Use the dedicated flag mRunning to check whether the DrawableAnimation
is running, rather than comparing mCurFrame with -1. This CL aims to
simplify the use of mCurFrame.
Bug: 17112962
Change-Id: I15f9e4c102f504b8c806f029949fe9ec872479a5
Bug: 20940526
Rather than throwing an exception on accessing a recycled()
bitmap let certain operations succeed and just return dummy
values. Apps appear to be relying on this.
Change-Id: I74df2efdc29d93facd8553ed31cda3addf0b28eb
We are passing the wrong parameters for setting hotspot bounds.
Bottom and right are in the wrong order, correct it.
Change-Id: I2762fc3a4c29f05ba8b7e71a5c6cad0be16c2ae0
Fix the issue where Bitmap requires two GC passes
to release its byte[] by using some questionable
ref-counting hacks to manage whether or not
native has a strong or weak ref to the byte[]
Change-Id: Ia90a883579f61c0b1904b5549a66bd0ef34b32c5
The docs are now really explicit about the layer's ID and how to set
or update the mask layer from code.
Bug: 20493831
Change-Id: I801f10cd08fd1b4bb226c63a1bdf3271229928ea
Binder APIs which wish to consume Bitmaps *and* drawable
resources can now do so by using Icon, a kind of union type
that accommodates each of these. Icon also accepts byte
arrays holding compressed Bitmaps (PNG, JPEG, etc), which
saves clients the additional memory cost of decoding and
sending full uncompressed bitmaps through Binder interfaces.
Receiving clients can call loadDrawable{,Async} and then
getDrawable to start immediately using the image in an
ImageView or other Drawable-hosting container.
Bug: 19609468
Change-Id: Ic1343711c2ac0b15876b46f0b6008b0108a49470
Switch a few places to using android::canvas
instead of SkCanvas as well which eliminated
some JNI
Change-Id: I8f98b56442a06362b82b984cd1bd3a92398d8dbc
Stop assuming that a Java Bitmap has a SkBitmap* that
has some externally managed lifecycle, and instead switch
a bunch of users to accessing the bitmap by providing
their own SkBitmap* on which to set the (ref counted!)
SkPixelRef* instead
Attempt #2 to land this, original issue was in getSkBitmap
and should be fixed
Change-Id: I0fd9e193968b41e5597784140d56b4885906864a
The developer used to be responsible for this and automatically
propagating the values breaks some use cases (ex. in Camera).
Bug: 20696311
Change-Id: Ia8ddc132c6d629bcc27d66f654baf30d9f078568
Addresses a regression in Calendar where a custom drawable that doesn't
implement intrinsic bounds is laid out incorrectly. Also fixes an issue
where ColorDrawable would be laid out incorrectly.
Bug: 20643614
Change-Id: I796f0d3e189e482cf6fa169a5dd13f654d72653c
For correct low-level drawing of low level text, a method that
includes context for shaping is necessary, and it's similarly useful
to provide the direction explicitly rather than running the BiDi
algorithm on the text. The drawTextRun method (in both char[] and
CharSequence variants) has provided this functionality for several
major releases but has been internal. This patch exposes the
methods publicly, and also improves the doc strings for both
the new method and some related ones.
Bug: 20193553
Change-Id: I9be33ca5ae3e7db2b69a56298400671d5ef8ad05
Failing to check the return result from intersect() can cause artifacts
because the developer may assume that the rect has been changed to the intersection.
In particular, when the two rects do not overlap, there is no change made to the
source rect. Instead, the method simply returns false. When developers do not
check that return value, they may use the source rect and get undefined results.
This change adds the @CheckResults annotation that will cause a Lint warning
when developers call this method without checking the return value.
Change-Id: I476367d74e712038c248c2379fb124734298fcc1
Adds explicit padding attributes to LayerDrawable, constrains the ripple
to fit within the container to avoid projection.
Bug: 19359934
Bug: 20083467
Change-Id: I46da941fba37552adeb2c1fe617b799f576df0c6
