| 4 Apr 2024 |
 jrick | i'm less sure about that PR now, doesn't seem like it would change what gets placed in the proxied module cache | 20:48:59 |
| 5 Apr 2024 |
 Paul Meyer (katexochen) | In reply to @jrick:zettaport.com
i'm less sure about that PR now, doesn't seem like it would change what gets placed in the proxied module cache Shouldn't the module cache just contain the dependencies of the main module anyway? The builder is running go mod download in $modRoot, so it (correctly) won't download deps of submodules. | 08:19:01 |
|  Frédéric Christ joined the room. | 10:57:51 |
| jrick | yeah it's fine enough to build, just happens to break go list -m all in this case if you run that from e.g. checkPhase. i'm a bit confused why it is copying all of the other files for those modules, but lacks the .info files | 13:03:38 |
| Frédéric Christ | Hi,
any successful experiences in building go modules with dependencies in private repositories?
buildGoModule obviously does not consider my local SSH configuration, so the build fails.
There's also a post in our discourse about that but their intermediate solution is go mod vendor. | 13:43:27 |
 qbit | Frédéric Christ: vendor will for sure be the least friction | 14:29:53 |
| qbit | you can add some stuff to nix.conf.. but that's not very reproducible across a team | 14:30:11 |
| qbit | look for access-tokens in man nix.conf | 14:30:53 |
| Frédéric Christ | In reply to @qbit:tapenet.org
you can add some stuff to nix.conf.. but that's not very reproducible across a team That would be a problem. I'd favour a solution that works on dev machines as well as in pipelines. | 14:43:44 |
| qbit | that'd be vendor afaik | 14:44:08 |
| Frédéric Christ | For rust, I think crane solved that very well. Haven't had any issues with private repositories there yet. | 15:27:00 |
| Frédéric Christ | * qbit: For rust, I think crane solved that very well. Haven't had any issues with private repositories there yet. | 15:27:06 |
| qbit | i think there are other tools for go to do similar.. but I am not familiar with them | 15:27:48 |
| Frédéric Christ | There's gomod2nix but it seems a bit unmaintained. | 15:30:23 |
 Diamond (it/she) | I believe it is possible to enforce ssh info via env vars but that's impure | 15:30:35 |
| Diamond (it/she) | In reply to @frederic:scs.ems.host
There's gomod2nix but it seems a bit unmaintained. I don't think this would've helped, it's really just another way to checksum dependencies and has less to do with fetching them somehow | 15:31:03 |
| Diamond (it/she) | perhaps this could be useful: https://discourse.nixos.org/t/git-buildgomodule-private-repositories/5167/7?u=diamondburned | 15:32:29 |
| Diamond (it/she) | I opened this Discourse back in 2019 | 15:32:42 |
| Diamond (it/she) | unrelated but kinda crazy to think I would see tomberek like 4 years after that without even remembering I made this post lol | 15:33:24 |
|  Alex Ou joined the room. | 17:53:37 |
| jrick | https://github.com/NixOS/nixpkgs/pull/301373 why was this merged directly into staging branch instead of master? | 18:18:51 |
| jrick | security policy maybe? | 18:19:48 |
 Sumner Evans | I think to avoid massive amounts of rebuilds | 18:21:16 |
| jrick | i see some other go updates were merged to master first, and i think staging is preferably used over master to reduce rebuilds | 18:35:23 |
| qbit | it all depends on the number of rebuilds | 19:36:20 |
| qbit | some of the other updates are for non-default versions of go.. so they only have like 10 things dependent on them | 19:36:38 |
| 7 Apr 2024 |
|  SYMYƧ changed their display name from symys to symyƨ. | 07:58:17 |
| SYMYƧ changed their display name from symyƨ to sγmγƨ. | 08:01:04 |
| SYMYƧ changed their display name from sγmγƨ to SYMYƧ. | 08:08:40 |
| 11 Apr 2024 |
|  @withoutwithin:matrix.org joined the room. | 21:26:50 |
