Skip to content

Record decision 32: resolve #54 (independent versioning, computed-at-cutover target)#56

Merged
enricopiovesan merged 1 commit into
mainfrom
claude/decision-log-32-version-strategy
Jul 18, 2026
Merged

Record decision 32: resolve #54 (independent versioning, computed-at-cutover target)#56
enricopiovesan merged 1 commit into
mainfrom
claude/decision-log-32-version-strategy

Conversation

@enricopiovesan

Copy link
Copy Markdown
Contributor

Summary

/brainstorm resolution of #54 (the traverse-registry 0.8.1 collision
recorded in decision-log entry 31). Key outcomes:

  • traverse-registry goes independently versioned post-extraction
    (not lockstep with traverse's workspace version).
  • The real extraction's publish version is computed at Extract traverse-registry crate + governing specs from traverse into this repo #9's cutover
    execution time
    (next minor above whatever's actually published then),
    not hardcoded as 0.8.0 — a hardcoded number today would risk the same
    staleness that just invalidated 0.8.0, since traverse keeps
    publishing traverse-registry releases until cutover.
  • traverse#627's existing "remove traverse-registry from
    publish_crates.sh atomically with the cutover PR" plan (decision 29e)
    is reaffirmed, not reopened — no standalone early fix was requested.
  • This repo's 0.0.1 placeholder stays un-yanked (harmless, no benefit to
    touching it).
  • No new/amending spec for spec 010's stale FR-006 — the decision log
    plus DoD text in Extract traverse-registry crate + governing specs from traverse into this repo #9/traverse#627 carry the correction; the substantive
    requirement (unchanged pipeline) still holds, only the illustrative
    number was wrong.

Full six-question brainstorm transcript (including a mid-session
self-correction after reading traverse#627's already-independent
discovery of the same collision) is in the entry itself.

Governing Spec

no-spec-needed (decision-log entry, not a governing spec change)

Project Item

Definition of Done

  • Decision 32 appended to docs/decision-log.md in the file's existing numbered style

Validation

  • Manual review: entry follows the same structure/density as decisions 26–31

…cutover target)

/brainstorm resolution of the traverse-registry 0.8.1 collision (entry 31):
independent versioning going forward, real-publish version computed at #9's
cutover execution time rather than hardcoded, traverse#627's atomic-removal
staging (decision 29e) reaffirmed unchanged, 0.0.1 placeholder left
un-yanked, no new spec drafted for stale FR-006.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
@enricopiovesan
enricopiovesan enabled auto-merge (squash) July 18, 2026 15:46
@enricopiovesan
enricopiovesan merged commit d4aeb2c into main Jul 18, 2026
6 checks passed
@enricopiovesan
enricopiovesan deleted the claude/decision-log-32-version-strategy branch July 18, 2026 15:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Decide traverse-registry version-number strategy for #9 — 0.8.0 target invalidated by traverse's own 0.8.1 publish

1 participant