The code handling the rendering of the text-decoration-line got moved
into its own function to reduce clutter.
The CSS property text-decoration-line now supports underline, overline
and line-through.
This change allows us to specify where on screen we'd like the Dialog
window to be drawn. By default it's set to CenterWithinParent which
may fall back to Center if parent window is unset or not visible on
screen.
- 'u' and 'd' now scroll up and down half a page
- Typing a number followed by 'j', 'k', 'return', 'up' or 'down' will
scroll that many lines in the appropriate direction
- Typing a number followed by 'g' or 'G' will scroll directly to the
line corresponding to that number
The previous behavior of mapping a missing value to the "inherit"
state is incompatible. Now, a missing value maps to the "true" state,
which is the expected behavior.
Apparently, some code points fit both categories, for example U+0345
(COMBINING GREEK YPOGEGRAMMENI). Handle this fact when determining if
a code point is a final code point in a string.
There's a fair amount of boilerplate when e.g. adding a new UCD file to
parse or a new enumeration to generate. Reduce the overhead by adding
helper lambdas. Also adds a couple missing spec links with UCD field
information.
Note that unlike the main property list, each code point has only one
word break property. Code points that do not have a word break property
are to be assigned the property "Other".
This reverts commit cfc9ee6f16.
This change was wrong: The parent *does* lose hover when the mouse
cursor enters a child widget. Hover is not hierarchical, there is only a
hovered window and a hovered widget within that window.
This fixes an issue with GUI::TabWidget buttons appearing hovered
despite the mouse cursor not being over the buttons.
This reverts commit a5a32fbcce.
This change was wrong: it's fine for a child widget to receive an enter
event despite its parent never getting one. Only the widget directly
under the mouse cursor is considered hovered for enter/leave purposes.
There's no point in using a keyring file we just downloaded from the
same file mirror to verify the authenticity of the binutils tarball.
If someone were to compromise the file mirror they could just as easily
replace the keyring file and we'd happily tell the user that their copy
of binutils is genuine.
