This CL also updates the documenation to make it clear that the API
returns in local space, not clipped to the size of the bitmap/device.
bug: 8747526
Change-Id: I389844672ce955341863f9940c3b401ab00dc1dc
For fullwidth latin (or, more generally, a font considered as a complex
script, but not supported by the font specified), and for a font other
than a system default font, we were getting a mismatch in baseGlyphCount
between the shaping and painting fonts.
This patch preserves the original typeface in this case. In that sense,
it is a slightly more general fix for bug 8127795.
Change-Id: Ibdbfdb34eb7dc39cc8f0784a67c93c9af4ce6ab2
There were many places where the native object was being
accessed improperly. Also some places where CloseGuard might
not be acquired or released correctly or where the generation
count might not be updated.
Fixed them all.
That said, Surface isn't intended to be used concurrently
so please don't do it. This is only intended to make
hard to find crashes less likely.
Bug: 8328715
Change-Id: I981ef33425823e0fd7ad6b64443f2ec9b0c8335e
Add the ES-relevant constants from the EGL_KHR_create_context
extension. These allow apps to check which EGLConfigs (if any) support
ES3 contexts. Otherwise, the app has to create a context with an
EGLConfig and then check whether it is an ES2 or ES3 context.
Also move eglSetPresentationTimeANDROID from EGL14 to EGLExt, since it's an
extension function. It's new in API 18.
Bug: 8678160
Change-Id: I3cba6e59ebb0a3a4c4012aa54a36b940c288bcec
This patch makes segmentation into script runs behave the same in RTL
as in LTR modes - so that inherited script characters are always
associated with the preceding run. Otherwise, for a sequence such as
u+0631 u+064d u+0020, it would get split after the first character,
which would lose the ability to correctly position the u+064d mark.
Change-Id: I3c12ba1b77d18334f55e707f518be1046e6b339b
The rendering code optimizes by rejecting drawing operations that
lie outside of the bounds of their views. This works in most
situations, but breaks down when containers have called
setClipChildren(false), because we reject drawing that is outside
of that container, but which should be drawn anyway.
Fix is to pass in the value of that flag to the DisplayList drawing
routines which take that flag into account when deciding whether
to quickReject any particular operation.
Issue #8659277 animation clipping
Change-Id: Ief568e4db01b533a97b3c5ea5ad777c03c0eea71
This change disables all atrace tracing in Zygote immediately after it is
initialized. This is necessary because Zygote has no way to receive
notifications that the enabled trace tags have been changed. Tracing is
re-enabled when other processes fork from Zygote.
Change-Id: If2983858fb0c4890ba9ab041849b1c4d98f66c13
TextLayoutShaper was defaulting to the Regular-style fallback chain
for all unknown scripts. This became problematic in that the
codepoint->glyphID enumeration stage of shaping was always using the
Regular-style fallback chain for unknown scripts, and there's no
guarantee that glyph indices are compatible between fallback chains.
Defaulting to a style-appropriate fallback chain addresses this
issue, and probably unreported but related issues, as well.
Bug: 8189208
Change-Id: I6ecf531c74d71a8e4a5359d23439ccc950b0cf80
1. Add uncalibrated gyros and magnetic field sensor.
2. Change max number of events from 3 to 16.
3. Add new APIs for trigger sensors.
Change-Id: Ifac5c0024c8e5f88b721e5cd97ff26afaaa36717
This change adds new APIs to enable applications to generate custom Systrace
begin/end events. Application-generated events use the ATRACE_APP_TAG tag,
which is enabled only if either the application has declared itself debuggable
in its manifest or ro.debuggable is set to 1 on the device.
Change-Id: I311d09e2e6ed1a30f5ffa84907f250e11cc0d48d
InputChannels are normally duplicated when sent to a remote process
over Binder but this does not happen if the recipient is running within
the system server process. This causes problems for KeyGuard because the
InputMethodManagerService may accidentally dispose the channel
that KeyGuard is using.
Fixed the lifecycle of InputChannels that are managed by the IME
framework. We now return a duplicate of the channel to the application
and then take care to dispose of the duplicate when necessary.
In particular, InputBindResult disposes its InputChannel automatically
when returned through Binder (using PARCELABLE_WRITE_RETURN_VALUE).
Bug: 8493879
Change-Id: I08ec3d13268c76f3b56706b4523508bcefa3be79
If the receiver or sender was not properly disposed, then
the underlying input channel might be leaked because the
native peer was holding a strong reference to the object.
Switched to using a weak reference.
Also updated Binder to use a new helper created for this purpose.
Change-Id: I19680bf96d0548777bff02aa1d91874d1e8e41da
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
Deprecated existing API which only supported passing buffer
and length in favor of a new one that also supports passing
the start.
Being able to pass a start index is very important because
the kernel imposes size limits on how much data can be transferred
at a time so we often need to transmit large buffers in chunks.
Change-Id: I3484b5e68f1ece61d4645ea04be8ee6a3b79169d
Disabled some error logs that occur when an input channel is
closed remotely by its peer. These can happen during normal
course of operations because the act of removing windows or
finishing IME sessions is asynchronous so both peers may react
to the change independently at different times.
The coordination with the input dispatcher is designed to
avoid logging these errors. However it's not possible to achieve
the same coordination with the IME so we might as well silence
the errors since they don't tell us anything useful.
If something truly bad happens then one of two things will occur:
1. The system will realize that the process crashed because
of a DeadObjectException or other error, so it will take
measures to clean up.
2. If the error was spurious and non-fatal (how?) then at worst
an ANR may occur because the consumer stopped reading from
the input channel. However this has never been observed
and I doubt it's even possible.
Change-Id: I11a05d6d75e63e91be003971a544069b3a0d77f4
The input method manager service now supplies an input channel for
communication while creating an IME session on behalf of the
application.
This change significanly reduces the overhead of IME event dispatch
by using a standard input channel to send input events rather than
using binder. This results in fewer thread context switches
and fewer object allocations.
What's more, the IME may perform additional batching of the motion
events that it receives which may help it catch up if it is
getting behind while processing them.
Bug: 7984576
Bug: 8473020
Change-Id: Ibe26311edd0060cdcae80194f1753482e635786f
It was possible to overrun this buffer, and even if you did fill
the buffer in a single read and exit the loop, the "string" still
wouldn't be terminated.
Bug: 8468088
Change-Id: Ia19f4d26dfd79e7b63807a7ec9540b932163d246
bug:8450062
- Fixes overdraw indication with DeferredDisplayList
- Fixes drawHardwareLayer called after flush
Additionally changes drawLayer to pass its paint to native via setLayerPaint
Wrap flush in save/restore so that reordering doesn't affect final
transform
Change-Id: I08befa42c28500da6387699eefd4be28aedf9f4c
