| 30 Nov 2024 |
 K900 | The package probably just needs to be fixed | 19:14:46 |
 kuruczgy | In reply to @kuruczgy:matrix.org
Reverting https://github.com/NixOS/nixpkgs/pull/220192 seems to solve this particular issue. So far the only non-straightforward fix is this one, because apparently the original PR was already trying to fix some cross-compilation, but I don't quite understand the argument given there. | 19:16:54 |
 Artturin | Redacted or Malformed Event | 21:51:23 |
| Artturin | In reply to @kuruczgy:matrix.org
Another one:
(import nixpkgs {
localSystem.system = "aarch64-linux";
crossSystem.system = "x86_64-linux";
}).pkgsHostTarget.lsb-release
for some reason has the wrong interpreter shebang.
(But please someone tell me if I am crazy, my cross settings are correct for aarch64 -> x86_64, right? I guess this is just very rarely tested.)
Since https://github.com/NixOS/nixpkgs/pull/300322 was closed
I think we should switch lsb-release to https://github.com/NixOS/nixpkgs/pull/301350 | 21:54:44 |
| kuruczgy | In reply to @artturin:matrix.org
Since https://github.com/NixOS/nixpkgs/pull/300322 was closed
I think we should switch lsb-release to https://github.com/NixOS/nixpkgs/pull/301350 Yep, makes sense, I can make a PR. Replace substituteAll with replaceVars, and use runtimeShell, that's it I guess? | 22:01:26 |
| kuruczgy | Actually... Where do I put the meta attr? replaceVars doesn't seem to accept it | 22:03:51 |
| Artturin | Use runCommand to copy the output of replaceVars to bin | 22:05:55 |
| Artturin | * Use runCommand to copy the output of replaceVars to bin | 22:05:59 |
| Artturin | * Use runCommandLocal to copy the output of replaceVars to bin | 22:06:20 |
| Artturin | * Use runCommand to copy the output of replaceVars to bin | 22:07:18 |
| kuruczgy | Hm something like this?
runCommand "lsb_release"
{
meta = with lib; {
description = "Prints certain LSB (Linux Standard Base) and Distribution information";
mainProgram = "lsb_release";
license = [ licenses.mit ];
maintainers = with maintainers; [ primeos ];
platforms = platforms.linux;
};
} ''
install -Dm 755 ${replaceVars ./lsb_release.sh {
inherit runtimeShell coreutils getopt;
}} $out/bin/lsb_release
''
| 22:13:12 |
| Artturin | Yep | 22:15:52 |
| kuruczgy | Here: https://github.com/NixOS/nixpkgs/pull/360576
Lemme also test if cross actually works now with it... | 22:18:35 |
| kuruczgy | Yep, it does | 22:23:46 |
| Artturin | In reply to @kuruczgy:matrix.org
Okay, thanks for looking into this. Not sure when I will have time to make a PR, but I guess the fix is simple enough, so no reason not to. https://github.com/NixOS/nixpkgs/pull/360602 | 23:40:44 |
| kuruczgy | In reply to @artturin:matrix.org
https://github.com/NixOS/nixpkgs/pull/360602 Oh cool, so this is the proper fix. Nice. | 23:44:00 |
| kuruczgy | Though I am not sure about eval performance, do you have tooling to measure that there is no significant regression? | 23:45:16 |
| Artturin | Nope and when I tried ofborg | 23:48:41 |
| Artturin | 
Download image.png | 23:48:43 |
| Artturin | * Nope and when I tried ofborg previously | 23:48:50 |
| Artturin | * Nope and when I tried ofborg previously for the splicing path | 23:49:17 |
| 1 Dec 2024 |
|  @turlando:matrix.org set a profile picture. | 12:38:34 |
|  Simon.Sieben@mailbox.tu-dresden.de joined the room. | 13:00:05 |
| @turlando:matrix.org left the room. | 14:03:52 |
|  @tanvir:hackliberty.org removed their profile picture. | 17:39:52 |
| @tanvir:hackliberty.org removed their display name 𝒕𝒂𝒏𝒗𝒊𝒓. | 17:42:27 |
| @tanvir:hackliberty.org left the room. | 17:48:02 |
| 2 Dec 2024 |
 kjeremy | Artturin: it's crazy to me how easily it is to get the evaluator to OOM with small changes on cross. I would expect memory usage to be linear wrt non-cross but it's definitely not | 15:26:27 |
 Mindavi | I think it instantiates nixpkgs x amount of times | 15:28:34 |
 hexa | and don't try python things 🙈 | 16:12:19 |
