Previously, it would always try to be as big as possible. This CL fixes a
problem with the search bar whereby the text wasn't centered. Now the Stack for
the text sizes its height properly and then becomes centered.
R=ianh@google.com
Review URL: https://codereview.chromium.org/1184483002.
This moves input.dart to editing2/, since that way we can define the layering as strictly unidirectional.
It also reorders a bunch of imports to fit the style guide.
I removed the old remnants of the widgets example, and put the fn2 examples into the examples/widgets/ directory, to parallel the framework directory names.
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/1177243002.
Previously, they always started at the center of the RenderInkWell. Now we
remember the local coordinates we computed during the hit test and provide that
information when handling the event.
R=ianh@google.com
Review URL: https://codereview.chromium.org/1180553002.
I'd do it now, but these files are still being used by fn.dart/stocks1, and it's not worth forking the whole animation directory just for this.
TBR=abarth
Review URL: https://codereview.chromium.org/1170033005
In general, we are hugely inconsistent in what onFoo/onFoo() means.
I'm cleaning out all the cases where onFoo() is a method.
Also, this fixes EditableText, which got missed in the recent syncFields() pass.
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/1175753003
Tweaked the layout in Scaffold a little: the Scaffold
accounts for the (Android) notification bar's height,
instead of the ToolBar. I think it would be best if
moving the app below the notification area were handled
at a lower level. Doing it in Scaffold seemed OK for the
moment, since one could put a ToolBar component anywhere
in an app.
Closes#208R=ianh@google.com
Review URL: https://codereview.chromium.org/1171173003.
This also removes one bit of magic to make it more obvious what on is
going on during a sync, which should hopefully help.
Components have to decide if they support being stateful or not. If
they do, then they must implement syncFields() and have mutable
fields; if they don't, then they must have final fields. This isn't
particularly enforced, though.
This also renames _willSync() to _retainStatefulNodeIfPossible(), for
clarity, and fixes some minor style issues and one typo that was
breaking the drawer.
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/1174023003
Previously, the ink splashes would disappear immediately because the button
highlight would change the structure of the component hierarchy and we weren't
smart enough to retain the underlying RenderInkSplash. This CL is a point fix
for that issue by avoiding changing the structure of the component hierarchy. A
more complete fix would be to make Container smarter about how it syncs its
subcomponents.
R=eseidel@chromium.org, ianh@google.com
Review URL: https://codereview.chromium.org/1168323003.
This CL replaces the (non-working) components2 InkWell with some code based on
the ink_well example. There are at least two issues with the implementation:
1) The ink splash always starts at the center of the well because we don't have
a facility for converting from global to local coordinates, which means we
can't tell where the tap occurred in the local coordinates we need to use
for painting.
2) When used inside a MenuItem, the in splash disappears shortly after
starting, presumably because the button starts highlighting, which causes a
component rebuild and somehow we lose the RenderInkWell instance.
I plan to address these issues in subsequent CLs.
R=ianh@google.com
Review URL: https://codereview.chromium.org/1172033003.
- constructors first
- have either a blank line at the top and bottom of each class, or no blank line in either place (it's still inconsistent about which we should do)
- hide clamp()
R=abarth@chromium.org
Review URL: https://codereview.chromium.org/1172173002
This CL implements a basic version of screen rotation where we just relayout at
the new size. In the future, we'll want a fancy version where we transition
between the two layouts.
R=ianh@google.com
BUG=
Review URL: https://codereview.chromium.org/1177563002.
We already know how to talk to the network_service from Dart
via fetch.dart. Might as well use that for Image loading
as well insetad of having ImageLoader do it.
As part of this I've renamed *ImageLoader to *ImageDecoder
and moved all the image loading logic into Dart. This required
me to teach the idl system about mojo handles so that I could
pass the resulting MojoHandle from fetch.dart up through to
ImageDecoder.
R=abarth@chromium.org, jackson@google.com, hansmuller@google.com
Review URL: https://codereview.chromium.org/1173703002.
