| 24 Feb 2026 |
 Sergei Zimmerman (xokdvium) | That doesn't count derivation outPath equality and size mismatches of attrs | 11:21:33 |
 emily | the pointer comparison is hideously impure so not a good reason to not reduce existing impurity | 13:49:32 |
 K900 | Do we really want to say that flakes are hard frozen? | 13:49:47 |
| emily | the ordering also matches attrNames/attrValues | 13:49:49 |
| K900 | I feel like it's fine to pick some one-line patches that fix things | 13:50:03 |
| emily | so attrsets already have a canonical ordering | 13:50:04 |
 piegames | @k900:0upti.me that depends on whether that change may cause conflicts in the migration, I don't know the state of that CC @raitobezarius:matrix.org | 13:51:18 |
 raitobezarius | I am sorry I am failing to understand the conversation | 13:51:46 |
| raitobezarius | I read things on pointer comparisons from emily and things about flakes | 13:51:54 |
| raitobezarius | Can someone TLDR me? | 13:51:57 |
| raitobezarius | or repeat what they want from me | 13:52:00 |
| emily | i.e. you can implement an ordered-by-attrname == in Nix today (modulo pointer equality) | 13:52:13 |
| K900 | https://github.com/NixOS/nix/commit/3180671cabeb6a6010057770731e12761ed5666c cppnix broke flakes again | 13:52:16 |
| K900 | We could cherry-pick | 13:52:27 |
| K900 | Also the nix profile add thing from the other day | 13:52:37 |
| raitobezarius | man | 13:52:39 |
| K900 | I think it's fine to pick these small-ish changes that should merge cleanly into the migration | 13:53:01 |
| raitobezarius | Yes, but this change is non trivial from a libexpr perspective | 13:53:16 |
| raitobezarius | Now url Value have context strings? | 13:53:26 |
| K900 | Yeah that one might be too much | 13:53:34 |
| raitobezarius | I need to think | 13:53:47 |
| K900 | I was thinking about the nix profile add one more | 13:53:47 |
| raitobezarius | the nix profile add is controversial because this change has no raison d'etre and I understand that some folks in the team are getting really pissed off by this irresponsible behavior | 13:54:41 |
 delroth | this was actively being discussed on zulip yesterday fwiw if you don't follow that medium | 13:54:47 |
| emily | also, lists already have short-circuiting equality with a canonical ordering | 13:55:18 |
| emily | so making attrsets inconsistent with that seems gratuitous | 13:55:29 |
| delroth | really the irresponsible behavior is 3rd party software deciding to build on an experimental interface, no? | 13:55:34 |
| raitobezarius | yes | 13:55:47 |
| raitobezarius | hence: https://github.com/nix-community/home-manager/issues/8786#issuecomment-3951682232 | 13:55:54 |
| delroth | nice, that was roughly my take on this whole thing as well | 13:56:25 |
