!tDnwWRNkmmYtMXfaZl:nixos.org

Nix Language

601 Members
Nix programming language123 Servers

Load older messages


SenderMessageTime
3 Dec 2022
@theliver:matrix.orgtheliverI'm guessing it's due to the channels having been updated in the meantime since the last build was done. So I guess if I can figure out which revision/commit the channels were on, on the last generation build, I can rollback to that and rebuild a new generation without the nix cache reference to hydra.iohk.io 14:47:19
@theliver:matrix.orgtheliver Well in that case it was just the last channel revision so it was easily solved by just running sudo nix-channel --rollback. 14:53:57
@noob_tea:matrix.orgtea joined the room.20:10:21
@kokern:matrix.orgshe/they/khinah joined the room.21:11:04
@vimalk:matrix.orgvimalk joined the room.22:47:12
@aktaboot:tchncs.deaktaboot joined the room.23:35:50
4 Dec 2022
@jez:petrichor.me@jez:petrichor.me left the room.16:51:01
5 Dec 2022
@lezhou8:matrix.org@lezhou8:matrix.org joined the room.03:28:00
@lezhou8:matrix.org@lezhou8:matrix.org left the room.03:28:36
@mon:tchncs.deribosomerocker changed their display name from mon_aaraj to ribosomerocker.22:45:11
6 Dec 2022
@federicodschonborn:matrix.orgFederico Damián Schonborn joined the room.13:00:38
7 Dec 2022
@desttinghim:matrix.orgdesttinghim changed their profile picture.07:22:35
@threedee:matrix.orgthreedee joined the room.20:15:50
8 Dec 2022
@vector1dev:vector1dev.ml@vector1dev:vector1dev.ml left the room.09:53:40
@farcaller:matrix.orgfarcaller hi o/ I don't seem to understand what are the inputs to a derivation. I'm writing a derivation that fetches files from internet and processes them; thus it's not contained and I have to specify outputHash-things. My expectation was that the environment is part of input, so if I change the environment that must trigger a re-build even if there is a derivation for the same hash. But e.g. given this derivation https://gist.github.com/farcaller/1180cff0f448d4cf89575b07566c6cd6 once I build it once and set the chartHash to the resulting hash, no matter if I change any of the input params the build result is the same. 17:58:16
@farcaller:matrix.orgfarcalleram I doing something totally wrong or am I doing something that nix as a build system cannot really handle?17:58:40
@raphi:tapesoftware.netraphithat is how fixed output derivations work: they are only identified by their output hash. changing the derivation attributes (except for the hash of course) will not cause a rebuild.18:00:44
@farcaller:matrix.orgfarcallerright. Is there any way around it?18:01:05
@rick:matrix.ciphernetics.nlMindaviChange the hash18:01:35
@farcaller:matrix.orgfarcallerwell; other than that :D it's easy to forget to and that will result in a potentially broken build18:01:58
@farcaller:matrix.orgfarcalleractually, it does depend on the name too, right?18:07:03
@farcaller:matrix.orgfarcallerI think that's my way out18:07:07
@rick:matrix.ciphernetics.nlMindaviThere is also impure-derivations, but I don't know if that is what you're looking for. It's experimental and will _always_ rebuild18:08:18
@rick:matrix.ciphernetics.nlMindaviUnless you let it eventually end up in a FOD, but then you go back to square one18:08:40
@farcaller:matrix.orgfarcallermy plan is to extract the download part. It's addressed by repo+name+version and I can store those in the name, making the download properly content-addressable18:09:51
@farcaller:matrix.orgfarcaller \o/ it works! thanks guys 18:22:19
@farcaller:matrix.orgfarcaller how do I remove a nested field from an attrset? e.g. given { a = 1; b = { c = 2; d = 3; }; } I want to get { a = 1; b = { c = 2; }; } 19:23:53
@infinisil:matrix.orginfinisil farcaller: The easiest is manually with attrs: attrs // { b = removeAttrs attrs.b [ "d" ]; } 19:50:44
@infinisil:matrix.orginfinisil If you need more complicated nested attribute set updates, I can recommend lib.updateManyAttrsByPath: https://github.com/NixOS/nixpkgs/blob/a8550d1553ff4b2deed7ff3e984cc958032b3eb7/lib/attrsets.nix#L100-L138 19:54:23
@cods:matrix.orgcods joined the room.21:13:20

There are no newer messages yet.


Back to Room ListRoom Version: 6