release: v1.3.0 — honest partial-emit surfaces (wardline-7924d67d3b)#89
Merged
Conversation
…cally only" Review P2 pair: 89563f1 taught the producers to preserve partial-emit truth (EmitResult.chunks_landed + accumulated counts/warnings on mid-batch failures; WriteResult.written preserved on mid-batch outage/403), but the CLI human rendering still branched on binary `reachable` and only rendered counts in the reachable branch — so a mid-emit transport drop printed "could not reach Filigree ... findings written locally only" after Filigree had already ingested chunks and run mark_unseen, and a partial Loomweave write printed "taint store not written" over committed facts. The unreachable branches now discriminate partial evidence (chunks_landed for Filigree, written for Loomweave — deliberate skips always carry written=0) and render the counts, failed tally, and producer warnings in every soft-failure sub-branch (auth / server error / transport). The transport case names "connection dropped mid-emit" rather than the first-contact "could not reach". Machine surfaces (agent-summary/MCP/ scan-job blocks) already carried this truth; the operator line no longer contradicts them (PDR-0023 at the human surface). Glossary line anchors for the scan summary/gate lines shifted +38 with the edit; _ANCHORS and finding-lifecycle-vocabulary.md citations updated. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…d3b residuals) filigree_disabled_reason now takes chunks_landed and stops flattening a mid-batch failure to "filigree unreachable" — the string scan_jobs persists verbatim as the terminal job `error` field. The transport rung becomes "connection dropped mid-emit (K chunk(s) landed before the failure)"; the 401/403/5xx rungs keep their diagnosis and append the landed-chunk truth. The result-based status builder threads EmitResult.chunks_landed through, so the CLI agent-summary, scan-job status.json, and scan_file_findings surfaces inherit the honest string; the MCP from-block path is unchanged (its block already carries the partial story in failures/warnings, and its golden bytes stay frozen). Zero-chunk ladder bytes are byte-identical. Also closes the coverage gap the issue named: chunks_landed is now asserted on the mid-batch 401/403/5xx parameterization, not only the transport-drop path. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…t partial-emit surfaces Minor bump: adds concurrent-writer detection (WLN-ENGINE-TREE-CHANGED-DURING-SCAN) and hardens every credential-bearing transport against redirects (fail-closed WeftRedirectError; see [1.3.0] Security). Fixes ship the PDR-0023 partial-emit honesty through the human surfaces (CLI stderr + scan-job error field, wardline-7924d67d3b), MCP boolean-arg coercion, scan-job terminal-status monotonicity, and the zero-files vacuous-PASSED gate. Also repairs the Keep-a-Changelog compare links (stale since v1.0.6). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Both can alter a previously-"green" run: an endpoint behind a 3xx now fails loudly instead of silently emitting credentials cross-origin as a GET, and a signing path with unreadable git state refuses instead of coercing to clean. Also retitles the v1.2 section now that 1.2.0 has shipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…y strict) 89563f1's arg-hardening test helper is annotated, so strict mypy checks its body — dispatch() returns dict | None (None only for id-less notifications) and the bare return failed [return-value]. The helper always sends id=1, so assert the invariant and return the narrowed dict. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: ba85f28390
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".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| status=result.status, | ||
| token_sent=result.token_sent, | ||
| url=result.url, | ||
| chunks_landed=result.chunks_landed, |
There was a problem hiding this comment.
This only passes chunks_landed on the result-based status path; the MCP scan path still builds a raw block that drops er.chunks_landed (src/wardline/mcp/server.py:116-134), and filigree_emit_status_from_block() then calls filigree_disabled_reason() with the default zero. In an MCP scan where the Filigree connection drops after one or more chunks landed, the response can show nonzero created/updated counts and partial failures but still report disabled_reason: "filigree unreachable ...", recreating the misdiagnosis this change is meant to remove for the agent-facing surface.
Useful? React with 👍 / 👎.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Cuts v1.3.0 from the consolidation line. The consolidation content itself (redirect hardening, concurrent-writer detection, MCP strict args) is already on main; this PR carries the final review-driven fixes plus the version bump:
Fixes (wardline-7924d67d3b — both P2 review findings)
connection dropped mid-emitinstead of the first-contactcould not reach Filigree.taint store partially written … N taint fact(s) committed before the failureinstead of the flatnot written.filigree_disabled_reasonis partial-aware (chunks_landedkwarg): the string scan-jobs persists as the terminal joberrorno longer flattens a mid-batch failure tofiligree unreachable. Zero-chunk ladder bytes unchanged; MCP from-block goldens stay frozen.chunks_landedasserted on the mid-batch 401/403/5xx test parameterization; glossary line anchors re-pinned after the CLI edit.Release
_version.py→ 1.3.0; CHANGELOG[Unreleased]rolled into[1.3.0] - 2026-07-03; Keep-a-Changelog compare links repaired (stale since v1.0.6).[1.3.0]adds concurrent-writer detection and the never-follow-redirects federation hardening.Verification: full suite 4664 passed / 1 skipped; ruff + format + mypy clean;
wardline scan . --fail-on ERRORgate PASSED. Tagv1.3.0will be placed on the merge commit to trigger the PyPI publish workflow.🤖 Generated with Claude Code