!VRULIdgoKmKPzJZzjj:nixos.org

Nix Hackers

899 Members
For people hacking on the Nix package manager itself188 Servers

Load older messages


SenderMessageTime
4 Aug 2021
@fzakaria:matrix.orgfzakaria joined the room.21:34:07
@fzakaria:matrix.orgfzakariaJust noticed this channel. Wouldn't mind discussing https://github.com/NixOS/nix/issues/5092 with someone that might have some insight21:34:24
@fzakaria:matrix.orgfzakariaTrying to get some context to work on a fix21:34:37
5 Aug 2021
@Ericson2314:matrix.orgJohn Ericsondoes the store-specific options not get put in the manual yet?17:38:15
@Ericson2314:matrix.orgJohn Ericson * do the store-specific options not get put in the manual yet?17:38:20
@theophane:hufschmitt.netRegnat No, I don’t think they are. They are only available with nix describe-stores 20:32:40
6 Aug 2021
@cw:kernelpanic.cafecw (? days since last shower) changed their display name from continuouswave to cw (? days since last shower).18:10:47
@roberthensing:matrix.orgRobert Hensing (roberth) changed their display name from roberth to Robert Hensing (roberth).19:46:50
@kity:kity.wtfash (it/its) 🏳️‍⚧️ changed their display name from ashkitten (it/its) 🏳️‍⚧️ to kity.21:02:52
@kity:kity.wtfash (it/its) 🏳️‍⚧️ changed their display name from kity to ash (it/its).21:03:40
@kity:kity.wtfash (it/its) 🏳️‍⚧️ changed their display name from ash (it/its) to ash (it/its) 🏳️‍⚧️.21:04:53
@fzakaria:matrix.orgfzakariaSmall fix for the issue i filed above22:37:17
@fzakaria:matrix.orgfzakariahttps://github.com/NixOS/nix/pull/510122:37:18
@tomberek:matrix.orgtomberek
In reply to @theophane:hufschmitt.net
No, I don’t think they are. They are only available with nix describe-stores
I’ve always found myself going through the code to find them. Wasn’t aware of this command. Thanks!
23:00:41
7 Aug 2021
@balsoft:balsoft.rubalsoft https://github.com/NixOS/nix/pull/5102 is a very simple change, but it empowers us to use nixConfig.flake-registry for automatically keeping input versions up to date. It should be safe. 11:44:33
@tomberek:matrix.orgtomberek balsoft: this changes the concept of a registry a bit. How would this work when a flake depends on another? or "follows"? What if they depend on different flakes? 18:31:20
@tomberek:matrix.orgtomberek * balsoft: this changes the concept of a registry a bit. How would this work when a flake depends on another? or "follows"? What if they depend on different registries? 18:31:50
@balsoft:balsoft.rubalsoft
In reply to @tomberek:matrix.org
balsoft: this changes the concept of a registry a bit. How would this work when a flake depends on another? or "follows"? What if they depend on different registries?
No, it doesn't change the concept of anything. Only top-level flake registry matters
19:57:31
@balsoft:balsoft.rubalsoft * Only top-level flake nixConfig gets applied 19:57:39
@balsoft:balsoft.rubalsoft * Only top-level flake nixConfig gets applied 19:57:58
@tomberek:matrix.orgtombereki see. This means any other input flakes would use the overridden registry?19:59:38
@balsoft:balsoft.rubalsoft
In reply to @tomberek:matrix.org
i see. This means any other input flakes would use the overridden registry?
Other flakes won't use the registry at all, unless they are missing a lockfile.
20:01:47
@balsoft:balsoft.rubalsoft In which case yes, they would be using that registry when you update the top-level flake. 20:02:04
@balsoft:balsoft.rubalsoft Besides, this PR doesn't actually change any of this behavior, that was done in https://github.com/NixOS/nix/pull/4969 20:03:04
@balsoft:balsoft.rubalsoft This just whitelists nixConfig.flake-registry so that it can be used in non-interactive settings, e.g. an automatic update bot (https://github.com/serokell/update-daemon) 20:03:41
@balsoft:balsoft.rubalsoft Currently we hack this by adding .local/share/nix/trusted-settings.json with the relevant value, but since flake-registry is safe to set we can avoid that. 20:04:36
@tomberek:matrix.orgtomberek okay, then the PR seems fine, but the underlying semantics are what I'm trying to understand. At first glance it seems an input adding or removing a .lock might be unexpected. Can you share about what the right mental model is for this? 20:06:23
@balsoft:balsoft.rubalsoft It is safe to set because it should be exactly the same as just setting inputs in the flake. 20:06:20
@balsoft:balsoft.rubalsoft
In reply to @tomberek:matrix.org
okay, then the PR seems fine, but the underlying semantics are what I'm trying to understand. At first glance it seems an input adding or removing a .lock might be unexpected. Can you share about what the right mental model is for this?
I'm not sure I understand the question.
20:06:49
@balsoft:balsoft.rubalsoft The flake registry is not the source of truth for most flakes, flake.lock is. 20:07:13

Show newer messages


Back to Room ListRoom Version: 6