Logs: freenode/#xmonad
| 2021-04-08 07:25:28 | <Solid> | so I just build everything once |
| 2021-04-08 07:25:46 | × | idhugo quits (~idhugo@87-49-147-45-mobile.dk.customer.tdc.net) (Ping timeout: 240 seconds) |
| 2021-04-08 07:27:42 | <mc47[m]> | Is stack smart enough so when I put the same resolver, it only builds things once? |
| 2021-04-08 07:29:02 | <Solid> | yes |
| 2021-04-08 07:30:13 | <mc47[m]> | That's cool |
| 2021-04-08 07:32:26 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 07:38:03 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 265 seconds) |
| 2021-04-08 07:47:55 | → | idhugo__ joins (~idhugo@130.225.16.16) |
| 2021-04-08 07:50:11 | × | idhugo_ quits (~idhugo@80-62-116-143-mobile.dk.customer.tdc.net) (Ping timeout: 248 seconds) |
| 2021-04-08 07:51:33 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 07:55:15 | × | materiyolo quits (~materiyol@112.204.174.249) (Quit: WeeChat 3.0.1) |
| 2021-04-08 07:56:41 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 240 seconds) |
| 2021-04-08 08:10:47 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 08:17:02 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 246 seconds) |
| 2021-04-08 08:17:24 | <Liskni_si> | I wish Stack was also "smart enough" to support using precompiled binaries somehow |
| 2021-04-08 08:18:10 | <Liskni_si> | what's the point of everyone rebuilding all the LTS packages from source, possibly again and again every week? |
| 2021-04-08 08:21:24 | <Solid> | you mean of the whole executable? |
| 2021-04-08 08:21:48 | <Liskni_si> | No, the deps. |
| 2021-04-08 08:22:44 | <Solid> | like stack pulls their own precompiled binaries from stackage instead of pulling the sources? |
| 2021-04-08 08:22:47 | <Liskni_si> | I just want to do "apt install libghc-pandoc-dev-lts-16" and never have to rebuild it every again. |
| 2021-04-08 08:22:58 | <Solid> | ah I see |
| 2021-04-08 08:24:07 | <Liskni_si> | I can do "apt install libghc-pandoc-dev" now, but then I'm limited to whatever version the debian maintainers happen to have packaged, and I can't use cabal/stack to build the rest |
| 2021-04-08 08:24:37 | <Liskni_si> | I wish there was an apt repo with (meta)packages of multiple LTS versions, and for stack/cabal or whatever tool to support building on top of that |
| 2021-04-08 08:25:32 | <Liskni_si> | shouldn't be all that hard to do the apt repo part actually, I'm a bit surprised it hasn't been done yet |
| 2021-04-08 08:25:51 | <Liskni_si> | maybe I should propose that on the haskell discourse or something |
| 2021-04-08 08:27:04 | <Solid> | oh that reminds me to set up proper mail integration for that discourse |
| 2021-04-08 08:30:47 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 08:34:28 | → | thc202 joins (~thc202@unaffiliated/thc202) |
| 2021-04-08 08:35:43 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 260 seconds) |
| 2021-04-08 08:48:43 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 08:49:17 | × | Buliarous quits (~gypsydang@unaffiliated/buliarous) (Read error: Connection reset by peer) |
| 2021-04-08 08:53:11 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 240 seconds) |
| 2021-04-08 09:06:18 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 09:11:11 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 240 seconds) |
| 2021-04-08 09:11:55 | <Solid> | Is there a module to pretty-print keymasks and -codes? |
| 2021-04-08 09:13:18 | <Solid> | Like a pseudo-inverse to EZConfig: `ppKey :: (KeyMask, Keycode) -> String`, `ppKey (0, 97) == "a"` or something |
| 2021-04-08 09:22:51 | × | ericsagn1 quits (~ericsagne@2405:6580:0:5100:d2e3:b034:3303:1a96) (Ping timeout: 260 seconds) |
| 2021-04-08 09:24:26 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 09:24:56 | → | Buliarous joins (~gypsydang@unaffiliated/buliarous) |
| 2021-04-08 09:29:55 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 248 seconds) |
| 2021-04-08 09:34:40 | MrElendig | is now known as evilop |
| 2021-04-08 09:35:08 | → | ericsagn1 joins (~ericsagne@2405:6580:0:5100:7cd7:690c:dd62:864c) |
| 2021-04-08 09:40:04 | × | idhugo__ quits (~idhugo@130.225.16.16) (Read error: Connection reset by peer) |
| 2021-04-08 09:41:25 | → | idhugo__ joins (~idhugo@80-62-116-143-mobile.dk.customer.tdc.net) |
| 2021-04-08 09:43:42 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 09:48:41 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 240 seconds) |
| 2021-04-08 10:02:48 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 10:04:13 | → | dariof4 joins (~dario@217.171.66.53) |
| 2021-04-08 10:09:41 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 240 seconds) |
| 2021-04-08 10:18:01 | → | materiyolo joins (~materiyol@112.204.174.249) |
| 2021-04-08 10:23:40 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 10:26:04 | → | cyr4x3 joins (~cyr4x3@139.47.125.51) |
| 2021-04-08 10:27:58 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 240 seconds) |
| 2021-04-08 10:32:14 | <psibi[m]> | pandoc seems to be shipping binaries, probably that can be used: https://github.com/jgm/pandoc/releases/tag/2.13 (Sorry if that was already discussed!) |
| 2021-04-08 10:33:08 | × | cyr4x3 quits (~cyr4x3@139.47.125.51) (Quit: WeeChat 1.6) |
| 2021-04-08 10:34:37 | <psibi[m]> | I guess if it's for compiling hakyll, there is no easy way. My personal website is built via hakyll and I really want to move away from it because of the huge compile time. |
| 2021-04-08 10:35:28 | × | xaltsc_ quits (~xaltsc@unaffiliated/xaltsc) (Quit: WeeChat 3.1) |
| 2021-04-08 10:37:32 | → | cyr4x3 joins (~cyr4x3@139.47.125.51) |
| 2021-04-08 10:41:16 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 10:43:07 | × | cyr4x3 quits (~cyr4x3@139.47.125.51) (Quit: WeeChat 1.6) |
| 2021-04-08 10:43:24 | → | cyr4x3 joins (~cyr4x3@139.47.125.51) |
| 2021-04-08 10:43:42 | × | cyr4x3 quits (~cyr4x3@139.47.125.51) (Client Quit) |
| 2021-04-08 10:44:18 | → | cyr4x3 joins (~cyr4x3@139.47.125.51) |
| 2021-04-08 10:44:18 | × | cyr4x3 quits (~cyr4x3@139.47.125.51) (Client Quit) |
| 2021-04-08 10:46:36 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 260 seconds) |
| 2021-04-08 10:46:43 | → | cyr4x3 joins (~cyr4x3@139.47.125.51) |
| 2021-04-08 10:50:29 | <Liskni_si> | psibi[m]: pandoc's binaries don't help us, hakyll compiles against pandoc as a library |
| 2021-04-08 10:50:36 | <Liskni_si> | (just as GenerateManpage does/did) |
| 2021-04-08 10:51:57 | <psibi[m]> | Ah, right. That's unfortunate. |
| 2021-04-08 10:52:33 | <Liskni_si> | (dependencies are huge pain in the ass in most language ecosystems, but in Haskell they happen to be even worse than in most other) |
| 2021-04-08 10:53:12 | × | cyr4x3 quits (~cyr4x3@139.47.125.51) (Quit: WeeChat 1.6) |
| 2021-04-08 10:54:20 | <Liskni_si> | which reminds me that these discussions probably make it easier for me to decide that I want genmanpage variant A :-) |
| 2021-04-08 10:55:46 | <Liskni_si> | I was a bit torn before, as peti's arguments made sense and both variants' implementations were ugly in one way or another, but this half an hour of compiling issue is just fucked up (and yeah, I feel like it might drive me off Haskell sooner or later) |
| 2021-04-08 10:58:02 | <Liskni_si> | oh and xscreensaver 6.0 was released, and it should no longer be crashable by faulty keyboards, so Solid can continue recommending it in the tutorial and geekosaur can restore his faith in jwz :-) |
| 2021-04-08 10:59:15 | <Solid> | \o/ |
| 2021-04-08 10:59:40 | → | ml| joins (~ml|@unaffiliated/ml/x-3958674) |
| 2021-04-08 10:59:48 | → | CosmosAtlas joins (~CosmosAtl@180.110.107.139) |
| 2021-04-08 10:59:53 | <Solid> | > (and yeah, I feel like it might drive me off Haskell sooner or later) |
| 2021-04-08 10:59:54 | <Solid> | D: |
| 2021-04-08 10:59:54 | <lambdabot> | error: Variable not in scope: yeah :: t9 Boolerror: |
| 2021-04-08 10:59:54 | <lambdabot> | Data constructor not in scope: |
| 2021-04-08 10:59:55 | <lambdabot> | I :: t0 |
| 2021-04-08 11:00:01 | <Solid> | lol |
| 2021-04-08 11:00:07 | <Solid> | I'm sorry lambdabot |
| 2021-04-08 11:01:10 | <Liskni_si> | maybe it's not really "drive away from Haskell" but rather "shift weights in the decision process whether use of Haskell is a good trade-off" |
| 2021-04-08 11:02:42 | <Liskni_si> | I woke up today with an irresistible urge to rewrite some of my little web hacks (rss generators and some api translators) from Haskell to Python because I can no longer be bothered to fight the build system |
| 2021-04-08 11:03:16 | <Liskni_si> | I needed a short break from xmonad anyway :-) |
| 2021-04-08 11:04:11 | <CosmosAtlas> | I might miss a bit of the earlier discussion. But how did you setup your Haskell environment? |
| 2021-04-08 11:04:53 | <CosmosAtlas> | I once had many problems with that too. Then I switched to the "suggested" way of setting up Haskell/Stack/Cabal such without using system package manager. Never had a problem since. |
| 2021-04-08 11:05:24 | <Liskni_si> | CosmosAtlas: the missing discussion is logged here: https://ircbrowse.tomsmeding.com/browse/xmonad |
| 2021-04-08 11:05:32 | <CosmosAtlas> | IIRC, the huge big problem at that time (a few years ago) was everything either changed from dynamically linked to statically linked (or the other way around) on ArchLinux |
| 2021-04-08 11:05:52 | <Solid> | well, haskell to python surely is a switch of paradigms =) |
| 2021-04-08 11:06:08 | × | ml| quits (~ml|@unaffiliated/ml/x-3958674) (Ping timeout: 268 seconds) |
| 2021-04-08 11:06:14 | <Liskni_si> | CosmosAtlas: tldr is that we were talking about using hakyll for the website which meant that rebuilding the website in github actions takes almost an hour |
| 2021-04-08 11:06:34 | <Liskni_si> | compared to jekyll which is integrated in github pages and rebuilds the site in seconds |
| 2021-04-08 11:07:06 | <CosmosAtlas> | Ahh. I guess my points are irrelevant then. Thanks for the tip. |
| 2021-04-08 11:07:17 | <Liskni_si> | but yeah, the static/dynamic is a factor in my decision: I recently configured stack to build dynamically linked stuff, but this means that I can no longer scp binaries to my server and run them there |
| 2021-04-08 11:07:59 | <Liskni_si> | and with just 2G RAM on it, GHC takes _ages_ there |
| 2021-04-08 11:08:14 | <Liskni_si> | sure, I can tell stack to build a static binary |
| 2021-04-08 11:08:26 | <Solid> | but it's a server so build times don't matter :> |
All times are in UTC.