| 3 Jun 2024 |
 jade_ | (this is a sneak peek, it still needs some operationalizing) | 08:08:18 |
| jade_ | i have done a backfill run. i now need to actually deploy it properly | 09:57:07 |
| jade_ | tomorrow problem | 09:57:10 |
 John Ericson | You all might want https://github.com/NixOS/nix/pull/10838 | 18:29:45 |
| jade_ | https://git.lix.systems/lix-project/lix/issues/373 | 21:45:42 |
| John Ericson | otoh, I gotta get it fixed for macos | 22:01:01 |
| John Ericson | maybe 24.05 will help | 22:01:05 |
| jade_ | O_o how is it broken on libc++/macOS? that's .. an accomplishment, but C++ can certainly manage such things | 22:03:03 |
| John Ericson | std::declval<std::set>() <=> std::declval<std::set>() | 22:03:49 |
|  emily joined the room. | 22:04:07 |
| John Ericson | that is not resolving | 22:04:53 |
| 4 Jun 2024 |
 jakehamilton | Hey there folks, is anyone familiar with import <nix/fetchurl.nix>? I just found out about this and was confused because I thought that <> referred to NIX_PATH values, but this seemed to work. Notably import <nix> does NOT work. As it turns out, this behavior is custom inside of Nix itself which maps certain known files such as: https://github.com/NixOS/nix/blob/master/src/libexpr/fetchurl.nix | 05:06:26 |
| jakehamilton | It seems to be this corepkgsFS: https://github.com/NixOS/nix/blob/bf72b78ef2110f4bda6105b8adff131dc9435bff/src/libexpr/eval.hh#L237 | 05:07:10 |
| jakehamilton | This is definitely weird, but I ended up needing this fetcher because it supports some things that the builtin one doesn't (such as marking files executable, unpacking, and output hash info). | 05:08:13 |
 @irenes:matrix.org | (and I suggested mentioning it here because it's such astonishing behavior that I thought we might like to look into how important it is) | 05:10:32 |
 K900 | corepkgs continues to haunt us | 05:47:00 |
| jade_ | In reply to @jakehamilton:auxolotl.org
Hey there folks, is anyone familiar with import <nix/fetchurl.nix>? I just found out about this and was confused because I thought that <> referred to NIX_PATH values, but this seemed to work. Notably import <nix> does NOT work. As it turns out, this behavior is custom inside of Nix itself which maps certain known files such as: https://github.com/NixOS/nix/blob/master/src/libexpr/fetchurl.nix it's mentioned in puck's talk :) | 06:19:03 |
| jakehamilton | In reply to @jade_:matrix.org
it's mentioned in puck's talk :) Oooh I haven't seen or heard that yet. Would you happen to have a link? | 06:19:47 |
| jade_ | https://media.ccc.de/v/nixcon-2023-36425-reinventing-the-wheel-with-zilch | 06:20:13 |
| jakehamilton | Oh perfect, thank you!! | 06:20:43 |
 aloisw | Right now it seems that the thing that closes issues when they are fixed triggers on forks too (for example, https://git.lix.systems/lix-project/lix/issues/365#event-4081). Is this easy to fix or should I stop pushing to my fork? | 06:25:45 |
| jade_ | what the heck forgejo | 06:26:10 |
| jade_ | 
Download image.png | 06:26:57 |
| jade_ | we don't close from non-default branches, but clearly it also shouldn't close from forks | 06:27:32 |
| aloisw | So I guess I should just rename my branch? | 06:28:55 |
| jade_ | oh, is it on main? | 06:29:05 |
| jade_ | try not having it on main and see what happens I guess? | 06:29:14 |
| aloisw | In reply to @jade_:matrix.org
oh, is it on main? It is, sorry for the complications this may have caused. I will push to a different branch in the future. | 06:31:14 |
| jade_ | no worries, this is arguably a forgejo bug | 06:31:24 |
| jade_ | it's very absurd behaviour imo | 06:31:32 |
