Olivier found the following steps that reproduce the crash with Skia:
1. Open the gallery and select the "Text Edit" tab.
2. In the first text edit widget, select the first three letters.
3. Click into the second text edit (while the first one shows the selected text)
4. At the beginning of the text, insert the letter "ù" a few times.
This sequence would result in the cursor_position property get out of
sync. Since it represents a byte offset in the utf-8 text, it would
point straight into a utf-8 sequence, resulting in a panic when the
renderer would try to create a slice of the selected text.
As a remedy, this patch renames the cursor_position and anchor_position
properties and provides wrapper getter accessors for use within the
TextInput to safely bound the stored offset.
Eventually we need to intercept changes to the text property and adjust
the cursor and anchor position properties accordingly, so that they
remain valid. However this requires that the compiler is aware that a
property is not accessible directly but has a setter that needs to be
called instead.
This likely/hopefully also fixes#1707.
Otherwise computing the fit_size will lead to NaN and then panic.
The size is empty in the wasm case when loading an html image and it
isn't loaded yet. We will be called again because the htmlimage has an
internal property that will be changed when it is loaded
Bump fontdb, foreign-types (used for skia/metal), ttf-parser and winit.
The winit bump is explicit on the patch version to ensure the IME crash
fix is included.
This removes the special code for the generated property getters and
ensures type safety in the run-time library for property value setting.
In the Rust generated code we continue to do arithmetic on the scalar
values, that means we immediately extract the scalar, do arithmetic and
rely on the compiler to only allow compatible units.
Danger zone alert: In the interpreter Value::Number can now be converted
to LogicalLength as-is.
- On the web, to return the image size, we need to use the natural size
of the image, and not its dom size, as the later get modified since
commit b727aba4a0
- The target size did not take in account the image fit, that's because
former version of resvg could only render by respecting the aspect
ratio. But since the web don't have this limitation, we now need to
take it into account. And new version of resvg can also scale with
any aspect ratio
Pros: the composition shows up as selected
Cons: te cursor is shown at the begining of the preselection, and
clicking on the field commit the selection at the wrong place
The test system may run tests in threads using the fluent style and the
testing backend. But because the Qt backend is compiled in, the
NativeStyleMetrics is going to be Qt's and it will potentially be
constructed several times in parallel from different thread, causing
warnings and crashes within Qt.
So don't initialize the StyleMetrics, it is only used for the dark style
anyway
SVGs are rendered using HTML image elements, that are converted to textures.
The size of the texture defaults to the SVGs viewbox, which may be small - despite
it being possible to render the SVG at a higher resolution with great quality.
Similar to the native code path, this patch also uses the target image size
and propagates it to the DOM HTML image element to instruct the browser to
render the SVG at a higher resolution.
Co-authored-by: Olivier Goffart <olivier.goffart@slint-ui.com>
This avoids accidental use of logical pixels, esp. for SVG, elimiates
one call to `to_untyped()` as well as an untyped scale factor in the
femtovg renderer.
The suggestions popup from the input method was off by the y-coordinate. This was due to
text_input_cursor_rect_for_byte_offset returning the y position of the line only.
However QTextLine::y() is relative to QTextLayout::position().y(),
which needs to be added to take any vertical alignment into account.
This comes with the same caveats as the winit one: The composition is
not committed when clicking with the mouse or generally aborting the
composition.
Known caveats:
- winit doesn't forward mouse events to the IME, so clicking
with the mouse while composing results in funny effects such
as the pre-edit text following the cursor.
- With FemtoVG there's no text decoration support, thus no underlining
of the preedit area.
This allows setting the RUST_FONTCONFIG_DLOPEN environment variable
to dlopen fontconfig at runtime rather than linking it at build
time. This is helpful for cross compiling to Linux, particularly
because fontconfig has lots of C dependencies. Building a vendored
copy of fontconfig does not work as expected:
https://github.com/slint-ui/slint/issues/88
Suppose we render an SVG at 100px x 100px with a screen scale factor of 2.
Consequently we should be rendering the SVG at 200phx x 200phx, not 100x100.
Do this by applying the scale factor, just like it is done in the femtovg renderer.
