Previously layout was exiting early because the frame view
was null. Now we actually compute some layout and paint the
text. This patch makes paragraph_builder_test pass.
This patch start down the road of implementing text layout and painting without
the DOM. We can construct a basic paragraph consisting of a single run of text
and we can get through layout without crashing.
Dismissable now only depends on GestureDetector.
Added a unit test that verifies that issue #1068 has been fixed. It's commented out for now.
Cleaned up VelocityTracker.cc et al a little.
Added FlingGestureRecognizer and exposed it in the GestureDetector class. FlingGestureRecognizer is based on the Android/Chromium VelocityTracker class which computes a velocity vector for for a list of X,Y,Time tuples.
The Scrollable classes now use the FlingGestureRecognizer. Dismissable and Drawer still need to be updated
Alternating scroll gestures would sometimes be ignored because _ScrollGestureRecognizer didn't always reset its _state when the pointer[s] went up.
A Dismissable dismiss triggered by a drag and then a fling could cause the next attempt to drag-dimiss to fail.
Fixed the definition of lerpColor().
Rather than using the DOM to upload text and styling information into the
engine, this patch begins sketching a more direct API that bypasses the DOM and
CSS. Currently, this API doesn't do anything, but it's a first step.
The approach is to have a ParagraphBuilder object that can record a tree of
style interior nodes and text leaves. The build() function then applies
container-level styling information (such as TextAlign) and returns a Paragraph
object that can undergo layout and paint.
The inputs to the builder process are immutable style objects constructed from
primitive values. These primitives are currently carbon-copies of the primitive
we use in the framework today. After this patch lands, I'll convert the frame
to re-expose these values instead of re-defining them.
Now that we've hoisted Paint into the VM, we can expose these values as normal
properties. After we've updated all the callers, I'll remove the old setters.
This patch just adds them to dart:sky. A future patch will actually use them in
the framework once we've published an updated sky_engine package. Also, add
dartdoc to the classes touched in this patch.
This updates to mojo commit d259eb58aa59 and limits the roll script to
only pull in the parts of //mojo that are currently being used. More
stuff will be dropped in the future.
We still use SkPicture as our rasterization backend, but now we're uploading
our layer tree to C++. A future patch will push the layer tree deeper into the
system and we'll eventually switch backends.
Now we print Dart_GetError only when we don't have a stack trace. Also, improve
the usability of the sky_shell command line tool by defaulting to "packages"
for the package-root and not crashing when we don't have a package-root until
we actually try to load a package. Finally, remove console spam when we fail to
bind the observatory port.
Fixes#689
- Add Canvas.getSaveCount()
- Make RenderClipRect call context.paintChildWithClip instead of doing the clipping itself
- Make ClipLayer take a Rect instead of a Size
- Make PaintingContext.canvas read-only
- Add PaintingContext.paintChildWithClip()
- Minor rearrangings of code and style tweaks
Now we have the ability to draw multiple PictureLayers. We still squash all the
pictures together into a single SkPicture for simplicity. In the future, we'll
submit them to C++ separately and composite them on the GPU thread.
This patch converts sky.Paint to be a pure Dart object, which means we don't
need to open a weak handle to sky.Paint. Avoiding the weak handle reduces the
amount of weak handle callbacks we need to process at the end of GC and
therefore reduces jank.
This patch simplifies the SkPicture we generate for Skia. Instead of drawing
everything into a nested SkPicture, we now draw everything into the top-level
picture, which requires us to apply the device scale factor in Dart.
