Commit by kelvie :: r1082522 phonon/trunk/kdesupport/phonon/xine/xinestream.cpp:

Phonon: check get_xine_stream_text's return value

Instead of returning trash, return an empty string on error.

Commit by kelvie :: r1082521 phonon/trunk/kdesupport/phonon/xine/xinestream.cpp:

Ask Xine for the current time after seek.

Commit by cguthrie :: r1082383 phonon/trunk/kdesupport/phonon/phonon/audiooutput.cpp:

phonon: Ensure the backend object exists before we use GlobalConfig().

As the backends now explicitly enable PA support on initialisation
we cannot use GlobalConfig() until our backend is initialised.

While this is a rather annoying prerequisit, it doesn't seem to
cause any problem in the kdebase-runtime code which also
uses GlobalConfig().

While this raises some doubt in my mind about how well this
all holds together I think that defaulting to not supporting
PA in the backend allows for safer development of backends
where unexpected PA integration in phonon itself is not triggered
unless the backend specifically says to do so.

Commit by cguthrie :: r1082382 phonon/trunk/kdesupport/phonon/ (5 files in 3 dirs):

pulsesupport: Make the backends explicitly state that they want PA support.

Now that we ask if pulse isactive and bypass asking the backends
for device lists etc if so, we cannot just assume that PA is supported
in the backend itself. So ensure that we enable pulseaudio support
when initialising the backends.

This was not needed before because we used to always go via the backend
for device information, even if it was just proxying on to the
PulseSupport class. This way it's more efficient and less
work to integrate into a backend.

Commit by cguthrie :: r1082381 phonon/trunk/kdesupport/phonon/gstreamer/backend.cpp:

gstreamer: Remove unnecessary PulseSupport calls now that the library no longer asks us for this when PA is active.

Commit by cguthrie :: r1082380 phonon/trunk/kdesupport/phonon/xine/backend.cpp:

xine: Remove unnecessary PulseSupport calls now that the library no longer asks us for this when PA is active.

Commit by cguthrie :: r1082379 phonon/trunk/kdesupport/phonon/phonon/ (globalconfig.cpp objectdescription.cpp):

pulse: Intercept the objectDescription(Indexes|Properties) calls prior to getting to the backend

At present the xine and gstreamer backends both check for and return results from PA
for these methods. It would be cleaner (and less disruptive on the backends) if
we simply didn't call them when we know PA is active.

This commit does this check. It will be followed by cleanup commits in
both xine and gstreamer backends to remove the unneeded checks/calls that
result from this change.

Commit by ingmarvanhassel :: r1081485 phonon/trunk/kdesupport/phonon/xine/CMakeLists.txt:

Update xine homepage

Commit by rkcosta :: r1079077 phonon/trunk/kdesupport/phonon/phonon/CMakeLists.txt:

Only call macro_ensure_version if PulseAudio was found.

Commit by cguthrie :: r1079070 phonon/trunk/kdesupport/phonon/gstreamer/backend.cpp:

gstreamer: Add some extra mime-types for ogg support.

This is a patch I wrote a while ago and have been carrying
in Mandriva packages. It was originally written to allow
the backend to claim it supports a more types relating to ogg
than those that come directly from gstreamer. This stems from
the fact that several different mime types seem to have become
commonly used for Ogg Vorbis.

Commit by cguthrie :: r1079067 phonon/trunk/kdesupport/phonon/ (11 files in 3 dirs):

pulsesupport: Make the PulseSupport header public/installed.

In order to use pulsesupport to build backends out of tree (e.g. vlc backend)
this header needs to be exported.

Discussed with Martin Sandsmark on #phonon IRC

Commit by cguthrie :: r1079059 phonon/trunk/kdesupport/phonon/phonon/ (CMakeLists.txt pulsesupport.cpp):

pulsesupport: Add support for older PA versions sans module-device-manager.

This patch was originally written by Harald Fernengel and tweaked by
Rex Dieter to work with version numbers.

I've shuffled a few things around since Haralds initial patch,
renaming a function to be more appropriate and adding a few more
connection log messages. Also moved where some of the code was
executed, but this doesn't change the semantics of the original
patch.

The lowest version is currently set to 0.9.15 but it should
work with earlier versions too. If some one tests this I'd
be happy to lower it, but not got older versions available
to test right now.

Commit by cguthrie :: r1078262 phonon/trunk/kdesupport/phonon/xine/backend.cpp:

xine: Don't list the PulseAudio output plugin when not in PA mode.

We do a specific check for PA support earlier in the code and if we've
already checked that PA is not running there is very little point in showing
it as an option anyway.

CCBUG: 213351

Commit by cguthrie :: r1078226 phonon/trunk/kdesupport/phonon/xine/ (audiooutput.cpp backend.cpp):

xine: Always ensure we create the correct xine port for audio output.

When integrating with pulseaudio it was possible that the port was not correctly created.
This could happen if phonon was started and the first device detected from PA was a capture device
rather than a playback device (which is unlikely) or if the first device detected
was a removable audio device which was subsequently removed before playback started.

There are possibly other cases where this could happen too.
In this case xine would create a null port.

This change reverts a previous 'fix' committed in r1051721. I now issue a special case
for dealing with PA support in the createPort() method. As a result I no longer need
to special case handle the audioDriverFor() method of the backend so this is
now removed.

CCBUG: 223662

Commit by cguthrie :: r1078189 phonon/trunk/kdesupport/phonon/gstreamer/ (mediaobject.cpp mediaobject.h):

gstreamer: Move the invalidateGraph() method from the header in the cpp file.

While debugging issues relating to m_resetAsNeeded, I found it confusing that this function
was implemented in the header. As it does more than a simple accessor method, I think
it belongs in the main cpp file.

Commit by cguthrie :: r1078188 phonon/trunk/kdesupport/phonon/gstreamer/mediaobject.cpp:

gstreamer: Fix a problem encountered when playing, stopping and playing again.

As reported by Harald Fernengel, a similar problem to that fixed in r1076454
also exists when you play->stop->play a media object.

This should fix the issue by marking a reset needed whenever we reach the stopped state.
Thanks to Harald for finding the problem and pointing the way to the fix.

Commit by cguthrie :: r1076455 phonon/trunk/kdesupport/phonon/gstreamer/mediaobject.cpp:

gstreamer: in setSource() m_loading is always true, so tidy code and remove unnecessary setting of this variable

Commit by cguthrie :: r1076454 phonon/trunk/kdesupport/phonon/gstreamer/mediaobject.cpp:

gstreamer: Fix a problem where the fact a reset was needed was lost.

If you create the path (createPath()) *before* setting the media source,
(setCurrentSource()) the gstreamer backend would forget the fact that a reset
was needed and ultimately end up in an error state.

This change simply does not wipe out the m_resetNeeded flag when
the source is set and leaves it as it is.

This fixes the test case application posted on
https://qa.mandriva.com/show_bug.cgi?id=56807

Commit by cguthrie :: r1076453 phonon/trunk/kdesupport/phonon/gstreamer/mediaobject.cpp:

gstreamer: Clarify state change by not checking someone we know to be true.

Commit by jmthomas :: r1072693 phonon/trunk/kdesupport/phonon/ (2 files in 2 dirs):

Use the QString.clear() method rather than assigning a null QString
Fixes some Krazy warnings