Original commit message from CVS:
2004-01-25 Julien MOUTTE <julien@moutte.net>
* gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup),
(gst_play_identity_handoff), (gst_play_set_location),
(gst_play_set_visualization), (gst_play_connect_visualization): Another
try in visualization implementation. Still have an issue with switch
blocking when pulling from video_queue and only audio comes out of
spider.
* gst/switch/gstswitch.c: (gst_switch_release_pad),
(gst_switch_poll_sinkpads), (gst_switch_class_init): Implementing pad
release method. And check if the pad is usable before pulling.
Original commit message from CVS:
2004-01-23 Julien MOUTTE <julien@moutte.net>
* gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup),
(gst_play_set_location), (gst_play_seek_to_time),
(gst_play_set_audio_sink), (gst_play_set_visualization),
(gst_play_connect_visualization), (gst_play_get_sink_element): Reworked
the pipeline from scratch. Visualization is back and switch went out as
i realized it was not possible to use the way i wanted.
* sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear),
(gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext
clearing in state change from READY to NULL. So that one can clean the
X ressources keeping the element.
* sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get),
(gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state),
(gst_xvimagesink_colorbalance_set_value),
(gst_xvimagesink_colorbalance_get_value),
(gst_xvimagesink_set_property), (gst_xvimagesink_dispose),
(gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state
change from READY to NULL and fixed some stupid bugs in colorbalance
get/set values. Also added the following feature : when nobody tries to
set some values to the colorbalance levels before the xcontext is
grabbed, then when creating channels list from Xv attributes we set the
internal values to the Xv defaults. This way we handle buggy Xv drivers
that set default hue values far from the middle of the range (Thanks
to Jon Trowbridge for pointing that issue).
* sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if
colorbalance levels have been set before xcontext is grabbed.
Original commit message from CVS:
2004-01-14 Benjamin Otte <in7y118@public.uni-hamburg.de>
* gst-libs/gst/play/gstplay.c: (gst_play_tick_callback):
Query the audio element to get the time, not the clock. We're
interested in the element's time here.
Original commit message from CVS:
2004-01-12 Julien MOUTTE <julien@moutte.net>
* gst-libs/gst/play/gstplay.c: (gst_play_get_sink_element): When
analyzing the pads of an element the bin is mostly in READY state so
no caps were negotiated. This helper function needs to work with
_get_caps directly then. I was not freeing them though, added that to
fix the mem leak.
Original commit message from CVS:
Remove all usage of gst_pad_get_caps(), and replace it with
gst_pad_get_allowed_caps() or gst_pad_get_negotiated_cap().
Original commit message from CVS:
* examples/gstplay/Makefile.am: Adding the interface library.
* gst-libs/gst/play/Makefile.am: Adding the interface library.
* gst-libs/gst/play/gstplay.c: (gst_play_set_video_sink): Connecting tothe XOverlay size signal instead of GstVideoSink.
* gst-libs/gst/play/gstplay.h: Including the XOverlay interface to check GST_IS_X_OVERLAY before signal connect.
* gst-libs/gst/video/gstvideosink.c: (gst_videosink_class_init):
Removing the have_video_size signal.
* gst-libs/gst/video/gstvideosink.h: Removing the have_video_size signal and associated public method.
* sys/ximage/ximagesink.c: (gst_ximagesink_handle_xevents),
(gst_ximagesink_sinkconnect): Using XOverlay public method to fire size
signal.
* sys/xvideo/xvideosink.c: (gst_xvideosink_sinkconnect),
(gst_xvideosink_xwindow_new): Using XOverlay public method to fire size
signal.
* sys/xvimage/xvimagesink.c: (gst_xvimagesink_sinkconnect):
Using XOverlay public method to fire size signal.
Original commit message from CVS:
* examples/gstplay/player.c: (got_time_tick), (got_stream_length),
(got_video_size), (main): using g_print instead of g_message.
* gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup): Fixing EOS
signal which was not emitted because of "switch" element added to the
bin but not connected. (Removing from the bin temporarily)
Original commit message from CVS:
* configure.ac: X_DISPLAY_MISSING is set to 1 if AC_PATH_XTRA fails to
find X development files. I don't understand the previous tests and
they fail on my debian/ppc unstable. This one works.
* examples/gstplay/player.c: (main): Set the pipeline to READY before
exiting.
* gst-libs/gst/play/gstplay.c: (gst_play_get_length_callback),
(gst_play_set_video_sink), (gst_play_set_audio_sink),
(gst_play_set_visualization): Add some safety checks in set_ methods
and state_change. This was throwing some ugly CRITICAL messages when
pipeline was getting disposed and casts were failing.
Original commit message from CVS:
Remove all config.h includes from header files, add it to each source file and remove duplicate config.h includes from several source files
Original commit message from CVS:
There was a floating doubt on the way visualization is written. We use 'z' everywhere now. Both in gst-player and libgstplay
Original commit message from CVS:
Changing have_xid to have_video_out to match the new GstVideoSink parent object.
Adding have_vis_size.
Fixing gst_play_connect_visualisation
Original commit message from CVS:
Fix another instance of audio/raw -> audio/x-raw-int and
video/raw -> video/x-raw-{yuv,rgb}. There probably should
be a audio/x-raw-float for jacksink but I didn't add that.
Original commit message from CVS:
Fixing error message when a plugin is missing (gst-player -> libgstplay).
Pipeline state change on error is now done directly in error callback, there's still something weird with oss /dev/dsp ressource busy error.
Original commit message from CVS:
Error handling is inside and works :)
Mutexes have been removed. That should fix UI responsiveness problems on query_length. Please heavily test to check if they were really needed.
Some fixes in GClosures
Original commit message from CVS:
Make vis_video_thread play when connected afterwards
This generates some segfaults in gst_thread but we need to fix that
Original commit message from CVS:
Made a theorical libgstplay which refs/unrefs elements before putting them in AsyncQueue.
Added a "pipeline_error" signal which will later allow the player and apps to detect that pipeline was unable to play and why...
This version is NOT STABLE AT ALL. it will need fixes in core but i commit it as is so that we fix those problems
Original commit message from CVS:
make Company happy : Changed visualisation pipeline structure.. audio sink is directly connected to tee so no queue between volume and audio sink...
Original commit message from CVS:
Incorporating visualisation in libgstplay
adding have_vis_xid signal and a new type of GstPlay object (VIDEO_VISUALISATION)
Original commit message from CVS:
gst_play_get_sink_element has been improved :
- Recursively searches in bins
- Able to find sink_element of a specified GstPlaySinkType (audio,video,any)
This now allow us to send very complex bins to GstPlay for audio/video sink without breaking length querying, eos signaling, etc..
Original commit message from CVS:
Code cleanup and reorganization
I ll now be able to make nice gtk-doc for this lib
Fixed latency bug on async_queue pop
X video window now comes immediatly
Original commit message from CVS:
added func gst_play_set_data_src
removed pipeline GST_PLAY_PIPE_VIDEO_THREADSAFE - it never worked and its just taking up space
Original commit message from CVS:
another batch of connect->link fixes
please let me know about issues
and please refrain of making them yourself, so that I don't spend double
the time resolving conflicts
Original commit message from CVS:
Fixed gst_play_dispose function so that GstPlay objects can be unrefed safely.
Optimized some g_idle_add for the signal poller.
