Release 1.0.119

Simplify chunk compression logic; lower Zstd level to 2

All chunks are now compressed unconditionally, removing checks for compression effectiveness. Compression level for Zstd is reduced from 3 to 2 for improved performance.

Release 1.0.118

Update EF Core, dependencies, and FK delete behavior

Upgraded project dependencies to latest versions, including Entity Framework Core, MailKit, AWS SDK, ZstdSharp.Port, and EasyExtensions packages. Changed foreign key delete behavior from Cascade to Restrict in both Postgres and Sqlite databases to prevent accidental data loss. Added new migrations to enforce these restrictions. Updated EasyExtensions.Crypto API usage in Module.cs. Refreshed test dependencies and removed license headers from model snapshots.

Release 1.0.117

Refactor Snapshots and SnapshotFiles pages to use DataGrid for improv…

Release 1.0.116

Merge branch 'main' of https://github.com/bvdcode/Octockup

Release 1.0.115

Update copyright information and fix title case in index.html

- Changed copyright format from "Vadim Belov | bvdcode | belov.us" to "Vadim Belov <https://belov.us>" in multiple files.
- Updated the title in index.html from "octockup" to "Octockup".

Release 1.0.114

Add configurable chunked encoding for S3 uploads

Added support for a useChunkEncoding parameter to control whether chunked encoding is used when uploading to S3. The setting is parsed during initialization and applied to PutObjectRequest, allowing for more flexible S3 compatibility.

Release 1.0.113

Merge branch 'main' of https://github.com/bvdcode/octockup

Release 1.0.112

Add additional file extensions to skip compression check

Release 1.0.111

Flush uploaded hashes after backup; remove chunk exists check

Added a flush of pending uploaded hashes in the backup job's cleanup process to ensure all hashes are persisted after execution. Removed the logic that checked for chunk existence in storage and the database before uploading, so chunks are now always uploaded without prior existence verification. Errors during hash flush are logged but do not interrupt cleanup.

Release 1.0.110

Improve chunk existence check to avoid false positives

Previously, an exception was thrown if a chunk existed in storage but was not found in the database. Now, the code first checks if the chunk's hash is actually recorded in the database before throwing an exception, ensuring more accurate detection and reducing unnecessary errors.