| 8 Mar 2025 |
 WeetHet | In reply to @emilazy:matrix.org
fwiw Nix 2.3 is also buggy. it just has different bugs. It's simpler | 18:20:43 |
 emily | and I suspect Nixpkgs will not support being evaluated with 2.3 for much longer | 18:20:45 |
| WeetHet | In reply to @emilazy:matrix.org
and I suspect Nixpkgs will not support being evaluated with 2.3 for much longer Small features can be easily added | 18:21:41 |
| WeetHet | And it's so much easier to fix bugs when touching every code path doesn't break some flake-related stuff | 18:22:20 |
| WeetHet | The other option would be to take Lix's code and tear everything flake-related out of it | 18:27:01 |
 raitobezarius | In reply to @weethet:catgirl.cloud
The other option would be to take Lix's code and tear everything flake-related out of it we push it to its own libflakes and then someday we will see how to split it up (while providing it as part of the some normal distribution of Lix obviously) | 18:28:06 |
| WeetHet | Nix 2.3 has 40k sloc, while 2.26 has 89k | 18:44:02 |
| WeetHet | And nix 2.26 has 7 times as many files as nix 2.3 | 18:45:03 |
| raitobezarius | and alas 2.3 is pretty much dead | 18:45:24 |
| raitobezarius | 
Download clipboard.png | 18:45:25 |
| WeetHet | I would expect it to be dead, doesn't mean it's unusable | 18:46:00 |
| WeetHet | And doesn't mean you can't use it as a starting point | 18:46:11 |
|  Qyriad changed their display name from qyriad to Qyriad. | 21:41:03 |
 Sergei Zimmerman (xokdvium) | https://github.com/NixOS/nix/pull/12615 | Enabling UBSAN in CI checks
I think that's something that lots of people really wanted to see and meson migration / per-component packaging overrides now make it possible. | 22:10:49 |
| 9 Mar 2025 |
 fzakaria | does ca-derivations use source: in the fingerprint or the fixed:output ? | 06:02:20 |
| fzakaria | i'm having trouble doing the hash calculation by hand | 06:08:21 |
 puck | it's the normal format for CA store paths | 08:55:04 |
| puck | tho you'll have to make sure you take the references and the -{output name} quirks in account, since FODs don't have those | 08:56:27 |
| puck | yeah, just checked, it's CA, hashed over the nar and sha256 | 09:01:49 |
|  puffnfresh joined the room. | 09:21:37 |
|  @hurdpublic:pub.solar joined the room. | 11:44:20 |
| @hurdpublic:pub.solar changed their profile picture. | 14:05:39 |
| @hurdpublic:pub.solar changed their profile picture. | 14:11:16 |
| @hurdpublic:pub.solar changed their profile picture. | 14:28:51 |
 connor (burnt/out) (UTC-8) | tomberek: I was working on a PR to use immutable lists (https://github.com/NixOS/nix/pull/11767) and was working on shrinking the size of the Value structure (https://github.com/ConnorBaker/nix/commits/feat/tagged-values/). Both of them are what made me want to give my talk.
I don’t have a PR for the primitives — I think the first step would be to survey the ways we most commonly interact (or want to interact) with the data structures. Then, choosing primitives that enable those interactions and compose well. Then, choosing implementations for the data structures that make those primitives fast. | 19:58:46 |
| 10 Mar 2025 |
|  deprecated, moved to @poz:poz.pet changed their display name from jacek to deprecated, moved to @poz:poz.pet. | 00:30:33 |
|  sinan changed their profile picture. | 02:45:54 |
|  @marie:marie.cologne joined the room. | 11:35:30 |
| 11 Mar 2025 |
| fzakaria | Redacted or Malformed Event | 01:53:56 |
| fzakaria | Redacted or Malformed Event | 01:56:21 |
