Skip to content

docs: close spec gaps (#45/#46) and correct WHITEOUT's Sepolia status (#44) - #53

Merged
JamesCarnley merged 2 commits into
mainfrom
claude/v1-spec-whiteout-status
Nov 12, 2026
Merged

JamesCarnley merged 2 commits into
mainfrom
claude/v1-spec-whiteout-status

Conversation

@JamesCarnley

@JamesCarnley JamesCarnley commented Oct 27, 2026 •

Copy link
Copy Markdown
Member

Summary

  • specs/02: adds the fifth anchor-name reject rule the contract already enforces (a needless escape such as %41 for A), plus a note that NFC is checked on the decoded name, not the escaped string.
  • specs/09: pins two readings the §8 pseudocode already implied and the SDK already ships: the lowest-UID tie-break picks from all of an attester's redirects of any kind (§5), and D_MAX bounds one follower walk, not the whole path (§3). Adds conformance vectors 13 and 14.
  • WHITEOUT: current-state docs now say "implemented in source, not deployed on Sepolia", backed by a recorded readback in docs/CHAINS.md. ADR-0055/0067 get Status-line notes only; their bodies are unedited.
  • Indexer (second commit): the live Sepolia EFSIndexer predates ADR-0068 and ADR-0066: MAX_ANCHOR_DEPTH() returns 32 (main: 256), and index() still propagates folder presence (main: discovery-only). CHAINS.md records the readback; specs/03, FUTURE_WORK and the ADR-0066/0068 Status lines are chain-qualified. It also records that the indexer must be upgraded before WHITEOUT is deployed.

Why

v1 stays live on Sepolia with no v2 date, so its specs have to match the deployed contracts.

Sepolia readback (block 12007898):

Check Result
WhiteoutResolver Safe-keyed CREATE3 address 0xE007…eBFa (same derivation reproduces EdgeResolver's 0xD664…7F1) no code
SchemaRegistry getSchema(0xfccc…541d) (the UID that deploy would register; formula checked against DATA's UID) not registered
whiteoutResolver() on the current and previous EFSRouter / EFSFileView reverts (views predate WHITEOUT wiring)
deployments/sepolia/WhiteoutResolver.json, Sepolia entry in deployedContracts.ts absent

No client code references WHITEOUT, so nothing infers it from the ABI. Deploying it stays a separate owner decision.

Permanence tier

Durable (spec/ADR prose). No contract, schema, deploy, or client behavior changes.

Specs / ADRs touched or checked

  • specs/02-Data-Models-and-Schemas.md (updated)
  • specs/09-redirect-resolution.md (updated; vectors 13–14 added)
  • specs/overview.md, specs/04-Core-Workflows.md (updated: WHITEOUT status)
  • docs/CHAINS.md, docs/SEPOLIA_FREEZE_TABLE.md, docs/DEPLOYMENT.md, docs/FUTURE_WORK.md, docs/decisions.md (updated)
  • ADR-0055, ADR-0067 (Status-line note per docs/adr/README.md amendment pattern; Decision bodies unedited)

Test plan

  • Docs only; no build or test impact.
  • Contract NatSpec (EFSIndexer._isValidAnchorName) checked against the new specs/02 bullet.
  • NFC examples and the encodeURIComponent claim checked with Python unicodedata and Node.
  • specs/09 picks checked against the SDK on main (reads/redirects.ts selectLensRedirect, reads/file.ts per-landed-anchor budget).
  • Sepolia readback above run with cast against publicnode.

Agents involved

  • Dev: Claude Opus 5.5
  • Review: none yet
  • Orchestrator: James

Fixes #44
Fixes #45
Fixes #46

🤖 Generated with Claude Code

specs/02 listed four anchor-name reject rules, but
EFSIndexer._isValidAnchorName also rejects a needless escape such as
%41 for "A". A client built from the prose alone (or on
encodeURIComponent) would mint names the resolver reverts. Add the
fifth rule, and note that NFC is checked on the decoded name.

specs/09 left two choices implicit that the SDK's conformant reader
had to make. Pin both to what the section 8 pseudocode already does:
the lowest-UID tie-break picks from all of an attester's redirects of
any kind, and D_MAX bounds one follower walk, not the whole path. Add
conformance vectors 13 and 14 for them.

Several docs called WHITEOUT "live" or "shipped". A Sepolia readback
(block 12007898) shows it is implemented in source only: no code at
the Safe-keyed CREATE3 address (the same derivation reproduces
EdgeResolver's address), no registered schema at the UID that deploy
would produce, and whiteoutResolver() reverts on every live router
and file view. Record that in CHAINS.md and make the current-state
docs say so. ADR bodies stay unedited; ADR-0055 and ADR-0067 get
Status-line notes per the docs/adr/README.md amendment pattern.

Docs only. No contract, schema, deploy, or client behavior changes.

Fixes: #44, #45, #46
Permanence-tier: Durable
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 27, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-27T22:43:30.156878Z 3f4e508 New commits
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

Specs and ADRs described the indexer on main as current, but the live
Sepolia EFSIndexer proxy still runs the freeze-time implementation:
MAX_ANCHOR_DEPTH() returns 32 (main: 256, ADR-0068), and index() still
propagates folder presence (main: discovery-only, ADR-0066). The
upgrade carrying both was merged on 2026-06-23/25 but never executed.
These are the only two indexer changes since the freeze.

Add the readback to CHAINS.md (with the storage-layout and ordering
notes the eventual upgrade needs: indexer before WHITEOUT, or index()
on a WHITEOUT could mark a folder present), and chain-qualify specs/03,
FUTURE_WORK and the ADR-0066/0068 Status lines. Docs only.

Refs: #44
Permanence-tier: Durable
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
@JamesCarnley
JamesCarnley merged commit 458fefa into main Nov 12, 2026
3 checks passed
JamesCarnley added a commit that referenced this pull request Nov 12, 2026
Resolve the two overlaps with #53: keep both new decisions.md entries,
and in specs/overview.md read-flow keep this branch's step 3 (segment
re-encoding) with main's step 4 (WHITEOUT terminal not yet live on
Sepolia).

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

2 participants