!djTaTBQyWEPRQxrPTb:nixos.org

Nixpkgs Architecture Team

231 Members
https://github.com/nixpkgs-architecture, weekly public meetings on Wednesday 15:00-16:00 UTC at https://meet.jit.si/nixpkgs-architecture53 Servers

Load older messages


SenderMessageTime
5 Aug 2022
@k900:0upti.meK900FWIW, I do think getting more people from the Nixpkgs side involved in Nix development (and vice versa) would be great for both sides, but I'm not really aware of the previous issues so I'll just stay out of this conversation14:23:28
@profpatsch:augsburg.oneprofpatsch Features like __impure or similar could be good to have in nixpkgs, but if no backwards compatibilty guarantees are defined, they might just happen by removing nix-build et al. 14:23:45
@profpatsch:augsburg.oneprofpatschSo it’s in the interest of the nixpkgs side of things to define what is expected14:24:13
@chris:mkaito.netmkaitoFairly obvious, but has this been brought up on the nix issue tracker for actual discussion?14:24:55
@profpatsch:augsburg.oneprofpatschEven without “grievances” (it’s a software project, nobody here did anybody wrong), just having formally defined borders is a good step towards a healthy project14:25:07
@profpatsch:augsburg.oneprofpatschSee e.g. Rust, which has a lot of formally defined structure and procedures14:25:24
@infinisil:matrix.orginfinisilYeah, I think it's very important to have backwards compat guarantees. Requiring features to be stable for 2 years sounds reasonable. It will also allow people with up to two year old NixOS installations to update to the latest one in a single step14:25:48
@profpatsch:augsburg.oneprofpatschSo it’s perfectly reasonable to focus on e.g. user experience for nixpkgs usage, even if that means putting some maintenance burden on nix development14:25:59
@chris:mkaito.netmkaitoI'm fine with the idea of requiring work from Nix to make things in nixpkgs possible or easier, of course.14:26:41
@profpatsch:augsburg.oneprofpatschchris: The nix issue tracker currently has 1500 issues open, with only 2500 closed, so I personally don’t see it as the place. Maybe the Discourse.14:26:58
@chris:mkaito.netmkaitothe number is meaningless IMO14:27:17
@profpatsch:augsburg.oneprofpatschI could open an issue, but the chance is that it would just be ignored. If the Nixpkgs Arch team made a statement of intent, that could not just be ignored.14:28:08
@profpatsch:augsburg.oneprofpatsch… Which is the entire reason for a team like that to exist imho14:28:40
@chris:mkaito.netmkaitoEither works. If we want the team to make a statement, we'll have to wait until we get to that point in the schedule (is __impure even on the schedule?) and reach formal consensus. You can also open the issue yourself. We'll figure out a way for it to not be ignored. We've all learned a lot since the 2.4 release.14:29:25
@chris:mkaito.netmkaitoWork on Nix itself might be explicitly out of scope for the team, but as I said above, and will stubbornly repeat until your ears fall off, "we are nix". We're all in this together, like it or not.14:30:27
@profpatsch:augsburg.oneprofpatschThat’s good to hear.14:30:47
@profpatsch:augsburg.oneprofpatschWell, I don’t think that’s entirely true (e.g. what is “it”? For many people it’s just setting up a nice environment and then doing something else)14:31:32
@chris:mkaito.netmkaito"it" is the nix ecosystem, at least in my mind14:31:49
@chris:mkaito.netmkaitonixpkgs is a major player in "it", obviously14:32:08
@profpatsch:augsburg.oneprofpatschI’d start by asking what the main use-cases for nixpkgs are. And then building some Personas, UX design style14:32:49
@profpatsch:augsburg.oneprofpatsche.g. “Will works for corp X and has a python dev environment. Margaret likes to merge nixpkgs Issues in her free time”14:33:19
@chris:mkaito.netmkaitoWe should put that on the agenda for next week. Sounds much more important than deconstructing mkDerivation IMHO14:33:24
@profpatsch:augsburg.oneprofpatschOnce there is a good set of those, one can figure out what their specific pain points are14:33:37
@profpatsch:augsburg.oneprofpatsch“Bertha just runs NixOS in a pretty default config on her system and does not program at all”14:35:34
@qyliss:fairydust.spaceAlyssa RossI am inclined to suggest there should be a Nix/Nixpkgs sync meeting14:36:20
6 Aug 2022
@ronnypfannschmidt:matrix.orgRonny joined the room.12:33:23
@wentasah:matrix.orgwentasah joined the room.22:15:29
@tomberek:matrix.orgtomberek
In reply to @qyliss:fairydust.space
As a Nixpkgs developer, I do not feel represented or listened to by the Nix developers. I tihnk profpatsch feels the same way.
I consider myself a Nix developer. Part of the reason I’m here is to ensure some cross-ecosystem consistency and coordination. I’ve done a few interview sessions with people to try to truly grok their point of view. I’d be happy to do that with either profspatsch or yourself.
23:20:11
7 Aug 2022
@srid:matrix.orgSridhar Ratnakumar joined the room.13:36:53
@sternenseemann:systemli.orgsterni joined the room.14:18:30

There are no newer messages yet.


Back to Room ListRoom Version: 9