Original commit message from CVS:
* gst/gst.defs:
parse_launch also owns the return ref
* gst/gstbin.override:
fix a leak in get_by_name()
* testsuite/common.py:
loop when garbage collecting; necessary when deleting, say,
a pipeline, that contains elements, to ensure the elements
also get collected
* testsuite/test_element.py:
* testsuite/test_pipeline.py:
test for garbage collection
Original commit message from CVS:
* gst/Makefile.am:
* gst/common.h:
* gst/pygstobject.c: (pygstobject_sink), (pygstobject_new),
(pygst_object_unref):
* gst/pygstobject.h:
* gst/gstmodule.c: (init_gst):
add/move some code to handle wrapping/refcounting of possible
GstObject
* codegen/argtypes.py:
* gst/gstbin.override:
* gst/gstbus.override:
* gst/gstelement.override:
* gst/gstpad.override:
* gst/interfaces.override:
use this reffing code
* gst/gst-types.defs:
* gst/gst.override:
add a __gstrefcount__ field to GstObject types
add tp_traverse, tp_dealloc and tp_clear, so we handle refcounting
properly related to garbage collection
* testsuite/test_element.py:
* testsuite/test_pad.py:
add more tests, add some refcount checks
Original commit message from CVS:
* gst/__init__.py:
require pygtk 2.0 when importing gst
it only works with 2.0, and it sets up the path correctly
when running against a pygtk compiled into a different prefix
Original commit message from CVS:
* gst/gst.defs:
* gst/gst.override:
add another registry method
* testsuite/Makefile.am:
add check-verbose
* testsuite/runtests.py:
check for VERBOSE env var
* testsuite/test_pad.py:
refactor a test and add one more for probes
Original commit message from CVS:
* gst/gstbuffer.override:
Increment the refcount when fetching GST_BUFFER_CAPS.
Python gurus, please fix if there's a better way this should be done.
Original commit message from CVS:
2005-09-21 Andy Wingo <wingo@pobox.com>
* gst/gst.defs: Update for probe removal api
* gst/gstpad.override (_wrap_gst_pad_remove_data_probe)
(_wrap_gst_pad_remove_event_probe)
(_wrap_gst_pad_remove_buffer_probe): Removed, upstream does the
right thing now.
(_wrap_gst_pad_add_data_probe, _wrap_gst_pad_add_event_probe)
(_wrap_gst_pad_add_buffer_probe): Indent.
Original commit message from CVS:
* gst/gst.defs:
* gst/gststructure.override:
* testsuite/test_struct.py:
add a .keys() method to gst.Structure, plus test
Original commit message from CVS:
2005-09-16 Andy Wingo <wingo@pobox.com>
* gst/gst.defs (registry_get_default): Renamed from
gst_registry_get_default.
Original commit message from CVS:
2005-09-15 Andy Wingo <wingo@pobox.com>
* gst/gst-types.defs:
* gst/gst.defs: Update for registry/plugin/pluginfeature API changes.
Original commit message from CVS:
* examples/vumeter.py:
update for property name change
* gst/gst.override:
* testsuite/test_element.py:
actually log on objects, yay
Original commit message from CVS:
* gst/gst.defs:
add gst.ghost_pad_new_notarget
* gst/gstpad.override:
change wrap_gst_pad_new a little and add logging
* testsuite/test_pad.py:
add tests for constructors of gst.Pad
Original commit message from CVS:
* gst/gst-types.defs:
* gst/gstbuffer.override:
* gst/pygstminiobject.c: (pygstminiobject_new_noref):
* testsuite/test_buffer.py:
add a .data to buffers; it feels more natural to me than doing
str(buffer) compared to the other properties.
make sub_buffer test for data
Original commit message from CVS:
fix a race condition in test_buffer.py
* gst/gst.override:
* gst/gstmodule.c: (init_gst):
add a pygst debug category for bindings themselves to use
* gst/gstbuffer.override:
add a repr method; add some assertions
* gst/pygstminiobject.c: (pygst_miniobject_init),
(pygstminiobject_register_wrapper), (pygstminiobject_new),
(pygstminiobject_new_noref), (pygstminiobject_dealloc),
(pygstminiobject_clear):
make the miniobjs hash private with an underscore
add debugging for inserting/removal in hash
fix pygstminiobject_clear - it also needs to remove
from the global hash. Fixes a nasty race problem in
test_buffer
* testsuite/test_buffer.py:
expand on the subbuffer test
Original commit message from CVS:
* gst/__init__.py:
del ltihooks properly; fixes loading of testhelper in testsuite
* gst/gst.defs:
wrap link_filtered
Original commit message from CVS:
2005-08-19 Andy Wingo <wingo@pobox.com>
* gst/interfaces.defs (MixerTrack): MixerTrack is a GObject. How
did this ever work before?
Original commit message from CVS:
2005-08-19 Andy Wingo <wingo@pobox.com>
* gst/interfaces.override (_wrap_gst_tuner_list_channels)
(_wrap_gst_tuner_list_norms, _wrap_gst_mixer_list_tracks)
(_wrap_gst_color_balance_list_channels): Catch programming errors
without segfaults, as glib would do.
Original commit message from CVS:
* gst/gstmodule.c: (init_gst):
changed debug category name from 'gst-python' to 'python'
* gst/gst.override: (pygst_debug_log):
gchar is way smaller than an int... resulting in negative line numbers
in debug.
Original commit message from CVS:
* gst/gst.override: (pygst_debug_log):
gchar is way smaller than an int... resulting in negative line numbers
in debug.
Original commit message from CVS:
* gst/gst.override: (pygst_debug_log):
printf("urlencoded%20space") != printf("%s", "urlencoded%20space")
Funny how a lot of problems have gone since that's fixed :)
Original commit message from CVS:
* gst/gst.defs:
* gst/gst.override:
* gst/gstmodule.c:
We now have debugging in gst-python ;)
gst.[log|debug|info|warning|error]()
Original commit message from CVS:
2005-08-03 Andy Wingo <wingo@pobox.com>
* gst/gstmessage.override (_wrap_gst_message_parse_error): Always
return a tuple of length 2.
Original commit message from CVS:
* gst/gst-types.defs: (GhostPad):
Added GhostPad type
* gst/interfaces.defs:
Added MixerTrackFlags, TunerChannelFlags and MixerOptions
Want the rest of the poem ? ;)
Original commit message from CVS:
* configure.ac:
* gst/Makefile.am:
* gst/interfaces.defs:
* gst/interfaces.override:
* gst/xoverlay.override:
It seems that, somewhere along the way, I forgot to make sure interfaces
were binded too :)
