| 9 Jun 2023 |
 infinisil | hexa: Is this rhetorical? Because that was just mentioned! See https://github.com/NixOS/rfcs/pull/140#issuecomment-1584882324 | 17:15:01 |
 hexa | uh, yes, it was rhetorical | 17:15:18 |
| hexa | I have not followed up on the thread, good that you mentioned it ig 🙂 | 17:15:41 |
| infinisil | hexa: The "not" in your sentence confused me a bit, I wasn't sure if you're saying it is a bad time and I shouldn't have brought it up again, or if it's a good time to bring it up :P | 17:17:19 |
| hexa | I mean, him choosing legacyPackages was much more confusing than either units or shards | 17:17:42 |
| hexa | so I don't get what the fuss is about | 17:17:57 |
 @piegames:matrix.org | In reply to @infinisil:matrix.org
piegames: So, the rfc process doesn't state who decides over whether the FCP needs to be cancelled:
In most cases, the FCP period is quiet, and the RFC is either merged or closed. However, sometimes substantial new arguments or ideas are raised, the FCP is canceled, and the RFC goes back into development mode.
Ack. I'm leaning on the "sometimes substantial new arguments", because while the opposition has been in parts strong, I'd argue that being for a proposal but against a name in it is not substantial | 17:20:05 |
| infinisil | phaer: j-k: growpotkin ( Alex Ameen ): piegames: As shepherds of RFC 140, do any of you think niksnut's recent criticism is a substantial new argument that requires canceling FCP and potentially changing the RFC? | 17:20:07 |
| @piegames:matrix.org | Ha, good timing ^^
IMO it is neither new nor substantial | 17:20:37 |
 Eelco | I don't think it requires cancelling the FCP | 17:21:53 |
| Eelco | As long as people don't think that RFC saying "pkgs/unit" can be used as a blanket override of any objections on an eventual PR | 17:27:42 |
| Eelco | 1 case of bad naming doesn't justify more bad naming. Quite the opposite. | 17:28:41 |
| hexa | certainly, to me sharding or partitioning captures the meaning better than unit | 17:29:45 |
 K900 | I wasn't 100% following the conversation | 17:30:14 |
| K900 | But was pkgs/simple proposed? | 17:30:26 |
| hexa | can you elaborate on the meaning? | 17:30:53 |
| hexa | * can you elaborate on the meaning of simple in that context? | 17:31:06 |
| infinisil | niksnut: The point of RFC's is to decide controversial things. If somebody has objections, now's the time to raise them. Once the RFC is accepted it's too late. I quote the RFC Readme:
In general though this means that the implementation will be merged as long as there are no substantial technical objections to the implementation.
| 17:31:23 |
| K900 | Simple as in simple package paths, also simple as in simple to add | 17:31:28 |
| infinisil | In reply to @k900:0upti.me
But was pkgs/simple proposed? I think that would imply that there's a complicated way to declare packages, which there currently is, but it's something we should get away from. Once we migrate everything, simple wouldn't mean anything anymore | 17:33:20 |
| Eelco | infinisil: The RFC process doesn't replace PR review. And "pkgs/unit" shouldn't get through PR review. | 17:34:17 |
| hexa | so at this time the RFC prefers a meaningless name over one with a meaning | 17:36:00 |
| @piegames:matrix.org | In reply to @hexa:lossy.network
so at this time the RFC prefers a meaningless name over one with a meaning Some people do, personally I am fine with a meaningless name because any name is better than no name. And so far no satisfactory meaningful name has been brought up IMO | 17:36:52 |
| infinisil | There's a decent argument for unit by Robert Hensing (roberth) here: https://github.com/NixOS/rfcs/pull/140#discussion_r1170362174 | 17:37:51 |
| infinisil | In reply to @niksnut:matrix.org
infinisil: The RFC process doesn't replace PR review. And "pkgs/unit" shouldn't get through PR review. Even if you created Nix, I expect you to respect the community's RFC process, especially for Nixpkgs. If you were to block/revert the implementation of the RFC as it is stated, I consider that a violation of the RFC process. | 17:40:18 |
| Eelco | In that case I withdraw my statement about not having to cancel FCP | 17:41:03 |
| infinisil | Well, it's up to the shepherd team to decide whether it needs to be canceled | 17:41:33 |
| infinisil | K900: Oh apparently simple was proposed, see this linked thread | 17:42:58 |
| Eelco | For instance, if a shepherd team makes some bad decision about Nix, I wouldn't feel that the Nix team would be required to implement it. That's not how open source works. You can't force maintainers to accept technical decisions that they can't get behind. | 17:43:13 |
| infinisil | niksnut: Fully agreed, but in this case it's not Nix, it's Nixpkgs, which is fully developed by the community, there's no official global Nixpkgs maintainers | 17:44:20 |
