Version 126.96.36.199 (June 2011 Update) was released on June 3, 2011.
This June 2011 Update is all about performance improvements. A new caching model has been added that extends and partly replaces the previous caching model. Document Toolkit now supports two caching scenarios. Raw document parts can be cached locally in memory or isolated storage by wrapping a package reader with the MemoryCacheReader or IsolatedStorageCacheReader. Instantiated fonts, pages and images are cached internally as well and configurable using the cache setting in the LoadXpsDocumentSettings class.
Image encoding and decoding functionality has been moved to the new FirstFloor.Documents.Imaging namespace. Document Toolkit now includes an pluggable image decoding API where custom image decoders can be added.
This release also introduces a new feature call image throttling. It helps reduce the overall memory footprint of document toolkit by limiting the size of images.
new: caching improvements
- new FirstFloor.Document.IO.Caching namespace
- document part caching with MemoryCacheReader and IsolatedStorageCacheReader
- improved page, resource dictionary, image and font caching
- caching behavior configurable in LoadXpsDocumentSettings
new: FirstFloor.Document.Imaging namespace for encoding and decoding images
- extensible image decoder API; add your own image decoder
- image throttling; limit the size of rendered images and reduce overall memory footprint
- image quality and maximum pixel size configurable in LoadXpsDocumentSettings
- fix: Path.Data mini-language parser extended to support all known path commands
- fix: image viewport offsetting
- fix: FixedPageViewer behaviors are no longer removed on unload
- fix: FixedPageAdorner behavior proper releases event handler to global selection/annotation
- fix: text selection now supports column based content
- fix: incorrect viewport calculation causes fitwidth mode to display vertical scrollbar
- fix: scrolling per page is 1px off
- fix: loading new document in DocumentViewer will jump to random page
- upd: hard-coded PDF watermark removed (still experimental)
- upd: FixedPage relations skipped when metadata loading is disabled, this further improves load performance