Now that we have the monorepo, we always have a checkout
of gstreamer and we don't need to explicitly clone it again.
cbuild in ci-templates will always clone the repository in
/tmp/clone and we can use that to initialize the cache.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/8994>
In the rare case when we get EOS on all pads before we get any buffer we
would finish the file without writing headers, i.e. the file would be
corrupt. This patch makes sure that a header is always written.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/9061>
The build files had quite a few things wrong:
* Not using the method: kwarg, which can cause the wrong Qt to be
used for building
* There was no way to enable the build for them
* Qt was being detected multiple times, differently
* Unnecessary check for libGL
* have_cxx was being used incorrectly
* Qt tool detection was outdated
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/9046>
Add a new API to control logging behavior, particularly for implementing
"log once" functionality and periodic logging. This helps avoid spamming
logs with repetitive messages.
The API provides:
- Static and dynamic context creation
- Configurable message identity calculation
- Periodic reset capability
- Context-aware logging macros
- Element message variants with context support
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6855>
As of Gst >= 1.24 we can just use GstVideoInfoDmaDrm APIs. Note
that SHM formats match DRM ones with only two exceptions.
No functional changes intended (for backporting) apart from
supporting a few more formats - those present in video-info-dma.c
but missing in the removed mapping.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/8983>
A new element wrapping the LiteRT (aka TensorFlow Lite) inference engine.
It currently supports only CPU.
Co-authored-by: Daniel Morin <daniel.morin@collabora.com>
Co-authored-by: Denis Shimizu <denis.shimizu@collabora.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/8523>
Add support to the ALSA device provider to enumerate PCM outputs that do
not correspond to a physical sound device i.e. they are "virtual" sinks,
like the plug, dmix, or softvol PCM outputs that can be setup in the ALSA
configuration files.
The main use-case for this is allowing usage of GstDeviceMonitor in setups
where there is no audio server and have custom ALSA audio configurations.
As those are likely to be uncommon, the feature is opt-in: a list of device
names and wildcard patterns separated by semicolons must be assigned to the
GST_ALSA_PCM_ALLOW environment variable before such PCM outputs will be
enumerated by the ALSA device provider. This allows either scanning all
PCM outputs, listing individual outputs, providing simple patterns with
'*' wildcards (which match only at the start or end of the name), or
a combination of them:
GST_ALSA_PCM_ALLOW=1 # Enable listing PCM outputs.
GST_ALSA_PCM_ALLOW='*' # Same, using a wildcard.
GST_ALSA_PCM_ALLOW='out_1;out_1' # Exact listing.
GST_ALSA_PCM_ALLOW='out_*' # Using a wildcard.
GST_ALSA_PCM_ALLOW='out_*;other_*;line_out' # Multiple items.
The main motivation for this patch is supporting enumeration of PCM outputs
in the WebKit GTK and WPE ports, which use GstDeviceMonitor to determine
which devices may be chosen for sound output. While on desktops typically
PulseAudio or PipeWire are used nowadays, on embedded devices it is often
desirable to avoid them and use custom configurations that perform audio
routing and processing using only ALSA.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/8831>
Add handling for the selectable query as an element query,
on top of the existing pad query handling. This is useful
for uridecodebin when handling stream collection messages
before any adaptivedemux source pads have been exposed.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/9002>
