| 5 May 2026 |
 Jens Petersen | Anyone good with meson? https://github.com/NixOS/nix/issues/15797 | 06:22:47 |
|  Notarin joined the room. | 14:51:41 |
| Notarin | Hi all. Where is pipe operators on the planner to stabalization. AFAICT last thing that happened to it was it being decided to be put in FCP and since Jan 23rd there's been nothing on it I can find. How long does a FCP typically last? | 14:57:21 |
|  whispers [& it/fae] joined the room. | 15:23:43 |
| 6 May 2026 |
|  Yogansh set a profile picture. | 03:32:56 |
 Robert Hensing (roberth) | Artemis Tosini (NixOS): since GitHub is broken at this time again, I'll bug you with review here :)
PR: https://github.com/NixOS/nix/pull/15793
tests/functional/dyn-drv/non-trivial.nix:
Are we phasing out $out for outputting derivations?
I'm leaning "good riddance" because this method had no purpose beyond a temporary stop gap for singleton derivations or perhaps small static graphs using multi-output derivations.
No need to hold on to that unnecessary code path.
If not, this should probably be a new test case instead of clobbering the old one.
Inside buildCommand:
Test that $out is unset?
src/libstore/unix/build/derivation-builder.cc/submitOutput:
- Refer to https://github.com/NixOS/nix/issues/12727 in the error message when not Opaque?
- Would be good to show the value of the previously submitted output
- Needs tests for both error scenarios
| 16:17:15 |
 Artemis Tosini (NixOS) | Oh, yes, I do definitely need more tests, I'll work on that. | 18:52:16 |
|  @rasmata:matrix.org joined the room. | 20:07:03 |
| @rasmata:matrix.org left the room. | 20:07:11 |
| 7 May 2026 |
|  Agahnim joined the room. | 12:02:35 |
|  e joined the room. | 21:28:02 |
| 8 May 2026 |
|  Aliaksandr joined the room. | 02:21:03 |
 raf | I'd like to add some new stuff to C API be used in my Rust bindings. Should I just send the PR, or is there a roadmap for the C API that I'm supposed to be looking at instead? | 08:22:50 |
| 10 May 2026 |
|  asp345 joined the room. | 11:11:34 |
|  Maxisthespy joined the room. | 19:15:57 |
| 12 May 2026 |
|  Awaiz Azam joined the room. | 00:18:27 |
|  Richard Tichý joined the room. | 11:28:09 |
| 13 May 2026 |
|  snowflake joined the room. | 13:42:28 |
| snowflake | hello everyone just wanted to know if the package of bisq would get updated? search.nixos.org/packages?channel=25.11&query=bisq | 13:44:12 |
| Notarin | This channel is for development on Nix itself, not nixpkgs. You're looking for #users:nixos.org | 20:16:45 |
| 14 May 2026 |
| snowflake | omg i'm sorry for messing that up! | 03:02:56 |
 tomberek | Mic92: lovesegfault thoughts on using CURLOPT_MAXAGE_CONN to un-revert https://github.com/NixOS/nix/pull/15522 ? | 04:37:53 |
 Mic92 | @tomberek:matrix.org: we can but we have to handle S3's custom XML errors same way as the SDK implementations | 04:43:03 |
| Mic92 | Afaik the Aws SDK also enables keep alive | 04:44:26 |
| tomberek | Other approach would be to conditionally set it. | 04:47:37 |
|  @tjni:matrix.org left the room. | 04:57:40 |
| tomberek | curl_easy_setopt(req, CURLOPT_TCP_KEEPALIVE, 1L);
curl_easy_setopt(req, CURLOPT_TCP_KEEPIDLE, 30L);
curl_easy_setopt(req, CURLOPT_TCP_KEEPINTVL, 30L);
curl_easy_setopt(req, CURLOPT_MAXAGE_CONN, 90L);
how about this. We may have been hitting a 60s timeout on the server side.. when matching ours there is a small race condition. Looks like 30s is fairly commonly used to prevent that.
| 05:00:29 |
| tomberek | For consideration: https://github.com/NixOS/nix/pull/15855 | 05:11:05 |
|  haz joined the room. | 17:01:42 |
|  kaustubh.desale joined the room. | 17:16:31 |
