mirror of
https://github.com/hanzoai/zap.git
synced 2026-10-30 18:17:49 +00:00
zeekayandHanzo Dev
a87d24a09f
The previous commit took `zap-proto`, which was measured from a LOCAL
checkout of ~/work/zap/spec that is 26 commits behind on a stale
`master` branch. The remote (zap-proto/spec@main) tells a different
story:
- its crate is now `zap-schema` 1.0.0, lib name `zap`; it no longer
publishes `hanzo-zap` at all — that half of the collision was
already fixed upstream;
- `reserve/zap-proto/` is a DELIBERATE name reservation, published by
its own release.yml (`working-directory: reserve/zap-proto`,
`cargo publish`), whose description points at `zap-schema`.
So `zap-proto` is owned and published by another repo, and taking it
would have recreated the exact two-codebases-one-name bug this work
exists to remove. `zap-schema`'s lib is also imported as `zap`, so the
lib name had to move too or the two could never share a dep graph.
package: zap-proto -> zapwire lib: zap -> zapwire
version: 1.0.1 -> 1.0.0 (zapwire is a NEW name: first
publish, not a bump)
`zapwire` is what this crate is: the wire runtime — header, data
segment, call envelope, transport. Verified free on crates.io and
absent from spec's reserve/ set.
Nothing to do in zap-proto/spec: it fixed itself. Judging it from the
stale local checkout would have "retired" 26 commits of live work,
including a Rust `cap` capability runtime.
make test: 27 lib + 10 crosswire + 2 live-Go interop + 1 doctest, green.
Co-authored-by: Hanzo Dev <[email protected]>