release: cut v0.8.1 to finish what v0.8.0 published #49
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "land/repo/v0-8-1-prep"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Why
v0.8.0 published every object, rewrote both managers, reported activation, and
left
stablenaming v0.7.0. Its lanes came from Plumb v0.26.0, which shiftedthe managers and never advanced the channel pointer (PerishLab/plumb#346, fixed
in v0.27.0).
That release cannot be completed, and each reason is independent:
a stamped point never moves, so the version cannot be re-pointed at arepaired head;
plumb release activateneeds the capsule that run built, and this forge nolonger serves the artifact.
Rebuilding that capsule by hand would need Plumb v0.26.0 and every original
input reproduced byte for byte.
verify::publishedchecks a capsule againstitself, not against the published seal, so a near miss would write a pointer
naming a seal digest nobody can match, with no gate to catch it. That is the
one property worth protecting here, and the tooling does not protect it.
So the repair is a version, not a rescue.
What
crates/cli0.8.0 -> 0.8.1.crates/coreis unchanged and stays 0.7.0.docs/CHANGELOG/v0.8.1/{en,zh}/{INDEX.md,MIGRATION.md}. The migration notescarry no new demand and send the reader to v0.8.0's, because anyone tracking
stablemoves from v0.7.0 straight to v0.8.1 and still owes that migration —the channel never named v0.8.0.
v0.8.0's objects stay published and sealed. An exact seal is create-only, and a
channel that never moved pointed nobody at them.
Tests
cargo fmt --check,clippy -D warnings,cargo check --release,cargo test --workspace,pnpm biome ci,pnpm -r test,ectropy .,plumb doctor .— all clean.Compatibility
None beyond v0.8.0's, which this release inherits in full.
88932b1badae25c64809