| 22 Nov 2025 |
 fzakaria | Doesn't MacOS have DTrace ? | 17:52:06 |
| fzakaria | https://8thlight.com/insights/dtrace-even-better-than-strace-for-os-x
I think the DTrace engineers even brought it to Linux also now too | 17:52:41 |
| fzakaria | so there's eBPF and DTrace | 17:52:45 |
 John Ericson | And right you were talking about that, thanks @joerg:thalheim.io: | 17:53:02 |
 Mic92 | They do have it, but it's not trivial to use and can freeze up your computer: https://lobste.rs/s/q50ecd/strace_macos_clone_strace_command_for | 19:02:13 |
| 23 Nov 2025 |
|  eveeifyeve joined the room. | 21:45:30 |
| eveeifyeve | Robert Hensing (roberth): is the sparseNodes already implemented? | 21:51:52 |
 Robert Hensing (roberth) | eveeifyeve: it's not been done yet | 21:53:45 |
| Robert Hensing (roberth) | it may make sense to play around with the idea in flake-compat first | 21:54:38 |
| Robert Hensing (roberth) | it's a more light weight environment for that kind of thing | 21:54:51 |
| Robert Hensing (roberth) | you'd have to generate the sparseNodes structure by hand, but I think it's a good way to explore what actually needs to be in there | 21:55:40 |
| Robert Hensing (roberth) | iirc flake-compat is written with one global context of one lock file in mind. Probably makes sense to make those functions a bit more independent first so that it's easier to turn it into something that works for multiple flake locks | 21:57:40 |
| Robert Hensing (roberth) | so I'm not exactly sure what needs to be in sparseNodes, but my idea is to treat nodes as a dumb storage for pins, ignoring the dependency injection related fields that are in nodes. I don't know if we even want those in the lock file at all because they're basically duplicated info from the easily readable flake.nix inputs. | 22:01:14 |
| Robert Hensing (roberth) | so I think the minimum we need is a mapping from the inputs structure to the nodes names that contain the corresponding pins | 22:02:21 |
| Robert Hensing (roberth) | and then in the final phase we put the pins in sparseNodes directly instead of a string that references the node | 22:02:57 |
| Robert Hensing (roberth) | btw this is about https://github.com/NixOS/nix/issues/7730 | 22:03:23 |
| Robert Hensing (roberth) | Reuse input lock files | 22:03:29 |
| 24 Nov 2025 |
|  amadaluzia changed their display name from amadaluzia (🇹🇷 til 25th) to amadaluzia. | 12:57:18 |
|  André Lima joined the room. | 21:01:04 |
| André Lima set a profile picture. | 21:10:25 |
| 25 Nov 2025 |
 Taeer Bar-Yam | RE https://github.com/NixOS/nix/issues/14642 regression
Here's my fix for master https://github.com/NixOS/nix/pull/14644, with tests as well.
CC: Sergei Zimmerman (xokdvium)
| 16:54:37 |
| Taeer Bar-Yam | i have to go for a couple of hours, but I'll check back in after to fix anything that needs fixing | 16:55:44 |
 Sergei Zimmerman (xokdvium) | Taeer Bar-Yam: is it fine if I push over those changes to squash/add tests and do a 2.32 fix? | 20:18:00 |
| Sergei Zimmerman (xokdvium) | Don't want to step on your toes | 20:18:11 |
| Taeer Bar-Yam | yeah, go for it | 20:40:15 |
| Taeer Bar-Yam | i'm so tired don't know how much more I can focus on this today in any case @_@ | 20:41:09 |
| Taeer Bar-Yam | Sergei Zimmerman (xokdvium) your changes look good to me. should I merge it? | 22:34:29 |
| Sergei Zimmerman (xokdvium) | If you think that we’ve sufficiently covered all regression then yeah. There’s also a 2.32 backport with revert | 22:36:30 |
| Taeer Bar-Yam | I imagine there's more tests we could cook up, but we can always add those later | 22:37:43 |
| Sergei Zimmerman (xokdvium) | Mhm probably. piegames did mention that they will have a bunch of tests we should yoink | 22:38:38 |
