* Add `_NET_WM_PID` atom to Linux X11 window
https://github.com/AvaloniaUI/Avalonia/issues/17444
* Also append WM_CLIENT_MACHINE
* Get the host name from uname method to avoid read the file
* Fix build fail on .NET Standard 2.0
* The render window is the child window. And it should not append pid atom.
* Using the atom from atoms
* Replace UtsName with gethostname
* Skia - Only reverse radial gradient stops if we cover the whole content bounds
* Skia - Only reverse radial gradient stops if we reach cover the whole content bounds
* Add some new lines
* Start working on Android explore by touch
* Start working on a more serious solution
* Reflectionless approach
* Allow multiple providers to be defined for the same AutomationPeer instance
* Implement EmbeddableControlRootAutomationPeer
* Garbage collection
* It's working!!
* Get better readouts
* Implement rest of providers and improve performance
* Some cleanup for the PR!
* Whoopsie!
* Better text readouts for more descriptive elements
* Fix bug with previous approach
* Some final tweaks
* Last tweak!
* Slight improvements
* Undo last change
* Fix bug where custom provider types would not be registered
* Better TextBox compatibility with screen readers & TalkBack
* Fix regression for LabeledBy tests
* Clean up provider code
* Final batch of fixes for TextBox behavior
* Append text instead of replacing it to fix buggy screen readers
* Even more fixes for buggy screen readers
* Remove english-specific state descriptions
* Code review improvements
---------
Co-authored-by: Jumar Macato <16554748+jmacato@users.noreply.github.com>
* Add failing tests for OneTime and null data context bindings
* Made OneTime bindings update on DataContext changes
Also allows null as a valid value for bindings without path
* Remove now obsolete test
* Ignore key modifiers on text editing if field is a password field.
It's not secure to rely on password field content when moving. It's should
give no information what so ever.
* Required changes
The modality of a window is decided at the time it is shown, and indicated by the `isDialog` parameter to `WindowImpl::Show`.
The code was also rather confusing in that `_isModal` was defined in `WindowBaseImpl` but `WindowBaseImpl.IsModal` always returned false. `WindowImpl.IsModal` then overrode `IsModal` to return WindowBaseImpl._isModal. That's messed up. Just define `_isModal` in `WindowImpl` not `WindowBaseImpl`.
* Fix BiDi algorithm and tests
* More adjustments
* Add EastAsianWidth trie and class
* Start implementing Unicode v15.1 Standard Annex #14
* Update Unicode data to V15.1 and initial Unicode® Standard Annex #14 implementation
* Some adjustments
* More fixes
* 543 to go
* 273 to go
* Almost done
* Finish up line breaker
* Some adjustments
* Tokenizer returns a ReadOnly Span instead of a string.
* Returning the old property for no API break changes
---------
Co-authored-by: Meloman19 <23280622+Meloman19@users.noreply.github.com>
* fix: add missing ref keyword to CombinedSpan3.CopyFromSpan
* fix: fix incorrect indent in IUITextInput.TextInRange and boundary condition for surroundingText
---------
Co-authored-by: Max Katz <maxkatz6@outlook.com>
* Add failing test for TextBlock
* Fix TextBlock re-measure in infinite container
* Fix outdated test
---------
Co-authored-by: Max Katz <maxkatz6@outlook.com>
Joseph Sawyer
lindexi
Benedikt Stebner
Isabelle Santin
Emmanuel Hansen
Maxwell Katz
BobLd
Nikita Tsukanov
Compunet
heartacker
Julien Lebosquain
Tom Edwards
Kieran Devlin
Dan Walmsley
Washi
metaone01
Stephen Monaco
c4llv07e
Steven Kirk
Jumar Macato
Meloman19
nickodei
laolarou
MizuKuma
Dong Bin
Alexander Marek