- Add more null safety
- Fix possible FilePathCache race condition
- Add MoveFilesToBooksDir progress reporting
- All metadata is now downloaded in parallel with other post-success tasks.
- Improve download resuming and file cleanup reliability
- The downloader creates temp files with a UUID filename and does not insert them into the FilePathCache. Created files only receive their final file names when they are moved into the Books folder. This is to prepare for a future plan re naming templates
The ZipFile sink could cause program hangs. Additionally, the only reason it was ever used was to package verbose AudibleApi account login errors, saving the returned Html page as a file. Otherwise, the zip file only contains a .log text file.
- Removed Serilog.Sinks.ZipFile
- Add Serilog configuration migration
- Added a custom destructure to handle logging files. If any files are logged, they will be written to "LogyyyyMM_AdditionalFiles.zip"
v12.3.0 caused a regression with contributors with a single word name, causing the name to be doubled. This was caused by using that name as both the first and last name, so swap the first name with the (blank) last name rather than duplicate them.
- Add template tag support for multiple series
- Add series ID and contributor ID to template tags
- <first author> and <first narrator> are now name types with name formatter support
- Properly import contributor IDs into database
- Updated docs
- Add My Music and Local Application Data to known directories
- Make %localappdata%\Libation the default settings folder on *nix machines
- Make %MyMusic%\Libation\Books the default books folder on *nix machines
LibraryCommands.GetCounts hits the file cache hard. The previous cache implementation was linear list, so finding an entry by ID was (n). When you consider that each book may have many files, the number of cache entries could grow to many multiples of the library size.
The new cache uses a dictionary with the ID as its key, and a CacheEntry list as its value.
