!UKDpaKNNsBpOPfLWfX:zhaofeng.li

Colmena

311 Members
A simple, stateless NixOS deployment tool - https://github.com/zhaofengli/colmena109 Servers

Load older messages


SenderMessageTime
14 Feb 2023
@keithhub:matrix.orgkeithhub joined the room.16:55:32
@rendakuenthusiast:imperishable.name@rendakuenthusiast:imperishable.nameare there any guides for using home-manager with colmena?17:13:11
15 Feb 2023
@cw:kernelpanic.cafeChinchilla Washington changed their display name from Chinchilla Washington to Chinchilla Wanker.17:37:32
@cw:kernelpanic.cafeChinchilla Washington changed their display name from Chinchilla Wanker to Chinchilla Washington.17:38:11
16 Feb 2023
@zachel:fairydust.space@zachel:fairydust.space changed their display name from zachel to zachel (Old).10:28:43
@sef:exotic.shsefidel joined the room.16:48:49
@zachel:fairydust.space@zachel:fairydust.space left the room.16:50:45
@sef:exotic.shsefidel 20:20:10
17 Feb 2023
@dstengele:pango.placeDennis Stengele joined the room.20:44:50
22 Feb 2023
@lgcl:lgcl.delgcl (she/they) changed their display name from lgcl to lgcl (they/them).20:33:42
@winterqt:nixos.devWinter (she/her) Zhaofeng Li: i don't see a channel for attic, so forgive me if there's a better place for this, but why would one want --no-closures for watch-store? i can't figure it out from the docs or the source :( 21:00:32
@zhaofeng:zhaofeng.liZhaofeng Li
In reply to @winterqt:nixos.dev
Zhaofeng Li: i don't see a channel for attic, so forgive me if there's a better place for this, but why would one want --no-closures for watch-store? i can't figure it out from the docs or the source :(
Right, it isn't really useful in practice and more for testing PushSession. I can probably hide the flag to avoid confusion.
21:45:11
@zhaofeng:zhaofeng.liZhaofeng Li(and I should probably create a room for Attic if enough people are interested, after I decide whether to go with the rename or not)21:47:38
@whentze:matrix.orgWanja HentzeYes, +1 to an attic room :) it's a seriously cool project and deserves one22:04:02
23 Feb 2023
@shapr:cofree.coffeeshaprCould I get a link to attic?18:11:44
@shapr:cofree.coffeeshaproh, I found it https://github.com/zhaofengli/attic18:12:07
24 Feb 2023
@dminuoso:matrix.orgdminuosoI have an IFD that appears to segfault, but it only ever happens when evaluated through colmena. All I ever get from colmena is:09:36:56
@dminuoso:matrix.orgdminuoso Evaluation failed: Child process was killed by signal 11 09:37:09
@dminuoso:matrix.orgdminuosoI am a bit at a loss how to even debug this09:37:17
@dminuoso:matrix.orgdminuosoWhat makes matters worse, the segfault occurs relatively rarely09:37:33
25 Feb 2023
@winterqt:nixos.devWinter (she/her)I bumped my Nixpkgs input and deployed using deploy-on-target, and Nix copied a patch file that's only used when building a derivation. However, this build was substituted from the cache -- why would Nix download/copy it, then? It definitely wasn't built on the server.20:21:53
@winterqt:nixos.devWinter (she/her)I presume there's no way I can actually figure out why, though 🙃20:23:00
@winterqt:nixos.devWinter (she/her)Just seems very weird.20:24:28
26 Feb 2023
@zhaofeng:zhaofeng.liZhaofeng Li
In reply to @dminuoso:matrix.org
Evaluation failed: Child process was killed by signal 11
Interesting. You can try evaluating the config with the lib.makeHive function in the flake (for non-flake, you can see how it calls the underlying evaluator) and see if you can reproduce the segfault
03:23:19
@winterqt:nixos.devWinter (she/her)
In reply to@winterqt:nixos.dev
I presume there's no way I can actually figure out why, though 🙃

@Zhaofeng Li: happen to have any clue if/how I can debug this? i saw your reaction so i doubt it but figured I'd ask directly.

as far as I can tell, the given derivation (and a bunch of others of course) were substituted... but the patch file is in my store, somehow. if I rollback, GC, and then try to build it again, I might be able to repro... but how could I actually tell why it's copying that file?

03:54:47
@zhaofeng:zhaofeng.liZhaofeng Li I guess it was because of --include-outputs which makes the recursive derivation copying also include existing outputs (e.g., the final patch from the patch derivation). 03:59:43
@zhaofeng:zhaofeng.liZhaofeng LiMaybe such copying still occurs even when the "final" derivation can be substituted directly04:00:14
@winterqt:nixos.devWinter (she/her)Hmm, that's a good theory, but wouldn't that mean e.g. copying all source derivations for every installed package?04:00:38
@zhaofeng:zhaofeng.liZhaofeng Li Not sure, we could rig up a small test to try. It can be something like: a.drv -> b.drv -> c.drv, where everything can be substituted via some cache 04:02:40
1 Mar 2023
@fabianhjr:matrix.orgFabián Heredia changed their display name from fabianhjr to Fabián Heredia.06:05:18

Show newer messages


Back to Room ListRoom Version: 6