Nixpkgs Architecture Team | 228 Members | |
| https://github.com/nixpkgs-architecture, weekly public meetings on Wednesday 15:00-16:00 UTC at https://meet.jit.si/nixpkgs-architecture | 53 Servers |
| Sender | Message | Time |
|---|---|---|
| 9 Jun 2023 | ||
| I guess the future of the NAT is working groups, for which we already create specific separate channels (the only current working group going on is #wg-pkgs-modules:matrix.org) | 20:14:46 | |
| So that works out :) | 20:14:51 | |
| 21:06:58 | ||
| 10 Jun 2023 | ||
In reply to @piegames:matrix.org strictly speaking, my insecurity comes from a place of wanting to respect hierarchy and responsibility. even if there were a separate RFC room, I would have had the same apprehension. personally speaking, I'm still very new to Nix/NixOS and moreover I've never explicitly expressed interest to anyone about wanting to contribute to the shape and structure of the project. I wanted to respect the folks who have already devoted a lot of time thinking about the problem, have devoted a lot of time discussing the problem, and I wanted to respect those who have stood up to be personally named and responsible for solving this problem. additionally, I know that some aspects of this RFC were a little protracted and maybe even a little heated. I also have no interest in contributing to the bikeshedding component in this RFC. I understand that names are often a source of trouble and that a well intentioned decision could lead to unintended consequences which could be difficult to revert or even potentially infeasible to revert so I sympathize with and respect the fact that so much time has been spent on the issue. I don't speak for anyone else when I say this, of course. this is just how I feel and how I feel is not the result of what the Nix/NixOS project has suggested to me in any way. the community has been very kind and welcoming and I respect all of you and I also feel respected. as for whether having separate rooms for discussing separate topics is a good thing, I think it can be. I think it can make discovery of discussion also more difficult. it would probably be useful to have a way to remind folks of ongoing discussions in the main channel as a way of inviting people to join on conversations they care about. case in point: I didn't even know this room existed until I saw the Summer of Nix lecture. one final comment I have on the RFC though, and I'm sorry to repeat my point again, but I do think this RFC has been made more difficult than it should have been because it has been decided that Nixpkgs should work around GitHub's UI/UX issues. I know GitHub is an important tool for the Nix/NixOS community but I do not think that the engineering and design of Nix/NixOS should be subject to arbitrary peculiarities of a UI. | 13:10:35 | |
| * strictly speaking, my insecurity comes from a place of wanting to respect hierarchy and responsibility. even if there were a separate RFC room, I would have had the same apprehension. personally speaking, I'm still very new to Nix/NixOS and moreover I've never explicitly expressed interest to anyone about wanting to contribute to the shape and structure of the project. I wanted to respect the folks who have already devoted a lot of time thinking about the problem, have devoted a lot of time discussing the problem, and I wanted to respect those who have stood up to be personally named and responsible for solving this problem. additionally, I know that some aspects of this RFC were a little protracted and maybe even a little heated. I also have no interest in contributing to the bikeshedding component in this RFC. I understand that names are often a source of trouble and that a well intentioned decision could lead to unintended consequences which could be difficult to revert or even potentially infeasible to revert so I sympathize with and respect the fact that so much time has been spent on the issue. I don't speak for anyone else when I say this, of course. this is just how I feel and how I feel is not the result of what the Nix/NixOS project has suggested to me in any way. the community has been very kind and welcoming and I respect all of you and I also feel respected. as for whether having separate rooms for discussing separate topics is a good thing, I think it can be. I think it can make discovery of discussion also more difficult. it would probably be useful to have a way to remind folks of ongoing discussions in the main channel as a way of inviting people to join on conversations they care about. case in point: I didn't even know this room existed until I saw the Summer of Nix lecture. one final comment I have on the RFC though, and I'm sorry to repeat my point, but I do think this RFC has been made more difficult than it should have been because it has been decided that Nixpkgs should work around GitHub's UI/UX issues. I know GitHub is an important tool for the Nix/NixOS community but I do not think that the engineering and design of Nix/NixOS should be subject to arbitrary peculiarities of a UI. | 13:35:10 | |
| The GitHub limit of 1000 files is admittedly a bit arbitrary, but also a lot of other software handles folders with many items poorly (mostly in terms of performance degradation). Therefore I don't think putting everything into one flat folder would be a good idea, even when putting GitHub aside. | 14:05:27 | |
| Brainstorming some more: What about `unsorted` or `uncategorized`? | 17:35:38 | |
| About `by-name`: to me it invokes the association of `/dev/by-*`, which has the caveat that ours is not a "view" onto the package set and therefore not exhaustive. But in the end it is not worse than unit so idc too much | 17:37:35 | |
In reply to @piegames:matrix.orgWhat about _? :DI mean, it has no intrinsic meaning yet conveys the idea of "we decided to move some packages here" quite well | 17:38:52 | |
_ conveys nothing at best. Normally we only use it for "I don't want to name this because it's not used here" | 19:01:00 | |
un* is slightly better than no information at all, but naming something after what it isn't isn't super helpful | 19:02:33 | |
we could call nixpkgs unsingular | 19:03:51 | |
| 11 Jun 2023 | ||
I'd like to hear if there are any people strongly opposing by-name, especially from those who didn't like unit | 10:08:29 | |
| Also, a meeting might be a good idea | 10:08:38 | |
I don't like unit and don't mind by-name FWIW | 10:09:18 | |
In reply to @piegames:matrix.orginfinisil: Do you know who suggested pkgs/_ first? Because it looks like you added this to the RFC before my message in here, and I don't see any force-pushes either | 10:14:28 | |
| it was K900 | 10:15:21 | |
| Nope | 10:15:27 | |
| Wasn't me actually | 10:15:30 | |
| oh | 10:15:31 | |
| by-what-else tho? 😄 | 10:15:37 | |
| It was on the RFC before me I believe | 10:15:38 | |
| by-prefix? | 10:15:40 | |
by-unit | 10:15:47 | |
In reply to @k900:0upti.meI understand, I wouldn't want to take credit for this either :p | 10:15:51 | |
Actually, _by-unit | 10:15:54 | |
| __no? | 10:16:00 | |
In reply to @k900:0upti.meThe ultimate compromise | 10:16:15 | |
In reply to @piegames:matrix.orgthe worst of all choices then? 😄 | 10:16:32 | |
| I don't know what unit means and by this point I'm afraid to ask | 10:16:33 | |