Use named parameter syntax in Vercel rewrite and add <base href="/"> tag to ensure assets load correctly from root when serving index.html through the API.
Instead of redirecting, serve the static index.html file directly. The Vercel rewrite preserves the /a/{naddr} URL, allowing client-side SPA routing to work correctly.
Browsers get 302 redirect to / where the SPA handles routing client-side with the original /a/{naddr} URL preserved. Crawlers/bots get the full HTML with OG meta tags.
- Remove unused readingPositions and markedAsReadIds from useBookmarksData
- Remove eventStore parameter from useBookmarksData call
- Add reads and links fields to MeCache interface
Resolved conflicts by keeping feature branch changes:
- Kept /me/reads and /me/links routes (not /me/archive)
- Kept ReadingProgressFilters component and readingProgressUtils
- Kept readsService, linksService, and readingDataProcessor
- Restored files that were renamed/deleted in master
- Remove unused state variables (readsMap, linksMap) by using only setters
- Move VALID_FILTERS constant outside component to fix exhaustive-deps warning
- Remove unused isReading variable in ReadingProgressIndicator
- Remove unused extractUrlFromBookmark function and IndividualBookmark import
- Fix type errors in linksFromBookmarks by extracting metadata from tags instead of non-existent properties
- Replace closure over tempMap with setState callback pattern
- Ensures we always work with latest state when merging progress
- Prevents stale closure issues that block state updates
- Apply same fix to both reads and links tabs
- Fixes reading progress not updating in UI
- Convert bookmark coordinates to naddr format in deriveReadsFromBookmarks
- Reading positions store progress with naddr as ID
- Using naddr format enables proper merging of reading progress data
- Simplify getReadItemUrl to use item.id directly (already naddr)
- Fixes reading progress not showing in /me/reads tab
- Extract metadata from tags same way BookmarkItem does (DRY)
- Add image tag extraction for article images
- Add summary tag extraction for article summaries
- Add published_at tag extraction for publish dates
- Images and summaries now display in /me/reads tab
- IndividualBookmark doesn't have separate title/event fields
- After hydration, article titles are stored in content field
- Simplified extraction logic to just use bookmark.content
- Use getArticleTitle, getArticleSummary, getArticleImage, getArticlePublished from Helpers
- Extract metadata from bookmark.event when available
- Fallback to bookmark fields if event not hydrated
- Fixes 'Untitled' articles in Reads tab
- Add deriveReadsFromBookmarks helper to convert 30023 bookmarks to ReadItems
- Add deriveLinksFromBookmarks helper for web bookmarks (39701) and URLs
- Update loadReadsTab to show bookmarked articles immediately, enrich in background
- Update loadLinksTab to show bookmarked links immediately, enrich in background
- Background enrichment merges reading progress only for displayed items
- Preserve existing pull-to-refresh and empty state logic
- Create src/config/kinds.ts with named Nostr kind constants
- Add streaming support to fetchAllReads and fetchLinks with onItem callbacks
- Update all services to use KINDS constants instead of magic numbers
- Add mergeReadItem utility for DRY state management
- Add fallbackTitleFromUrl for external links without titles
- Relax validation to allow external items without titles
- Update Me.tsx to use streaming with Map-based state for reads/links
- Fix refresh to merge new data instead of clearing state
- Fix empty states for Reads and Links tabs (no more infinite skeletons)
- Services updated: readsService, linksService, libraryService, bookmarkService, exploreService, highlights/fetchByAuthor
Removed empty state messages like "No articles in your reads" and
"No links yet" - now just show loading skeletons until data arrives.
This is simpler and prevents showing empty states while data is still
being fetched in the background.
Users will only see:
- Skeletons when no data (loading or truly empty)
- "No articles/links match this filter" when filtered out
- Actual content when data is available
The bug was that showSkeletons checked if ANY tab had data, so if you
had highlights or bookmarks, it would never show skeletons for reads/links
even while they were still loading.
Fix: Each tab now checks its own loading state (loading && tabData.length === 0)
instead of using the shared showSkeletons variable.
This makes the logic simple and clear:
1. If loading AND no data → show skeletons
2. If not loading AND no data → show empty state
3. If has data but filtered out → show no match message
4. Otherwise → show content
- Create readingProgressUtils.ts with filterByReadingProgress function
- Create readingDataProcessor.ts with shared processing functions:
- processReadingPositions
- processMarkedAsRead
- filterValidItems
- sortByReadingActivity
- Refactor readsService.ts to use shared utilities
- Refactor linksService.ts to use shared utilities
- Eliminate 100+ lines of duplicated code
- Simplify Me.tsx filter logic to 2 lines
Benefits:
- Single source of truth for reading progress filtering
- Easier to maintain and modify
- Less code duplication across services
- More testable with isolated utility functions
- Reads: Only Nostr-native articles (kind:30023)
- Links: Only external URLs with reading progress
- Create linksService.ts for fetching external URL links
- Update readsService to filter only Nostr articles
- Add Links tab between Reads and Writings with same filtering
- Add /me/links route
- Update meCache to include links field
- Both tabs support reading progress filters
- Lazy loading for both tabs
This provides clear separation between native Nostr content and external web links.
- Exclude Nostr articles without event data (can't fetch title)
- Exclude external URLs without proper titles
- Prevents cluttering Reads with items that have no meaningful title
- Only shows items we can properly identify and display
- External URLs with 0% progress are now filtered out
- External URLs only appear if readingProgress > 0 OR marked as read
- Nostr articles still show even at 0% (bookmarked articles)
- Keeps Reads tab focused on actual reading activity for external links
- Replace spinner in highlights tab with 'No highlights yet' message
- Replace spinner in reading-list tab with 'No bookmarks yet' message
- Only show these messages when loading is complete and arrays are empty
- Remove unused faSpinner import
- Consistent with skeleton placeholder pattern used elsewhere
- Add loadedTabs state to track which tabs have been loaded
- Create tab-specific loading functions (loadHighlightsTab, loadWritingsTab, loadReadingListTab, loadReadsTab)
- Only load data for active tab on mount and tab switches
- Show cached data immediately, refresh in background when revisiting tabs
- Update pull-to-refresh to only reload the active tab
- Show loading skeletons only on first load of each tab
- Works for both /me (own profile) and /p/ (other profiles)
This reduces initial load time from 30+ seconds to 2-5 seconds by only fetching data for the active tab.
- Remove separate loadingReads state
- Keep single loading state true until ALL data is loaded
- Matches existing pattern used in other tabs
- Keeps code DRY and simple
- Exclude items without readingTimestamp or markedAt from reads
- Prevents 'Just Now' items from appearing in the reads list
- Only show reads with valid activity timestamps
- Add separate loadingReads state to track reads fetching
- Show skeletons during the entire reads loading period
- Set loading=false after public data (highlights/writings) completes
- Prevents showing 'No articles match this filter' while reads are being fetched
