Cachix #459
Cachix #459
Conversation
|
Thanks for the suggestions @cole-h. I've marked this ready for review now ;) |
|
cc @grahamc |
Co-Authored-By: Cole Helbling <cole.e.helbling@outlook.com>
Co-Authored-By: Cole Helbling <cole.e.helbling@outlook.com>
| info!("Done!"); | ||
| } | ||
| } | ||
|
|
||
| fn cache_push(cfg: &CachixConfig, log: Vec<String>) -> bool { | ||
| if !log.is_empty() { | ||
| let store_path = log.last().unwrap(); |
There was a problem hiding this comment.
Something I just thought of that should probably be addressed: what about when there were multiple builds, e.g. here: https://logs.nix.ci/?key=nixos/nixpkgs.82982&attempt_id=62265253-63b3-44e2-aa81-3fb8464be6ff
/nix/store/vb15nidh4wbsg9lrf93iv859mvkbn0pl-ripgrep-12.0.1
/nix/store/vywixxgrbrq76nbdra96h3nzrlnagq5m-zcash-2.1.1-1
/nix/store/v9c8cyl9rrspkv18w0f6b56ydb0350gf-lsd-0.17.0
In this case, I believe only the lsd path will be pushed to the cache.
There was a problem hiding this comment.
Good point! We can either call cache_push for each path or pipe them to Cachix via stdin. Just need to identify all the paths - maybe something like get all the last few consecutive lines matching /nix/store prefix?
There was a problem hiding this comment.
Or, get the last n lines, where n = number of built attrs? We gate this on a successful build anyways, so this should be a safe bet (assuming that any failing build sets status to be not-BuildStatus::Success -- I'll have to read through the source again to be entirely certain).
Or both, as sanity checks for the other?
I'll do some brainstorming.
|
I decided I don't care. |
|
It would be great if binaries were pushed and available to users for testing. I'm happy to sponsor a cache for ofborg if someone manages to get integration working. |
I have no idea what I'm doing.