Compressed assets larger than one megabyte are now decompressed on demand
rather than being decompressed in their entirety and held in memory. Reading
the data in order is relatively efficient, as is seeking forward in the stream.
Seeking backwards is supported, but requires reprocessing the compressed data
from the beginning, so is very inefficient.
In addition, the size limit on compressed assets has been eliminated.
Change-Id: I6e68247957e6c53e7e8ba70d12764695f1723bad
- The linear cost associated with mSampleSizes.size() call causes the
CPU load to increase overtime as more and more audio/video samples
are recorded.
- Other Lists used in the MP4 file writer may have similar issues, but
the size() call is not made for each output audio/video output sample,
or the size of the list is bounded (not linear to the total number
of audio/video output samples). As for now, we can live with the small
cost (tested with 30 minutes long recording).
Change-Id: I23bd93ea4256cb0be0c1649760e39e6809eb4946
When a temporary container is created, its existence should be
ephemeral. However, if there is an error that causes system_server to be
killed during the process of creating a finalized ASEC, delete the stale
containers on the next start-up of system_server.
Change-Id: I7be8f94638a824295474c9a95960594848cb726b
getTempContainerId() would always return "smdl2tmp1" unless you had
MAX_CONTAINERS number of SD card SDKs, because of an array sort that put
all the zeros at the beginning.
Switch from trying to find a hole in the series of numbers to just
getting a number that's one larger than the previous. This reduces the
algorithmic complexity and the memory requirements.
Bug: 2832580
Change-Id: I32dc75ef5a6645f594ea47b032d7402e8860ebcd
Refactored the input reader so that each raw input protocol is handled
by a separate subclass of the new InputMapper type. This way, behaviors
pertaining to keyboard, trackballs, touchscreens, switches and other
devices are clearly distinguished for improved maintainability.
Added partial support for describing capabilities of input devices
(incomplete and untested for now, will be fleshed out in later commits).
Simplified EventHub interface somewhat since InputReader is taking over
more of the work.
Cleaned up some of the interactions between InputManager and
WindowManagerService related to reading input state.
Fixed swiping finger from screen edge into display area.
Added logging of device information to 'dumpsys window'.
Change-Id: I17faffc33e3aec3a0f33f0b37e81a70609378612
Merge commit '17876aa586cc9acfb3e5b909c14b9e73537a1a8d' into gingerbread
* commit '17876aa586cc9acfb3e5b909c14b9e73537a1a8d':
Minor changes to ContentProvider javadoc to improve consistency.
- Don't store timestamp for each output sample
- Don't store timestamp for statistical data collection if the collection of statistical data is not requested
TODO:
1. Reduce CPU load by elimnating the list cost associated with List.size() call.
Change-Id: I590bc17176596a65952c982574b82ee3b15b7d1c
- default interleave duration is set to 1 second
This can dramatically reduce the memory usage
by the MP4 file writer.
Change-Id: Ia3ff202cabfcd2d3f183065d31e4596617c2dded
The type of the cmd, cmdSize and *pReplySize parameters of the effect control interface command()
function have been modified from int to uint32_t. This is more consistent with their role.
Change-Id: I84d289fc262d6753747910f06f485597dfee6591
this situation happened when the last buffer needed to be resized
(or allocated, the first time). the assumption was that the buffer
was in use by SF itself as the current buffer (obviously, this
assumption made no sense when the buffer had never been allocated, btw).
the system would wait until some other buffer became the "front" buffer.
we fix this problem by entirely removing the requirement that the
buffer being resized cannot be the front buffer. instead, we just
allocate a new buffer and replace the front buffer by the new one.
the downside is that this uses more memory (an extra buffer) for a
brief amount of time while the old buffer is being reallocated and
before it has actually been replaced.
Change-Id: I022e4621209474ceb1c671b23deb4188eaaa7285
Merge commit '86c035f0d176be9cb06b1e4f2390c25701417586' into gingerbread
* commit '86c035f0d176be9cb06b1e4f2390c25701417586':
COMMENT ONLY change to clarify ContentProvider documentation.
Gets a little more specific about thread behavior, and makes
pointed comments about not doing too much work in onCreate().
Change-Id: I682f0eb7d7559babee901ed26642751a6ba0a1ea
This is already mentioned at the top of Service.java's javadoc, but
I'm seeing people get confused (myself included the other day) after
getting used to incoming service stub binder calls being on their own
threads and forgetting that startCommand() didn't work that way.
Change-Id: Ib79f48d9368ac0ba2d37b7e4dc98d53804bcd59e
Merge commit '8c65ee2d509db7dcb50ce4530d52eb5bdca3f917' into gingerbread
* commit '8c65ee2d509db7dcb50ce4530d52eb5bdca3f917':
Add a method to let a properly permissioned app directly
a single call.
Addional changes:
- Fixed simulator build
- Use effect interface UUIDs from OpenSL ES includes when available
- Added cleanspec rules to remove now obsolete test effect libraries
- Fixed bug in AudioEffect JNI setParameter function.
Change-Id: Ic25ddb135e2cec5a68c181d727321f5ac7a1ab6b
rework how our EGL wrapper manages EGLConfig:
- we now store the EGLConfig with the EGLSurface and EGLContext
so that we can have easy access to it from eglQueryContext
and eglQuerySurface.
- EGLConfig now are an index into a sorted table of egl_config_t,
we use a binary search to retrieve our EGLConfig (the index) from
the implementation's EGLConfig.
- egl_config_t keeps track of the implementation's index,
EGLConfig and CONFIG_ID as well as our CONFIG_ID.
In many ways, this implementation is simpler and more robust, as it doesn't
assume anything about the number of implementations nor what EGLConfig is
made of (the previous code assumed EGLConfig didn't usem more than 24-bits).
Change-Id: Id5abe923aacb6e1fd2b63bd8c15d7b04ae824922
