* Replace LibraryThing cover download plugin with a new plugin to download covers from Amazon
* Kindle driver: Add an option to turn off sending page number information. Also add an option to use a more accurate but slower algorithm to calculate page numbers.
* E-book viewer: Maintain a list of recently opened books (to see the list click the arrow next to the open button
* Improve performance of device connection with huge libraries.
* Add a button ot the search preferences to clear search history.
* Device driver customization: Allow user to tell calibre to send any ebook format to the device rather than just the list of formats that calibre thinks the device will support
* Metadata search and replace: Allow case changes and also fix search and replace when the destination field is a tags like field
* FB2 Input: Add support for super and sub-scripts
* When converting non english texts to english for filenames, use the users currect calibre interface language. This allows Japanese/Korean/Vietnamese characters to be correctly converted. Previously they were assumed to be Chinese.
* Content server: Make the authors/series/tags/publisher/etc. in the book details box clickable so that similar books can be found easily.
* Dragging from book list: Only start drag if user clicks on an already selected item. Makes multi-selecting with the mouse easier
* Move the grouped search terms option from tweaks to the Search Customization prefrences. Also allow grouped search terms to be displayed in the Tag browser
* When adding .txt files, if they reference images, autoconvert them to a txtz file containing all referenced images. This is simliar to the way HTML files are converted to ZIP
* MOBI Output: Partially support CSS vertical-align length values, converting them to or tags. Also support vertical-align for images
* TXT Input: New paragraph-type option (off) to disable modifying the paragraph structure.
* Device driver for the Kendo/Yifang M7 and the Wolder Mibuk Life
Bug Fixes
* Fix main memory and storage card for Cybook Orizon being swapped with some firmwares
* Fix regression that broke deleting items from the Manage X dialogs.
* Disable driver for Huawei android tablet/phones as Huawei for some reason seems to think that using the same USB ids for devices and USB sticks is a good idea. This would cause calibre to try to connect to the USB stick as a device continuously
* MOBI Output: Normalize unicode strings when writing metadata to MOBI files as the Kindle cannot handle non-normalized unicode.
* Fix regression that broke completion of tags with special characters
* Conversion pipeline: Discard invalid CSS generated by Microsoft products before parsing. Should improve performance when converting HTML from Word in particular.
* Fix completion popups in the edit metadata dialog partially obscuring text in windows 7
* Allow the left panel of the tweaksdialog to be resized
* Fix reported number of books in library sometimes incorrect when using restrictions
* PML Output: Don’t write Cn tags when we are within a heading. Put images in DropBook required image subfolder. Remove Cn tags written inside of x and Xn tags.
* Kobo driver: Handle Kobo databases that contain corrupted strings that cannot be decoded from UTF-8
* Adding from ISBN: Show a progress dialog for record creation as well as metadata download.
* PDB Input: Fix some TOCs not being detected
* Heuristics: Fix italicize patterns sometime breaking HTML tags.
* When trying to detect the encoding of html, do not use more than the first 50KB so that detection is not too slow
* Fix regresion that caused calibre to loose track of whether the content server is running if the content server is started from the preferences
* Smarten punctuation: Handle non breaking spaces used as indents when converting opening quotes
* Catalog generation: Do not abort on author sort mismatches when generating an EPUB catalog
* When editing emtadata by clicking on the book list, remove leading and trailing whitespace
* Fix typo that removed the date format tweaks from the new style tweaks interface
* TXTZ Input: Fix images being lost
* TXT Input: Rename none formatting-type to plain to correspond to the output option.
* Conversion pipeline: Correctly handle align attribute on img tags.
* Fix regression that caused completion in authors/series/tags fields on OS X to return extra text.
* Heuristics processing: Fix bug in italicize regeps that could cuase a scenebreak consisting of multiple underscores to be prefixed by the word None.
