtime
<pie_> which is fine
<pie_> but then i want to actually call the funcitons somehow, and its not clear to me how i can write a wrapper that just takes a variable number of arguments of whatever types and then calls the function pointer
<pie_> should i write a format string interpreter for function pointers? :V
<MichaelRaskin> pie_: do you know how to do it with chicken for a fixed number of arguments?
<pie_> MichaelRaskin, so far it looks like just calling inline C
<pie_> MichaelRaskin, my current mess that works https://bpaste.net/show/0ed724810243
<MichaelRaskin> I would just use macros to generate known-argument-list calls
<pie_> mayb eim not thinking far enough ahead but i want to be able to call arbitrary c functions at runtime
<pie_> ^i mean at whatever you just said
<MichaelRaskin> Hm. Runtime compilation is probably indeed too costly
<pie_> at a real low level my understanding is that i "just" need to push the appropriate number of arguments on the stack
<pie_> well and or something about argument numbers and registers but...
<pie_> lol i dont really want to construct my own calling convention handlers in asm
<eyJhb> infinisil:don't tease the people that have to worry about that :p
<MichaelRaskin> pie_ : maybe you could use some proper Chicken FFI facilities
<pie_> MichaelRaskin, yeah im peeking through https://wiki.call-cc.org/man/5/Module%20(chicken%20foreign) and http://wiki.call-cc.org/eggref/5/bind#bind-type but i havent seen anything for this
<pie_> i think it all results in generated C code, which is what chicken compiles to
<pie_> man i dont know if its just my search bubble but scheme is f***ing hard to get relevent google results for
<MichaelRaskin> Maybe use libffi?
<pie_> hm sounds like an idea
<pie_> MichaelRaskin, coincidentally, https://wiki.call-cc.org/eggref/4/lazy-ffi
<pie_> This facility uses libffi and dynamic loading (via dlopen(3)) to load shared libraries and construct foreign calls at runtime and in interpreted code.
<pie_> its for the last major versin tough
<pie_> wonder how hard it would be to port
<pie_> hmm its also got an unported dependency >.>
__monty__ has quit [Quit: leaving]
pie_ has quit [Ping timeout: 258 seconds]
endformationage has joined #nixos-chat
drakonis_ has quit [Ping timeout: 244 seconds]
drakonis_ has joined #nixos-chat
pie_ has joined #nixos-chat
emily has quit [Ping timeout: 250 seconds]
<infinisil> I finally detached the last couple hoses from apple
<infinisil> Using self-hosted mail now, self-hosted contacts, calendars and such
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.2)]
<infinisil> (which I still need for my iPhone)
<infinisil> Well I still depend on apple hardware, but at least not icloud anymore
<infinisil> The only thing left is that the apple email forwards everything to my new one
<infinisil> Eventually I should be able to turn that off too and delete my icloud account (if that's even possible..)
<eyJhb> *and throw out the iPhone too
<eyJhb> ? :p
<infinisil> eyJhb: I'm waiting on my Librem 5, so yes!
<infinisil> Eventually
<MichaelRaskin> Could we please go back to thick 5'' smartphones?
* infinisil wouldn't mind
<eyJhb> infinisil: will you give me to me too??
<infinisil> eyJhb: error: parse error on input 'me to me'
<infinisil> (I don't get that question)
<eyJhb> infinisil: bad parser.
<eyJhb> Will you give one to me too**
<infinisil> Ahh
<infinisil> I only got one, so nope :)
<eyJhb> Does.. Does that make sense? I am not quite sure I work anymore
<infinisil> But you can still order them
<infinisil> I think
<eyJhb> I just ordered my trip to Cape Town with Uni. I have to live off oatmeal and air pretty much :p
<infinisil> I sure like to have air around me
<infinisil> You could say I am a fan of air
<infinisil> Hah
<eyJhb> I kinda feel like it invades my personal space :(
<infinisil> I feel ya
<joepie91> [18:23] <infinisil> You could say I am a fan of air
<joepie91> then you are about to become a valuable commodity in the Netherlands...
<joepie91> the run on air-cooling devices has started
<joepie91> (forecasted 39 celsius next friday)
drakonis has joined #nixos-chat
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos-chat
drakonis_ has quit [Ping timeout: 248 seconds]
<manveru> yeah, such fun weather next week, just when i get back from holidays :P
<manveru> just setup my portable air-conditioner again to hopefully make my office survivable, living under a roof sucks :(
<aanderse> manveru: i took a brief look at the bundlerApp stuff for redmine and became confused...
<manveru> anything i can help with?
<aanderse> any examples with bundlerApp that src is provided?
<aanderse> i didn't see any examples of bundlerApp with that
<manveru> i don't think that's a thing
<aanderse> oh
<aanderse> i don't think we can use bundlerApp for redmine then
<manveru> you use bundlerApp to expose a single executable of a built Gemfile.lock
<aanderse> ah
<aanderse> ok
<manveru> (or more)
<aanderse> thanks for clarifying then
<manveru> i'm not familiar with redmine, but usually it should work with that too
<aanderse> redmine has a gemfile but the gemfile doesn't provide the actual source code for redmine, just the deps (i think)
<aanderse> not a ruby guy... so could be wrong
<manveru> yeah, that's usual for rails apps for example
<manveru> you'd still use bundlerApp to execute `rails` for example inside the source directory
<manveru> inside the service definition
<manveru> i think gitlab uses that for example
<aanderse> ok, i'll take a look at gitlab
<aanderse> thanks
<manveru> hm, no, they still use bundlerEnv
<aanderse> ah
<manveru> that makes my eyes hurt :(
<aanderse> ha ha ha
<manveru> looking at the redmine module atm
<manveru> `bundle exec` is never acceptable :)
<aanderse> help from ruby pro appreciated!
<manveru> if you use bundlerApp you can directly call `rake` or `rails`
<aanderse> ok, so use bundlerApp to package rails, not redmine?
<manveru> like this
<manveru> then add `cfg.package.rubyApps` to `path` in `systemd.services.redmine`
<manveru> maybe a better name, but well :)
<aanderse> thanks :)
<aanderse> trying
<manveru> i really should go through our modules and clean them up some day... i just don't use any of them myself :|
<manveru> oh, also you won't need the `buildInputs` in the redmine derivation
<aanderse> mhm
<aanderse> after reading a few other things i'm starting to think a few other things could be dropped here too
<aanderse> i'll bug you about that after i get this going, though :)
<manveru> likely :)
<aanderse> i think this was my first or second nixos module to work on, and it was long abandoned when i picked it up
<aanderse> so... it isn't the greatest quality
<manveru> things tend to get messy because people try so many different ways to get ruby apps going
<manveru> it's not very clear what's going on
<manveru> guess we could drop 3.x with 19.09?
<cocreature> you can do static linking with the GHC distributed in arch just fine
<cocreature> oops I was scrolled up, sry
<manveru> aanderse: i like the tmpfiles thing, first time i see that
<aanderse> manveru: yeah that has been a moreso recent thing in nixos that i've been doing
<manveru> usually i just symlink the temporary directories from the source dir
drakonis has quit [Read error: Connection reset by peer]
<aanderse> tmpfiles is misleading as a name. the files don't need to be temporary
drakonis_ has joined #nixos-chat
<manveru> aye
<aanderse> we're trying to get upstream to change/alias the name
<manveru> i figured :)
<manveru> i'm still tinkering on bundix to make it more user-friendly and support gemset.json
<aanderse> getting error "called without required argument 'pname'" for the bundlerApp bit
<aanderse> so i try pname
<aanderse> few more errors
<aanderse> called with unexpected argument 'groups'
<manveru> hmm
<aanderse> we discussed that a week or so ago
<aanderse> i was saying that i need that 'groups' bit, otherwise i have to modify the Gemfile even more
<manveru> ah, i guess that's not passed through :(
<manveru> well, you can stay with `bundlerEnv` and still use the same way
drakonis has joined #nixos-chat
<manveru> the only benefit of `bundlerApp` is that it won't expose other executables
<manveru> (more or less)
<manveru> so it avoids some collisions that might happen, but mostly useful when used for systemPackages
drakonis_ has quit [Ping timeout: 259 seconds]
<aanderse> you were saying just use "rake" instead of "bundle exec rake", right?
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos-chat
Guanin has joined #nixos-chat
<eyJhb> Anyone a JS, UX expert?
<eyJhb> ... The person who did this `Fegge-/Agger-/Sebbersundvej 9220 Aalborg Øst`, is not very parse Friendly... Any ideas?
drakonis has quit [Ping timeout: 250 seconds]
drakonis has joined #nixos-chat
* joepie91 appeareth
<joepie91> eyJhb: define "parse friendly"
<eyJhb> Extracting the actual street name from that, so I can put it into Google and get latlng
<joepie91> eyJhb: does this only need to work for Sweden(?), or other countries also?
<eyJhb> joepie91: only for this list https://termbin.com/7ci0
<eyJhb> And.. Denmark :p
<eyJhb> This is after a rough... Parsing... Feel like I need to implement a couple of checks, and call it a day
<eyJhb> E.g. tried to just take everything after the last /, buuut.. THen we have `Stjernepladsen 19/27`
<eyJhb> Bastards
<joepie91> eyJhb: right, so, the short answer is "don't try to parse address data, throw the entire thing into Google and let it figure it out"
<joepie91> address data is a nightmare to parse
<joepie91> even in the best of circumstances
<eyJhb> Of course, if I had any alternative, I would not do this
<eyJhb> Not like I think it is fun :p
<joepie91> eyJhb: right, so what is the concrete issue you're running into here?
<joepie91> from a glance at that code, the address format looks fairly consistent
<eyJhb> joepie91: parsing a website, to display students apartments up for rent on Google Maps, with better filtering etc.
<eyJhb> Yeah, but then you get the edge cases
<joepie91> eh, at that list *
<eyJhb> Which I don't manually want to tune each time
<joepie91> the edge cases don't seem awful in that list
<eyJhb> As they could add/remove apartments
<eyJhb> Nope, should be fairly doable
<eyJhb> Just need to implement the checks...
<joepie91> eyJhb: so, rough approach: every slash followed by 2+ alpha characters constitutes an address separator
<joepie91> eyJhb: detect end of the address section by detecting the start of the postal code
emily has joined #nixos-chat
<eyJhb> Postal code+city I am in control of, everything after ;, is postal code + city
<joepie91> dash seems to fairly consistently mean house number range
<joepie91> even for extensions like letters
<joepie91> comma being a list of numbers
<joepie91> that covers about all the edge cases, no?
<joepie91> oh, I missed the +
<joepie91> which is equivalent to comma
<joepie91> and for the ranges of house numbers, you look at whether the second component starts with a number or a letter to determine whether it is a number range or an extension range
<joepie91> eyJhb: anything I missed? :P
<joepie91> actually, lemme roll some code that parses this real quick, for funsies
<MichaelRaskin> Good thing you can ignore appartment numbers sometimes being floor+direction
<eyJhb> joepie91: I ended up with this, seeing as it just needs to be really rough :p - https://termbin.com/wb12
<eyJhb> It is just to give a idea of the location of the apartments
<eyJhb> It can be off by meters, and it will be fine :p
<eyJhb> But kilometers, not fine anymore :p
<joepie91> eyJhb: ah, right, I assumed you also needed to know the affected house numbers :P
<joepie91> since right now you're treating what is probably a block of houses as a single one
* joepie91 resumes his attempt to parse this Properly
<eyJhb> joepie91: yeah I am :p No need for anything more precise :D
<eyJhb> https://i.imgur.com/IO6Tehf.png basically something like this
<eyJhb> Just, to get a quick overview, of where you might want to live, and where e.g. supermarkets are, university, fitness, etc.
<joepie91> eyJhb: what does 'og' mean?
<joepie91> Snebærvej 6-34 og 21-31
<joepie91> without?
<eyJhb> og => and
<eyJhb> :D
<joepie91> oh
<joepie91> ah, even/odd numbering, I get it
<eyJhb> joepie91: which language?
<eyJhb> *code
<joepie91> JS :P
<eyJhb> Ahh.. :p
<eyJhb> This is what I am going to use in the JS frontend -> https://termbin.com/yelo not pretty, but, not that big
<joepie91> Fegge-/Agger-/Sebbersundvej;9220 Aalborg Øst
<joepie91> I assume this works line in Dutch?
<joepie91> where the - is replaced with the last part of the last item?
<eyJhb> What?
<eyJhb> Yeah
<joepie91> Foo-/Bar-/Bazthing = Foothing/Barthing/Bazthing
<eyJhb> Yeah
<joepie91> right
<eyJhb> Buuut, knowing when Sebber and sundvej starts/ends
<joepie91> yeah, not bothering with that lol
<joepie91> wasn't planning on implementing natural language processing
<joepie91> :P
<eyJhb> DO SOME AI!
<eyJhb> :p
<eyJhb> Now... *sigh* for the frontend
<joepie91> eyJhb: nearly there: https://i.imgur.com/0ERjlDN.png :P
<joepie91> eyJhb: "50E, 2. sal" -- means "50E, second room", right?
<joepie91> eyJhb: I think I've covered every single edge case now: https://gist.github.com/joepie91/6adc1ee7faed4be9d1da0ca8fcc79d9f?ts=4 -- code at the bottomn
<joepie91> bottom*
drakonis has quit [Ping timeout: 252 seconds]
drakonis has joined #nixos-chat
Guanin has quit [Remote host closed the connection]
drakonis1 has joined #nixos-chat
drakonis_ has joined #nixos-chat
drakonis2 has joined #nixos-chat
drakonis has quit [Ping timeout: 252 seconds]
drakonis_ has quit [Ping timeout: 257 seconds]
<ldlework> I've been writing a web app in ASP.NET Core and it's been pretty nice.
day|flip has joined #nixos-chat
jasongrossman has joined #nixos-chat
day|flip has quit [Remote host closed the connection]