For some reason having |packages| in this .gitignore file is causing Atom to
grey out all the files in the Sky package. Moving the entry up to the root of
the git repo fixes the issue.
We need to compute whether a RenderObject has a composited descendant so that
we can decide whether to use canvas.saveLayer or to create a new composited
layer while walking down the tree during painting.
The compositing update walks the tree from the root only to places where the
tree's structure has been mutated. In the common case during an animation loop,
we won't need to visit any render object beyond the root.
Separating these notions makes them easier to work with because offset is
relative to the parent layer whereas size is intrinsic to the layer itself.
This patch fixes the underpainting bugs when compositing the stocks example.
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.
Adds CicularProgressIndicator and LinearProgressIndicator
Setting a ProgressIndicator's value to null (the default) makes it an "indeterminate progress" or activity indicator.
The indeterminate animations for both kinds of progress bars are essentially the same and wrong vis the Material Design spec, http://www.google.com/design/spec/components/progress-activity.html. I'll improve conformity with the visual design in a future CL.
Setting a ProgressIndicator's value to null (the default) makes it an "indeterminate progress" or activity indicator.
The indeterminate animations for both kinds of progress bars are essentially the same and wrong vis the Material Design spec, http://www.google.com/design/spec/components/progress-activity.html. I'll improve conformity with the visual design in a future CL.
This patch prepares for introducing the compositing system by repurposing the
relevant parts of the createsNewDisplayList system. This patch also removes the
no-longer-relevant parts of that system.
1. We would assert if you tried to start an animation from within an
animation status callback. This is a common pattern, so I fixed this
assert (in Ticker._tick).
2. We would assert for animations with duration under a millisecond.
Fixed.
Also removed the workarounds assocated with #1.
Previously we were doing the bounds testing for hit tests in the parent, but
that doesn't work if the child paints at a location other than 0.0, 0.0. Now we
do the bounds check in the child. This also simplifies Scaffold's hit testing.
Fixes#558
Set the scrollOffset of a widget's Scrollable ancestor so that the
widget is centered within the scrollable. A future CL will add
support for specifying exactly where the widget appears. The scroll
can be animated by specifying the animation: parameter.
Changed the duration Scrollable.scrollTo() parameter from a Duration
to an AnimationPerformance so that one can configure all aspects of
the animation. The caller may also listen to the animation to schedule
other work while it updates or when its status changes.
complete
This introduces the concept of an Anchor, which you can use to link
transitions together. I've used this in the Fitness and Stocks apps to
link the FAB and SnackBar to animate together by sharing the
SlideTransition.
I also fixed the Scaffold hit testing code to apply sub-widget
transforms, so it works with Transformed nodes.
Now it wont crash on save, but it still doesn't display
the error message in a snackbar and this also doesn't
wire up the enter/go key to actually do anything.
We'll need to implement performEditorAction on our
InputConnection implementation to catch this key and
pass it along to dart.
Our Input control needs to get more powerful to handle things
like this, including filtering of input.
The code was crashing by triggering a bug in the SnackBar
which didn't assert for actions, but crashed when they
were not null.
Partial fix for https://github.com/domokit/sky_engine/issues/543
@abarth
Now that we get sky.Image callbacks synchronously, we were actually figuring in
the width and height of images the second time the drawer opened in the Stocks
app. That exposed a bug in our RenderImage layout code whereby it would ignore
the _width and _height properties when the image was non-null.
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.
