When the Android runtime starts, the system preloads a series of assets
in the Zygote process. These assets are shared across all processes.
Unfortunately, each one of these assets is later uploaded in its own
OpenGL texture, once per process. This wastes memory and generates
unnecessary OpenGL state changes.
This CL introduces an asset server that provides an atlas to all processes.
Note: bitmaps used by skia shaders are *not* sampled from the atlas.
It's an uncommon use case and would require extra texture transforms
in the GL shaders.
WHAT IS THE ASSETS ATLAS
The "assets atlas" is a single, shareable graphic buffer that contains
all the system's preloaded bitmap drawables (this includes 9-patches.)
The atlas is made of two distinct objects: the graphic buffer that
contains the actual pixels and the map which indicates where each
preloaded bitmap can be found in the atlas (essentially a pair of
x and y coordinates.)
HOW IS THE ASSETS ATLAS GENERATED
Because we need to support a wide variety of devices and because it
is easy to change the list of preloaded drawables, the atlas is
generated at runtime, during the startup phase of the system process.
There are several steps that lead to the atlas generation:
1. If the device is booting for the first time, or if the device was
updated, we need to find the best atlas configuration. To do so,
the atlas service tries a number of width, height and algorithm
variations that allows us to pack as many assets as possible while
using as little memory as possible. Once a best configuration is found,
it gets written to disk in /data/system/framework_atlas
2. Given a best configuration (algorithm variant, dimensions and
number of bitmaps that can be packed in the atlas), the atlas service
packs all the preloaded bitmaps into a single graphic buffer object.
3. The packing is done using Skia in a temporary native bitmap. The
Skia bitmap is then copied into the graphic buffer using OpenGL ES
to benefit from texture swizzling.
HOW PROCESSES USE THE ATLAS
Whenever a process' hardware renderer initializes its EGL context,
it queries the atlas service for the graphic buffer and the map.
It is important to remember that both the context and the map will
be valid for the lifetime of the hardware renderer (if the system
process goes down, all apps get killed as well.)
Every time the hardware renderer needs to render a bitmap, it first
checks whether the bitmap can be found in the assets atlas. When
the bitmap is part of the atlas, texture coordinates are remapped
appropriately before rendering.
Change-Id: I8eaecf53e7f6a33d90da3d0047c5ceec89ea3af0
- fix DrawableContainerState.getChangingConfigurations() to take care about its children
- make Resources.verifyPreloadConfig() return false when the changing configuration
contains layout direction bits (this is when a Drawable is having different version
for LTR and RTL layout directions)
- use constant state instead of the resource type value for checking if we can
preload the drawable
- fix typo
Change-Id: Idd64caf0fbe0f5cfd5ffe09343e84bafa9446ea5
We've a number of native functions in the text layout path that take
a bidiFlags argument. We've a number of callers of those functions
passing in SkPaint::Flags in that slot. This completely breaks text
directionality for the affected functions, as
SkPaint::kAntiAlias_Flag happens to share values with kBidi_RTL,
resulting in anti-aliased SkPaints measuring text as if it were RTL,
and non-anti-aliased SkPaints measuring text as if it were LTR,
regardless of the actual text directionality. Oops!
To address the issue, this commit replaces erroneous calls to
SkPaint.getFlags() with the value of Paint.mBidiFlags, and includes
the necessary plumbing to get that value where it needs to be.
Bug: 8471481
Change-Id: I2d04b70defed3130fc1ad13f4c9098f5fce4ffde
Drawables added to a view's Overlay will now cause the Overlay to
be invalidated via the normal drawable-invalidation mechanism. That is,
changes to any of the drawables in the overlay should cause invalidation of
the proper area of the overlay and thus the hostView, causing the appropriate
area to be redrawn.
Also, fixed a bug in drawable invalidation so that bounds changes will now
correctly invalidate both the old and new bounds areas.
Issue #8350510 Add APIs needed for future animation capabilities
Change-Id: Icae5fa0e420232ee17dc39be10084345bae8dbd8
- remove the ICU related methods and update the methods using the "reserved" argument
- update to CTS in another CL too
Change-Id: I5509736568c342d9d17bfeafc17951117ab5d3cc
Drawable has setAlpha(int), but no getAlpha() (although some subclasses have added the
method). This makes it more tedious to use the property. For example, animations that wish to
animate this property must explicitly give it a start value since this value cannot be queried
from the object.
The trick is that setAlpha(int) is abstract, only implemented by subclasses. We cannot take this
approach for getAlpha(), as we would break all subclasses of Drawable until they implemented the
method. Instead, we'll add a default method which returns an invalid value, making it easier for
clients of the method to detect whether the value is valid.
All subclasses of Drawble in frameworks have been changed to add an override of getAlpha() when
appropriate.
Issue #7485875 Drawables is missing getAlpha()
Change-Id: I06b6e35f1a56d202838eca44759c85c82595020a
A common source of layout bugs we're seeing these days involves the
output of measureText() being fed into StaticLayout's constructor.
measureText() returns subpixel-accurate values, but StaticLayout only
takes integral bounds, resulting in the subpixel portion of the
bounds being truncated. This leaves StaticLayout with insufficient
space to layout the text that was just measured, causing all manner
of unexpected line breaks.
This could be causing issues elsewhere, as well. Until our text
pipeline is fully subpixel-perfect, it's best that measureText
guarantee that the value it returns will be sufficient to contain
the text, even if cast to int.
Cherry-pick of Ib84947f0d0a1229287f5b19b99e7efd40f5317f7
Bug: 8164205
Change-Id: I69795ba5770112c0a3d0e752158076a8d1646b1a
When provided with a new bitmap device the canvas will retain
certain information such as it's matrix and clip.
bug: 8167188
Change-Id: I20ac8a24eebc85fb7c147504a103ce9a95b530b3
bug:8037003
-Merges replay methods
-Bounds checking for DrawBitmapMesh, DrawRects and DrawDisplayList
-Use clip as bounds for otherwise unbounded draw operations
Clip-as-bounds is correct for drawColor and functor, but other draw
operations (pos text, text on path, and layers) still need true bounds
calculation
Change-Id: I5d5149d2c624f01e3fe46628bf156e835e69b9d5
# Via Android (Google) Code Review (1) and Victoria Lease (1)
* commit 'c69bce2e41c70da35e1b6a2ab5ec44469ab14c80':
fix argument mismatches in Paint JNI
Paint.getTextRunCursor() no longer has a "flags" argument on the Java
side. The native side, however, still had the argument, and was being
called with misaligned arguments, causing all manner of madcap fun.
Also, the version of Paint.getTextRunCursor() that took String as an
argument needed to lose the "flags" argument, as well, to prevent an
infinite loop in the CharSequence version of the function, which was
supposed to be calling the String version but was actually calling
itself.
Bug: 8201224
Change-Id: Iad0dabaf81185f29a082566cc64590f2ba9bc31c
This reverts commit 6c0307dd0a, reversing
changes made to a2cd828b74.
Conflicts:
packages/SystemUI/res/values-sv/strings.xml
Change-Id: Ia178efe8b14751583d47b2826bfe3d3d5463dd2e
# Via Android (Google) Code Review (1) and Michael Jurka (1)
* commit 'c24b14481527e2f87ddb3a156b8217d32c645b10':
Don't mark recycled bitmaps as immutable
