fix(replies): address PR #59 review — feed reply count, route-by-type, tests, docs#60
Merged
Merged
Conversation
…, tests, docs Follow-up to #59 (merged). Addresses the should-fix findings from the multi-agent review. Web (stores/freenet.ts): - Feed reply-count badge now updates. onRepliesUpdated already carries the authoritative thread-shard reply list, so it also sets post.replies = replies.length on the feed post (same cadence as like/quote aggregates). Previously contractPostToUiPost hardcoded replies: 0 and nothing ever updated it — the headline #12 surface read blank. - Route delegate responses by the authoritative response type: SignedReply -> completeReply, Signed -> completePublish. Replaces the try-completeReply-then-fall-back-to-completePublish probing, which logged a misleading "no matching pending draft" for a genuinely-failed reply. Tests: - web vitest (freenet-api.test.ts): cover the new reply path — dropPendingReply nonce matching, completeReply success (Replies delta + refresh), completeReply foreign-nonce no-op, and onRepliesUpdated emission sorted oldest-first. Mirrors the existing like/quote coverage. 44 pass. - node-e2e (reply-thread.spec.ts): the cross-session reload test exercises the #50 single-node reload-GET seam, which is not guaranteed on a single node. Gated behind test.fixme(!E2E_RUN_CROSS_SESSION) so it no longer redden CI; set E2E_RUN_CROSS_SESSION=1 to enforce it. Same-session test stays a hard gate. Title de-hedged; TODO(#50) for marker-based post targeting. Docs: - README roadmap: replies checked off; SignReply added to the delegate message list; stale "reply signing not yet wired" prose removed. - ADR-0001 Phase 4: document the SignReply variant + byte-identical SignPost. Gate: clippy -D, fmt-check, tsc, vitest (44) all green; node-e2e same-session reply passes (cross-session skipped by design). Claude-Session: https://claude.ai/code/session_01Q6rfcR7TYVpMCTELqr6kWv
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.
What
Follow-up to #59 (merged). Addresses the should-fix findings from the multi-agent
/pr-reviewposted on #59.Fixes
Feed reply-count badge never updated (was the headline #12 surface — read blank)
contractPostToUiPosthardcodedreplies: 0and nothing updated it.onRepliesUpdatedalready carries the authoritative thread-shard reply list, so the store handler now also setspost.replies = replies.lengthon the feed post — same cadence as the like/quote aggregates (refreshes when that post's thread shard is read).Route delegate responses by response type (
web/src/stores/freenet.ts)SignedReply → completeReply,Signed → completePublish. Replaces the try-completeReply-then-fall-back-to-completePublish probing, which logged a misleading "no matching pending draft" when a reply genuinely failed (thefalsereturn was overloaded: "not my nonce" vs "was my reply but the update threw"). The delegate already tags the response authoritatively.Web vitest for the reply path (
web/src/freenet-api.test.ts, 4 new tests, 44 pass)dropPendingReplynonce matching ·completeReplysuccess (Replies delta + refresh) ·completeReplyforeign-nonce no-op ·onRepliesUpdatedemission sorted oldest-first. Mirrors the existing like/quote coverage that previously had no reply analog.Cross-session E2E no longer silently reddens CI (
web/tests/node-e2e/reply-thread.spec.ts)The reload test exercises the #50 single-node reload-GET seam, which isn't guaranteed on a single node. Gated behind
test.fixme(!process.env.E2E_RUN_CROSS_SESSION)— skipped in normal CI, runnable withE2E_RUN_CROSS_SESSION=1. The same-session test stays a hard gate. Title de-hedged;TODO(#50)left for marker-based post targeting.Docs
README.md: replies checked off the roadmap;SignReplyadded to the delegate message list; stale "reply signing not yet wired" prose removed.docs/adr/0001-implementation-notes.mdPhase 4: document theSignReplyvariant + byte-identicalSignPost.Not in scope
git rm --cached web/dist/index.html— the tracked-but-gitignored release snapshot the review flagged. Deferred (touches the release/verify-published-contract path; warrants its own change).thread_root/ nested threads — Threads / replies: schema, validation, thread view, reply compose #12 stays open until that + this land.Verification
cargo make clippy(-D warnings) ✅ ·fmt-check✅ ·tsc --noEmit✅vitest— 44 pass (incl. 4 new reply tests) ✅cargo make test-ui-node-e2e— same-session reply passes; cross-session skipped by design; the only flaky spec is the pre-existing live-node Discover round-trip (Node-backed E2E test tier (browser ↔ live node ↔ delegate ↔ contract) #50), unrelated.Addresses review feedback on #12.
https://claude.ai/code/session_01Q6rfcR7TYVpMCTELqr6kWv