* win32 ime wip
* ime window starts tracking the cursor, but coords are wrong
* fix win32 ime cursor coord
* win32-ime lang-specific behaviors
* track language id in WindowImpl
* lowercase dllimport
* create initial ime on window creation
* InputMethodManager: connect to client even if im is absent at the moment
* proposal: IKeyboardDevice.NotifyInputMethodUpdated
* finalizing
* ime: allow client to request active state change
* remove backward incompatible ActiveState.
* InputMethodManager: NotifyInputMethodUpdated: filter the window of current focused element
* [IME] [Windows] ability to enable/disable IME for any InputElement
* [IME] [Windows] Refactor Imm32InputMethod - create a single one for dispatcher. Also change a method of enabling/disabling IME to work like in WPF.
* [IME] [Windows] Fix IME after dialog show not working - active window context is not applied.
* [IME] [Windows] fix intermediate input position
* [IME] [Windows] PreEdit font size is applied
* [IME] [Windows] Make MoveImeWindow code to be exact like in chrome - fix a lot of possible issues. Added comments. Minor Refactoring
* [IME] [Windows] Refactor caret management, improve deactivation, remove comments
* [IME] [Windows] Remove redundant api changes (request from @kekekeks)
* Fix .sln and ApiCompatBesaline.txt redundant changes.
* [Windows] [IME] move IsInputMethodEnabled subscription to InputMethodManager, Move check for IsInputMethodEnabled before TextInputMethodClientRequestedEvent query
* [IME] [Windows] remove redundant SetActive(false) call, because it's called in Client setter
* remove redundant change
Co-authored-by: Yatao Li <yatli@microsoft.com>
Co-authored-by: Max Katz <maxkatz6@outlook.com>
And no longer target net461 in unit tests as we've not had a test pass in net core and fail in netfx in years, and it bloats our CI time; EXCEPT Avalonia.Markup.Xaml.UnitTests where we're still testing against net4.7 as MSIL generation is runtime-sensitive.