V0.5.6

Changes

• Build Flavors: Added separate standard and ai app flavors so the default build works without AI-specific dependencies.
• AI Summaries: Reworked summarization behind a SummaryEngine abstraction with a disabled fallback for standard builds and an AI-backed implementation for the ai flavor.
• AI Initialization: Fixed concurrent initialization in LlmEdgeSummaryEngine so multiple requests share one initialization instead of racing or loading the model multiple times.
• CI Updates: Split GitHub Actions validation into standard debug, AI debug, and release-specific workflows to avoid requiring signing secrets during normal validation.
• Release Builds: Updated CI and documentation to use explicit flavor-specific Gradle tasks such as assembleStandardDebug, testStandardDebugUnitTest, and assembleAiDebug.
• Security: Added safer DNS resolution for network requests to block localhost, private networks, link-local addresses, metadata IPs, and DNS rebinding-style redirects.
• Web Fetching: Removed unsafe Jsoup.connect fallback paths so HTML fetching goes through the shared OkHttp client and security policy.
• Cloudflare WebView: Hardened WebView navigation and settings, blocked unsafe schemes/hosts, and ensured the WebView is destroyed when dismissed.
• Reading Progress: Refactored reading progress into ReaderProgressController, added explicit progress update semantics, and improved lifecycle-safe progress persistence.
• Database Safety: Enabled Room schema export, added migration coverage across supported schema versions, removed persisted transient selection state, and removed destructive migration fallback.
• Library State: Moved item selection out of persisted database state and kept it as transient UI/ViewModel state.
• PDF Support: Added a bounded PDF document handle pool, document-scoped PDF image caching, deterministic image filenames, and cache cleanup for extracted PDF images.
• Image Loading: Refactored image caching and downloading into dedicated ImageCache and ImageDownloader components.
• Image Downloads: Streamed large web images to disk with byte limits and temp-file handling instead of buffering entire images in memory.
• Image Prefetching: Added priority-aware image downloads so speculative prefetches no longer poison user-requested image loads.
• Prefetching: Replaced unbounded prefetch retry behavior with explicit, bounded retry limits.
• EPUB Images: Added Coil-based EPUB image loading through a custom fetcher, reducing manual bitmap lifecycle management.
• Testing: Added and reorganized tests for migrations, AI initialization, URL security, progress persistence, image downloading, WebView navigation, and reader state.
• Benchmarks: Moved benchmark-style tests out of the normal unit-test path so PR tests stay deterministic.
• Dependencies: Centralized dependency declarations through the Gradle version catalog and updated AI integration to use the Maven-published llmedge dependency.
• Project Cleanup: Renamed the Gradle root project to EasyReader, updated package namespace references, removed stale UI dump artifacts, and cleaned up documentation.

Fixes

• Fixed clean-clone builds by removing the required local llmedge-release.aar from the default build path.
• Fixed PR CI incorrectly depending on AI flavor and release signing.
• Fixed possible duplicate AI model initialization under concurrent summary requests.
• Fixed possible data loss risk from Room destructive migration fallback.
• Fixed stale nullable reading progress fields that could not previously be cleared.
• Fixed PDF handle pooling so active and idle handles are bounded together.
• Fixed leaking randomly named PDF image cache files.
• Fixed full-image response buffering that could cause memory spikes on large images.
• Fixed speculative image requests causing visible image loads to fail.
• Fixed WebView URL validation false positives caused by substring-based scheme checks.
• Fixed image dimension sniffing to handle partial reads correctly.
• Fixed duplicate or misplaced AI unit test source sets.
• Fixed misleading README build commands after adding product flavors.

Full Changelog: V0.5.5...V0.5.6

V0.5.5

Changes :

• PDF Support: Added resume page support, preloading, and optimized content extraction.
• Image Viewer: Added image subsampling, optimized bitmap decoding, enhanced zoom, and animated loading UI.
• Long-Strip & Webtoons: Improved rendering and capped strip group sizes to maximize scroll performance.
• Text & HTML: Implemented TextFormattingPipeline and refactored HtmlParser for better sentence merging and selection.
• Gestures & Navigation: Added scroll gesture handling, simplified scroll boundary checks, and prevented false edge navigation triggers.
• Caching Strategy: Overhauled caching and request handling in WebContentLoader and ChapterListSheet.
• Chapter Management: Improved unread chapter selection, auto-delete logic, and chapter normalization.
• Content Resolution: Refactored ContentRepository to improve content type resolution.
• UI Updates: Debounced searches, updated navigation to select the latest chapter via UI snapshots, and hid finished novels in drawers.
• Performance: Optimized build times, network, memory, and CPU usage.
• Testing: Added unit tests for reading lifecycle, UI logic, library view models, and ContentRepository.
• Dependencies: Updated llmedge-release.aar and added R8 dontwarn rules for llmedge/Ktor dependencies.
• Fixes: Corrected PDF loading state handling, fixed AbstractITextEvent errors, fixed BottomNavigationBar slider thumb size, and enforced minimum height in ReaderImageView.

Full Changelog: V0.5.4...V0.5.5

V0.5.4

What's Changed

• Performance Optimization: Eliminate Repeated Regex Compilation in TextUtils by @Aatricks in #59
• Optimize repeated Regex compilation in string replacement utilities by @Aatricks in #60
• Optimize O(N^2) Lookup in NovelFireSource Search Fallback by @Aatricks in #61
• Optimize Library Updates Concurrency by @Aatricks in #62
• Optimize Regex Compilation in SummaryService by @Aatricks in #63
• Optimize O(N^2) list lookup in NovelFireSource with HashSet by @Aatricks in #64
• Refactor TextUtils to use static Regex instances by @Aatricks in #65
• Optimize text formatting string allocations in formatText by @Aatricks in #66
• Added support for image parsing in pdfs by @Aatricks

Full Changelog: V0.5.3...V0.5.4

V0.5.2

What's Changed

• Secure Backup Configuration by Whitelisting Safe Domains by @Aatricks in #46
• Fix Arbitrary Local File Access via Intent by @Aatricks in #47
• Optimize background image caching with bounded worker pool by @Aatricks in #50
• Fix Zip Bomb vulnerability in EPUB processing by @Aatricks in #51
• Optimize library updates by skipping inactive novels by @Aatricks in #52
• Optimize PDF page loading by caching PdfDocument by @Aatricks in #53
• Fix SSRF via Redirects in NetworkModule by @Aatricks in #54
• Optimize EPUB loading with ZipFile and caching by @Aatricks in #55

Full Changelog: V0.5.1...V0.5.2

V0.5.1

What's Changed

• Limit concurrency in NovelFireSource to prevent rate limiting by @Aatricks in #40
• Limit concurrent background image downloads in ContentRepository by @Aatricks in #35
• Optimize PDF text extraction by using constant Regex patterns by @Aatricks in #36
• Optimize cache size calculation using NIO walkFileTree by @Aatricks in #37
• Optimize adjustChapterUrl with static Regex patterns by @Aatricks in #38
• Optimize LibraryRepository.updateReadingMode by @Aatricks in #41
• Optimize LibraryRepository.updateNovelInfo with direct SQL update by @Aatricks in #42
• Optimize ReaderViewModel auto-deletion concurrency by @Aatricks in #43
• Optimize library updates with batch processing by @Aatricks in #44

Full Changelog: V0.5...V0.5.1

V0.5

What's Changed

• Optimize EPUB image extraction using ZipFile random access by @Aatricks in #33
• Add pinch-to-zoom and double-tap zoom support for Reader images by @Aatricks in #24
• Include database in data extraction rules for backup by @Aatricks in #25
• Disable Cleartext Traffic to Enforce HTTPS by @Aatricks in #26
• Prevent redundant image request enqueueing in ReaderScreen by @Aatricks in #27
• Security Fix: Remove insecure 'Ignore SSL Errors' option by @Aatricks in #28
• Implement lazy loading for PDF content to improve performance by @Aatricks in #29
• Fix global hostname verification vulnerability by @Aatricks in #30
• Fix WebView Mixed Content Vulnerability by @Aatricks in #31
• Fix Intent URL Validation & SSRF Vulnerability by @Aatricks in #32
• Optimize library update performance with parallel requests by @Aatricks in #34

Full Changelog: V0.4.8...V0.5

V0.4.8

Changes :

• Add selection buttons, auto deletion and decimal chapter numbers by @Aatricks
• Add chapter count to ExploreItem and display in ExploreScreen by @Aatricks
• Fixed an issue leading to mangabat novel not being set on the first chapter but on the welcome page by @Aatricks

Full Changelog: V0.4.7...V0.4.8

V0.4.7

What's Changed

• Ssl ignore by @Aatricks in #23

Full Changelog: V0.4.6...V0.4.7

V0.4.6

What's Changed

• fix: improve EPUB image parsing and path resolution by @Aatricks in #22

Full Changelog: V0.4.5...V0.4.6

V0.4.5

What's Changed

• Make 'New' badge clickable and clear updates when reading by @google-labs-jules[bot] in #20
• Refactor by @Aatricks in #21

Full Changelog: V0.4.4...V0.4.5