| 31 Mar 2026 |
 Lisanna | Jon Ericson and I have been working on getting a Store refector in to cppnix that separates building into its own class, which is the first step towards making implementing custom building logic more ergonomic. Once this is finalized and merged, would the Lix team be interested in accepting a port of this work for Lix?
https://github.com/NixOS/nix/pull/15542
| 20:05:31 |
| Lisanna | * John Ericson and I have been working on getting a Store refector in to cppnix that separates building into its own class, which is the first step towards making implementing custom building logic more ergonomic. Once this is finalized and merged, would the Lix team be interested in accepting a port of this work for Lix?
https://github.com/NixOS/nix/pull/15542
| 20:05:42 |
 raitobezarius (DECT: 7248) | I am not sure any port of this nature can happen without more involvement with the development team, we have a bunch of plans of our own regarding this topic and we have noticed we do not share the same technical vision with CppNix (at least, based on public RFC information we were asked to give our opinion on) -- I wouldn't want to waste anyone's time on a port that would be -2 architecturally after a lot of efforts were made
What would make more sense to me is to merge it and perhaps discuss your concepts in Lix's Zulip and see if there are interests to take the same architectural decisions, with this information, a "port" or doing the work would make sense IMHO | 20:20:57 |
| Lisanna | Alright ^-^ | 20:22:27 |
| raitobezarius (DECT: 7248) | Thanks for the kind offer nonetheless! | 20:25:11 |
| Lisanna | I'll put together some more information on use cases once it's finalized and join the zulip to discuss the idea | 20:32:57 |
| 4 Apr 2026 |
|  Lotte (it/its)/Cinny (she/her) θΔ& changed their profile picture. | 10:58:32 |
 antifuchs | lix in nixpkgs deprecation-warning about nixpkgs's export of lib.or is pretty funny though. | 14:55:46 |
| antifuchs | some other deprecation warnings have been very helpful in finding very screwed-up regexes (or things that aren't actual regexes), so overall a fantastic release, thanks everyone who worked on getting these improvements out! | 15:05:31 |
 piegames | lib.or should have died over a decade ago. lib.or is the reason why the precedence is fucked around or as an operator | 16:55:38 |
| antifuchs | Anything that speeds its demise is very welcome. What a weirdo identifier. | 17:10:10 |
| antifuchs | Also, it is objectively hilarious | 17:10:32 |
| 5 Apr 2026 |
| Lotte (it/its)/Cinny (she/her) θΔ& changed their profile picture. | 16:57:20 |
| 6 Apr 2026 |
| Lotte (it/its)/Cinny (she/her) θΔ& changed their profile picture. | 13:47:41 |
| Lotte (it/its)/Cinny (she/her) θΔ& changed their profile picture. | 13:48:49 |
| 7 Apr 2026 |
|  @vigress9:matrix.org removed their display name V. 🏳️⚧️. | 17:46:19 |
| @vigress9:matrix.org left the room. | 17:47:05 |
| 8 Apr 2026 |
 K900 | Has anyone ever proposed a, like, --retry 3 option for nix-build? | 20:30:18 |
| K900 | I feel like I might have | 20:30:22 |
| K900 | But my brain is also mush | 20:30:25 |
 Sergei Zimmerman (xokdvium) | In reply to @k900:0upti.me
Has anyone ever proposed a, like, --retry 3 option for nix-build? As in build attempts? That used to be a thing (though broken) in old nix versions I think | 21:04:20 |
| K900 | Yes | 21:04:37 |
| K900 | This message in particular is brought to you by me trying to get flaky scipy tests to run overnight | 21:04:56 |
| Sergei Zimmerman (xokdvium) | Yeah I think there are even remnants of it in the daemon protocol | 21:05:06 |
| Sergei Zimmerman (xokdvium) | It was it about doing the build multiple times to check for determinism. Not sure | 21:08:01 |
| Sergei Zimmerman (xokdvium) | * Or was it about doing the build multiple times to check for determinism. Not sure | 21:08:09 |
| 9 Apr 2026 |
| antifuchs | Think buildbot had an option for that too | 01:27:07 |
|  ElvishJerricco joined the room. | 20:30:34 |
| Lisanna changed their profile picture. | 21:59:43 |
| Lisanna changed their profile picture. | 22:00:59 |
