| 29 Aug 2025 |
 toonn | Well, yeah, but that's going beyond what Hackage and the package maintainers explicitly support. | 20:23:08 |
| 30 Aug 2025 |
 sterni (he/him) | toonn: Non-breaking change. Otherwise, if only new bindings, types, classes, non-orphan instances or modules (but see below) were added to the interface, then A.B MAY remain the same but the new C MUST be greater than the old C. | 01:55:58 |
| sterni (he/him) | (below is only about leaking instances) | 01:56:09 |
 maralorn | sterni: I thought the point is: When you import a package unqualified then you need to have an upper bound on the .C to conform with PVP. | 02:08:57 |
| maralorn | * sterni: I thought the point is: When you import a package unqualified then you need to have an upper bound on the .C of that package to conform with PVP. | 02:09:12 |
| sterni (he/him) | they don't even use an uppercase MAY or RECOMMENDED for that | 02:12:21 |
| sterni (he/him) | they say not doing that is “slightly risky” | 02:12:52 |
| sterni (he/him) | but you are technically conforming when everything's fine when uploading | 02:13:03 |
| maralorn | Weird. All my years reading CLC discussions had made me think that | 09:10:31 |
| toonn | Are you though? Because the PVP as written is for versioning a package, not for how to bound dependencies, no? | 09:24:25 |
 Teo (he/him) | Out of interest, how many packages did this break in nixpkgs? I would've expected everything to be updated by now. I guess it's things not on Stackage? | 10:15:42 |
|  Ember Void changed their profile picture. | 15:30:24 |
