Hi, I’m new to Julia package development.
I’m interested in developing a package similar to Bash’s command-not-found, but for Julia. In my initial design, the package would include a platform-independent dataset stored in either an SQLite or compressed TOML file. Ideally, this binary file would be mirrored by the official Julia storage server to improve accessibility, as GitHub may be slow in some regions.
I’m wondering if an artifact file—when associated with a package registered in the General Registry but not part of an Yggdrasil-built JLL package—would be stored on the Julia Storage Server (accessible via https://pkg.julialang.org/artifacts/xxxx).
So far, I haven’t found any documentation explicitly describing the relationship between the Julia Artifact system and the storage server. However, I did find some clues in a comment from a Yggdrasil member, which states that “JLL packages created with BinaryBuilder don’t provide any advantage compared to having an Artifacts.toml file.” This makes me suspect that artifact files might be uploaded to the storage server when a package is newly registered or updated in the General Registry.
Since I can’t register an incomplete package just for testing, I’d appreciate it if someone could clarify this.