Post B1say5ThMORd0d2GCu by Pol@mathstodon.xyz
 (DIR) More posts by Pol@mathstodon.xyz
 (DIR) Post #B1sZQkBvNn0xTmG6Lo by nobody@mastodon.acm.org
       2026-01-02T15:56:33Z
       
       0 likes, 0 repeats
       
       Flakes
       
 (DIR) Post #B1say5ThMORd0d2GCu by Pol@mathstodon.xyz
       2026-01-02T16:13:43Z
       
       0 likes, 0 repeats
       
       @nobody It is funny to see all these userland initiatives. While flakes are still experimental, nothing prevents anyone from implementing the same specifications. And if flakes ever stabilise, switching from one implementation to another should be fairly trivial... theoratically.
       
 (DIR) Post #B1sd2d44cxa8F8LKmO by nobody@mastodon.acm.org
       2026-01-02T16:37:00Z
       
       0 likes, 0 repeats
       
       @PolUhm no. Flakes are already "stabilized" by 2 of 3 major implementors and they are already incompatible, just as varying version numbers within 1 implementation you have been getting incompatible and irreproducible behaviours too. The imlementation should have never been moved out of the userland (flake-compat) in the 1st place
       
 (DIR) Post #B1se4UvTxPfjXEmOAq by Pol@mathstodon.xyz
       2026-01-02T16:48:31Z
       
       0 likes, 0 repeats
       
       @nobody Those "implementors" are proprietary companies looking for businesses and I pretend they don't exist since I will never use or recommend them to anyone. For the rest, I believe `flake-compat` is existing just to bridge the Nix "flake" world to the Nix "legacy" world, so, IMO it doesn't implement flakes, it's just a compatibility layer relying on flake (since it requires a `flake.lock`).
       
 (DIR) Post #B1sg5RbTzno3Qc0J3Q by nobody@mastodon.acm.org
       2026-01-02T17:11:08Z
       
       0 likes, 0 repeats
       
       @PolLix is not proprietary
       
 (DIR) Post #B1siS9i2tLjwV11z7o by Pol@mathstodon.xyz
       2026-01-02T17:37:36Z
       
       0 likes, 0 repeats
       
       @nobody Ooh IIUC, they diverged (in a backward incompatible way) from the official flake implementation ? Note, I am talking about flake only, not about the Nix code (I'm already aware of this https://discourse.nixos.org/t/lix-introduces-integer-coercion-in-string-interpolation/65024/4).
       
 (DIR) Post #B1snaZoiCxQJYdz4OO by nobody@mastodon.acm.org
       2026-01-02T18:35:11Z
       
       0 likes, 0 repeats
       
       @PolThere's no "official flake implementation" though
       
 (DIR) Post #B1spChJqrZQHmfI5B2 by Pol@mathstodon.xyz
       2026-01-02T18:53:15Z
       
       0 likes, 0 repeats
       
       @nobody OK fair enough... to me the official Flake implementation is the current state of things in Nix.
       
 (DIR) Post #B1stdbvPQUiCgnwk1A by nobody@mastodon.acm.org
       2026-01-02T19:42:58Z
       
       0 likes, 0 repeats
       
       #flakes