 | Nix Flakes | 859 Members |
| 177 Servers |
| Nix Flakes | 859 Members |
| 177 Servers |
| Sender | Message | Time |
|---|---|---|
| 9 Mar 2025 | ||
 emily | flakes have reliably had more breaking changes than any other part of the language really | 15:01:14 |
| emily | lots of weird stuff around subflakes too | 15:01:30 |
 WeetHet | In reply to @elikoga:matrix.orgTbf one could argue it became this way due to widespread propaganda of flakes as an entry point into the nix world which they imho shouldn't be | 15:17:50 |
 elikoga | What the fuck else should I be using to lock my lockfile? npins? Pinning by hand? | 15:21:01 |
| elikoga | Should I set NIX_PATH myself like a Victorian era peasant? | 15:21:37 |
| elikoga | Should I write my own script to do the fetching? | 15:21:51 |
| elikoga | What is the path of least resistance to getting a lock file? Using the built in lock file management. I must admit, the shilling can go on my nerves too, but as far as I can tell, that's why there is a low lack of understanding accross the entire nix ecosystem. Getting users acquainted with configuration.nix first may be very easy but it can lead to a large population of users unfamiliar with what they are doing | 15:24:53 |
 hexa joined the room. | 15:26:44 | |
| elikoga | * What is the path of least resistance to getting a lock file? Using the built in lock file management. I must admit, the shilling can go on my nerves too, but as far as I can tell, that's why there is a lack of understanding accross the entire nix ecosystem. Getting users acquainted with configuration.nix first may be very easy but it can lead to a large population of users unfamiliar with what they are doing | 15:27:41 |
| elikoga | I am sorry for my crass language yet the other side can call the feature "a stain" on the language I am allowed to swear a little bit when complaining about a lack of solutions given | 15:29:53 |
| emily | I think it's not productive to discuss this further unless you want to retract "all the arguments I can see going against this are bad-faith, non-users, agitators, that deliberately misrepresent the technical state", which seems to have been based more on your own ignorance of the technical state than mine | 15:33:26 |
| emily | I haven't participated in any mud-slinging myself, so I don't think you have any reason to be like this when I'm trying to engage against my better judgement | 15:34:03 |
| hexa | https://discourse.nixos.org/t/experimental-does-not-mean-unstable-detsyss-perspective-on-nix-flakes/32703/2 is another post that summarizes existing issues with flakes | 15:37:33 |
| elikoga |
Actually this smells like the collusion and intimidation I know from university politics. This post has been quoted many times, yet I believe that all of the points are not release-blockers and deliberately use emotional language like "is awful and a clear downgrade from stable Nix." while not considering any of the technical tradeoffs I think I'll write up something longer and much more calmer on this in a bit or defer that to the future since I don't see going around in small messages going anywhere. Thank you all for taking the time to discuss | 15:42:06 |
| emily | I literally gave a link and a quote when you asked? | 15:43:20 |
| emily | I spent several minutes digging them up for you, even though the quote was from a forum thread you've already been participating in… | 15:43:53 |
| emily | "collusion and intimidation" – I am speaking entirely for myself and have not participated in any of these forum threads | 15:44:16 |
| hexa | "collusion and intimidation" 🤦♂️ | 15:44:39 |
| emily | I talk to people on both the Nix and Lix team. it seems like you are angry and just want to assign me to a side so you can dismiss my substantive responses to you | 15:45:05 |
| hexa | yes, feel free to write something up, since this discourse is not really working out for you | 15:45:08 |
| hexa | * yes, feel free to write something up, since this discourse obviously is not working out for you | 15:45:42 |
| hexa | * yes, feel free to write something up, since this discourse is obviously not working out for you | 15:45:49 |
| emily | anyway, I think a lot of rhetoric getting thrown around is dumb, so if you assume I agree with all of it then I understand that my position would look disingenuous or silly | 15:46:02 |
| emily | I do not think we should immediately destroy flakes and break everything forever, or that Nix before flakes was a crystalline jewel of bug-free perfection and no weird path-dependent backwards compatibility concerns, or anything | 15:46:33 |
| emily | I do think that flake internals are weird and buggy enough, and the number of breaking changes resulting from bug fixes that would be considered unacceptable for a stable feature but that would have been awful has been high enough, and continues climbing enough, that were the ecosystem to fork over, say, a declared stability guarantee for exactly one semantics of flakes (already different to the semantics we had one stable version ago), it would be very bad for all parties | 15:48:09 |
| emily | and if you think that both the Nix team, the Lix team, and the Tvix team are colluding to arrive at the position of "stabilizing flakes in a way that doesn't cause huge problems down the line is going to be a delicate process" then… well, anyone managing to establish a robust conspiracy from that coalition might as well be a god, but also I wonder how much time you have actually spent with the codebase? | 15:50:10 |
 raitobezarius joined the room. | 15:50:45 | |
 Sandro 🐧 | In reply to @hexa:lossy.networkThat's pretty well written 👍 | 15:51:20 |
| emily | there are political problems of course… for instance, you can probably guess at a reason why the Nix team's communications on this matter are quiet | 15:51:23 |
| elikoga | Enough to feel silly when seeing infinisil complain about coupling of the eval-cache mechanism with flakes. I can understand criticizing that the name isn't "flake-eval-cache" but as it stands that proposal is not coherent as infinisil said. I don't know about that not mud-slinging defense Also I'm just going to call a spade a spade and both lix and detnix are hostile forks of github:nixos/nix by definition | 15:54:41 |