Logs: liberachat/#haskell
| 2021-07-08 23:48:01 | <orzo> | i'm using an old laptop as my main computer |
| 2021-07-08 23:48:09 | <orzo> | that i do all my dev on |
| 2021-07-08 23:48:31 | <orzo> | there's no reason why projects can't build on this machine |
| 2021-07-08 23:48:49 | <orzo> | i'm particularly annoyed with the latest android tools |
| 2021-07-08 23:48:52 | <orzo> | not haskell related |
| 2021-07-08 23:49:19 | <oak-> | I have also done Android development in the past in various companies |
| 2021-07-08 23:49:29 | <orzo> | i wish i could write android in haskell and use only ghc to make an apk file |
| 2021-07-08 23:49:39 | <oak-> | Like at the operating system level, built the whole Android image from the source |
| 2021-07-08 23:49:49 | <orzo> | oh right |
| 2021-07-08 23:49:53 | <orzo> | that's what i wish i could avoid |
| 2021-07-08 23:50:12 | <orzo> | but just using the sdk/studio seems to be almost as much resources as the whole os |
| 2021-07-08 23:50:15 | <oak-> | Even with the latest hardware the clean build seems to take ~1 hour |
| 2021-07-08 23:50:44 | <oak-> | Last Android I was building was the 8.1 or something, I haven't tried to build the latest 11 version |
| 2021-07-08 23:51:47 | <oak-> | There's just massive amounts of Java code on the Android operating system, especially in the frameworks/base |
| 2021-07-08 23:51:59 | <oak-> | Yeah, isn't the Android Studio based on IntelliJ IDEA or something |
| 2021-07-08 23:52:19 | <orzo> | i'd like a way to just build apk files from java source and resources using gnu make or something |
| 2021-07-08 23:52:22 | <koz> | oak-: Yep, I believe so. |
| 2021-07-08 23:52:38 | <orzo> | we shouldn't need a huge sdk |
| 2021-07-08 23:52:46 | <orzo> | android is just a linux distro ultimately |
| 2021-07-08 23:54:33 | <oak-> | Last time I did Android App development I just used NeoVim from the shell and built with Gradle from the shell |
| 2021-07-08 23:54:39 | <oak-> | It still uses quite much resources though |
| 2021-07-08 23:55:47 | <orzo> | is there some reason why community people cant maintain an alternate build system for android? |
| 2021-07-08 23:56:10 | <orzo> | it seems simple in principle, just pack .dex files and .res files into an .apk zip |
| 2021-07-08 23:56:37 | <orzo> | gradle would be an improvement over the studio maybe, but i'm currently annoyed with it too |
| 2021-07-08 23:57:18 | <orzo> | i think letting google maintain all the dev tools for android is not working out |
| 2021-07-08 23:57:21 | <oak-> | Doesn't Studio use Gradle in the background? |
| 2021-07-08 23:57:24 | <orzo> | yes |
| 2021-07-08 23:57:31 | <orzo> | but i don't know how to use gradle |
| 2021-07-08 23:57:41 | <orzo> | i'm coming back to android after a long hiatus |
| 2021-07-08 23:57:41 | → | cjb joins (~cjb@user/cjb) |
| 2021-07-08 23:57:48 | <orzo> | back in my day it was eclipse and ant |
| 2021-07-08 23:58:41 | <oak-> | Gradle seems fine as long at it works :D |
| 2021-07-08 23:58:46 | <orzo> | i have good makefile skills, so its a little annoying to have to work with these super ambitious build systems that ultimately just disable incremental builds |
| 2021-07-08 23:59:47 | <oak-> | I tried Eta (compiler which compiles Haskell to Java bytecode) a year ago, and it was pain to work around all the tooling |
| 2021-07-09 00:00:07 | <orzo> | i trie dto build that recently with a ghc 7.10.3 binary i downloaded |
| 2021-07-09 00:00:14 | <orzo> | it fails with a lot of link issues |
| 2021-07-09 00:01:13 | <orzo> | one thing i really like about ghc is that you don't even need cabal if you are compiling a single file haskell program |
| 2021-07-09 00:01:23 | <orzo> | you just pass it to the compiler and it makes a binary |
| 2021-07-09 00:01:24 | <orzo> | very simple |
| 2021-07-09 00:01:32 | <orzo> | i want that for android |
| 2021-07-09 00:01:44 | <orzo> | hello world android shouldn't even need a build system |
| 2021-07-09 00:01:48 | <oak-> | I've been looking forward to try to use Reflex to build native Android apps, that should use regular GHC to build native code |
| 2021-07-09 00:02:01 | <orzo> | i'm not familiar with Reflex |
| 2021-07-09 00:02:03 | → | wroathe joins (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) |
| 2021-07-09 00:02:35 | <oak-> | That is a set of libraries for Haskell to implement the functional reactive programming paradigm |
| 2021-07-09 00:02:39 | <orzo> | even deb files can be packed manulally from components |
| 2021-07-09 00:03:04 | <orzo> | is that how you pack your debs on your haskell-mac-to-ubuntu system? |
| 2021-07-09 00:03:31 | <oak-> | I think it needs to be packaged because of the Java's class system |
| 2021-07-09 00:04:30 | <oak-> | I'm just building the binary with stack, and then generating the .deb package with fpm |
| 2021-07-09 00:05:20 | <orzo> | i wonder if fpm can be made to support android |
| 2021-07-09 00:05:43 | <oak-> | fpm makes it much easier to generate simple .deb packages (and also simple .rpm packages) |
| 2021-07-09 00:05:44 | × | DNH quits (~DNH@2a02:8108:1100:16d8:99cf:ff96:88:4a77) (Ping timeout: 255 seconds) |
| 2021-07-09 00:06:00 | <orzo> | i'm pretty familiar with dpkg-deb and what not |
| 2021-07-09 00:06:12 | <orzo> | i make a lot of casual deb packaging for my own purposes |
| 2021-07-09 00:06:36 | <oak-> | Yeah it's just more complicated, you need to write like 5 files to control what kind of .deb package is created :D |
| 2021-07-09 00:06:37 | <orzo> | i use debian |
| 2021-07-09 00:06:49 | <orzo> | hm |
| 2021-07-09 00:06:49 | × | wroathe quits (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) (Ping timeout: 246 seconds) |
| 2021-07-09 00:06:59 | × | Deide quits (~Deide@user/deide) (Quit: Seeee yaaaa) |
| 2021-07-09 00:07:35 | <orzo> | fpm doesn't seem to be in th edebian archives |
| 2021-07-09 00:08:22 | <orzo> | if it was, i'd install it and maybe try it out |
| 2021-07-09 00:08:32 | <orzo> | but i cant be bothered to clone it and install from source |
| 2021-07-09 00:08:59 | <oak-> | It can be installed from rubygems |
| 2021-07-09 00:11:02 | × | acidjnk_new3 quits (~acidjnk@p200300d0c72b9563e5efeb4dbf0edadc.dip0.t-ipconnect.de) (Read error: Connection reset by peer) |
| 2021-07-09 00:11:27 | <orzo> | i have an old android app that i screwed up the gradle files for |
| 2021-07-09 00:11:33 | <orzo> | i don't know how to get it to build |
| 2021-07-09 00:11:49 | → | wroathe joins (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) |
| 2021-07-09 00:12:00 | <orzo> | i recently got a fresh wizard-created app to build and run from the studio |
| 2021-07-09 00:12:20 | <orzo> | but my old eclipse/ant app is borked |
| 2021-07-09 00:15:59 | <orzo> | every binary in the sdk seems to consist of a bash script with over 80 lines of code to set up the enviro |
| 2021-07-09 00:16:01 | → | chris-the-slurpa joins (~chris@81.96.113.213) |
| 2021-07-09 00:16:46 | × | Lord_of_Life quits (~Lord@user/lord-of-life/x-2819915) (Read error: Connection reset by peer) |
| 2021-07-09 00:16:48 | <oak-> | It shouldn't be that hard but you probably need to match the SDK version you initially used to build the app |
| 2021-07-09 00:19:08 | <oak-> | Also It starts to be a bit offtopic if we start to debug your Android app building in the #haskell channel :P |
| 2021-07-09 00:20:04 | → | Lord_of_Life joins (~Lord@user/lord-of-life/x-2819915) |
| 2021-07-09 00:21:18 | <orzo> | yeah sorry |
| 2021-07-09 00:30:21 | → | P1RATEZ joins (piratez@user/p1ratez) |
| 2021-07-09 00:33:25 | → | dajoer joins (~david@user/gvx) |
| 2021-07-09 00:36:29 | × | Atum_ quits (~IRC@user/atum/x-2392232) (Quit: Atum_) |
| 2021-07-09 00:36:38 | <ephemient> | Bazel and Buck both seem to have teams maintaining their android app building capabilities |
| 2021-07-09 00:37:19 | <ephemient> | probably not gonna be any lighter on orzo's computer than Gradle, though |
| 2021-07-09 00:40:55 | → | hmmmas joins (~fidnc@183.217.200.144) |
| 2021-07-09 00:43:17 | <orzo> | everyone makes huge hungry build enviros these days |
| 2021-07-09 00:44:29 | <orzo> | there should be an effort to make building accessible to cheap low power devices |
| 2021-07-09 00:44:51 | <orzo> | i'm on an i5 anyway with 8gig ram, it should be sufficient |
| 2021-07-09 00:45:18 | × | Vajb quits (~Vajb@hag-jnsbng11-58c3a1-224.dhcp.inet.fi) (Read error: Connection reset by peer) |
| 2021-07-09 00:45:51 | <orzo> | one might suspect they don't want people to build their own binaries |
| 2021-07-09 00:46:03 | → | Vajb joins (~Vajb@hag-jnsbng11-58c3a1-224.dhcp.inet.fi) |
| 2021-07-09 00:53:40 | → | fizbin joins (~fizbin@c-73-33-197-160.hsd1.nj.comcast.net) |
| 2021-07-09 00:53:51 | × | cjb quits (~cjb@user/cjb) (Ping timeout: 246 seconds) |
| 2021-07-09 00:57:39 | × | epolanski quits (uid312403@id-312403.brockwell.irccloud.com) (Quit: Connection closed for inactivity) |
| 2021-07-09 00:58:14 | × | fizbin quits (~fizbin@c-73-33-197-160.hsd1.nj.comcast.net) (Remote host closed the connection) |
| 2021-07-09 00:59:51 | × | Feuermagier quits (~Feuermagi@user/feuermagier) (Remote host closed the connection) |
| 2021-07-09 01:04:51 | <exarkun> | what force is there exerting pressure on the software development process not increase its resource usage, apart from the absolute limit on what resources are actually available? |
| 2021-07-09 01:05:18 | <exarkun> | and lacking such a force, why would resource usage not continuously expand to that limit? |
| 2021-07-09 01:08:28 | <davean> | exarkun: I find making it complex harder, and I don't need to show I'm making a tangled mess to keep my job, so I can do it simply? |
| 2021-07-09 01:09:09 | <orzo> | the force is randos like me longing for simpler tools |
| 2021-07-09 01:10:56 | × | sheepduck quits (~sheepduck@user/sheepduck) (Remote host closed the connection) |
| 2021-07-09 01:12:05 | <exarkun> | davean: Is _making_ it complex harder? Or is it harder to deal with once it's been made complex? |
| 2021-07-09 01:12:23 | <exarkun> | And then once it's hard to deal with, what's the easiest thing to do? Well, add some more complexity of course... |
All times are in UTC.