Skip to content

KEP-0004: record Phase 2 as shipped - #55

Merged
baijum merged 2 commits into
mainfrom
claude/kep-0004-phase2-shipped-stacked
Aug 28, 2026
Merged

KEP-0004: record Phase 2 as shipped#55
baijum merged 2 commits into
mainfrom
claude/kep-0004-phase2-shipped-stacked

Conversation

@baijum

@baijum baijum commented Aug 28, 2026

Copy link
Copy Markdown
Member

Stacked on #53 — review that one first; this PR's diff is against its branch, and merging #53 will retarget this to main.

#53 audited the implementation plan and found Phase 2 in an odd state: the gate had cleared on 2026-07-14 (both #1487 and #1489 closed, KEP-0002 fully shipped in v0.15.0/v0.16.0), but nobody had written the identifier — "this KEP's one unblocked open work item". kaappi#2402 writes it. This PR updates the KEP to match.

Changes on top of #53

  • Phase 2 flips from "gate cleared; identifier still unshipped" to shipped, keeping all of KEP-0004: record the cleared Phase 2 gate and resolve UQ1/UQ2 #53's dating and context. The new material is what landed and why it is placed where it is: the identifier sits on the non-wasm branch of platform_features beside kaappi-threads, not in the always-on base array, because promotion requires OS threads and §5 of this KEP already records that notify is a no-op on wasm32-wasi that nothing ever calls. Test coverage follows Phase 1's pattern, with a note that features.sh derives its checks from the built binary — so the three surfaces are held in agreement by construction rather than by hand-kept lists.
  • Status field drops "but identifier unshipped".
  • Unresolved question 2KEP-0004: record the cleared Phase 2 gate and resolve UQ1/UQ2 #53 resolved it as moot, since both candidate gates had passed. That is right, and this sharpens why: what actually set the date was neither candidate but the Implementation plan's own stricter phrasing, "Phase 3 on main with its review findings resolved". Phase 3 merged while #1489 was still open, so the findings, not the merge, are what the identifier waited on. Worth recording, because the stricter phrasing is the one that behaved correctly.

Phases 3 and 5 untouched — KEP-0003 is still Draft, the runtime-variance predicates still unstarted.

Companions: kaappi#2402, kaappi.github.io#39.

🤖 Generated with Claude Code

baijum and others added 2 commits August 28, 2026 09:53
Stacked on the amendment in #53, which found the gate cleared and the
identifier unwritten. kaappi#2402 writes it, so Phase 2 flips from the
open work item that PR identified to shipped: the non-wasm placement and
its KEP-0002 §5 rationale, the test coverage, and the docs row that no
longer has to describe a gap.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Baiju Muthukadan <baiju.m.mail@gmail.com>
…2-shipped-stacked

# Conflicts:
#	keps/0004-discoverable-deviations.md
@baijum
baijum changed the base branch from claude/kep-0004-review-b98651 to main August 28, 2026 06:08
@baijum
baijum merged commit d905faf into main Aug 28, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

1 participant