| 9 Jul 2025 |
 K900 | But I have a sneaking suspicion that I know how to make it fast | 19:17:26 |
| K900 | OK that did not help much lol | 19:26:21 |
| K900 | But I can get 3 minutes now! | 19:26:26 |
| K900 | On 6 workers | 19:26:43 |
| K900 | And if I get 16 workers, it's closer to 2 | 19:26:47 |
| K900 | I will now try 24 workers and OOM myself | 19:27:04 |
| K900 | The nice thing is that the actual thingy now takes up like 5% CPU | 19:27:38 |
| K900 | Aaaand yep oom | 19:28:04 |
| K900 | I wonder if it'll help if I give it more read buffer | 19:28:52 |
| K900 | Nope | 19:31:06 |
| K900 | OK giving it even more pipe buffer helps | 19:42:13 |
| K900 | OK holy shit | 19:45:31 |
| K900 | I gave it even more buffer | 19:45:36 |
| K900 | Now I need to setcap it | 19:45:38 |
| K900 | But I can do an eval with 16 workers in under 2 minutes | 19:45:44 |
 raitobezarius | Good | 19:49:54 |
 Dreamgineer | Where can I find the schema for `nix build --log-format internal-json`? | 20:47:50 |
| K900 | You don't, it's internal for a reason | 20:48:25 |
| Dreamgineer | Isn't it the same thing nix-output-monitor use | 20:49:16 |
| Dreamgineer | Or maybe I could just parse the bar output | 21:02:51 |
 maralorn | It is. | 21:10:57 |
| maralorn | That’s even worse. | 21:11:05 |
| maralorn | I reconstructed it from reading nix source. | 21:11:18 |
| maralorn | If you don’t want to do that you can also read the nix-output-monitor source of the parser. 😄 | 21:12:03 |
| maralorn | Actually I basically started nix-output-monitor because I didn’t know about the internal-json. It was more meant as a joke to parse the human readable output of nix. 😄 I only later switched to using the internal-json. Which has changed a tiny bit with the years but until now has been quite stable. | 21:14:07 |
| Dreamgineer | Thank you for your response. I just roughly read through the parser and it looks quite understandable. | 21:19:16 |
|  lambadada joined the room. | 23:46:29 |
| 10 Jul 2025 |
|  Mark N3MSC changed their display name from mcornick to Mark N3MSC. | 00:47:44 |
 Laguito | Redacted or Malformed Event | 03:43:55 |
| Laguito | In reply to @charles:computer.surgery
because if both of those point to a copy of the exact same nix expressions then the result will be the exact same because that's kind of the point of nix What kind of magic did you do? | 03:46:30 |
