Update dependency @adobe/remark-gridtables to v3.0.13 - autoclosed#4
Closed
renovate[bot] wants to merge 1 commit intomainfrom
Closed
Update dependency @adobe/remark-gridtables to v3.0.13 - autoclosed#4renovate[bot] wants to merge 1 commit intomainfrom
renovate[bot] wants to merge 1 commit intomainfrom
Conversation
|
Hello, I'm the AEM Code Sync Bot and I will run some actions to deploy your branch and validate page speed.
Commits
|
renovate
Bot
force-pushed
the
renovate/adobe-remark-gridtables-3.x-lockfile
branch
from
8813fd9 to
ef5da98
Compare
aem-code-sync
Bot
had a problem deploying
to
renovate/adobe-remark-gridtables-3.x-lockfile
renovate
Bot
changed the title
renovate
Bot
changed the title
anfibiacreativa
pushed a commit
that referenced
this pull request
Apr 16, 2026
Only send main.innerHTML to saveToDa
kmurugulla
added a commit
that referenced
this pull request
Apr 29, 2026
Fixes final PR #394 high-priority and minor issues: High #5: Async message handler race condition - Changed worker.onmessage from async to sync - Extracted token refresh to separate async function handleTokenRefresh() - Prevents message ordering issues and concurrent listener stacking Minor #14: Watchdog timeout never resets on activity - Added resetWatchdog() helper function - Reset watchdog timer on 'progress' and 'progressive' messages - Prevents killing long-but-active builds at 30min timeout - Watchdog now tracks inactivity, not total duration Minor #12: getMediaLibraryAppHref drops environment params - Deep links from plugin mode now preserve ?nx=local, ?da-admin=stage, ?da-etc=local - Fixes broken dev/stage workflows when inserting media from plugin - Preserves nx, da-admin, da-etc query parameters Note on claimed issues that are ALREADY FIXED in previous commits: Critical #1 (env localStorage): FIXED in 22e47c8 - constants.js:143-151 handles ?da-admin=reset (clears) and ?da-admin=stage (persists) Critical #2 (dead code): NO ISSUE FOUND - No unreachable if(indexing) branches exist in current code Critical #3 (external media dedup): ALREADY IMPLEMENTED - linked-content.js:184-230 has full deduplication logic - Purges invalid entries, removes obsolete ones, updates/adds current ones High #4 (watchdog): ALREADY IMPLEMENTED in previous commit - build.js:106-115 has BUILD_MAX_DURATION_MS timeout All medium/minor issues #6-#11: ALREADY FIXED in commits 22e47c8 and 7a54dd9 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
kmurugulla
added a commit
that referenced
this pull request
Apr 30, 2026
Fixes final PR #394 high-priority and minor issues: High #5: Async message handler race condition - Changed worker.onmessage from async to sync - Extracted token refresh to separate async function handleTokenRefresh() - Prevents message ordering issues and concurrent listener stacking Minor #14: Watchdog timeout never resets on activity - Added resetWatchdog() helper function - Reset watchdog timer on 'progress' and 'progressive' messages - Prevents killing long-but-active builds at 30min timeout - Watchdog now tracks inactivity, not total duration Minor #12: getMediaLibraryAppHref drops environment params - Deep links from plugin mode now preserve ?nx=local, ?da-admin=stage, ?da-etc=local - Fixes broken dev/stage workflows when inserting media from plugin - Preserves nx, da-admin, da-etc query parameters Note on claimed issues that are ALREADY FIXED in previous commits: Critical #1 (env localStorage): FIXED in 22e47c8 - constants.js:143-151 handles ?da-admin=reset (clears) and ?da-admin=stage (persists) Critical #2 (dead code): NO ISSUE FOUND - No unreachable if(indexing) branches exist in current code Critical #3 (external media dedup): ALREADY IMPLEMENTED - linked-content.js:184-230 has full deduplication logic - Purges invalid entries, removes obsolete ones, updates/adds current ones High #4 (watchdog): ALREADY IMPLEMENTED in previous commit - build.js:106-115 has BUILD_MAX_DURATION_MS timeout All medium/minor issues #6-#11: ALREADY FIXED in commits 22e47c8 and 7a54dd9 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
kmurugulla
added a commit
that referenced
this pull request
May 1, 2026
* feat(media-library): implement worker-based indexing for full and incremental builds Migrate media library indexing from main thread to web workers to prevent UI blocking and improve performance. Both full and incremental builds now run entirely in workers with proper authentication and data handling. Changes: - Created index-worker/ directory with worker-safe implementations: - index-worker.js: Main worker entry point with postMessage interface - full-indexer.js: Full build using Status API (extracted from build.js) - worker-incremental.js: Incremental build using auditlog (extracted from build.js) - worker-fetch.js: Worker-safe API functions with runtime token injection - worker-bulk-status.js: Worker-safe Status API operations - worker-linked-content.js: Process PDFs, SVGs, fragments, external media - worker-parse.js: buildUsageMap extracted verbatim from parse.js - worker-utils.js: Extracted utility functions (sortMediaData, path helpers) - worker-admin-helpers.js: Extracted helper functions from admin-api.js - Modified load.js: - Added runWorkerBuild() to create worker with blob URL (handles ?nx=local CORS) - Worker receives runtime context: imsToken, daOrigin, daEtcOrigin, IndexConfig - Handles progress/success/error messages from worker - Modified parse.js: - Made admin-api.js and params.js imports lazy to avoid window access at module level - Modified core/constants.js: - Added AEM_ORIGIN, DA_ORIGIN, DA_ETC_ORIGIN as fixed constants - Avoids importing from public/utils/constants.js (has window.location) - Modified core/paths.js and indexing/admin-api.js: - Updated imports to use constants from core/constants.js instead of public/utils/constants.js Key implementation details: - Extracted code verbatim from build.js (lines 168-527 for incremental, 529-978 for full) - Only modified: imports (use worker-fetch.js) and runtime context injection - Business logic unchanged from production main branch - Full build uses Status API to discover ALL pages - Incremental build uses auditlog for changed pages since lastFetchTime - Proper index merging: loads existing index, adds new entries, removes orphans - Sorting by modified timestamp before saving (newest first) - Lock mechanism prevents concurrent builds 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix(media-library): preserve page references in incremental builds Fix critical bug where incremental builds lost page references for unchanged pages. Root cause: - existingIndex is loaded from deduplicated media sheet (no doc fields) - buildUsageSheet(updatedIndex) only sees changed pages with doc fields - Unchanged page references were discarded on each incremental save Solution: - Maintain usageMap from loaded usage data (has all page→hash mappings) - After processing changes, update usageMap for changed pages only - Save usageMap directly as usage sheet instead of buildUsageSheet(updatedIndex) - Preserves references from unchanged pages + updates changed pages This bug also exists in original build.js:474 (pre-worker implementation). 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix(media-library): preserve references when medialog is incomplete Fix issue where incremental builds incorrectly removed media references when a page was previewed but medialog showed no new entries. Root cause: - User uploads images before lastFetchTime (medialog entry at T1) - User adds images to page and previews (auditlog entry at T3) - Incremental build looks for medialog entries since lastFetchTime (T2) - Doesn't find the upload from T1 (it's before T2) - Incorrectly assumes page has no media → removes old references Solution (conservative approach): - When page is previewed but medialog shows 0 new entries, preserve old entries - Don't remove references without definitive proof they're gone - This prevents false removal when media was uploaded before lastFetchTime Also enhanced worker-linked-content.js to call buildUsageMap for incremental builds (for PDFs/SVGs/fragments/external media). This bug affects both worker and original build.js implementations. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * feat(media-library): parse markdown for images in incremental builds Fixes reference count dropping to 0 for images uploaded before lastFetchTime. Previous behavior: - Incremental builds relied solely on medialog for page→image mappings - Medialog only shows operations since lastFetchTime - Images uploaded before lastFetchTime but added to pages were missed - References incorrectly removed when page previewed New behavior: - Parse markdown of changed pages to extract actual image references - Match parsed images with medialog for metadata when available - Use page timestamp as fallback for images not in medialog - Remove images no longer in markdown, add images found in markdown Changes: - Added extractImageAndVideoUrls() to extract jpg/png/gif/webp/mp4/etc - Modified buildUsageMap to include images: new Map() in usageMap - Added processMarkdownParsedImages() to merge parsed images with index - Reverted conservative "preserve old entries" fix in medialog.js - Worker-incremental now calls buildUsageMap for changed pages 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix(media-library): add worker-safe versions of detectMediaType and computeCanonicalMetadata Fixes "window is not defined" error in incremental builds. Issue: Importing detectMediaType and computeCanonicalMetadata from parse.js triggered module loading chain that eventually accessed window. Solution: Created worker-safe versions in worker-parse.js: - detectMediaType: Duplicated from parse.js:187-192 - computeCanonicalMetadata: Duplicated from parse.js:165-185 - isHashLikeName and extractBestFilename: Helper functions Now worker-incremental.js imports these from worker-parse.js instead of parse.js, avoiding the window access issue. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: use createMedialogEntry instead of duplicating logic Removes duplicate detectMediaType and computeCanonicalMetadata from worker-parse.js. Uses createMedialogEntry from parse.js (same as full-indexer.js does). This is cleaner and avoids code duplication while maintaining worker-safety. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix(media-library): use worker-safe buildUsageMap in worker-linked-content Fixed "window is not defined" error in incremental builds. Issue: worker-linked-content.js was importing buildUsageMap from parse.js instead of worker-parse.js, triggering window access when parsing linked content (PDFs/SVGs/fragments). Solution: - Import buildUsageMap from worker-parse.js - Add context parameter to processLinkedContent - Pass context to buildUsageMap call 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: use medialog timestamp batching, remove markdown parsing for internal images * fix: remove unused imports and functions * fix: extract pathname from originalFilename URLs for proper folder paths * feat: restore buildUsageMap in full builds for parity - Add buildUsageMap call to full-indexer.js to parse markdown for image truthing - Import buildUsageMap from worker-parse.js and createLinkedContentEntries from parse.js - Set actual perf.markdownParse values instead of hardcoded zeros - Pass usageMap to processLinkedContent instead of null - Matches main branch full build behavior for data quality This restores markdown parsing in full builds that was skipped for performance. The parsing validates page-image relationships from medialog against actual markdown content. * fix: process all Status API files including unreferenced SVGs Previously worker-linked-content.js only processed files referenced in usageMap (PDFs/SVGs/fragments found in markdown parsing), missing standalone uploads that were previewed but not used in pages. Main branch's linked-content.js has critical allLinkedPaths logic that processes ALL files from Status API, not just referenced ones. Added: - filesByPath processing from files parameter (Status API events) - deletedPaths tracking and removal - allLinkedPaths collection (Status API files + usageMap paths) - Logic to create standalone entries (doc: '') for unreferenced files This restores parity with main branch, capturing ~13,240 standalone SVGs that were missing from worker builds. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * feat: add image truthing to validate references against markdown After buildUsageMap parses all pages to extract image references, filter the index to remove stale image entries that don't appear in any markdown. This ensures only images actually referenced in pages remain in the index. Image truthing logic: - Collects all image paths found during markdown parsing (usageMap.images) - Filters index entries: images with doc refs must exist in usageMap - Keeps standalone images (no doc ref) and non-image entries unchanged - Logs removed count and timing in perf metrics This catches cases where medialog shows an image was previewed on a page but the image was later removed from the markdown. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * feat: add reference-style markdown link support and HTML escaping * feat: use cached site token with expiry check for worker builds Export getAemSiteToken from admin-api.js and use it in worker bootstrap to fetch a fresh, cached site token instead of reading stale token from localStorage. This leverages the same token caching and expiry logic as the main branch, ensuring the worker starts with a valid token. While this doesn't implement full 401/403 retry within the worker (which would require message-based token refresh protocol), it significantly reduces the likelihood of token expiry during worker execution by: - Using the centralized token cache (aemSiteTokenCache) - Checking token expiry before starting the worker - Automatically refreshing expired tokens via getAemSiteToken This brings worker token handling closer to parity with page-based builds. * feat: decouple display from indexing with mode-aware initialization Add mode parameter to initService() to control initialization behavior: Plugin mode (embedded in iframe/sidekick): - Only starts polling for index updates - Does NOT auto-trigger builds if index is missing - User must manually trigger builds via UI App mode (standalone, default): - Starts polling for index updates - Auto-triggers builds if index is missing or needs refresh - Maintains current behavior for backward compatibility Detection: Uses window.self !== window.top to detect embedded context. This decouples display initialization from indexing logic, allowing the media library to be embedded without forcing expensive index builds. * refactor: remove dead build code after worker migration Deleted buildFullIndex and buildIncrementalIndex functions (913 lines) that are no longer called now that worker-based indexing is active. Only retained getIndexStatus and checkReindexEligibility which are still used by coordinator.js to determine when to trigger builds. File reduced from 982 lines to 68 lines (93% reduction). 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix: restore auto-trigger indexing in app mode and optimize worker performance - Fix app mode detection using URL pathname check (/apps/media-library) - Restore auto-trigger build when index is missing in app mode - Fix linting error: declare timeoutId before use in worker-fetch.js - Optimize worker-incremental.js metrics: single-pass counting replaces redundant filter() calls - Export clearCachedAemSiteToken for worker token refresh - Add neutral indexing event interface for future display/indexing decoupling 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: decouple indexing from display with event-based architecture 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: move progressive deduplication to bridge and app policy to app layer 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: deduplicate parse functions into core/parse-utils.js Moved 9 duplicated functions from parse.js and worker-parse.js into core/parse-utils.js (runtime-neutral utilities that work in both main thread and web worker contexts). Eliminated 447 lines of duplicate code. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: consolidate duplicated utilities into core/parse-utils.js Moved 4 runtime-neutral utility functions from worker-utils.js to core/parse-utils.js: - getCanonicalMediaTimestamp - sortMediaData - normalizeSitePath - getContentPathFromSitePath Updated core/utils.js and core/paths.js to re-export these functions for backward compatibility. Deleted worker-utils.js as it's no longer needed. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: simplify worker file naming by removing redundant prefixes Renamed index-worker/ → worker/ and removed worker- prefixes from all files: - full-indexer.js → full.js - worker-incremental.js → incremental.js - index-worker.js → worker.js (entry point) - worker-admin-helpers.js → admin-helpers.js - worker-bulk-status.js → bulk-status.js - worker-fetch.js → fetch.js - worker-linked-content.js → linked-content.js - worker-parse.js → parse.js Updated all imports within worker files and load.js to reference new paths. The indexing/ directory context already implies worker execution, making both the directory name and file prefixes redundant. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: rename build.js and relocate parse-utils.js for better organization 1. Renamed build.js → index-status.js - File only contains getIndexStatus() and checkReindexEligibility() - New name accurately reflects its focused purpose 2. Moved core/parse-utils.js → indexing/parse-utils.js - All 7 consumers are in indexing/ directory - Utilities are indexing-specific, not general-purpose - Maintained backward compatibility via re-exports in core/utils.js and core/paths.js Updated all import paths across codebase and tests. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: separate display and indexing layers with proper architecture Split load.js (602 lines) into three focused modules aligned with mldecouple architecture: **New Structure:** - display/data.js (295 lines) - Load & process index for UI display - indexing/build.js (280 lines) - Orchestrate index builds via workers - indexing/locks.js (100 lines) - Manage index build locks **Renamed:** - core/indexing-bridge.js → core/indexing-adapter.js (clearer name) **Benefits:** - Clear separation: display layer reads index, indexing layer builds it - No circular dependencies - Easier testing and maintenance - Aligns with event-based architecture (adapter translates events → UI) **Architecture:** ``` display/data.js → reads published index for UI core/indexing-adapter.js → translates indexing events → UI state indexing/build.js → builds index, emits events indexing/locks.js → manages build locks ``` All 580 tests passing ✓ 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: make display layer truly read-only by removing lock coupling **Problem:** display/data.js was checking locks and returning lock status, creating backwards coupling (display telling indexing about indexing state). **Solution:** - Removed lock checking from loadMediaSheet (display layer) - Moved all lock logic to indexing/locks.js (indexing layer) - Coordinator (indexing) now checks locks separately - Display layer is now purely read-only (just loads index files) **Changes:** - display/data.js: 295→228 lines (removed lock checks, saveMediaSheet) - indexing/locks.js: Added checkIndexLock, isFreshIndexLock functions - coordinator.js: Imports lock functions from locks.js, not display **Clean separation achieved:** ``` display/data.js → reads index (no lock awareness) indexing/locks.js → manages locks (write + read) coordinator.js → checks locks, loads display data separately ``` All 580 tests passing ✓ 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * refactor: consolidate UI layer into single ui/ directory * refactor: align test folder structure with code (features → ui) * fix: update asset URL paths for views moved to ui/views * fix: use lighter background (gray-200) for media preview section * refactor: remove obvious inline comments * feat: add dual polling strategy with app/plugin mode support - Renamed polling functions for clarity: - startPolling → startCheckingIndexChanges - pausePolling → pauseCheckingIndexChanges - resumePolling → resumeCheckingIndexChanges - Added outward polling for incremental builds: - startCheckingChanges (120s interval) - pauseCheckingChanges - resumeCheckingChanges - Mode-specific behavior: - Plugin mode: No polling (load data once) - App mode: Both checkIndex (60s) + checkChanges (120s) polling - Added comprehensive perf logging with debug=perf: - Service initialization (mode and polling setup) - Polling start/pause/resume for both types - Poll execution when changes detected - Uses checkIndex/checkChanges terminology 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * chore: add comment to force cache refresh * fix: correct CSS path for topbar after ui/ refactor * fix: address all PR #394 review comments Fixes all blockers, correctness issues, and implements all suggestions: 🔴 Blockers Fixed: - Worker-safe origin resolution (DA_ORIGIN, AEM_ORIGIN, DA_ETC_ORIGIN) - Created resolveDaOrigin/resolveDaEtcOrigin/resolveAemOrigin functions - Main thread passes locationData to worker instead of resolved origins - Supports ?da-admin=stage|local and ?da-etc=local correctly - Fixed dead UI state code (indexing/lockFresh destructure) - Moved initService before loadMediaData for parallel lock check - Prevents empty state flash when lock is held 🟠 Correctness Fixed: - External media deduplication in incremental builds - Added purgeInvalidExternalMediaEntries and proper dedupe logic - Prevents unbounded growth of duplicate entries 🟡 Suggestions Implemented: - Worker watchdog timeout (BUILD_MAX_DURATION_MS: 30min) - Fixed async message handler concurrency (shared listener + Map) - Moved polling intervals to IndexConfig with better names - INDEX_POLLING_INTERVAL_MS (was INWARD) - LOGS_POLLING_INTERVAL_MS (was OUTWARD) - Fixed sticky module state (service key-based lifecycle) - Replaced string error detection with typed MediaLibraryError - Removed duplicate state update in indexing-adapter 🧹 Additional Improvements: - Removed redundant inline comments - Removed unused exports (deprecated functions, internal helpers) - Fixed misleading poll log messages (loaded vs detected changes) - Fixed image sort order (don't bump on re-preview, only on new refs) All 583 tests pass. No new lint errors. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * feat : adding plugin mode , insert via DA SDK, open-app link when index missing * fix : plugin mode link insertion * fix : fixing the link insert behavipr in plugin mode * fix auto close plugin after insert action * fix: auto close plugin fixes * fix : showing loading instead of discovering in plugin mode * fix : donot show broken images * fix: address PR #394 review comments (typed errors, event factories, config constants, worker env resolution) Fixes 4 items from PR review that were claimed but not fully implemented: 1. String-based error detection → typed error codes - coordinator.js: Use `error?.code === ErrorCodes.LOCK_HELD_BY_OTHER` instead of `error.message?.includes` 2. createLockDetectedEvent factory usage - coordinator.js: Use `createLockDetectedEvent(ownerId, timestamp, true)` instead of hand-built object 3. IndexConfig constants usage - coordinator.js: Import IndexConfig and use for all timing constants (INDEX_POLLING_INTERVAL_MS, LOGS_POLLING_INTERVAL_MS, LOCK_CHECK_INTERVAL_MS) - Removed local CONFIG block 4. Module-level sticky state - coordinator.js: Replaced pollingStarted boolean with currentServiceKey tracking - Service lifecycle now uses (sitePath, mode) tuple as key - Prevents state pollution across mode switches (app ↔ plugin) Additional fixes: 5. Env-resolution edge cases - constants.js: Handle ?da-admin=reset (clears localStorage) - constants.js: Handle ?da-admin=stage (persists to localStorage) - Matches canonical behavior from public/utils/constants.js::getDaEnv 6. Worker localStorage mismatch - build.js: Resolve origins on main thread (has localStorage access) - build.js: Pass resolved origin strings (daOrigin, daEtcOrigin, aemOrigin) to worker - worker.js: Receive pre-resolved origins instead of locationData - Fixes: User with localStorage.da-admin=stage but no URL param hitting stage on main thread but prod in worker 7. indexLockedByOther race condition - media-library.js: Remove `indexLockedByOther: false` from loadMediaData updateAppState - Prevents clobbering LOCK_DETECTED event from parallel initService lock check 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix: address remaining PR #394 high-priority issues (SDK check, i18n, perf) Fixes 3 remaining high-priority issues from PR review: 1. Remove unused sendText check (export.js:67) - Only sendHTML is used, sendText was never called - Unnecessarily rejects SDKs that only expose sendHTML 2. Hardcoded 'Copy' bypasses i18n (mediainfo.js:797) - Added UI_COPY_BUTTON to messages.js - Changed 'Copy' → t('UI_COPY_BUTTON') 3. Multiple isMediaLibraryPluginMode() calls per render (mediainfo.js:791-797) - Each call re-reads window.location.pathname (3x per render) - Memoized into const isPluginMode at top of render() Note: PR review claimed issues #2 and #3 from previous commit were not fixed, but they ARE fixed in commit 22e47c8: - Worker localStorage mismatch: Origins resolved on main thread, passed as strings (build.js:73-75, worker.js:45-47) - Env param persistence: ?da-admin=reset clears localStorage, ?da-admin=stage persists (constants.js:143-151) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix: watchdog reset on activity + preserve env params in deep links Fixes final PR #394 high-priority and minor issues: High #5: Async message handler race condition - Changed worker.onmessage from async to sync - Extracted token refresh to separate async function handleTokenRefresh() - Prevents message ordering issues and concurrent listener stacking Minor #14: Watchdog timeout never resets on activity - Added resetWatchdog() helper function - Reset watchdog timer on 'progress' and 'progressive' messages - Prevents killing long-but-active builds at 30min timeout - Watchdog now tracks inactivity, not total duration Minor #12: getMediaLibraryAppHref drops environment params - Deep links from plugin mode now preserve ?nx=local, ?da-admin=stage, ?da-etc=local - Fixes broken dev/stage workflows when inserting media from plugin - Preserves nx, da-admin, da-etc query parameters Note on claimed issues that are ALREADY FIXED in previous commits: Critical #1 (env localStorage): FIXED in 22e47c8 - constants.js:143-151 handles ?da-admin=reset (clears) and ?da-admin=stage (persists) Critical #2 (dead code): NO ISSUE FOUND - No unreachable if(indexing) branches exist in current code Critical #3 (external media dedup): ALREADY IMPLEMENTED - linked-content.js:184-230 has full deduplication logic - Purges invalid entries, removes obsolete ones, updates/adds current ones High #4 (watchdog): ALREADY IMPLEMENTED in previous commit - build.js:106-115 has BUILD_MAX_DURATION_MS timeout All medium/minor issues #6-#11: ALREADY FIXED in commits 22e47c8 and 7a54dd9 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * test: add coverage for environment resolution and deduplication logic Add test coverage for critical paths in media library refactor: Environment Resolution (constants.test.js): - resolveDaOrigin with ?da-admin=stage/local/reset - localStorage persistence across page loads - da.page vs da.live origin handling - resolveDaEtcOrigin with custom endpoints - Full workflow tests (set → persist → use) External Media Deduplication (linked-content.test.js): - Verify update-or-add deduplication pattern exists - Verify obsolete entry removal logic - Verify invalid entry purging (wrong operation, no media type) - Verify linkedPages processing from usage map - Verify incremental vs full build handling - Verify PDF/SVG/fragment deduplication logic Tests use source code verification approach to validate implementation patterns without requiring complex worker context setup. All 22 new tests passing. Addresses PR #394 review feedback about missing test coverage for new worker/event/plugin architecture. 🤖 Generated with Claude Code Co-Authored-By: Claude <noreply@anthropic.com> * fix: replace weak source-text tests with real unit tests for toExternalMediaEntry Replace string-matching tests with actual behavioral tests for external media deduplication logic. Tests now verify: - Entry creation for YouTube, Vimeo, external PDFs - Null returns for internal/non-media URLs - Missing/null/empty timestamp handling - URL-encoded display name decoding - Consistent hash generation for deduplication - YouTube URL normalization (youtu.be → youtube.com/watch) These are real unit tests that will fail if the deduplication logic breaks, unlike the previous source-text assertions. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * test: add coordinator event factory tests Add comprehensive unit tests for event factory functions that the coordinator uses to emit events to the display layer. Tests verify: - Event structure and field names (BUILD_STARTED, BUILD_PROGRESS, etc.) - Timestamp inclusion (only BUILD_STARTED and LOCK_DETECTED have them) - Optional field handling (itemsProcessed, batchIndex excluded when null) - Error event persistent vs transient classification - Lock detection with owner ID and freshness - Index missing/loaded events with data validation - Consistent event type and error code enumerations These tests cover the event-based architecture that decouples coordinator (indexing orchestration) from display layer (UI state/notifications). All 35 new tests passing (640 total). Addresses PR #394 review feedback about missing event/coordinator tests. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * test: add behavioral test for processLinkedContent dedupe logic Addresses PR #394 review feedback about incremental external-media deduplication. Tests the core update/remove/avoid-duplicates behavior: - Updates existing entry when page still references URL - Removes obsolete entry when page no longer in usage map - Adds new entry for newly-referencing page - Prevents duplicate rows for same URL+page combination This directly validates the incremental dedupe logic in processLinkedContent() (worker/linked-content.js:188-230). 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix: preserve video references in incremental builds - Fixed incremental builds orphaning entries from unparsed pages - Added parsedPages check to only affect pages actually parsed in build - Fixed linked content (PDFs, SVGs, fragments, videos) removal logic - Fixed external media (YouTube, Vimeo, etc.) removal logic - Collect external URLs from existing entries to properly remove obsolete refs - Added test for preserving unparsed page references - Fixed ESLint violations (line length, unused imports, formatting) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * fix: reduce Status API partition size from 20K to 10K to prevent server-side timeouts on large sites --------- Co-authored-by: Claude <noreply@anthropic.com>
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
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.
This PR contains the following updates:
3.0.6->3.0.13Release Notes
adobe/remark-gridtables (@adobe/remark-gridtables)
v3.0.13Compare Source
Bug Fixes
v3.0.12Compare Source
Bug Fixes
v3.0.11Compare Source
Bug Fixes
v3.0.10Compare Source
Bug Fixes
v3.0.9Compare Source
Bug Fixes
v3.0.8Compare Source
Bug Fixes
v3.0.7Compare Source
Bug Fixes
Configuration
📅 Schedule: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined).
🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.
♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.
🔕 Ignore: Close this PR and you won't be reminded about this update again.
This PR was generated by Mend Renovate. View the repository job log.