| 18 Sep 2021 |
| George Wyatt left the room. | 00:09:46 |
| 19 Sep 2021 |
| Artturin changed their display name from artturin to Artturin. | 04:46:52 |
| shadowninja55 left the room. | 13:05:48 |
| 22 Sep 2021 |
| bew joined the room. | 22:24:11 |
bew | Hello, I'm reading https://github.com/wmertens/rfcs/blob/master/rfcs/0017-intensional-store.md and there's something I don't understand: In the part A note about reproducibility, it says:
There is no need for a given $out to always generate the same $cas. [...] There is no obligation that a single $out only stores a single $cas entry.
| 22:26:23 |
bew | How is that possible? | 22:26:33 |
bew | How can a single $out generate multiple $cas ? | 22:27:35 |
bew | (if there's a better room for this question, please tell me 🙏) | 22:28:07 |
tomberek | bew: Basically that is saying that the builds don't have to be reproducible for the intentional store to work. (of course it helps) | 23:07:25 |
tomberek | You and I can build the same $out, but get different $cas's. This isn't ideal, but that's why there is a mapping from 1 $out to possible multiple $cas's. | 23:09:12 |
| 23 Sep 2021 |
afontain | what's $cas ? | 10:14:18 |
afontain | or is it $cas's ? | 10:14:31 |
nrdxp | Content Addressed Store | 14:16:22 |
nrdxp | tl;dr: It's an experimental feature which places packages in the nix/store based on a hash of their output rather than their input | 14:16:51 |
| 24 Sep 2021 |
afontain | ok, thanks | 15:31:36 |
| 25 Sep 2021 |
| trofi joined the room. | 08:28:33 |
| 26 Sep 2021 |
| lyderichti59 joined the room. | 13:13:22 |
| 27 Sep 2021 |
| yusdacra changed their profile picture. | 19:50:42 |
| 29 Sep 2021 |
j-k | in a non-nix context on your typical distro is it possible to have a reproducible binary that has dynamic links?
For achieving that reproducability I guess it'd be safest/easiest to use a statically linked gcc? If you're not using a statically linked gcc I guess you would need to keep track of your gcc version and glibc version? | 11:25:10 |
andi- | It is possible. The closure might not be entirely reproducible if you are running e.g. Distro verison A but try to reproduce version B. The binary might be the same but e.g. the dynamic linker (even thought it is at the same spot) might be different. | 11:44:55 |
j-k | Ok cool, that's in line with what I thought. Just needed to clarify 😅 | 11:46:57 |
| 30 Sep 2021 |
| Robby O'Connor joined the room. | 01:18:27 |
| matrixforever joined the room. | 02:16:18 |
| matrixforever left the room. | 04:38:12 |
| Robby O'Connor left the room. | 05:49:56 |
| 1 Oct 2021 |
| etrigan63 joined the room. | 10:13:15 |
| 4 Oct 2021 |
| Julio Castilla joined the room. | 03:46:18 |
Julio Castilla | Does anyone know how to dump the entire nix store using nix store --export?
| 03:46:22 |
Julio Castilla | My goal is to export and then import using a dockerfile | 03:46:41 |
| bew left the room. | 07:01:57 |