| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixed problems:
- Fix sound stucks on multiple resets/starts of QAudioSink.
It was possible to reproduce on playback position change in
mediaplayer.
- Improve audiosink stop (reduce waiting time). The optimization is
based on the fact that it's possible to call AudioOutputUnitStop from
the thread where it was started.
- add some auto test + imrove errors logging in tests. Tests work
fine locally but still need some tune on CI
Task-number: QTBUG-111567
Change-Id: I0eb5c32af4c12dfc0694ee8f5967b4960a0b4ab2
Reviewed-by: Doris Verria <doris.verria@qt.io>
(cherry picked from commit 919b3d308b711c0b267808c783327f2c95233428)
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The NVidia conversion element is necessary for the QT videoSinkBin to
be able to work together with the Nvidia formats used by the
nvv4l2decoder.
Fixes: QTBUG-110995
Change-Id: Icb88caf91eeaf5beeeed18e7ac5b3971ef5a83ca
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit 7d0563ae811f3a6783c068a039a840db4c8400ce)
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The camera view in portrait mode was displayed as landscape.
After 579a0423b8b3c62c582a706d338a9eb1bcd788a3 commit, the resolution is
not transposed if it is not in supported preview sizes. That is the
expected way for camera preview size, but video output resolution need
to be transposed anyway.
This commit rotates the resolution of the video output in portrait mode,
even if the camera resolution is not rotated due to unsupported size.
Fixes: QTBUG-110975
Fixes: QTBUG-111266
Fixes: QTBUG-111421
Fixes: QTBUG-107173
Change-Id: I773fb8d0df1b616c6af8db9a8d7ad785da8295ee
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
(cherry picked from commit 8d5683e85c0d85e31b6ea6d4caf1769737e558b8)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If a slot connected to stateChanged deletes the sink (ie. because the
sink went idle and the source has no more data), then we must not access
the errorState member anymore.
Fixes a crash seen occasionally in the Qt TextToSpeech test.
Change-Id: I5f348844d73100ad4ca5f6e15b3493707f33746e
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Doris Verria <doris.verria@qt.io>
(cherry picked from commit 7ae01fd8b637b59d0c1a0664e27ffd120c1f2deb)
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
| |
Change-Id: I46d17875433e01ad520ffbe5b5f5b5b85efd9c55
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
| |
Fixes: QTBUG-104849
Change-Id: I82afd862722dc212800d8f69f1a4110c11810a38
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit 5677aa3337a8925304c52e0c750c8c91f16f8f8c)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When recording audio, we are setting the audio channel count to the
one specified in the encoder settings.
If this is not set, we were setting it to 2 channels by default.
However, this was causing sometimes problems as the internal
AudioConverter could not always convert from the audio capture device's
actual format to one with 2 channels. To avoid this, we should default
to setting the channel layout configuration settings based on the
channel layout of the audio device. This is safer and avoids unnecessary
format conversions.
Fixes: QTBUG-110844
Change-Id: I3d78ec8b37d6780c7d96147f81ce15bd6425b448
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit a40b9fed50ef75504cec128d4acfb1062e90df98)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
CaptureSession has a functionality of route audio to an output device by
assigning an AudioOutput object to the audioOutput property. It was
missing in android-backend.
This commit add implementation for streaming audio from chosen input
device to chosen output device.
Fixes: QTBUG-109659
Change-Id: I07dfe2188dd9e98a740ccaa6188daaf8e34ca0df
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit b5e0c68b965bd9e6ea99e1c20994d33194324220)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: I11a0bcf1d73354927b4db4f9b2a2ae0e3b830463
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After commit: d4e02d9efd6e8e41aa1a83c0f3d6534d138e7ae3, deadlock from
QAndroidTextureVideoOutput was removed. Necessary cleanup for m_rhi
member (which was created in thread of AndroidTextureThread) was also
added.
After that, during the clean up, we can still get a crash with log
error:
F libapptestVideo_arm64-v8a.so: Cannot make QOpenGLContext current in a
different thread
F libc : Fatal signal 6 (SIGABRT), code -1 (SI_QUEUE) in tid 28160
(qtMainLoopThrea), pid 28130
To avoid that, we need to be sure, that m_rhi was cleared before
m_surfaceThread was stopped. That is why clearSurfaceTexture is called
as BlockingQueuedConnection.
Fixes: QTBUG-110797
Change-Id: I35aa72c8416a99c5372f93d455bc23963fa182e4
Reviewed-by: Samuel Mira <samuel.mira@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit d3f0b75e3c1e98988eec439bf51945e4047ff2af)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The case is reproduced if QAudioInput is set to QMediaCaptureSession
but the platform doesn't have audio inputs.
The case is to be unit tested after implementation of
comprehensive mock of MediaDevices
Change-Id: I372c8c1260cb3fa0d70ea48a31c315bd9580cb06
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit c56de3e60cd3c86524b1767277a69d6e02e1350d)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
| |
Add filter element in gstreamer queue and update size of
captured image when image settings are changed.
Task-number: QTBUG-98506
Change-Id: I87d8ed20cf5326f9f5db76626e476499a2e1d491
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit f27ba7f8759db070525ef4cbaae757a34d80a306)
|
|
|
|
|
| |
Change-Id: I534e1fc73356261b1860aaee25d8ca9bc8a2ff14
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
| |
Fixes: QTBUG-109504
Change-Id: I117eed1ccb6886349eb21185a5c7fefb03bbc561
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
(cherry picked from commit 4088598ca61aef8793d1b255409a58532fb463f9)
|
|
|
|
|
| |
Change-Id: Ib2f281361cdfe72f71aa9ad7c3b865fc5479fe58
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's not clear how to reproduce the crash, but we have crash reports.
The assumption is the crash is a side effect of used deleteLater
for a few objects in the windows backend design. In theory, objects
can have enqueued events that are handled between deleteLater() and
delete, at this time QAudioSink is already deleted.
As a result, I propose using qpointer in order to check
if the object exists.
Task-number: QTBUG-108693
Change-Id: I0c974b1faa6f0a434e95781f22984b532671daa4
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit 365ef2fcc34ae3a269dd019b232e3dc9c9550f60)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
| |
The example is located in spatialaudio/ directory.
Fixes: QTBUG-109956
Change-Id: I1d8a3630722cafa4f664f9ceca74dc5b48bd1ed9
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
(cherry picked from commit 2fd190a23e1d24a26de1196d96d926df7358beb2)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In the main thread, in AVFVideoSinkInterface, we were updating the
texture cache by freeing it and replacing it with a newly created one.
However, in AVFVideoBuffer::textureHandle, called by the render thread,
we were accessing this texture cache to create a new texture.
This led to a race condition, where the texture cache may happen to
be invalid after being freed. To avoid, add a mutex to the
AVFVideoSinkInterface to protect access to the the texture cache.
Fixes: QTBUG-110131
Change-Id: If3a4f17f51b5cb38174815b54f03961584b756b9
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit f4afbf6c573f8d863f456f25b6cc9ac37e4de5c7)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Regression appeared after bd2e80867e4e04fff277cf7340747630d2ea55a4
commit. Before this change, qandroidvideooutput did not create
AndroidSurfaceTexture without QRhi. That's why creation was postopned
and allows qandroidmediaplayer to try create surfaceTexture again during
the Media setting.
From one hand we need to allow to create surfaceTexture without QRhi.
From the other hand we need to allow to recreate surfaceTexture if
QVideoSink has an associated QRhi.
To handle mentioned case shouldTextureBeUpdated() method was added to
QAndroidTextureVideoOutput.
Task-number: QTBUG-109731
Change-Id: Ic845647ab2a9823d02cfc7edb9c735499c3e563b
Reviewed-by: Samuel Mira <samuel.mira@qt.io>
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit f29bf9e033add62da509708e72bfb6359f1e3602)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: I3633c1db1f67edc02dbe4cbbfb611e6b8b553aca
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: Ie6260389461da5b41d217c0b7848798e0e73d701
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The leak is trivial, just not deleted object.
The crash is not trivial for reproducing; the fix is based on
the assumption that some events can be enqueued for an object
between deleteLater and real deleting.
Task-number: QTBUG-109149
Change-Id: I256f29a03e7f01d27923b183b63d9a6b6a4f21b9
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit de266fd951d56dd937d63e79c6adf1702fbe0ec9)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: Ifbdbf95a0f29053ef295a68d078c5d847f049c20
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
| |
Task-number: QTBUG-110453
Change-Id: Iaf7f8dce951174aa1bf46e1fe261b49ae336db33
Reviewed-by: Jani Heikkinen <jani.heikkinen@qt.io>
(cherry picked from commit 6220c1036e576d232d3a9cc3e71d89367b838f1c)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
| |
Task-number: QTBUG-110453
Change-Id: I3be638d836368ada78ce64d9771995e158a6baa9
Reviewed-by: Jani Heikkinen <jani.heikkinen@qt.io>
(cherry picked from commit bc43e99c57f84e8d8fe6773c71de1181d4a173e1)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: I7c4fda22b6fd4fa6d7ddcf9616233136d9fbf004
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
| |
Change-Id: Ib02b6d0bd4562e574462f5e63e1e91d6c34c94a1
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit 6a58cdd1249e34f4d35f68f17a001e083a2487f7)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When using the gstreamer backend, the RTSP latency is set to 2000ms by
default. This latency creates issues for cases where near realtime
camera feedback is critical for directing the user.
This change adds the ability to turn 3 knobs in the RTSP jitter buffer
pipeline using the following environment variables:
1. QT_MEDIA_RTSP_LATENCY
An unsigned integer value expressed in ms. This dictates the
high water mark of the latency buffer.
Default: 2000 ms
2. QT_MEDIA_RTSP_DROP_ON_LATENCY
May be set to 1 for on or 0 for off
When turned on, the jitter buffer will drop packets that arrive
after the high water mark of the buffer has been filled. No new
packets will be accepted until the high water mark recedes
sufficiently to allow new packets to be stored in the buffer.
Default: off
3. QT_MEDIA_RTSP_DO_RETRANSMISSION
May be set to 1 for on or 0 for off
When set to true, the jitter buffer will request a re-send of
dropped packets.
Default: off
This change is strictly for a customer project, but may be helpful in
understanding what additions to the public API would be helpful in the
future.
Change-Id: Icede93161d72936970f4835c6443e369789b4c38
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
Reviewed-by: Doris Verria <doris.verria@qt.io>
(cherry picked from commit 2fcf1212ff78f02e7c7d408c2719a0369eab5bad)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
| |
The crash was reproduced with OBS virtual camera
which doesn't provide kCVImageBufferYCbCrMatrixKey
Change-Id: I9c13de41e7d892b4d58aa151f914b450dd28716e
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit a5127186701df965dd94f712c64981eacd022b0a)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: I5408504c3c5abaadf8883910d8e55e099ef8db2d
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: I184fec634196c814c33e767ca8c5f5dba6e8704f
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: I2b89b71aabcd2f3bb70252cc83ecdd159011968b
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
| |
Video element is not visible unless width and height are set explicitly.
Change-Id: I4ad2e1c815139869ca771c48acef0d8198c3b0f2
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit 8d98e2553e65dfa2a102a210c4237615239e007f)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: Ibf6f5261be38629a14ea667d0efd5dc7b912970b
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After commit: 2abfa1e5cd3cbb148305dc0fe85671488a542397, there was a
thread race in QAndroidCameraSession. Calling setVideoSize put
setFrameSize method to queue (which resets a viewport). It happened from
time to time that source texture coordinates was calculated with not
ready viewport.
To avoid this situation setFrameSize will be queued as
BlockingQueuedConnection.
Fixes: QTBUG-109561
Fixes: QTBUG-109391
Change-Id: Ie6a7bbd332fb42f1aba4fea53449ecac082aaf89
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit e08d5348459fa9fbb67404ffd0c3c6ba818ebe75)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Both ffmpeg and gstreamer plugins define the same logging category,
which breaks application linking when using static linking:
error: /usr/lib/plugins/multimedia/libgstreamermediaplugin.a(qgstreamerimagecapture.cpp.o): multiple definition of 'qLcImageCapture()'
/usr/lib/plugins/multimedia/libffmpegmediaplugin.a(qffmpegimagecapture.cpp.o): previous definition here
Fixes: QTBUG-109613
Change-Id: I9dd573fe8412ec0ea9144ec4a56f5680060cec16
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
(cherry picked from commit 24eb3e2a8859eeff73c4954ff2a3d366a7ab0511)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In case no audio backend is available we should provide some default
implementation for QPlatformDevices otherwise the code crashes on
an attempt list audio devices.
As a bonus clean some leftovers of video devices interface.
Fixes: QTBUG-108221
Change-Id: I8d7e5ebe2ec98ba9381fae429f359b8ffaea3bb6
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit ab49644bea444d105e6fa6ca246f21eccd7a67e0)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
FFmpeg doc says that the following case is possible:
AVFrame.linesize[i] < 0 and AVFrame.data[i] points to the end of data,
it's a weird case of flipped pixels direction, only one video with
such an encoding has been found for now.
Function sws_scale solves the problem so we should call it
in the cases of flipped videos.
I'm going to add a test after finding a way of getting such files.
Task-number: QTBUG-108403
Change-Id: I72f45781e2c3983d6a1991db2b833b78184b90dd
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit 1e8653062d6d23e4dc026661c485e74750c9d4c2)
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Android backend was different from the rest of the backends and it
automatically started the camera when it was attached to media capture
session. This patch changes to set it as the rest of the backends.
Fixes: QTBUG-109415
Change-Id: I88342e9db068bec05404f5bfd91f73ea5015bb96
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit f18be9c59cb38fb3cc5046932c5587997af35cf5)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add error message to Devices example for cmake and qmake.
It will allow to avoid building by unsupported operating systems in the
future.
Fixes: QTBUG-109539
Change-Id: Id8d63e8b6c2a9f066466cbd718510007b55019e1
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit aba7bc72e1ed1b964959e6e358ea3c77db3586c3)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
| |
Change-Id: I2ea4f8c21838abf8fcb727d95b9c8c24b0ac6d05
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit 15c88c4b6ef08e0112b12b5883e481a2372329c0)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
* There is no MediaRecorder.RecordingStatus constant, so change to
MediaRecorder.RecordingState.
* There's no status prop, so change this to recorderState.
Fixes: QTBUG-97265
Change-Id: Iad40b161f86a0c386f745ce7d293a81718f90c6f
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit 108cda751aab4a753c1bfac73d2941f02abdfa49)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
| |
The QLocale supports ISO639-2 language tags. That is why we should start
to use QLocale::codeToLanguage in our implementation.
Task-number: QTBUG-97166
Change-Id: I8f62d6b3d7c8be555249792bfe0c6d6f82af217b
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit 4b976b27e1a879fd89d786ff0776a275eb5e1938)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: I05a3382720297175c32174bf30decb869a77f98c
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
'QAndroidTextureVideoOutput' destructor waited for 'm_surfaceThread' to
be finished but seems that this thread was nowhere requested to finish.
In addition, 'AndroidTextureThread' created 'std::unique_ptr<QRhi> m_rhi;'
in it's own thread but didn't reset it. Thus, deleting the whole
'AndroidTextureThread' from another thread, started deleting it and
failed.
Task-number: QTBUG-109168
Change-Id: I9e12228c579457710c4fe5d8185bffabf9f3755a
Reviewed-by: Samuel Mira <samuel.mira@qt.io>
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit d4e02d9efd6e8e41aa1a83c0f3d6534d138e7ae3)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
gst_device_get_caps and gst_pad_query_caps require decreasing of
reference count when caps are not longer needed. This was not done
with QGstCaps. This patch adds reference counting on QGstCaps and
remove QGstCapsMutable as it was based on a missconception. GstCaps can
be mutated only when refernce count is 1. This was not ensured with
QGstCapsMutable.
Change-Id: Id058a0a916b5f76eef758a099791bec7cc917af9
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit f0ae787ec3d087a8afac45b1473a22c535a44dc3)
|
|
|
|
|
|
|
|
|
|
|
|
| |
The v4l2 camera checkup can properly handly only discrete types.
In case ioctl reports anything else, make sure we increase the index
to avoid getting stuck on infinite loop.
Fixes: QTBUG-109535
Change-Id: I7e669b8167361875b75d4eece5c0c9a92a76c331
Reviewed-by: Artem Dyomin <artem.dyomin@qt.io>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
(cherry picked from commit 8897d36de57db3ce9d75fbd87bcec96bbbd4d69f)
|
|
|
|
|
| |
Change-Id: I746d678a09e489e75b2ed040693948a12511b035
Reviewed-by: Qt Submodule Update Bot <qt_submodule_update_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Functions AudioObjectGetPropertyDataSize and AudioObjectGetPropertyData
have been wrapped into util functions that can print warnings with
info that might help us.
Task-number: QTBUG-108176
Change-Id: I787d5a93ad395d2156df6b542c290949c61daaa3
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Doris Verria <doris.verria@qt.io>
(cherry picked from commit 23e4f614b3ab22300b2ce58c50b1e9a6d4ab5011)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
| |
Change-Id: I0588f613a261451eb8f9ce897d8a0cbed2c1b63d
Reviewed-by: Jani Heikkinen <jani.heikkinen@qt.io>
|