The "guard against zombie callbacks" subtest passed its asserts but leaked async
activity after the test ended (node:test flags that as failure). Root cause: it
mocked modal.generateTorrentMetadata / resolveUploadIdentifier, but
handleVideoSelection drives the upload through this.mediaUploader.uploadVideo —
a DIFFERENT object with its own helpers — so the mediaUploader still ran real
createTorrentMetadata()/hashing on a synthetic {name:"video.mp4"} file ("invalid
input type" async) and captureVideoMetadata() ran URL.createObjectURL(file)
("must be an instance of Blob"). Not intermittent — a consistent wrong-boundary mock.
Fix: mock at the real DI boundary — modal.mediaUploader.uploadVideo = () =>
delayedUpload (controllable pending promise) — and stub captureVideoMetadata.
Hermetic, deterministic (3/3 exit 0, # pass 3). Mutation-verified: breaking the
videoUploadId zombie guard makes the test fail. Production code untouched. See
TEST_INTEGRITY.md.
NOTE: the runner is fail-fast, so the full suite surfaces pre-existing failures one
at a time. tests/unit/hashChangeHandler.test.mjs is next (consistent AssertionError
on "loads default view if hash is empty" — likely the same stale-default bug the
sibling tests/app/hash-change-handler.test.mjs already had when TRENDING became the
logged-out default). Enumerating full scope next.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Fully closes #11b's un-quarantine work: the QUARANTINE map is now EMPTY.
dm-block-filter.test.mjs (node:test) PASSED all subtests then hung, so the
node:test runner never exited. Diagnosed via process.getActiveResourcesInfo():
the "integration with loaded user block list" test's manager.loadBlocks() opens a
LIVE block-list subscription through nostrClient's real SimplePool aimed at the
test's fake relay URLs, leaking 4 TCP sockets + reconnect timers. It is NOT the
decrypt-retry timer (verified absent). Nulling nostrClient.pool did not help —
loadBlocks calls ensurePool() (userBlocks.js:967) to re-create+connect a real
pool when pool is falsy.
Fix (node:test-SAFE, no force-exit): make the test hermetic — (1) a TRUTHY stub
pool so ensurePool() is skipped, (2) a no-op getSubscriptionManager so
startBlockListSubscription opens nothing. The block-merge logic under test still
runs fully through the already-stubbed fetchListIncrementally. Restored in finally.
Verified BOTH directions: a clean run exits 0 (# pass 12), AND a deliberately
broken assertion still exits 1 (# fail 1) — failures are NOT masked. A force
process.exit in an after() hook was rejected precisely because it masked failures
(process.exitCode isn't set when the hook fires). See TEST_INTEGRITY.md.
NOTE: the full suite still has one SEPARATE, pre-existing red unrelated to #11b —
tests/ui/uploadModal-reset.test.mjs (TODO #11): its "zombie callbacks" subtest
feeds a plain object where webtorrent expects a Blob ("must be an instance of
Blob") and leaks async activity after the test ends. Untouched here; tracked
under #11 (harden flaky tests), not the #11b quarantine work.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Closes the #11b un-quarantine work: the three target files now pass and run in
CI. A full-suite run with a per-file timeout then surfaced a FOURTH, pre-existing
hang that had never been in the map (dm-block-filter) — quarantined with a triage
note rather than left to hang CI.
user-blocks.test.mjs — post-completion HANG. The decrypt-timeout section installs
a never-resolving nip44Decrypt whose background decrypt-retry timer outlives the
run (manager.reset() races it). Added the repo's established bare-assert
`setTimeout(() => process.exit(0), 50)` exit after the final assertion.
nostr-publish-rejection.test.mjs — post-completion HANG (was NOT in the map, so it
had been silently hanging local `npm run test:unit`, which has no per-file timeout
unless UNIT_TEST_TIMEOUT_MS is set). All assertions pass, then the imported app
stack (app.js + subscriptions + userBlocks) keeps the loop alive. Same
`process.exit(0)`-after-cleanup fix. (The dev-mode + NIP-71 spec fixes were
already in the file.)
nwc-client.test.mjs — (1) mock shadow: readToolkitFromScope prefers the bootstrap's
frozen canonical toolkit over window.NostrTools, so per-section mocks never took
effect and real @noble rejected the fake keys. Remove the scope canonical so the
resolver falls through to each mock. (2) spec correction: one section expected a
rejection when the toolkit "lacks nip44", but mergeWithCanonical ALWAYS backfills
the canonical nip44 (even explicit `nip44:null`) — nip44_v2 is never absent. The
wallet now advertises a genuinely-unsupported scheme (nip44_v3) to exercise the
same UNSUPPORTED_ENCRYPTION path, and asserts error.code too. See TEST_INTEGRITY.md.
dm-block-filter.test.mjs — QUARANTINED (pre-existing, node:test). All subtests pass,
then it hangs: manager.loadBlocks() opens a live block-list subscription through
nostrClient's real SimplePool pointed at the test's fake relay URLs, leaking TCP
sockets + reconnect timers that node:test can't drain. Needs a node:test-safe fix
(inject a mock pool so no real socket opens) — a force process.exit in after() is
UNSAFE (verified: process.exitCode isn't set yet, so it masks failures).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Triaged two more silently-excluded test files back into CI.
- nostr-boost-actions: STALE SPEC. Test required buildRepostEvent to throw
/missing-event-relay/ with no relay hint, but a relay hint is a NIP-18 SHOULD
(not MUST) — the builder degrades to a two-element `e` tag so a user can always
repost. Replaced with an exact assertion of the valid output (kind 16 + e/p/k).
- admin-list-store: STALE HARNESS (no production change, no assertion weakened).
Community curator lists are fetched via the batched SubscriptionManager
(cold-start relay-storm fix) the harness never mocked; the mock "hex" curator
values started with "npub" so parseCommunityBlacklistReferences decoded them
twice; and createListEvent omitted the `d` tag selectNewestEventsForReferences
matches on. Fixed all three to be faithful to production.
Removed both from the QUARANTINE map; TEST_INTEGRITY.md + TODO updated.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Triaged the monolithic tests/view-counter.test.mjs and got the whole file green,
then removed it from the QUARANTINE map:
- Pinned nostrClient.getSubscriptionManager() to null so the test exercises the
legacy per-pointer hydrate/subscribe path it mocks. The batched
SubscriptionManager path (added with the relay-storm work) was silently
shadowing the mocked legacy APIs.
- Made the dedupe-window test bucket-aligned (it used Date.now(), so the
+window/2 offset could straddle a fixed-bucket boundary and flake by time of
day).
- Corrected the hydrateVideoHistory root-timestamp assertions to the refactored
unit boundary: hydrateVideoHistory populates the rootCreatedAtByRoot MAP; the
video object's rootCreatedAt FIELD is applied by syncActiveVideoRootTimestamp.
reactionCounter.js audit (kind-7 likes): accurate, no fix needed. It does not use
NIP-45 COUNT; it lists kind-7 events and dedupes via applyReactionToState keyed
by pubkey (one reaction per user, latest-wins) with per-content totals — no
raw-count inflation. Kind-7 isn't replaceable, so listing+dedupe is the only
accurate model.
Spec corrections documented in TEST_INTEGRITY.md (SCN-view-counter-legacy-path,
SCN-hydrate-root-timestamp).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Record this session's fixes (delete/zombie, NIP-71 mirror teardown,
watch-history bucketing, the zap chain, the test-runner quarantine work)
and the test spec-corrections.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
On cold login the app opened WebSockets to a user's full NIP-65 read set
(in prod often ~20 relays, most dead). The flood of failing connections
churned the main thread and starved the nip-07 extension's postMessage
round-trips, so encrypted lists (blocks, subscriptions, hashtags) were
fetched but timed out during decryption ("message channel closed",
"Decryption timed out").
Bound every read/subscribe path to MAX_SUBSCRIBE_RELAYS=8 via a shared,
test-aware capReadRelays helper, applied at the chokepoints:
- client.js applyRelayPreferences (this.relays / this.readRelays)
- relayBatchFetcher.js fetchListIncrementally (blocks, subs, hashtags)
Writes stay uncapped (publish reach preserved).
capReadRelays is user-relays-first (a user's own relays are where their
data authoritatively lives, and small fast-path limits would otherwise
crowd them out — the regression that reverted the first attempt) while
reserving RESERVED_DEFAULT_RELAY_SLOTS=2 for reliable default aggregators
so the "~20 personal relays, all dead" case still reaches a live relay.
Under test the reserved core resolves to the test-relay override, keeping
harnesses isolated.
Tests: new relay-subscribe-cap coverage (bounded + reserved-defaults +
writes-uncapped); spec-corrected user-blocks, subscriptions-manager and
nostr/client fetch tests to bounded (not exact) relay fan-out while
keeping all behavioral/security assertions (per-relay coverage, correct
decrypt, block hydration, dedup) intact. See TEST_INTEGRITY.md and
KNOWN_BUGS #0/#4. Full unit + dm:unit suites green.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
performProfileFetch now issues ONE deduped, health-gated list() across all
relays via the L1 SubscriptionManager (the pool merges per-relay results),
replacing the manual per-relay fan-out. Newest-per-author selection unchanged;
legacy per-relay path retained as the manager-less fallback.
Spec correction: tests/app-batch-fetch-profiles.test.mjs asserted one query
PER relay (implementation detail); updated to assert one BATCHED query across
all relays. All behavioral assertions (cached hydration, newest profile
rendered, failing relay tolerated) unchanged; added a relays-targeting
assertion. Documented in TEST_INTEGRITY.md (SCN-profiles-batched).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
