| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=774612
|
|
|
|
|
|
| |
Regardless of whether we have something to display, we need to invalidate
the content on any change. Otherwise we miss the invalidation on the
frame->no-frame transition and the previous will remain visible.
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=773810
|
|
|
|
|
|
|
| |
This happens when shutting down a pipeline. This was also supposed to be
handled, just go it wrong.
https://bugzilla.gnome.org/show_bug.cgi?id=773810
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
The BT.2020 YUV to RGB conversion matrix was wrong.
https://bugzilla.gnome.org/show_bug.cgi?id=770799
|
|
|
|
|
| |
We need the version and enumeration header files when generating the
introspection data.
|
|
|
|
| |
This avoids a compiler warning.
|
|
|
|
|
|
|
|
| |
Drop backwards compatibility with long since deprecated versions of
automake.
The recent work done to support non-srcdir builds also allows us to drop
the configuration variable for Continuous to pick up.
|
|
|
|
| |
Don't use deprecated variables and arguments.
|
|
|
|
|
|
| |
Drop $(srcdir) and $(builddir) from the file lists, and add the
$(srcdir) prefix only where needed. This is similar to what was done in
the clutter-gst-2.0 branch.
|
|
|
|
|
|
|
|
|
| |
Since Clutter-GStreamer does not use Makefile.introspection to generate
the GIR and typelib files, we need to try and keep the build rules as
close as possible, in the hope that the reason why Continuous breaks
whenever generating the introspection data for GStreamer and
GStreamer-based libraries lies in how the build rules for generating
introspection data are defined.
|
|
|
|
|
| |
Include the compiler flags from the dependencies, as well as declare the
`gst_init()` function before calling it.
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=767626
|
| |
|
| |
|
|
|
|
|
|
| |
and not transparent black.
https://bugzilla.gnome.org/show_bug.cgi?id=763367
|
|
|
|
|
|
| |
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=758768
|
|
|
|
|
|
| |
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=758768
|
|
|
|
|
|
| |
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=758768
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
- Set <TITLE> to Types
- Ensure all annotated types and functions are in html
- Remove orphan </SECTION>
- Single <SUBSECTION Standard>
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=758768
|
|
|
|
|
|
|
|
|
|
| |
- Replaces external links to foldoc
- Fixes build warnings e.g.:
In gtk-doc.xsl: For acronym (transfer none) no value found!
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=758768
|
|
|
|
|
|
|
|
| |
Generate index sections by adding name-space to MKDB_OPTIONS
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=758768
|
|
|
|
|
|
|
|
|
| |
Option introduced in GTK-Doc 1.9
https://developer.gnome.org/gtk-doc-manual/stable/modernizing-gtk-doc-1-9.html
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=758768
|
|
|
|
|
|
|
|
| |
Required to allow Python to use method via GObject Introspection
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=758770
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is fixed by making the uploading source equen priority to
the redraws. This is important, as Texture2D upload on slow memory
system can take a lot of time. You still want your UI to be
responsive (being able to show controls, resize, etc.) and not spend all
the time uploading. Ideally, we want exactly one upload per render but
we don't have that much control. A lower priority makes the video jitter
when there is overlayed animation happening.
https://bugzilla.gnome.org/show_bug.cgi?id=759208
|
|
|
|
|
|
|
| |
We need to cleanup all our run-time resources when we go back to
ready state. Otherwise we endup holding buffers in the renderers.
https://bugzilla.gnome.org/show_bug.cgi?id=759207
|
|
|
|
|
|
| |
Use RGBx/BGRx instead of RGBX and BGRX
https://bugzilla.gnome.org/show_bug.cgi?id=759206
|
|
|
|
|
|
|
|
| |
Currently we leak the kid by not removing it from the
content. Remove the kid sink from the content when removing
it from the pipeline.
https://bugzilla.gnome.org/show_bug.cgi?id=759205
|
|
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=758615
Signed-off-by: Chris Mayo <aklhfex@gmail.com>
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=758604
|
| |
|
| |
|
| |
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=756257
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In codecs like H264, the encoder can work with memory frames larger than the
actual video frames. H264 deals with this by specifying both sizes within
its bitstream. Failure to take into account a difference between the two
sizes will likely result in garbage being displayed at the bottom/right of
the video.
In GStreamer these informations are given by the caps and the GstVideoFrame
structure. The caps gives the actual video size, while GstVideFrame
indicates the memory layout.
This patch adds a 2x2 matrix uniform used to reconciliate these 2 different
sizes by scaling the video textures in the sink's shader.
|
| |
|
| |
|
| |
|
| |
|