| 24 May 2021 |
|  berberman joined the room. | 12:02:42 |
|  nf joined the room. | 12:45:47 |
 srid | Is anyone familiar with these linker errors on macOS?
impure path `/private/tmp/nix-build-ema-0.2.0.0.drv-0/ghc81651_0/ghc_32.filelist' used in link
...
`ld' failed in phase `Merge objects'. (Exit code: 1)
https://github.com/srid/emanote/issues/15#issuecomment-846738416
| 16:10:04 |
|  benley joined the room. | 16:41:40 |
| srid | User just confirmed that re-installing Nix fixed it. ^^ | 17:25:43 |
| srid | MACOS! | 17:25:57 |
 pnotequalnp | Can I easily generate a cabal freeze file based on my flake.lock? So that cabal v2 commands will use the same dependency versions as nix? | 17:32:11 |
|  Edward Tjörnhammar joined the room. | 18:55:28 |
|  carter joined the room. | 19:02:04 |
 maralorn | In reply to @pnotequalnp:matrix.org
Can I easily generate a cabal freeze file based on my flake.lock? So that cabal v2 commands will use the same dependency versions as nix? Easily? Don‘t think so. In a normal setup your flake.lock contains no information about individual Haskell dependencies only the pinned nixpkgs. | 19:14:35 |
| maralorn | Of course, the information does exist. And I think you might have luck to insert a sneaky "cabal freeze" command in the build phase of the package build so that you get an additional freeze file as an output of your derivation. | 19:16:51 |
| pnotequalnp | Ooh yes that sounds like a good idea. | 19:20:02 |
|  Dusk joined the room. | 19:42:32 |
|  jul1u5 removed their profile picture. | 20:05:12 |
|  infinisil joined the room. | 20:30:14 |
|  immae (he/him) joined the room. | 22:17:08 |
| 25 May 2021 |
|  pasukon joined the room. | 02:53:59 |
|  pacc6 joined the room. | 07:41:39 |
| maralorn | I don‘t get it. github has bumped this PR now twice into my notifications, since I merged it. But everytime I go there I don‘t see that anything has happened. https://github.com/NixOS/cabal2nix/pull/495 | 10:38:55 |
 sterni (he/him) | cdepillabout: removed the darwin builds in staticHaskellPackages for now the libiconv failure really doesn't need to concern us | 10:42:18 |
| sterni (he/him) | well maybe if someone has some extra time and energy | 10:42:31 |
| maralorn | I have just disabled the hls-class-plugin check on darwin. I didn’t dare to open an upstream bug, because there are obvious steps to debug this first which I cannot do, because I don‘t have a darwin machine. | 11:03:21 |
 cdepillabout | sterni (he/him): maralorn Thanks! | 11:21:19 |
| cdepillabout | In reply to @maralorn:maralorn.de
I don‘t get it. github has bumped this PR now twice into my notifications, since I merged it. But everytime I go there I don‘t see that anything has happened. https://github.com/NixOS/cabal2nix/pull/495 Yeah, I got that as well. I wonder why it is happening. | 11:22:35 |
| maralorn | I am by the way not 100% satisfied with the result of that PR but I decided that I will not under any circumstances piss off someone who contributes docs.^^ | 11:30:38 |
| cdepillabout | Yeah, that is completely understandable. | 11:31:35 |
| maralorn | I really think that there could be gained a lot by collecting all the various blog posts people created about setting up their Haskell projects with nix and distilling the best practices in our documentation. | 11:32:54 |
| maralorn | It’s so crazy how often I come by some random config by someone and am like "wow, that’s a nice twist to improve your config that I hadn’t seen before". | 11:33:46 |
| cdepillabout | I agree. The Haskell infrastructure is quite difficult to document though, given how many different ways you can use it. | 11:34:48 |
| cdepillabout | Sometimes I think about trying to write the "canonical" documentation, but I always give up after considering the scope. | 11:35:17 |
