Newest at the top
| 2025-02-03 22:46:53 +0100 | <dminuoso> | Fair enough. Ill give it a try with haskell.nix another day, but not now - I must sleep. |
| 2025-02-03 22:46:19 +0100 | ash3en | (~Thunderbi@2a03:7846:b6eb:101:93ac:a90a:da67:f207) (Quit: ash3en) |
| 2025-02-03 22:46:01 +0100 | <Athas> | And not "trying". It works! |
| 2025-02-03 22:45:41 +0100 | <Athas> | dminuoso: yes. |
| 2025-02-03 22:45:35 +0100 | <dminuoso> | Athas: Is the thing you are trying to statically link futhark itself? |
| 2025-02-03 22:42:15 +0100 | <Athas> | I sort of understand why Nixpkgs' Haskell setup was built the way it is, but I think it might not be the best way to do it. |
| 2025-02-03 22:41:55 +0100 | gorignak | (~gorignak@user/gorignak) gorignak |
| 2025-02-03 22:41:24 +0100 | gorignak | (~gorignak@user/gorignak) (Quit: quit) |
| 2025-02-03 22:39:31 +0100 | <Athas> | That is indeed a thing I miss in Nixpkgs. |
| 2025-02-03 22:38:15 +0100 | <dminuoso> | So there's no differece between nix and out-of-nix builds in terms of versions. |
| 2025-02-03 22:37:54 +0100 | <dminuoso> | The big thing I like about haskell.nix is that it uses the same build plans that cabal-install does. |
| 2025-02-03 22:37:22 +0100 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 248 seconds) |
| 2025-02-03 22:37:03 +0100 | lxsameer | (~lxsameer@Serene/lxsameer) (Ping timeout: 246 seconds) |
| 2025-02-03 22:36:58 +0100 | <dminuoso> | But it also introduces trust issues, and availability issues like you pointed out. |
| 2025-02-03 22:36:43 +0100 | target_i | (~target_i@user/target-i/x-6023099) (Quit: leaving) |
| 2025-02-03 22:36:34 +0100 | <dminuoso> | Dont get me wrong, the IOG cache is fine *if* you follow their update cadence. That is, you will have the past few months on cache. |
| 2025-02-03 22:35:19 +0100 | <dminuoso> | Its not large, just a matter of whether you dont mind building a bunch of GHCs every now and then. |
| 2025-02-03 22:35:03 +0100 | <Athas> | I do use cachix however, so at long as the needed cache isn't too large, maybe I can cope. |
| 2025-02-03 22:34:35 +0100 | <Athas> | My least favourite thing is having to change something because of what goes on in a computer somewhere else in the world. |
| 2025-02-03 22:34:11 +0100 | <Athas> | Oh, right. Ugh. |
| 2025-02-03 22:33:58 +0100 | <dminuoso> | While IOG provides a cache, they evict it regularly |
| 2025-02-03 22:33:51 +0100 | <dminuoso> | Athas: So the way everything works, they naturally have their own GHC closure, which of course requires some 3 or 4 other GHCs to bootstrap (I think its one additional from what pure GHC needs outside of nix) |
| 2025-02-03 22:33:50 +0100 | tnt2 | tnt1 |
| 2025-02-03 22:33:50 +0100 | tnt1 | (~Thunderbi@user/tnt1) (Ping timeout: 252 seconds) |
| 2025-02-03 22:33:21 +0100 | tnt2 | (~Thunderbi@user/tnt1) tnt1 |
| 2025-02-03 22:32:58 +0100 | <Athas> | "own CI"? What do you mean? |
| 2025-02-03 22:32:51 +0100 | <lxsameer> | Athas: Space is like workspace in matrix |
| 2025-02-03 22:32:36 +0100 | <dminuoso> | Disclaimer: Be ready to have your own CI or follow the update cadence of IOG. |
| 2025-02-03 22:31:56 +0100 | <dminuoso> | (I mean the hackage versions of course) |
| 2025-02-03 22:31:44 +0100 | gorignak | (~gorignak@user/gorignak) gorignak |
| 2025-02-03 22:31:40 +0100 | <Athas> | I can probably live without that if necessary. That's more nice-to-have than need-to-have. |
| 2025-02-03 22:31:17 +0100 | <dminuoso> | Unsure whether there are tools for getting nixpkgs pins, though. |
| 2025-02-03 22:31:13 +0100 | gorignak | (~gorignak@user/gorignak) (Quit: quit) |
| 2025-02-03 22:30:55 +0100 | <Athas> | dminuoso: I will have to try that out one day. |
| 2025-02-03 22:30:39 +0100 | <dminuoso> | And *that* worked for our static builds. |
| 2025-02-03 22:30:31 +0100 | <dminuoso> | Athas: at any rate, for haskell.nix getting a static build amounts to just using `pkgs.pkgsCross.musl64.haskell-nix.project` instead of `pkgs.haskell-nix.project` |
| 2025-02-03 22:30:25 +0100 | <Athas> | What is "Nix space"? |
| 2025-02-03 22:30:00 +0100 | <lxsameer> | I'm pretty much interested in statically compiling my stuff |
| 2025-02-03 22:29:53 +0100 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn |
| 2025-02-03 22:29:50 +0100 | glguy_ | glguy |
| 2025-02-03 22:29:40 +0100 | <lxsameer> | Athas: if you are not already, join the haskell channel on Nix space |
| 2025-02-03 22:29:27 +0100 | glguy_ | (glguy@libera/staff/glguy) glguy |
| 2025-02-03 22:27:59 +0100 | <Athas> | I've been pretty happy about distributing statically linked binaries. |
| 2025-02-03 22:27:24 +0100 | remedan | (~remedan@62.245.108.153) remedan |
| 2025-02-03 22:26:34 +0100 | <dminuoso> | Luckily we gave up on statically linked binaries a long time ago. |
| 2025-02-03 22:26:20 +0100 | <Athas> | I think I am just not good enough at Nix to follow, or passionate about Nix to become good enough to follow. |
| 2025-02-03 22:26:02 +0100 | remedan | (~remedan@62.245.108.153) (Quit: Bye!) |
| 2025-02-03 22:25:59 +0100 | <Athas> | I think I have indirectly benefited from nh2's work, but I must admit I have had trouble following their user-oriented documentation. |
| 2025-02-03 22:25:48 +0100 | <dminuoso> | Just poke Niklas on matrix. |
| 2025-02-03 22:25:19 +0100 | <dminuoso> | I think its mostly done? Unclear. |