| 17 Jan 2026 |
 eveeifyeve | Even with how windows handles absolute paths. It causes an issue. | 00:37:05 |
 Sergei Zimmerman (xokdvium) | In reply to @eveeifyeve:matrix.org
Even with how windows handles absolute paths. It causes an issue. Not on master, we did merge a fix a while back | 00:37:29 |
| eveeifyeve | Mingw handles absolute paths very differently. See https://matrix.to/#/!GmJrQRmEoTfulqyDge:nixos.org/$wgpJvzKLbFEP9H1BGsAUd2NW1q4hMz4p_VDOmJzZ55E?via=nixos.org&via=matrix.org&via=tchncs.de | 00:38:12 |
| eveeifyeve | It tries to return a base error however error isn't initialized so it panics. | 00:39:30 |
| eveeifyeve | * It tries to return a base error however error isn't initialized so it panics with no message. | 00:39:38 |
| Sergei Zimmerman (xokdvium) | In reply to @eveeifyeve:matrix.org
Mingw handles absolute paths very differently. See https://matrix.to/#/!GmJrQRmEoTfulqyDge:nixos.org/$wgpJvzKLbFEP9H1BGsAUd2NW1q4hMz4p_VDOmJzZ55E?via=nixos.org&via=matrix.org&via=tchncs.de I’m aware, please see https://github.com/NixOS/nix/commit/675656ffbaca8b9c7982a3a5301b517b14f5cfa7 | 00:40:41 |
| eveeifyeve | Do you want me try to update https://github.com/msys2/MINGW-packages/pull/22499 and try running it again? | 00:41:48 |
| Sergei Zimmerman (xokdvium) | In reply to @eveeifyeve:matrix.org
It tries to return a base error however error isn't initialized so it panics with no message. Once again, that is an issue with a global initializers not printing the exception info before main. Not much we can do now | 00:41:49 |
| eveeifyeve | Is this backported? | 00:44:38 |
| eveeifyeve | Sergei Zimmerman (xokdvium): Compiled it, however still failing to launch. | 02:03:53 |
| eveeifyeve | * Sergei Zimmerman (xokdvium): Compiled it, however still failing to launch. I am going to look into a debug build to try and debug it. | 02:04:08 |
|  Theo Paris joined the room. | 02:04:16 |
| eveeifyeve | Should I move this discussion to #windows:nixos.org? | 02:15:33 |
| eveeifyeve | * Should I move this discussion to https://matrix.to/#/#windows:nixos.org? | 02:16:08 |
|  @normalcea:matrix.org removed their profile picture. | 04:09:07 |
| @normalcea:matrix.org removed their display name jasi 🏳️⚧️. | 04:09:15 |
| @normalcea:matrix.org left the room. | 04:09:23 |
| 18 Jan 2026 |
 Robert Hensing (roberth) | I would much appreciate your feedback on https://github.com/NixOS/rfcs/pull/196: Self-describing store
This RFC proposes that stores provide a nix-store.json file declaring their properties, enabling:
- Discovery of store settings (storeDir, encoding, etc.)
- Forward compatibility through explicit feature requirements
- A unified replacement for the ad hoc nix-cache-info format
| 16:55:25 |
| Robert Hensing (roberth) | (best discussed over there) | 16:56:38 |
|  losvedar joined the room. | 17:37:54 |
|  isabel changed their profile picture. | 20:43:31 |
| 19 Jan 2026 |
 Eelco | looks good to me in principle | 10:01:39 |
 raboof | is it expected that // is not lazy? i.e. it evaluates the LHS even if the RHS holds the thing I'm taking from it? | 14:26:57 |
| raboof | * is it expected that // is not lazy? i.e. it evaluates the left parameter even if the right parameter holds the thing I'm taking from it? | 14:27:44 |
 Alyssa Ross | nix-repl> ({ foo = throw "oh no"; } // { foo = 1; }).foo
1
| 14:28:09 |
| Alyssa Ross | looks lazy to me? | 14:28:13 |
| raboof | nix-repl> ((throw "oh no") // { foo = "foo"; }).foo
error: oh no
| 14:31:30 |
 ma27 | attr-names are generally not lazy in Nix, hence the throw is being evaluated. | 14:32:02 |
| raboof | I read // as a function (in infix notation) with two parameters, perhaps that's not how this works? | 14:32:41 |
| raboof | right, so // creates its set of attr-names at first access. that makes some sense I guess. | 14:40:42 |
