Thibault Saunier authored 6 years ago and Nicolas Dufresne committed 6 years ago

And let it the oportunity to get its other pad linked

Example:

```
$ gst-launch-1.0 uridecodebin uri=file:///home/thiblahute/gst-validate.save/gst-integration-testsuites/testsuites/../medias/defaults/flv/819290236.flv caps=audio/x-raw expose-all-streams=FALSE ! fakesink
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
ERROR: from element /GstPipeline:pipeline0/GstURIDecodeBin:uridecodebin0/GstDecodeBin:decodebin0/GstFlvDemux:flvdemux0: Internal data stream error.
Additional debug info:
../subprojects/gst-plugins-good/gst/flv/gstflvdemux.c(2760): gst_flv_demux_loop (): /GstPipeline:pipeline0/GstURIDecodeBin:uridecodebin0/GstDecodeBin:decodebin0/GstFlvDemux:flvdemux0:
streaming stopped, reason not-linked (-1)
ERROR: pipeline doesn't want to preroll.
Setting pipeline to NULL ...
Freeing pipeline ...
```

Christopher Snowhill authored 6 years ago and Nicolas Dufresne committed 6 years ago

Modify the caps string to allow width and height greater than 4096.
There is no need to restrict it since the matroska format allows the
width and height values to be up to eight bytes long, and this also
applies to the webm subset of the format.

Fixes #550

Tested with cross-ios-arm64 and cross-ios-x86, since those two are the
only archs shipped with the official Qt binaries.

This was originally added for fixing conflicting definitions between
Android and Qt, but times have changed and now this breaks the build
on iOS:

 [...]/OpenGLES.framework/Headers/ES3/gl.h:1006:48: error: unknown type name 'gst_qt_GLsync'
 GL_API void           GL_APIENTRY glGetSynciv (GLsync sync, GLenum pname, GLsizei bufSize, GLsizei* length, GLint* values)   OPENGLES_DEPRECATED(ios(7.0, 12.0), tvos(9.0, 12.0));
                                               ^
 ../ext/qt/gstqtgl.h:49:16: note: expanded from macro 'GLsync'
 #define GLsync gst_qt_GLsync
                ^
 6 errors generated.

Instead, we simply avoid defining GLsync ourselves if we're using Qt.

Nicolas Dufresne authored 6 years ago and Olivier Crête committed 6 years ago

When multiple nals are aggrgated, the marker bit should be associated only
with the last NAL of the packet. Otherwise we may break rendering in with
AU alignment.

Nicolas Dufresne authored 6 years ago and Olivier Crête committed 6 years ago

Only forward the marker for the last NAL of the STAP-A. Otherwise each NAL
endup being assumed to be a full frame which may break rendering.

Fixes 557

Nicolas Dufresne authored 6 years ago and Olivier Crête committed 6 years ago

We are missing a sample to test this, but a fix has been made, so add a
todo.

Nicolas Dufresne authored 6 years ago and Olivier Crête committed 6 years ago

Related to #557

This is especially never available on iOS.

Allow fallback to orc subproject if any.
Additionally 'dependencies' keyword is removed from find_library,
because it's invalid keyword for find_library.

Mathieu Duponchelle authored 6 years ago and Sebastian Dröge committed 6 years ago

It makes no sense for a decoder to output a BYTES segment, and
many elements one would plug downstream of a video decoder assume
the segments they receive are in TIME format, for example this fixes:

gst-validate-1.0 filesrc location=opacity01.svg ! gdkpixbufdec ! \
videobalance ! videoconvert ! fakesink

In that case, videobalance was emitting an assertion when trying to
call gst_object_sync_values()

Tested with Cerbero.

Needs gnustl for C++ STL support, which is the GNU STL on Android API
19 and older, and is a wrapper for the llvm-libc++ STL on newer APIs.

QtGui C++ templates use GL functions, so GLESv2 is needed at link time

Vincent Penquerc'h authored 6 years ago and Olivier Crête committed 6 years ago

regardless of whether they're positioned, since positioning
with a 1 channel stream doesn't change anything.

Nirbheek Chauhan authored 6 years ago and Nirbheek Chauhan committed 6 years ago

The Qt Android integration is now signalled with HAVE_QT_ANDROID

See: !86

There was a mismatch between the .pc files generated by
autotools and by meson that would lead to meson not detecting
that opengl api is available even though it is, if -base was
built with autotools. The mismatch has now been rectified in
-base, so we need to update for that.

This is mostly for consistency, this problem didn't seem
to affect anything in -good.

See gstreamer/gst-plugins-bad#871

Fixes FIXME.

And rename x11 option to ximagesrc.

Fixes #553

The android code path is slightly different than the EGLFS one,
so I added previously a HAVE_QT_ANDROID define for use with qmake.
Here I also add it in meson, although I expect nobody will ever use
meson to build this, as it's complicated.

pkg-config should do it's job here, this is unnecessary and implies using cerbero

it is perfectly legal to use the <module/class> style of includes with Qt
and it avoids the need for having the module's include dir in the include path

Victor Toso authored 6 years ago and Nicolas Dufresne committed 6 years ago

More false positives as both of them are initialized in the line
before they are used, wrapped with fail_unless() check.

Victor Toso authored 6 years ago and Nicolas Dufresne committed 6 years ago

False positive for the three variables but some warnings like:

   ../tests/check/elements/matroskamux.c:875:10:
    warning: 'chapters_offset' may be used uninitialized in this function [-Wmaybe-uninitialized]
   *index = chapters_offset;
   ~~~~~~~^~~~~~~~~~~~~~~~~

The above is false positive as there is a gboolean to check if it was
initialized or not (found_chapters_declaration).

Arun Raghavan authored 6 years ago and Arun Raghavan committed 6 years ago

This reverts commit 4d67d1bd.

Using the extended API for the capture path depends on a fix in
PulseAudio
(pulseaudio/pulseaudio!49).

Until then, let's go back to the standard API.

Fixes: #552

The canonical name for the binary is qmake-qt5, and qmake is the
generic name that can also be a qt4 qmake.

Guillaume Desmottes authored 6 years ago and Nicolas Dufresne committed 6 years ago

The cdata structure was freed but not its caps.
It was already done in gst_v4l2_video_dec_subclass_init() and
gst_v4l2_video_enc_subclass_init().

Binding the vertex array to 0 will unbind everything else already.

In the previous order older versions of the Intel GL driver caused
errors to be printed for every single call when disabling the vertex
attrib arrays after binding the vertex array to 0.