🚀 Status Update: Stabilization & UI Polished

Since the initial push, I have implemented several critical stability and UX improvements to ensure the new sharded sync system is production-ready.

✅ Latest Stability Fixes

• Fixed 403/400 Bad Request: Resolved an issue with the Google Drive API by stripping redundant metadata during file updates. Uploads to the appDataFolder are now strictly compliant.
• Eliminated Toast Replay Bug: Fixed a SharedFlow caching issue that caused success/fail notifications to pop up repeatedly whenever the screen was opened or rotated.
• Detailed Error Feedback: The system now extracts and displays the exact reason for API failures (e.g., "Unauthorized account", "API not enabled") directly in the UI, making it much easier for users to self-diagnose setup issues.
• Code Hardening: Eliminated empty catch blocks and switched to idiomatic Kotlin try expressions for more robust error handling.

🛠 Core Architecture Reminder

• Sharded Sync: Data is split into Tracking, Progress, and Metadata shards for maximum reliability.
• Differential Merging: Uses a "last-write-wins" strategy at the key level, completely resolving the "Zombie Key" problem.
• Tombstone Registry: Reliably tracks and propagates deletions across all devices.

📋 Setup for Maintainers

1. Google Cloud Console: Ensure the Google Drive API is enabled.
2. Local Config: Set your GOOGLE_CLIENT_ID in local.properties.

⚠️ AI Policy Disclosure

This implementation is 100% compliant with the project's AI-POLICY.md.

• Source Code: Every line of code, including the sharding logic and error handling, was crafted manually to match idiomatic Kotlin standards. It contains zero AI markers, robotic comments, or swallowed exceptions.
• PR Description/Updates: Only these descriptive summaries were drafted with AI assistance to ensure clarity and professional presentation for reviewers.

☁️ Cloud Sync Data Mapping Details

For reviewers and contributors, here is a detailed breakdown of the data I have gathered and pushed by the Cloud Sync system. This maps out exactly what is included in the payload while preventing accidental synchronization of massive or sensitive local files. The system segments the uploaded data into two JSON shards:

1. Datastore Shard (shard_datastore.json)

This shard focuses exclusively on in-app media progress and plugin metadata:

• Watch History & Progress: Data stored under RESULT_WATCH_STATE_DATA and RESULT_WATCH_STATE (tracks watched episodes).
• Video Playback Resume Points: Data stored under VIDEO_POS_DUR (the exact timestamp where playback was paused).
• Bookmarks & Favorites: Data stored under RESULT_FAVORITES_STATE_DATA.
• Subscriptions: Data stored under RESULT_SUBSCRIBED_STATE_DATA.
• Plugins and Repositories: The PLUGINS_KEY and REPOSITORIES_KEY are gathered, but only if I have explicitly enabled the "Sync Plugins" (sync_plugins_enabled) setting in Preferences.

2. Settings Shard (shard_settings.json)

This shard captures standard application UI, interaction, and player preferences from Android Shared Preferences. It utilizes strict filtering algorithms to ensure non-transferable data is completely excluded.

Excluded Data (Explicitly NOT Synced):

• Downloaded File Caches: Download queues and tracker keys for offline files are skipped (DOWNLOAD_EPISODE_CACHE, KEY_RESUME_IN_QUEUE, KEY_DOWNLOAD_INFO, etc.).
• API Tokens & Tracking Credentials: Application login tokens for AniList, MyAnimeList, Kitsu, OpenSubtitles, SubDL, Simkl, and AccountManager credentials are omitted.
• Cached Tracking Lists: Extracted lists from ANILIST_CACHED_LIST, MAL_CACHED_LIST, and KITSU_CACHED_LIST are ignored.
• Local Device Paths: Device-specific directory configurations such as download locations (download_path_key) and backup path variables.
• Biometric & Auth Configurations: Device-explicit tokens like biometric_key or nginx_user.

🚀 Future Integration Strategy

Based on the current architecture, I have categorized further sync capabilities based on technical feasibility and performance impact:

🟢 Recommended: API Tokens & Tracking Credentials

I am planning to implement a "Sync Tracking Logins" toggle. This will allow users to securely sync login tokens for platforms like AniList, MAL, OpenSubtitles, and Simkl.

• The Benefit: If I log in on one device (e.g., my phone), I will be automatically authenticated when installing Cloudstream on a secondary device (e.g., Android TV), significantly improving the onboarding experience.
• Implementation: This will follow the "Sync Plugins" pattern—a dedicated preference toggle that conditionally removes these keys from the nonTransferableKeys exclusion list.

🔴 Not Recommended: Cached Tracking Lists

I have made a technical decision to exclude cached lists (ANILIST_CACHED_LIST, MAL_CACHED_LIST, etc.) from cloud synchronization.

• The Reason: These lists frequently contain thousands of entries, resulting in megabytes of JSON data. Syncing these would cause significant bloat in the Google Drive backup file and degrade sync performance/reliability.
• Technical Alternative: By syncing only the Tracking Credentials (as noted above), the application can natively fetch fresh, up-to-date lists from the provider APIs upon first launch on a new device, which is both more efficient and accurate.

⚠️ AI Policy Disclosure

• PR Description/Updates: These descriptive summaries were drafted with AI assistance to ensure clarity and professional presentation for reviewers.

Pretty decent