25 Apr 2024 |
fricklerhandwerk | In reply to @marcusramberg:matrix.org Should the perl bindings also be changed to use the C bindings? 🤔 It would fit the agenda of getting the C API a baseline adoption in order to iron out the kinks. But there's a fundamental resource constraint at the moment: who's gonna do the work and who's gonna merge the PRs? There's a quick way out of this: saying whoever does the job gets commit access and ownership of that code, all strings attached. I don't think this has consensus though, but I'd like to be surprised. | 07:01:39 |
Marcus | In reply to @fricklerhandwerk:matrix.org It would fit the agenda of getting the C API a baseline adoption in order to iron out the kinks. But there's a fundamental resource constraint at the moment: who's gonna do the work and who's gonna merge the PRs? There's a quick way out of this: saying whoever does the job gets commit access and ownership of that code, all strings attached. I don't think this has consensus though, but I'd like to be surprised. I've pinged stigtsp about this as well, since he's more in touch with his Perl roots than I am these days 🙂 I quickly looked at the existing bindings, seems they use XSLoader towards the c++ code. | 07:10:10 |
jlesquembre | Could an admin add a link to https://github.com/NixOS/nix/labels/c%20api in the channel description? Notice that it's a list with all open issues/PRs with the "c api" label instead of a link to the milestone (https://github.com/NixOS/nix/milestone/52) because we have some PRs that aren't in the milestone | 08:14:33 |
Mic92 | In reply to @marcusramberg:matrix.org Should the perl bindings also be changed to use the C bindings? 🤔 Potentially. I think the perl api was already an ok abstraction on top of Nix. So maybe it could be used as a starting point? | 09:30:02 |
Mic92 | In reply to @marcusramberg:matrix.org Should the perl bindings also be changed to use the C bindings? 🤔 * Potentially. I think the perl api was already an ok abstraction on top of Nix c++ api. So maybe it could be used as a starting point? | 09:30:14 |
Mic92 | The API I was using in harmonia was inspired by the perl api. | 09:30:39 |
| Erin joined the room. | 13:38:30 |
| infinisil changed the room topic to "[Nix C bindings](https://github.com/NixOS/nix/labels/c%20api), language-specific bindings, and their use cases" from "Nix C bindings, language-specific bindings, and their use cases". | 19:50:48 |
| infinisil changed the room topic to "Nix C bindings (https://github.com/NixOS/nix/labels/c%20api), language-specific bindings, and their use cases" from "[Nix C bindings](https://github.com/NixOS/nix/labels/c%20api), language-specific bindings, and their use cases". | 19:51:00 |
infinisil | It would be great to have a https://wiki.nixos.org/ page to keep an ongoing summary of what's happening relating to bindings :D | 19:51:35 |
| David Mell (zraexy) joined the room. | 23:07:56 |
| David Mell (zraexy) changed their display name from David Mell to David Mell (zraexy). | 23:51:56 |
26 Apr 2024 |
Mic92 | Or even just a Nix issue or is this to messy? | 05:55:37 |
| Rosario Pulella joined the room. | 06:02:15 |
fricklerhandwerk | Tracking issue sounds resonable | 08:46:38 |
fricklerhandwerk | * Tracking issue sounds reasonable | 08:46:44 |
| Philip Taron (UTC-8) joined the room. | 15:50:20 |
Robert Hensing (roberth) | We have a start here, but issue may be better. https://github.com/NixOS/nix/pull/8699#issuecomment-2071151896 | 16:12:46 |
| Robert Hensing (roberth) changed the room topic to "Nix [C API](https://nixos.org/manual/nix/unstable/c-api), language-specific bindings, and their use cases" from "Nix C bindings (https://github.com/NixOS/nix/labels/c%20api), language-specific bindings, and their use cases". | 16:13:39 |
| Robert Hensing (roberth) changed the room topic to "Nix C API https://nixos.org/manual/nix/unstable/c-api, language-specific bindings, and their use cases" from "Nix [C API](https://nixos.org/manual/nix/unstable/c-api), language-specific bindings, and their use cases". | 16:14:01 |
Robert Hensing (roberth) | Path values are represented by absolute paths, so I think that's what's expected by that function | 16:58:39 |
Robert Hensing (roberth) | Not a great name, and insufficient docs though | 16:58:51 |
Robert Hensing (roberth) | I think it needs to be improved: https://github.com/NixOS/nix/issues/10613 | 17:02:05 |
27 Apr 2024 |
| pbsds joined the room. | 21:44:41 |
28 Apr 2024 |
| a-kenji joined the room. | 09:59:50 |
29 Apr 2024 |
| NixOS Moderation Botchanged room power levels. | 15:30:00 |
| c4lliope joined the room. | 18:20:13 |
1 May 2024 |
| NixOS Moderation Botchanged room power levels. | 15:06:16 |
2 May 2024 |
| Taeer Bar-Yam joined the room. | 15:10:51 |
Robert Hensing (roberth) | Does anyone know a reason to use void * instead of opaque pointers like Value * (after struct Value; )? I think we want to just change it: https://github.com/NixOS/nix/issues/10561 | 16:47:55 |