fix(release): correct lock-sync comments and gate stale lockfile #29
Loading…
Reference in a new issue
No description provided.
Delete branch "fix/FJMCP-18-release-lock-comments"
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?
What
Corrects the
create-releaserecipe comments imported by PR #28 (a sync from the multi-crateforgejo-cli) and adds a fail-fast lockfile gate.Why
PR #28 copied
forgejo-cli's recipe wording into forgejo-mcp. forgejo-mcp is a single crate, so the imported comments were false: there are no "three workspace members moving in lockstep", the version lives in the rootCargo.toml[package]table, and the recipe usescargo update --offline --package forgejo-mcp --precise, notcargo update --workspace. The comments now match the code and reference FJMCP-18.Functional state (confirmation)
The lock sync itself already worked. Release commit
0c58328(v0.2.2), cut after the lock-sync fix landed in1962ad4, includesCargo.lockalongsideCargo.toml. The release that omitted the lock was v0.2.1 (752ccfe), cut before the fix. FJMCP-18 itself was the v0.2.0 tag-build failure (cargo build --release --lockedrejecting an out-of-sync lock).New gate
After the version bump,
cargo metadata --offline --lockedruns and aborts the release ifCargo.lockis still out of sync with the bumped manifest. This mirrors CI'scargo build --release --lockedso the FJMCP-18 failure surfaces at release-creation time (loud, local) instead of at tag-build time (silent until CI). Verified both directions: passes on a synced lock, blocks on a deliberately desynced one.just pre-commitpasses (fmt, clippy, build, test).