The optimistic download_transfer was not setting the required flag to not
perform glReadPixels on subsequent map (READ). resulting in glReadPixels
happening twice.
The MPEG standard (ISO-13880-1) says the reserve bits need to be set
to one (2.1.64). This is causing transport streams to fail validation
on some systems.
https://bugzilla.gnome.org/show_bug.cgi?id=760127
After commit 64080e632, configure checks for all the header files that
should be available in OpenCV 2.3 and later. If any of these files isn't
there the OpenCV elements won't be part of the build.
No need to recheck for opencv2/legacy/legacy.hpp again in
gstpyramidsegment.h. Minimum supported OpenCV version must have this header
and configure already checks for it. Removing check.
Run the transform function of pyramidsegment in place, reusing the image
data as both source and destination in cvPyrSegmentation. This avoids
copying the image back and forth and the extra memory.
Fixed adaptivedemux seeking without flushing that just wants
to update stop position. This required protecting the segment
variables with a new mutex so that the seeking thread and the
download threads could safely manipulate the segment and
events related to it.
This contention is only locked/unlocked when starting a new
download, when the first fragment of a segment is received and
when seeking so, hopefully, it won't damage performance.
This works usually in this place, unless the compiler optimizes things in
interesting ways in which case it causes stack corruption and crashes later.
The compiler in question here is clang with -O1, which seems to pack the stack
a bit more and causes writing to the guint as pointer to overwrite map.memory,
which then later crashes during unmapping of the memory.
Some operations are unnecessary when running with only a single GL
context.
e.g. glFlush when setting a fence object as the flush happens on wait.
API: gst_gl_context_is_shared
Avoids downloading and pushing a full segment just to get 1 nanosecond
of data. This happens frequently when seeking is done with flags
that adjust to boundaries or when the start is aligned with segment
starts. The later is common when segment durations is a multiple of
a second.
Properly handle snap flags during reverse seeking. In this case
the before/after are also reversed, so handle those as such.
For example: with a sequence of 1s fragments:
|-- 0 --|-- 1 --|-- 2 --|-- 3 --|
If you seek to 1.5s it is inside fragment 1. With reverse and
snap-before: should play from the end of fragment 1
snap-after: should play from the end of fragment 0
For reverse, set position to segment.stop when starting and also
don't set the position to fragment end timestamp when it finishes,
just leave it at the fragment start.
When doing GLMemory avfvideosrc negotiates UYVY. This change allows avfvideosrc
! tee name=t ! ... ! glimagesink t. ! ... ! gldownload ! vtenc_h264 ! ...
to do GLMemory and 0-copy with the encoder (with the CV meta).
Change texture format from BGRA to NV12. This allows a pipeline like avfvideosrc
! tee name=t ! ... ! glimagesink t. ! ... ! gldownload ! vtenc_h264 ! ... to
negotiate GLMemory. This makes the glimagesink branch much faster (obviously)
and triggers the 0-copy path between avfvideosrc and vtenc (using the CV meta).
Combined this results in a huge perf improvement on iOS (25-30% of CPU time in a
pipeline like the one above).
Note that this doesn't introduce a new shader conversion in the sink, since BGRA
textures had to be copied/converted from format=BGRA,texture-target=RECTANGLE to
format=RGBA,texture-target=2D anyway.
1. Various elements/base classes only perform a subset check on accept-caps
2. Some GL elements have texture-target in their pad template
3. When checking subsets, only the caps to check are allowed to contain extra
fields. If the 'template' caps have extra fields, the subset fails.
Thus without texture-target on the caps, various accept-caps implementations
were failing.
Also, add some convenience functions for setting and retrieving
texture targets to/from GValue.
https://bugzilla.gnome.org/show_bug.cgi?id=759860
In very few cases the simple version was actually needed and having the
parameters hidden by a _full() version caused application that actually needed
it to not use it.
Avoids an unlikely crash.
Arguably, if allocation fails we have no chance of
recovering but nonetheless, RTMP_Alloc can fail and
librtmp's RTMP_init() (called next) assumes a non-NULL
pointer is passed without checking.
Additionally, unify exit path on error.
Avoids an unlikely crash.
Arguably, if allocation fails we have no chance of
recovering but nonetheless, RTMP_Alloc can fail and
librtmp's RTMP_init() (called next) assumes a non-NULL
pointer is passed without checking.
Additionally, unify exit path on error.
