Mark Nauwelaerts [Fri, 2 Dec 2011 12:19:38 +0000 (13:19 +0100)]
identity: unlock clock wait when appropriate
... notably FLUSH and state change to READY.
Stefan Sauer [Thu, 1 Dec 2011 15:25:07 +0000 (16:25 +0100)]
bus: use GST_MESSAGE_SOURCE_NAME() which also takes care of src=NULL.
Sebastian Dröge [Wed, 30 Nov 2011 12:59:46 +0000 (13:59 +0100)]
basetransform: Always intersect the suggested sink caps with the peer caps
This makes sure that we get correct and complete caps. The suggested caps
could be incomplete, e.g. video/x-raw-rgb without any fields, and by
intersecting with the peer caps we get something usable.
Fixes bug #662199.
Tim-Philipp Müller [Wed, 30 Nov 2011 00:24:32 +0000 (00:24 +0000)]
event: warn and fail instead of creating newsegment events in GST_FORMAT_UNDEFINED
Vincent Penquerc'h [Mon, 28 Nov 2011 12:30:15 +0000 (12:30 +0000)]
gstelement: add gst_element_class_add_pad_template_from_static
This function helps ensure the pad template is unreffed
without having to complicate the calling code.
https://bugzilla.gnome.org/show_bug.cgi?id=662664
Tim-Philipp Müller [Mon, 28 Nov 2011 01:10:20 +0000 (01:10 +0000)]
queue2: fix refactoring of draining-on-eos, munge flow return to FLOW_OK
Tim-Philipp Müller [Thu, 3 Nov 2011 10:34:49 +0000 (10:34 +0000)]
queue2: add bufferlist support
We want to maintain buffer lists if possible.
Tim-Philipp Müller [Thu, 3 Nov 2011 13:02:36 +0000 (13:02 +0000)]
queue2: split out draining of queue on FLOW_UNEXPECTED into separate function
Tim-Philipp Müller [Thu, 3 Nov 2011 08:55:20 +0000 (08:55 +0000)]
queue2: pass item type enum to _enqueue instead of simple isbuffer boolean
Avoids some unnecessary GST_IS_EVENT()
Stefan Sauer [Sun, 27 Nov 2011 19:36:31 +0000 (20:36 +0100)]
warnings: avoid set-but-unused warnings with load-save disabled
Tim-Philipp Müller [Sat, 26 Nov 2011 19:45:48 +0000 (19:45 +0000)]
uri: fix wrong G_GNUC_MALLOC
_get_protocols() points to const memory in 0.10
despite the non-const return value.
Tim-Philipp Müller [Sat, 26 Nov 2011 18:58:38 +0000 (18:58 +0000)]
libs: sprinkle some G_GNUC_MALLOC
Maybe gcc can do something clever with that, or at least
warn us if we don't save the return value somewhere.
Tim-Philipp Müller [Sat, 26 Nov 2011 18:57:44 +0000 (18:57 +0000)]
gst: sprinkle some G_GNUC_MALLOC
Maybe gcc can do something clever with that, or at least
warn us if we don't save the return value somewhere.
Tim-Philipp Müller [Wed, 16 Nov 2011 01:04:45 +0000 (01:04 +0000)]
buffer: add gst_buffer_{set,get}_qdata()
Allows people/us to attach arbitrary metadata to buffers.
https://bugzilla.gnome.org/show_bug.cgi?id=664720
API: gst_buffer_set_qdata()
API: get_buffer_get_qdata()
Sebastian Dröge [Thu, 24 Nov 2011 10:15:29 +0000 (11:15 +0100)]
pad: Remove g_warning() if pad accepted caps that are not a subset of the pad caps
This check is correct but unfortunately it's impossible to implement
in a threadsafe way because the caps could have changed in the meantime.
Fixes bug #659606.
Vincent Penquerc'h [Mon, 3 Oct 2011 11:34:20 +0000 (12:34 +0100)]
multiqueue: check filled state of queues even if another one is empty
This will avoid a case where overrun is never signalled if some
stream never produces any data, causing playbin2 to not end preroll.
https://bugzilla.gnome.org/show_bug.cgi?id=660778
Sebastian Dröge [Wed, 23 Nov 2011 12:42:56 +0000 (13:42 +0100)]
baseparse: Return template caps instead of other side's peer caps if get_sink_caps vfunc is not implemented
Using gst_pad_proxy_get_caps() breaks backwards compatibility with old
parsers because it will propagate the other side's fields like "parsed"
and "framed" and also breaks parser/converters.
Fixes bug #664221.
Tim-Philipp Müller [Sat, 12 Nov 2011 16:42:14 +0000 (16:42 +0000)]
gst-inspect: print current value as default value
Instead of printing separate 'Current' and 'Default' values
(the former obtained via g_object_get() and the latter from
the property GParamSpec), simply print the Current value as
the Default value. This is the right thing to do for almost
all elements and avoids confusion if a subclass of a base
class chooses a different default than the base class.
Tim-Philipp Müller [Sat, 12 Nov 2011 14:55:07 +0000 (14:55 +0000)]
gst, controller: replace g_list_prepend + reverse with GQueue
Stefan Sauer [Wed, 9 Nov 2011 10:05:59 +0000 (11:05 +0100)]
Android: build audio controller example
Add buildsystem hooks for building the audiocontroller example
with the NDK.
Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@collabora.com>
Vincent Penquerc'h [Tue, 8 Nov 2011 18:09:28 +0000 (18:09 +0000)]
gstvalue: consider lists and ranges equal if they hold the same set
This fixes caps operations when different elements advertise some
of their caps' properties differently (eg, for audio channels, either
a range from 1 to 2, or a list of 1 and 2).
https://bugzilla.gnome.org/show_bug.cgi?id=663643
Stefan Sauer [Wed, 9 Nov 2011 10:24:26 +0000 (11:24 +0100)]
tests: add a subset test for structure
Stefan Sauer [Fri, 28 Oct 2011 11:39:58 +0000 (13:39 +0200)]
docs: also fix wrong call order for controller in manual
René Stadler [Tue, 8 Nov 2011 11:22:21 +0000 (12:22 +0100)]
tests: pad: add test to verify flushing behaviour
Seems like a trivial case, but this was actually broken in 0.11 recently.
Tim-Philipp Müller [Tue, 8 Nov 2011 00:32:36 +0000 (00:32 +0000)]
docs: fix Since: markers for API added after 0.10.35
Tim-Philipp Müller [Tue, 8 Nov 2011 00:15:49 +0000 (00:15 +0000)]
win32: update .def file for new API
API: gst_caps_is_strictly_equal()
Vincent Penquerc'h [Thu, 27 Oct 2011 11:59:57 +0000 (12:59 +0100)]
gstcaps: remove unneeded use of gint64
https://bugzilla.gnome.org/show_bug.cgi?id=662777
Vincent Penquerc'h [Thu, 27 Oct 2011 11:24:13 +0000 (12:24 +0100)]
gstvalue: quicker version of intersection when we do not need the result
https://bugzilla.gnome.org/show_bug.cgi?id=662777
Vincent Penquerc'h [Thu, 27 Oct 2011 11:02:43 +0000 (12:02 +0100)]
gststructure: simplify return statement in gst_structure_can_intersect
https://bugzilla.gnome.org/show_bug.cgi?id=662777
Vincent Penquerc'h [Thu, 27 Oct 2011 10:41:30 +0000 (11:41 +0100)]
gststructure: early out when we know a value cannot be a subset
If two values can be ordered, but are unequal, they are
necessarily distinct, thus one cannot be a subset of the other.
https://bugzilla.gnome.org/show_bug.cgi?id=662777
Vincent Penquerc'h [Thu, 27 Oct 2011 09:35:53 +0000 (10:35 +0100)]
gstvalue: quicker test for substraction emptiness
When we do not care about the actual resulting set,
but only whether it is empty of not, we can skip a fair bit
of GValue juggling.
Add a function that does so, since we cannot just pass NULL
to the existing API as it may be part of the API contract.
https://bugzilla.gnome.org/show_bug.cgi?id=662777
Vincent Penquerc'h [Thu, 27 Oct 2011 08:45:41 +0000 (09:45 +0100)]
gststructure: rejig test ordering for speed
https://bugzilla.gnome.org/show_bug.cgi?id=662777
Sjoerd Simons [Thu, 3 Nov 2011 15:36:59 +0000 (15:36 +0000)]
basetransform: Only use the cached transform on strictly equal caps
https://bugzilla.gnome.org/show_bug.cgi?id=663333
Sjoerd Simons [Thu, 3 Nov 2011 15:35:32 +0000 (15:35 +0000)]
caps: Add gst_caps_is_strictly_equal
Peteris Krisjanis [Sat, 15 Oct 2011 19:52:25 +0000 (22:52 +0300)]
introspection: add Value annotations for GST_SECOND, GST_MSECOND, GST_USECOND, GST_NSECOND constants
gobject-introspection won't parse them properly otherwise.
Still need to force the right type though (either GstClockTime or
guint64), but Type: xyz has no effect for me here, so someone with
a newer g-i needs to test this.
Some other defines are also missing, e.g. GST_CLOCK_TIME_NONE.
Tim-Philipp Müller [Fri, 4 Nov 2011 00:03:05 +0000 (00:03 +0000)]
configure: suppress unused variable warnings if gst debugging is disabled
https://bugzilla.gnome.org/show_bug.cgi?id=662952
Tim-Philipp Müller [Fri, 4 Nov 2011 00:02:07 +0000 (00:02 +0000)]
configure: fix typo around GLIB_EXTRA_CFLAGS in GST_ALL_CXXFLAGS
Tim-Philipp Müller [Thu, 3 Nov 2011 23:08:52 +0000 (23:08 +0000)]
queue: use statically allocated GQueue
Tim-Philipp Müller [Thu, 3 Nov 2011 22:58:50 +0000 (22:58 +0000)]
bin: use statically allocated GQueue
Because we can.
Tim-Philipp Müller [Thu, 3 Nov 2011 22:51:57 +0000 (22:51 +0000)]
collectpads2: use flags enum instead of guint in structure
Tim-Philipp Müller [Thu, 3 Nov 2011 08:47:20 +0000 (08:47 +0000)]
queue2: use statically allocated GQueue
Nicolas Dufresne [Tue, 25 Oct 2011 21:26:50 +0000 (17:26 -0400)]
ghostpad: Don't cache internal proxy pad target
The internal proxy pad target is simply a cache of the internal proxy pad
peer. This patch uses the well implement GstPad peer handling to obtain the
target. This fixes issues with target not being set in both direction when
two ghostpads are linked together (empty bin).
https://bugzilla.gnome.org/show_bug.cgi?id=658517
Tim-Philipp Müller [Tue, 1 Nov 2011 14:17:21 +0000 (14:17 +0000)]
adapter: simplify gst_adapter_take_list()
Use a stack-allocated GQueue to assemble our GList.
Tim-Philipp Müller [Mon, 31 Oct 2011 17:45:44 +0000 (17:45 +0000)]
configure: make GLIB_EXTRA_CFLAGS overwritable
Make 'make GLIB_EXTRA_CFLAGS=...' work.
Tim-Philipp Müller [Sun, 30 Oct 2011 21:54:36 +0000 (21:54 +0000)]
filter: deprecate gst_filter_run()
It's not really used outside of core at all, and has
serious namespace issues. If anyone feels the need to
revive this one, please use a less generic name space.
API: deprecate gst_filter_run()
API: deprecate GstFilterFunc
Tim-Philipp Müller [Sun, 30 Oct 2011 21:39:36 +0000 (21:39 +0000)]
registry: don't use soon-to-be-deprecated gst_filter_run()
Lines-of-code savings are negligible anyway.
Tim-Philipp Müller [Sun, 30 Oct 2011 21:21:59 +0000 (21:21 +0000)]
pluginfeature: deprecate gst_plugin_feature_type_name_filter()
It's only used internally anyway and the helper struct
has namespace issues.
API: deprecated gst_plugin_feature_type_name_filter()
API: deprecated GstTypeNameData
Tim-Philipp Müller [Sun, 30 Oct 2011 09:58:16 +0000 (09:58 +0000)]
taglist: add to_string and new_from_string functions
We want to make GstTagList opaque and not have people use
GstStructure API on it.
API: gst_tag_list_to_string()
API: gst_tag_list_new_from_string()
Tim-Philipp Müller [Sun, 30 Oct 2011 00:46:22 +0000 (01:46 +0100)]
taglist: avoid pointless tag name -> quark lookups
We never get a tag name quark from a caller, it's always a
string, from which we'll try to look up our tag info in the
hash table, so change the hash table key from quark to string.
Avoids a bunch of pointless string => quark lookup in the
global quark table. We need to do an extra string => quark
conversion now when we copy a taglist, but in that case we're
in a slow path anyway.
Tim-Philipp Müller [Sat, 29 Oct 2011 23:44:44 +0000 (00:44 +0100)]
taglist: add gst_tag_list_is_equal()
API: gst_tag_list_is_equal()
Tim-Philipp Müller [Sat, 29 Oct 2011 22:52:22 +0000 (23:52 +0100)]
structure: identical structures are definitely equal
Sjoerd Simons [Fri, 28 Oct 2011 09:53:32 +0000 (11:53 +0200)]
basetransform: Also fush the cache when changing the upstream caps suggestion
Sjoerd Simons [Fri, 28 Oct 2011 09:50:23 +0000 (11:50 +0200)]
basetransform: Add debug output when returning a cached transform
Sebastian Dröge [Fri, 28 Oct 2011 08:54:19 +0000 (10:54 +0200)]
collectpads2: Fix refcount handling if a buffer was dropped due to clipping
Sebastian Dröge [Fri, 28 Oct 2011 08:37:21 +0000 (10:37 +0200)]
collectpads2: Merge the clip and prepare_buffer function into one
Sebastian Dröge [Fri, 28 Oct 2011 08:17:06 +0000 (10:17 +0200)]
collectpads2: Merge clipping API from old collectpads
Tim-Philipp Müller [Fri, 28 Oct 2011 08:19:29 +0000 (09:19 +0100)]
basesink: make default query function name show up in gst-inspect
Mark Nauwelaerts [Tue, 27 Sep 2011 13:48:52 +0000 (15:48 +0200)]
collectpads2: avoid hanging in case of sparse newsegment events
... in the extent that a non-waiting pad (so indicated by newsegment)
turns out to provide the best buffer, which is then forced to waiting
for book-keeping purposes, but that should only be temporary.
See bug #415754.
Sebastian Dröge [Fri, 28 Oct 2011 07:38:35 +0000 (09:38 +0200)]
collectpads2: Use G_DEFINE_TYPE instead of GST_BOILERPLATE
Sebastian Dröge [Fri, 28 Oct 2011 07:35:50 +0000 (09:35 +0200)]
collectpads2: Add to the documentation
Sebastian Dröge [Fri, 28 Oct 2011 07:26:11 +0000 (09:26 +0200)]
win32: Add new collectpads2 API
Sebastian Dröge [Fri, 28 Oct 2011 07:18:55 +0000 (09:18 +0200)]
base: Add collectpads2
This handles muxing of sparse/subtitle streams and has
lots of cleanup. Still missing is special support for
live streams but this can be added later without breaking
API/ABI.
Based on the version from the videomixer plugin.
https://bugzilla.gnome.org/show_bug.cgi?id=415754
Jan Schmidt [Thu, 27 Oct 2011 12:08:14 +0000 (23:08 +1100)]
basetransform: Fix refcount leak
Don't leak peercaps and a ref to the basetransform when returning
the cached caps.
Stefan Sauer [Thu, 27 Oct 2011 08:38:10 +0000 (10:38 +0200)]
controller: fix wrong order of calls in the docs.
Stefan Sauer [Wed, 26 Oct 2011 10:23:07 +0000 (12:23 +0200)]
event: clarify docs for step event
Tim-Philipp Müller [Tue, 18 Oct 2011 22:19:47 +0000 (23:19 +0100)]
registry: add support for GST_REGISTRY_REUSE_PLUGIN_SCANNER=no
This will make sure we spawn a new plugin scanner helper for each plugin
to be introspected, which helps with making sure we don't load too many
shared objects (libs, plugins) at the same time on systems where there
is a hard limit like on Android.
A better version might re-use the scanner for up to N times, though
it's not clear whether that would actually improve things dramatically.
https://bugzilla.gnome.org/show_bug.cgi?id=662091
Thiago Santos [Fri, 21 Oct 2011 13:23:04 +0000 (10:23 -0300)]
caps: No need to require writable caps for _append/_merge second caps
The second caps ownership is transfered, no need to require it to
be writable from the caller function. Instead, _append and _merge
make it writable on their own.
Discovered because of an assertion on encoding-profile.c in
_get_input_caps using _merge but not passing writable caps.
Sebastian Dröge [Thu, 20 Oct 2011 14:59:01 +0000 (16:59 +0200)]
baseparse: Fix documentation, it's pre_push_frame and not pre_push_buffer
Vincent Penquerc'h [Thu, 20 Oct 2011 13:02:44 +0000 (14:02 +0100)]
basetransform: cache transformed caps where appropriate
Speeds up negotiation a fair bit on a contrived pipeline
with a dozen colorspace conversions.
Hopefully clears out the cache every time it ought to.
https://bugzilla.gnome.org/show_bug.cgi?id=662291
Reynaldo H. Verdejo Pinochet [Thu, 20 Oct 2011 13:45:30 +0000 (10:45 -0300)]
Fix broken android building due to name change in
01d87250a845e55d6
Sebastian Dröge [Thu, 20 Oct 2011 06:31:18 +0000 (08:31 +0200)]
basebarse: Add detect vfunc to allow subclasses to do format detection before anything else
API: GstBaseParseClass::detect()
This is called with the first buffers until the subclass has finished detection
and only afterwards the original buffers are handled as before. The vfunc allows
detection of the stream format without breaking the upstream framing.
Stefan Sauer [Thu, 20 Oct 2011 06:57:57 +0000 (08:57 +0200)]
pad: not only describe conditions in the docs, also check them in the code
When blocking pads, check if the pad is in the rigt direction. Log some info
for the developer and return FALSE, instead of just locking up.
Thiago Santos [Tue, 18 Oct 2011 22:21:58 +0000 (19:21 -0300)]
message: protect against null message sources
Message sources can be null, check if it is before trying to
access its name.
Thiago Santos [Tue, 18 Oct 2011 11:48:37 +0000 (08:48 -0300)]
baseparse: add since doc to new getcaps function
Stefan Sauer [Tue, 18 Oct 2011 12:08:19 +0000 (14:08 +0200)]
logging: more logging and prefer human readable details over memory locations
Stefan Sauer [Tue, 18 Oct 2011 12:05:37 +0000 (14:05 +0200)]
debugutils: improve display of ghost- and proxypads
Handle virtual links between ghost and proxypads when iterating pads instead of
when linking. Besides using less code this provides a more accurate picture.
Thiago Santos [Mon, 17 Oct 2011 17:42:08 +0000 (14:42 -0300)]
baseparse: add getcaps function
Adds a getcaps function to the sink pad to make parsers propagate
downstream caps restrictions to upstream.
The pipeline "audiotestsrc num-buffers=100 ! faac ! aacparse !
"audio/mpeg, version=(int)4, stream-format=(string)adts" ! filesink"
wouldn't work because aacparse wouldn't propagate the adts restriction
upstream to faac.
This patch adds a default getcaps to the sink pad to simply proxy
downstream caps and also adds a 'get_sink_caps' function pointer
to GstBaseParseClass for subclasses that need more refined getcaps.
https://bugzilla.gnome.org/show_bug.cgi?id=661874
Wim Taymans [Tue, 18 Oct 2011 10:39:03 +0000 (12:39 +0200)]
basesrc: also update the stream time
Sebastian Dröge [Tue, 18 Oct 2011 08:58:14 +0000 (10:58 +0200)]
baseparse: Fix handling of queued frames
gst_base_parse_push_frame() already frees the frame, no need to
do it another time again.
Tim-Philipp Müller [Sun, 16 Oct 2011 20:12:07 +0000 (21:12 +0100)]
bus: give watch source a name
Give our GSource a meaningful name. Source names can be
used for debugging and profiling, for example with
systemtap or gdb.
Edward Hervey [Fri, 14 Oct 2011 07:35:09 +0000 (09:35 +0200)]
gstbuffer: Add transfer annotations for gst_buffer_join()
Stefan Sauer [Thu, 13 Oct 2011 15:33:06 +0000 (17:33 +0200)]
bus: fix typo in the docs
Stefan Sauer [Thu, 13 Oct 2011 14:48:02 +0000 (16:48 +0200)]
debugutils: show if an element is state-locked
Stefan Sauer [Thu, 13 Oct 2011 14:42:10 +0000 (16:42 +0200)]
logging: use _OBJECT variants more
Sebastian Dröge [Thu, 13 Oct 2011 08:19:34 +0000 (10:19 +0200)]
buffer: Use an inline function instead of a macro for gst_buffer_replace()
This gives us type checks by the compiler and more useful compiler errors.
Sebastian Dröge [Thu, 13 Oct 2011 08:18:51 +0000 (10:18 +0200)]
event: Use an inline function instead of a macro for gst_event_replace()
This gives us type checks by the compiler and more useful compiler errors.
Wim Taymans [Wed, 12 Oct 2011 15:17:09 +0000 (17:17 +0200)]
basesrc: properly adjust start time
When we do a non-flushing seek and closed the current segment,
make sure that we open the next segment from where we closed.
Tim-Philipp Müller [Mon, 10 Oct 2011 18:41:33 +0000 (19:41 +0100)]
utils: catch invalid instance sizes in gst_type_register_static_full()
Add guards to catch overly large instance sizes.
https://bugzilla.gnome.org/show_bug.cgi?id=660955
Tim-Philipp Müller [Sat, 8 Oct 2011 13:17:25 +0000 (14:17 +0100)]
gstparse: prefix generated parser functions so they don't get exported
Don't export those 35-something random _gst_parse_yy* symbols. These were
never in any header files and also blacklisted from our .def files, in
case anyone wonders.
Thiago Santos [Fri, 7 Oct 2011 12:45:20 +0000 (09:45 -0300)]
pipeline: Use pipeline category for one more log message
Makes debugging easier.
Robert Swain [Fri, 7 Oct 2011 14:36:10 +0000 (16:36 +0200)]
GST_PTR_FORMAT: Add GstBuffer ptr format and use in GST_SCHEDULING
GstBuffer pointers can now be printed using GST_PTR_FORMAT. This is used
in the very useful GST_SCHEDULING debug logs in gstpad.c and allows for
easier and more information tracking of buffer progress through a
pipeline with just debug logging.
Mark Nauwelaerts [Thu, 6 Oct 2011 12:34:09 +0000 (14:34 +0200)]
baseparse: send duration message when updating internal duration
Stas Sergeev [Mon, 26 Sep 2011 10:36:46 +0000 (14:36 +0400)]
baseparse: Return success if optional start/stop method is not provided
This allows to not implement the optional start/stop methods.
Edward Hervey [Mon, 3 Oct 2011 07:29:10 +0000 (09:29 +0200)]
gstquery: Make debugging message more informative
For all the newcomers out there who still don't know the values of
GstQueryType enum by heart...
... and old-timers who've got better things to do :)
Tim-Philipp Müller [Fri, 30 Sep 2011 14:25:20 +0000 (15:25 +0100)]
baseparse: make estimating the position in query handler actually work
No point estimating if we don't set the result afterwards.
Vincent Penquerc'h [Mon, 26 Sep 2011 12:14:42 +0000 (13:14 +0100)]
baseparse: answer position query in stream time and try upstream first
Let the demuxer have first say as well.
https://bugzilla.gnome.org/show_bug.cgi?id=659485
Tim-Philipp Müller [Fri, 30 Sep 2011 13:52:01 +0000 (14:52 +0100)]
gst.h: include header for atomic queue
Tim-Philipp Müller [Fri, 30 Sep 2011 13:50:51 +0000 (14:50 +0100)]
tests: add minimal test for GstAtomicQueue
Just new + free.
Vincent Penquerc'h [Mon, 26 Sep 2011 18:52:13 +0000 (19:52 +0100)]
basetransform: send delayed events earlier
Some elements (such as videorate) might push buffers early,
for instance in in transform_ip. We want events (and in particular
any NEWSEGMENT event) to be pushed before that.
This fixes transmageddon wedging on converting a file starting
with a non zero offset to Ogg.
https://bugzilla.gnome.org/show_bug.cgi?id=660165
Raimo Järvi [Sun, 25 Sep 2011 21:30:47 +0000 (00:30 +0300)]
gst: Fix compiler warnings on 64 bit mingw-w64
Fixes bug #660083.