11-11-2020, 10:40 AM
- Implemented CSS ‘position’ attribute values ‘absolute’ and ‘relative’.
- Implemented CSS attribute background-image: url("path/to/image.ext").
- Forced the read-aloud language code to always be just two letters (so EN-AU becomes EN for instance), which avoids the issue that a possibly reasonable language is getting left off the list of available languages.
- ‘text-to-speech applies emphasis’ is an optional setting, default no. Because Microsoft’s default voices produce an idiotic drawl when reading emphasised text.
- ‘text-to-speech pronounces image descriptions’ is an optional setting, default yes. Because sometimes a book has all sorts of rubbish in images’ alt-text, and we don’t want to hear it read out.
- Implemented ‘advanced rendering’ behaviour for TXT and FB2 books, so that the CFI-saving of bookmarks works for them too (I will eventually want to get rid of the legacy rendering mode altogether – and need to be forward-compatible with this change).
- For Android only (because Windows UWP apps don't have multiple TTS engines), added a dropdown to choose which text-to-speech engine will be used, if there are several of them installed.
- PDF pages can optionally have their margins trimmed (i.e. cut down to the rectangle containing non-white pixels).
- Implemented a special ‘settings to use for this book only’ preset. It appears in the preset list, with a name like "for current book" which you can save settings to, or load (if you wanted to restore some setting you'd changed). When you come to re-open the book, Freda looks to see whether you've previously saved a "for current book" preset for it, and if so, Freda will load it. [potential gotcha – there’s no easy way to make this preset ‘unload’ when you subsequently aren’t reading this particular book]
- Image viewer: Now provides a ‘rotate 90 degrees’ button, and on Windows, a full-screen view and add accessible controls / buttons for zoom in/out and pan.
- Added a setting to tell text-to-speech to 'skip' some characters (noting that ~ * _ seem to be particularly annoying).
- When a ‘folder’ or ‘storage card’ book source is created, it now gets a default label that’s localised into the right language. [note: storage card exists only on Android]
- Bug Fix: The code for squeezing table columns to fit their content is not working (it's giving up when floats were present, failing to account for margin/padding on image elements … )
- Bug Fix: (Android) Windows-1251 and 1252 encodings are not working (it needed a change to the .NET build options for the project, to make it work in Release builds)
- Bug Fix: Read-aloud is failing to proceed to the next chapter in some cases (in fact, when the chapter begins with a paragraph consisting only of non-selectable ‘formatting’ images.
- Bug Fix: The margin of left-floating elements is applied wrongly to the ‘float formatting context’ meaning that it can over-write the content that is meant to be floating round it to the right.
- Bug Fix: The width of a floated image should default to the actual width of the image in the file, rather than the layout-width of the block containing it.
- Bug Fix: (Windows) When GoogleDrive redirects to an app-specific URL in its auth workflow, it uses the same URL for Freda, Freda+ and Freda-Chinese. So if more than one of them is installed, the results will be random, and not very nice. The three apps should use three different uri schemes.