Check and cap the write size in pull mode to handle the case
where user returned an invalid write size through QIODevice
Change-Id: Ie0610a63f1d5400fba87f32a99bdc38479e0e7e8
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
Reviewed-by: Jonas Rabbe <jonas.rabbe@nokia.com>
At each userFeed, m_bytesAvailable is not updated properly in pullmode,
so the pull data size become irrelevent to the size we can actually write.
This problem is fixed by checking the actual writable size each time
before pulling the data and limit the size of the pulled and written data.
Change-Id: I6f53e6348693ddf4e3c79e90d3c3d0c1ffc713aa
Reviewed-by: Jonas Rabbe <jonas.rabbe@nokia.com>
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
Broken by qtbase dfa24768a3243eda68d91f573050bf23f96da2dd.
Change-Id: Ie0d79eaee5c6fedfaacae69800f6aec9190f83db
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
Because the Q_WS_* defines were removed, the Quicktime backend
stopped working. Replacing them with Q_OS_* works. Additionally,
the Quicktimer Player Service has to depend on QtWidgets.
Change-Id: I687b352191c38a91edaaf2a687ac13ed118fcf88
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
Reviewed-by: Dmytro Poplavskiy <dmytro.poplavskiy@nokia.com>
Patch from Sami Nurmenniemi.
Emits searchingChanged(false)
when scanning is aborted because it was initiated too close to
the limits of the frequency range.
Change-Id: I15a1334c8a7058dd8eb413c2742f8f12aed7bed5
Reviewed-on: http://codereview.qt-project.org/5570
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
It's necessary to temporarily disable show-preroll-frame of video sink,
load pipeline to paused state, seek to requested position,
and after seeking is finished (position updated) playback is started
with show-preroll-frame restored.
Task-number: MOBILITY-3030
Reviewed-by: Jonas Rabbe
Change-Id: Ide1e6d909dd53f670229c293bc6be496a54e8626
(cherry picked from commit 625cce87e28fc7b5ec8785824affb3129fd3607b)
Reviewed-on: http://codereview.qt-project.org/5500
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jonas Rabbe <jonas.rabbe@nokia.com>
Includes some behavior for all standard methods.
Also fixes a typo in qradiotunercontrol.h and a couple of minor
bugs in the radio example that came to light using this new
backend.
Change-Id: I65b8b8715a46f0fd702f9630ea81f7d5df527055
Reviewed-on: http://codereview.qt.nokia.com/3619
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Dmytro Poplavskiy <dmytro.poplavskiy@nokia.com>
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
Moved controls specific bus/sync messages handling from
player/camera/capture session to corresponding controls.
Reviewed-by: Michael Goddard
Change-Id: Ieb67976ed335b0ef1cde87dc60e8ad8da3409526
Reviewed-on: http://codereview.qt.nokia.com/2535
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
Waiting for start() to be invoked in the main thread may block
if gstreamer blocks the main thread until this call is finished.
This situation is rare and usually caused by setState(Null)
while pipeline is being prerolled.
The proper solution to this involves controlling gstreamer pipeline
from other thread than video surface.
Currently start() fails if wait() timed out.
Task-number: QTMOBILITY-1663
Reviewed-by: Michael Goddard
Change-Id: Ib95e589a814e53efb9b4c454ef9f233658ff8c6a
(cherry picked from commit ac9762e2dbc06d696c6c74825ee22ac1fc176d9c)
Reviewed-on: http://codereview.qt.nokia.com/2073
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
The surface format stride calculation in directshow backend is wrong which results in memory access violation.
Change-Id: I80da5affc9a727513bad9c8d74a9f49d0c1a6c0d
Task-number:QTMOBILITY-1772
Reviewed-by:Michael Goddard
(cherry picked from commit 0b010e781634d3b33750fcead445fc7bd3a6f828)
Reviewed-on: http://codereview.qt.nokia.com/2070
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
Add the list of tests to sync.profile, and make sure each test only happens
on the platforms it's intended for.
Change-Id: Id022a6b85b349f4a8a649d3449b622e4ada7f4f4
Reviewed-on: http://codereview.qt.nokia.com/1910
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jonas Rabbe <jonas.rabbe@nokia.com>
qRound of an integer type never made sense.
Since 2dcd3939a8bd5ff743e4c87f87b2d81b1a101467 in qtbase, it won't
compile any more.
Change-Id: Ia30334bc92236d0fef9b6ac99d65bae4fe79dc57
Reviewed-on: http://codereview.qt.nokia.com/1905
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Michael Goddard <michael.goddard@nokia.com>
Docs still need some fixing, but there are some legitimate usages of
it there.
Change-Id: I94c7b68788ce97829711a4384884d7e52fe84162
Reviewed-on: http://codereview.qt.nokia.com/1736
Reviewed-by: Jonas Rabbe <jonas.rabbe@nokia.com>
When closing a QT7PlayerSession, it sets the movie to 0. There is no
need to check if a NULL movie supports the core pixel formats, and
that check lead to a segmentation fault with QtSDK 1.1.1.
Task-number: QTMOBILITY-1746
Reviewed-by: Dmytro Poplavskiy
(cherry picked from commit b5d9964aab96ef1a02ca1af411178f05c42c2bf8)
Change-Id: I7cdddbd8fc1b8d8d667747aec3567aeb857accd6
Reviewed-on: http://codereview.qt.nokia.com/984
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Michael Goddard
