Changes `loadBlocks` to fetch `kind: 10000` with `d=user-blocks` instead of `kind: 30002` for block lists, while keeping the generic `kind: 10000` fetch for mute lists.
Updates `applyEvents` to separate block lists (tagged or legacy 30002) from mute lists (untagged 10000) and merge them instead of prioritizing one over the other.
Updates tests to reflect the merged behavior.
Co-authored-by: PR0M3TH3AN <[email protected]>
- Updated `NostrClient.fetchListIncrementally` to prioritize user's NIP-65 read relays, cap concurrent requests, and temporarily mark failing relays as unreachable.
- Added `markRelayUnreachable` to `NostrClient` for temporary relay exclusion.
- Updated `SubscriptionsManager`, `HashtagPreferencesService`, and `UserBlockListManager` to use `relayManager.getReadRelayUrls()` for selecting relays, ensuring consistency with user preferences.
- Added structured logging for relay selection in `fetchListIncrementally`.
- Updated unit tests to mock `relayManager` behavior correctly.
Co-authored-by: PR0M3TH3AN <[email protected]>
- Prioritize Kind 10000 events with `d=user-blocks` over plain mute lists in `loadBlocks`.
- Add `isTaggedBlockListEvent` helper for differentiation.
- Add regression test case to `tests/user-blocks.test.mjs` verifying that tagged lists override newer plain lists.
This commit modifies `js/userBlocks.js` to correctly parse Kind 10000 events that have empty content but populated `p` tags (used by clients like Primal). Previously, these events were treated as empty block lists, causing the user's block list to be cleared.
Changes:
- Refactored `loadBlocks` to check for empty content and fallback to extracting pubkeys from tags using `extractPubkeysFromTags`.
- Updated `loadBlocks` to use a sequential promise strategy (`Promise.any` then `Promise.allSettled`) to ensure eventual consistency without race conditions or double-processing.
- Added extensive tests in `tests/user-blocks.test.mjs` covering tag-only lists, whitespace content, and concurrency scenarios.
