platform/upstream/gstreamer.git
13 years ago0.10.31.3 pre-release
Tim-Philipp Müller [Tue, 11 Jan 2011 18:39:06 +0000 (18:39 +0000)]
0.10.31.3 pre-release

13 years agopo: update translations
Tim-Philipp Müller [Tue, 11 Jan 2011 18:38:37 +0000 (18:38 +0000)]
po: update translations

13 years agocontroller: Fix headers to use < > in #include statements
Benjamin Otte [Tue, 11 Jan 2011 18:26:40 +0000 (19:26 +0100)]
controller: Fix headers to use < > in #include statements

The behavior of " " in include statements is implementation-defined -
see the C standard, section 6.10.2 or
http://stackoverflow.com/questions/21593

13 years agogstnet: Fix --c-include for gir generation
Benjamin Otte [Tue, 11 Jan 2011 18:14:28 +0000 (19:14 +0100)]
gstnet: Fix --c-include for gir generation

Previously it was - probably due to copy/paste error - looking for
gstbase headers.
It's changed now to only include the one public header for gstnet.h

13 years agogobject-introspection: make g-ir-scanner try harder to find the right libgstreamer
Tim-Philipp Müller [Tue, 11 Jan 2011 13:47:38 +0000 (13:47 +0000)]
gobject-introspection: make g-ir-scanner try harder to find the right libgstreamer

It's not quite clear to me why g-ir-scanner doesn't get this info from
the pkg-config file, nor why libtool doesn't get it from the .la.

https://bugzilla.gnome.org/show_bug.cgi?id=637549

13 years agopkgconfig: also add libdir to -uninstalled .pc files
Tim-Philipp Müller [Tue, 11 Jan 2011 13:05:12 +0000 (13:05 +0000)]
pkgconfig: also add libdir to -uninstalled .pc files

This way we can find the paths to pass e.g. g-ir-scanner for
uninstalled setups.

https://bugzilla.gnome.org/show_bug.cgi?id=639039

13 years agoAutomatic update of common submodule
Stefan Kost [Tue, 11 Jan 2011 13:49:23 +0000 (15:49 +0200)]
Automatic update of common submodule

From e572c87 to f94d739

13 years agovalve: fixe the property link and the sice docs
Stefan Kost [Tue, 11 Jan 2011 13:10:32 +0000 (15:10 +0200)]
valve: fixe the property link and the sice docs

13 years agogstpoll: fix compiler warning with MingW
Raimo Järvi [Tue, 11 Jan 2011 12:46:31 +0000 (12:46 +0000)]
gstpoll: fix compiler warning with MingW

gstpoll.c: In function 'gst_poll_get_read_gpollfd':
gstpoll.c:692:10: warning: assignment makes integer from pointer without a cast

https://bugzilla.gnome.org/show_bug.cgi?id=638900

13 years agotest: outputselector: Add another negotiation test
Thiago Santos [Mon, 10 Jan 2011 17:37:41 +0000 (14:37 -0300)]
test: outputselector: Add another negotiation test

Adds an unit test to check that the output-selector works
when negotiating before srcpads are requested

13 years agooutputselector: Improve get and set caps functions
Thiago Santos [Mon, 10 Jan 2011 17:19:17 +0000 (14:19 -0300)]
outputselector: Improve get and set caps functions

Improve sink pad getcaps and setcaps by handling the case where
no src pads exist yet

13 years agoAutomatic update of common submodule
Tim-Philipp Müller [Mon, 10 Jan 2011 16:35:25 +0000 (16:35 +0000)]
Automatic update of common submodule

From ccbaa85 to e572c87

13 years agoAutomatic update of common submodule
Tim-Philipp Müller [Mon, 10 Jan 2011 14:52:40 +0000 (14:52 +0000)]
Automatic update of common submodule

From 46445ad to ccbaa85

13 years agotests: selector: unref peer pad
Thiago Santos [Mon, 10 Jan 2011 12:59:19 +0000 (09:59 -0300)]
tests: selector: unref peer pad

Do not forget to unref peer's pad on output-selector negotiation
tests

13 years agobasesrc: make sure we wait and release the live lock
Wim Taymans [Mon, 10 Jan 2011 12:18:16 +0000 (13:18 +0100)]
basesrc: make sure we wait and release the live lock

Make sure we release the live lock and wait in all cases when we need to wait
for the playing or flushing state change.

Fixes #635785

13 years agoregistry: Don't replace valid existing plugins by blacklisted ones
Edward Hervey [Sat, 8 Jan 2011 13:52:27 +0000 (14:52 +0100)]
registry: Don't replace valid existing plugins by blacklisted ones

Only replace existing plugins by blacklisted ones if they correspond
to the exact same plugin. If they're not the same, keep the existing
valid one.

Fixes #638941

13 years agoconfigure, gst-uninstalled: remove a few bashism
Tim-Philipp Müller [Sat, 8 Jan 2011 02:19:31 +0000 (02:19 +0000)]
configure, gst-uninstalled: remove a few bashism

https://bugzilla.gnome.org/show_bug.cgi?id=638961

13 years ago0.10.31.2 pre-release
Tim-Philipp Müller [Fri, 7 Jan 2011 12:13:37 +0000 (12:13 +0000)]
0.10.31.2 pre-release

13 years agoindexers: fix two small leaks
Tim-Philipp Müller [Fri, 7 Jan 2011 02:18:37 +0000 (02:18 +0000)]
indexers: fix two small leaks

element factory plugin_names are interned strings these days.

13 years agotests: never disable g_assert() and cast checks for the unit tests
Tim-Philipp Müller [Fri, 7 Jan 2011 00:53:11 +0000 (00:53 +0000)]
tests: never disable g_assert() and cast checks for the unit tests

The unit tests are riddled with g_assert() and friends, make sure we
don't disable assert and cast checks for the unit tests even if
this has been specified for the rest of the code base, e.g. via
--disable-glib-asserts.

13 years agoregistry: ignore old libgstvalve.so and libgstselector.so plugins to minimize upgrade...
Tim-Philipp Müller [Thu, 6 Jan 2011 19:40:27 +0000 (19:40 +0000)]
registry: ignore old libgstvalve.so and libgstselector.so plugins to minimize upgrade pain

Ignore plugins which have been moved into coreelements, so it's
still possible to just upgrade GStreamer core without having to
upgrade the whole stack.

13 years agoRevert "padtemplate: allow disablinbg the template name conformance checks"
Tim-Philipp Müller [Thu, 6 Jan 2011 19:21:56 +0000 (19:21 +0000)]
Revert "padtemplate: allow disablinbg the template name conformance checks"

This reverts commit f9039c22040e1a38a3691bc4d299af427c963255.

We use -DG_DISABLE_ASSERTS for releases and pre-releases, but
don't want to disable pad name checking for releases in general,
I think. Need a better solution here. Fixes pad unit test in
pre-release/release mode.

13 years agooutput-selector: Add pad-negotiation-mode property
Thiago Santos [Tue, 4 Jan 2011 15:42:50 +0000 (12:42 -0300)]
output-selector: Add pad-negotiation-mode property

Adds getcaps/setcaps to output-selector and adds a property
to select which type of negotiation should be done.

The available modes are:
 * none:   no negotiation (current behavior), getcaps return ANY and
           setcaps aren't set on any of the peers
 * all:    use all pads (default), getcaps returns the intersection of
           peer pads and setcaps is set on all peers
 * active: getcaps and setcaps are proxied to the active pad

https://bugzilla.gnome.org/show_bug.cgi?id=638381

13 years agodocs: mention extra input-selector pad properties
Tim-Philipp Müller [Thu, 6 Jan 2011 18:18:29 +0000 (18:18 +0000)]
docs: mention extra input-selector pad properties

https://bugzilla.gnome.org/show_bug.cgi?id=638381

13 years agopo: update translations
Tim-Philipp Müller [Thu, 6 Jan 2011 17:47:38 +0000 (17:47 +0000)]
po: update translations

13 years agoinfo: avoid unnecessary malloc/free for each log function call on MSVC
Tim-Philipp Müller [Thu, 23 Sep 2010 11:49:59 +0000 (12:49 +0100)]
info: avoid unnecessary malloc/free for each log function call on MSVC

Avoid unnecessary malloc/free to get the file basename on MSVC to avoid
unnecessary overhead when doing GST_DEBUG=foo:5 or so (since it would be
done before the category log level filtering).

13 years agoUpdate spec file with latest changes and enable GIR
christian schaller [Thu, 6 Jan 2011 17:29:46 +0000 (17:29 +0000)]
Update spec file with latest changes and enable GIR

13 years agoobject: Fix creation of default name
David Schleef [Wed, 5 Jan 2011 20:59:48 +0000 (12:59 -0800)]
object: Fix creation of default name

Change the fixed allocation (!) to g_strdup_printf().

13 years agoGstElement: Add a more flexible way to get request pads.
Edward Hervey [Mon, 20 Dec 2010 12:30:43 +0000 (13:30 +0100)]
GstElement: Add a more flexible way to get request pads.

The new request_new_pad_full vmethod provides an additional caps field,
which allows elements to take better decision process.

Also, add a gst_element_request_pad() function to allow developers to be
able to specify which pad template they want a pad of.

Convert gstutils to use that new method instead of the old one when more
efficient.

This is useful for being able to request pads in a more flexible way,
especially when the element can provide pads whose caps depend on
runtime configuration and therefore can't provide pre-registered
pad templates.

API: GstElement::request_new_pad_full
API: gst_element_request_pad

https://bugzilla.gnome.org/show_bug.cgi?id=637300

13 years agoinputselector: remove "select-all" property
Tim-Philipp Müller [Wed, 5 Jan 2011 15:53:28 +0000 (15:53 +0000)]
inputselector: remove "select-all" property

select-all mode is a bit broken (e.g. newsegment event
handling), so remove that for now. The funnel element
in farsight provides similar functionality.

https://bugzilla.gnome.org/show_bug.cgi?id=539042
https://bugzilla.gnome.org/show_bug.cgi?id=638381

13 years agogst: remove safety check for GLib < 2.8
Tim-Philipp Müller [Wed, 5 Jan 2011 15:36:55 +0000 (15:36 +0000)]
gst: remove safety check for GLib < 2.8

Don't worry about accidentally using GLib < 2.8 at runtime anymore.

13 years agogstdatetime: Disable usage of GDateTime on MacOSX
Edward Hervey [Wed, 5 Jan 2011 10:26:13 +0000 (11:26 +0100)]
gstdatetime: Disable usage of GDateTime on MacOSX

GLib's GDateTime doesn't handle properly the gmt offset. Therefore
use our own internal version instead on MacOSX.

See bug #638666

13 years agodesign: more updates for the progress messages
Wim Taymans [Wed, 5 Jan 2011 09:56:37 +0000 (10:56 +0100)]
design: more updates for the progress messages

13 years agopwg: Fix link to GObject documentation
Edward Hervey [Tue, 4 Jan 2011 17:36:41 +0000 (18:36 +0100)]
pwg: Fix link to GObject documentation

Better have gnome.org than viagra :)

13 years agopluginloader: Always mark reception as complete after EXIT
Edward Hervey [Mon, 3 Jan 2011 19:32:23 +0000 (20:32 +0100)]
pluginloader: Always mark reception as complete after EXIT

Avoids waiting forever on gst_poll_wait when using the select
backend.

Fixes #637057

13 years agogstinfo: don't use printf extensions if GLib isn't using the system printf
Tim-Philipp Müller [Tue, 4 Jan 2011 00:48:15 +0000 (00:48 +0000)]
gstinfo: don't use printf extensions if GLib isn't using the system printf

Might help with https://bugzilla.gnome.org/show_bug.cgi?id=638599

13 years agogstpoll: Fix for (p)select backend
Edward Hervey [Mon, 3 Jan 2011 19:19:05 +0000 (20:19 +0100)]
gstpoll: Fix for (p)select backend

We need to reset the revents field of each pollfd when reading the results
from select else we'll end up with stray info from previous calls to
select.

13 years agogstvalue: make new gst_value_list_merge() work properly
Tim-Philipp Müller [Mon, 3 Jan 2011 01:06:06 +0000 (01:06 +0000)]
gstvalue: make new gst_value_list_merge() work properly

Fix freeing of partially-inited list value when both values
passed are equal and we want to return a single non-list
value as result. Fixes unit test. Also fix up docs a bit.

https://bugzilla.gnome.org/show_bug.cgi?id=637776

13 years agotags: don't produce duplicated entries when merging same value twice
Stefan Kost [Tue, 21 Dec 2010 21:03:12 +0000 (23:03 +0200)]
tags: don't produce duplicated entries when merging same value twice

Add a variant of gst_value_list_concat() that skips duplicates and use that when
merging taglists.

API: gst_value_list_merge()

13 years agobasesrc, basesink: add some FIXMEs for the type of the blocksize property
Tim-Philipp Müller [Sun, 2 Jan 2011 16:58:39 +0000 (16:58 +0000)]
basesrc, basesink: add some FIXMEs for the type of the blocksize property

13 years agoinputselector: make pad's get_type function thread-safe
Tim-Philipp Müller [Fri, 31 Dec 2010 12:08:19 +0000 (12:08 +0000)]
inputselector: make pad's get_type function thread-safe

13 years agocoreelements: remove unused files
Tim-Philipp Müller [Fri, 31 Dec 2010 11:37:56 +0000 (11:37 +0000)]
coreelements: remove unused files

13 years agoinputselector: Protected g_object_notify() calls for the active-pad with a recursive...
Sebastian Dröge [Fri, 31 Dec 2010 11:27:45 +0000 (12:27 +0100)]
inputselector: Protected g_object_notify() calls for the active-pad with a recursive mutex

This works around a thread safety problem in GLib < 2.26.0 and should
be removed when we depend on 2.26.0.

Fixes bug #607513.

13 years agotests: fix typo
Zhang Wanming [Fri, 31 Dec 2010 08:52:54 +0000 (16:52 +0800)]
tests: fix typo

Signed-off-by: Zhang Wanming <Wanming.Zhang@tieto.com>
13 years agogstfdsink: fix typo
Zhang Wanming [Fri, 31 Dec 2010 08:50:57 +0000 (16:50 +0800)]
gstfdsink: fix typo

Signed-off-by: Zhang Wanming <Wanming.Zhang@tieto.com>
13 years agodocs: fix typo
Zhang Wanming [Fri, 31 Dec 2010 08:50:17 +0000 (16:50 +0800)]
docs: fix typo

Signed-off-by: Zhang Wanming <Wanming.Zhang@tieto.com>
13 years agotests: remove output-selector test which needs elements from -base
Tim-Philipp Müller [Fri, 31 Dec 2010 01:09:40 +0000 (01:09 +0000)]
tests: remove output-selector test which needs elements from -base

Move it to -base instead.

13 years agochecks: enable input-selector and output-selector unit tests after move
Tim-Philipp Müller [Fri, 31 Dec 2010 00:59:53 +0000 (00:59 +0000)]
checks: enable input-selector and output-selector unit tests after move

13 years agodocs: add input-selector and output-selector to docs
Tim-Philipp Müller [Fri, 31 Dec 2010 00:48:55 +0000 (00:48 +0000)]
docs: add input-selector and output-selector to docs

13 years agocoreelements: GST_BOILERPLATE already sets parent_class
Tim-Philipp Müller [Fri, 31 Dec 2010 00:45:54 +0000 (00:45 +0000)]
coreelements: GST_BOILERPLATE already sets parent_class

13 years agoinput-selector, output-selector: minor clean-ups
Tim-Philipp Müller [Fri, 31 Dec 2010 00:43:37 +0000 (00:43 +0000)]
input-selector, output-selector: minor clean-ups

13 years agocoreelements: move input-selector and output-selector to core
Tim-Philipp Müller [Thu, 30 Dec 2010 18:57:13 +0000 (18:57 +0000)]
coreelements: move input-selector and output-selector to core

Moved to core from gst-plugins-bad.

https://bugzilla.gnome.org/show_bug.cgi?id=614306

13 years agooutput-selector-test: don't hardcode videosinks and use more colorspace conv.
Stefan Kost [Wed, 24 Nov 2010 10:22:01 +0000 (12:22 +0200)]
output-selector-test: don't hardcode videosinks and use more colorspace conv.

Use autovideosink instead of hardcoded sinks. Use an additional colorspace
converter between videotestsrc and timeoverlay.

13 years agooutputselector: Avoid losing the last_buffer when switching
Thiago Santos [Fri, 17 Sep 2010 12:52:12 +0000 (09:52 -0300)]
outputselector: Avoid losing the last_buffer when switching

This patch makes outputselector take an extra ref when pushing
the last_buffer to avoid it losing it during the switch function.

This makes resend-latest properly work if the active-pad is changed
during the switch function buffer pushing (on a pad probe, for example).

https://bugzilla.gnome.org/show_bug.cgi?id=629917

13 years agooutputselector: Recheck pending switch after pushing buffer
Thiago Santos [Fri, 17 Sep 2010 12:44:02 +0000 (09:44 -0300)]
outputselector: Recheck pending switch after pushing buffer

This patch makes output-selector always recheck if there's a
pending pad switch after pushing a buffer, preventing that
it pushes a buffer on the 'wrong' pad.

https://bugzilla.gnome.org/show_bug.cgi?id=629917

13 years agoinputselector: log times in human readable form
Stefan Kost [Mon, 1 Nov 2010 21:04:44 +0000 (23:04 +0200)]
inputselector: log times in human readable form

13 years agoinputselector: move reoccuring logs to LOG and remove a double info
Stefan Kost [Mon, 1 Nov 2010 20:40:36 +0000 (22:40 +0200)]
inputselector: move reoccuring logs to LOG and remove a double info

Less debug spew in DEBUG category. No need to log pad again if we use
GST_LOG_OBJECT(pad,...).

13 years agovarious (gst): add missing G_PARAM_STATIC_STRINGS flags
Stefan Kost [Tue, 19 Oct 2010 10:43:14 +0000 (13:43 +0300)]
various (gst): add missing G_PARAM_STATIC_STRINGS flags

Canonicalize property names as needed.

13 years agoinputselector: sync with copy in -base
Stefan Kost [Mon, 6 Sep 2010 08:03:07 +0000 (11:03 +0300)]
inputselector: sync with copy in -base

Use _get_caps_reffed to avoid copies.

13 years agooutputselector: move the debug init to the boilerplate macro
Stefan Kost [Tue, 24 Aug 2010 08:50:47 +0000 (11:50 +0300)]
outputselector: move the debug init to the boilerplate macro

13 years agoinputselector: use GST_BOILERPLATE macro
Stefan Kost [Tue, 24 Aug 2010 08:50:09 +0000 (11:50 +0300)]
inputselector: use GST_BOILERPLATE macro

13 years agoAdd -Wwrite-strings
Benjamin Otte [Mon, 22 Mar 2010 12:16:33 +0000 (13:16 +0100)]
Add -Wwrite-strings

and fix its warnings

13 years agoAdd -Wmissing-declarations -Wmissing-prototypes to configure flags
Benjamin Otte [Sun, 21 Mar 2010 20:39:18 +0000 (21:39 +0100)]
Add -Wmissing-declarations -Wmissing-prototypes to configure flags

And fix all warnings

13 years agogst_element_class_set_details => gst_element_class_set_details_simple
Benjamin Otte [Thu, 18 Mar 2010 16:30:26 +0000 (17:30 +0100)]
gst_element_class_set_details => gst_element_class_set_details_simple

13 years agoRevert "inputselector: Protect g_object_notify() with the object's mutex"
Sebastian Dröge [Mon, 25 Jan 2010 11:21:34 +0000 (12:21 +0100)]
Revert "inputselector: Protect g_object_notify() with the object's mutex"

This reverts commit 7e067615ffb5614f068f7753c10dde99afe49c3c, it's causing
deadlocks with playbin2.

13 years agoinputselector: Protect g_object_notify() with the object's mutex
Kipp Cannon [Sun, 24 Jan 2010 19:53:00 +0000 (20:53 +0100)]
inputselector: Protect g_object_notify() with the object's mutex

This works around the thread unsafety of g_object_notify()

Fixes bug #607513.

13 years agoinputselector: Remove useless variables and fix a uninitialized variable compiler...
Sebastian Dröge [Mon, 9 Nov 2009 10:49:15 +0000 (11:49 +0100)]
inputselector: Remove useless variables and fix a uninitialized variable compiler warnings

Merged from gst-plugins-base, dfd51aa82a9e1c9924375183796eab70e574a231.

13 years agoinputselector: Make sure that running_time->timestamp calculation never becomes negative
Sebastian Dröge [Mon, 9 Nov 2009 10:48:39 +0000 (11:48 +0100)]
inputselector: Make sure that running_time->timestamp calculation never becomes negative

Merged from gst-plugins-base, f3653854585864a09d35e037853407332ea6901f.

13 years agoinputselector: Use the start time (i.e. timestamp) as the last stop
Sebastian Dröge [Mon, 9 Nov 2009 10:48:00 +0000 (11:48 +0100)]
inputselector: Use the start time (i.e. timestamp) as the last stop

Using the end time makes it impossible to replace buffers, which is
a big problem for subtitles that could have very long durations.

Merged from gst-plugins-base, 27034be4611231cc55fa3d3e253baa40c6bff41d.

13 years agoinputselector: Improve debugging
Sebastian Dröge [Mon, 9 Nov 2009 10:47:15 +0000 (11:47 +0100)]
inputselector: Improve debugging

Merged from gst-plugins-base.

13 years agoRevert "inputselector: use get_caps_reffed()"
Wim Taymans [Thu, 5 Nov 2009 12:53:46 +0000 (13:53 +0100)]
Revert "inputselector: use get_caps_reffed()"

This reverts commit 49ec4f796a0e3f88a851708782c853baf4b05724.

We can't use this new function yet.

13 years agoinputselector: use get_caps_reffed()
Wim Taymans [Thu, 5 Nov 2009 12:29:55 +0000 (13:29 +0100)]
inputselector: use get_caps_reffed()

13 years agoinputselector: also add inline to the proto to fix the build
Stefan Kost [Tue, 3 Nov 2009 22:21:19 +0000 (00:21 +0200)]
inputselector: also add inline to the proto to fix the build

13 years agogst: Remove dead assignments and resulting unused variables
Edward Hervey [Tue, 3 Nov 2009 17:14:12 +0000 (18:14 +0100)]
gst: Remove dead assignments and resulting unused variables

Merged from gst-plugins-base, 8cd1b5209b68944e1be56ca8bb69e46d4abb7a34.

13 years agoinputselector: Use the same iterate internal links function as in gst-plugins-base
Sebastian Dröge [Tue, 3 Nov 2009 17:12:21 +0000 (18:12 +0100)]
inputselector: Use the same iterate internal links function as in gst-plugins-base

13 years agoinput-selector: Remove GST_DEBUG_FUNCPTR where they're pointless
Tim-Philipp Müller [Tue, 3 Nov 2009 17:11:13 +0000 (18:11 +0100)]
input-selector: Remove GST_DEBUG_FUNCPTR where they're pointless

There's not much point in using GST_DEBUG_FUNCPTR with GObject
virtual functions such as get_property, set_propery, finalize and
dispose, since they'll never be used by anyone anyway. Saves a
few bytes and possibly a sixteenth of a polar bear.

Merged from gst-plugins-base, 6f4c1ac58341ee189225d313ddfe9ae24a65c88c.

13 years agoinput-selector: Remove Ronald Bultje from Authors field
David Schleef [Tue, 3 Nov 2009 17:09:55 +0000 (18:09 +0100)]
input-selector: Remove Ronald Bultje from Authors field

Replaced with "GStreamer maintainers
<gstreamer-devel@lists.sourceforge.net>" or just removed,
depending on the number of other authors.

Merged from gst-plugins-base, 0e9bc5125aca546d773ed1002df573dd8e2dc136.

13 years agoinputselector: set output caps before pushing
Wim Taymans [Tue, 3 Nov 2009 17:08:05 +0000 (18:08 +0100)]
inputselector: set output caps before pushing

Set the output caps on the srcpad before pushing the buffer because else core
will do a rather expensive check to see if we can actually accept those caps on
the srcpad.

Merged from gst-plugins-base, bdfb4b46d746ef298fcf44260879c342af4cafa3.

13 years agoinputselector: install an acceptcaps function
Wim Taymans [Tue, 3 Nov 2009 17:06:11 +0000 (18:06 +0100)]
inputselector: install an acceptcaps function

Install a custom acceptcaps function instead of using the default expensive
check. We accept whatever downstream accepts so we pass along the acceptcaps
call to the downstream peer.

Merged from gst-plugins-base, 5b72f2adf996739036e8d9b5f91005455d1fface.

13 years agoRemove executable bits from non-executable files.
Michael Smith [Tue, 27 Oct 2009 18:51:05 +0000 (11:51 -0700)]
Remove executable bits from non-executable files.

13 years agoinputselector: Use new single iterator for the internally linked pads
Sebastian Dröge [Fri, 25 Sep 2009 09:07:02 +0000 (11:07 +0200)]
inputselector: Use new single iterator for the internally linked pads

This fixes a deadlock and removes some useless code.

13 years agooutputselector: make GST_FORMAT_TIME the default segment format
Havard Graff [Mon, 31 Aug 2009 17:31:57 +0000 (19:31 +0200)]
outputselector: make GST_FORMAT_TIME the default segment format

13 years agoinputselector: Use iterate internal links instead of deprecated get internal links
Sebastian Dröge [Wed, 19 Aug 2009 15:05:32 +0000 (17:05 +0200)]
inputselector: Use iterate internal links instead of deprecated get internal links

13 years agooutputselector: check for pending srcpad in _get_property()
Tommi Myöhänen [Fri, 31 Jul 2009 08:27:03 +0000 (11:27 +0300)]
outputselector: check for pending srcpad in _get_property()

If there is a pending srcpad, return it instead of active srcpad
in gst_output_selector_get_property() function.

13 years agooutputselector: do the pad_alloc for the pad that is pending and have a fallback
Stefan Kost [Fri, 12 Jun 2009 07:14:27 +0000 (10:14 +0300)]
outputselector: do the pad_alloc for the pad that is pending and have a fallback

We should do the pad_alloc for the pending pad if any, as we will switch to that
pad on next _chain() call. Also do a fallback alloc, if there is no output yet to
not fail state transitions in dynamic pipelines.

13 years agooutput-selector: serialize setting and actual changing of new active pad
Lasse Laukkanen [Mon, 1 Jun 2009 13:31:42 +0000 (16:31 +0300)]
output-selector: serialize setting and actual changing of new active pad

13 years agooutput-selector: unref latest buffer also when resending has been disabled
Lasse Laukkanen [Mon, 4 May 2009 09:29:54 +0000 (12:29 +0300)]
output-selector: unref latest buffer also when resending has been disabled

13 years agooutput-selector: keep ref to buffer for resending only if explicitly requested
Lasse Laukkanen [Thu, 16 Apr 2009 14:32:03 +0000 (17:32 +0300)]
output-selector: keep ref to buffer for resending only if explicitly requested

13 years agoinputselector: don't leak pads in iterator
Wim Taymans [Thu, 4 Jun 2009 17:08:16 +0000 (19:08 +0200)]
inputselector: don't leak pads in iterator

13 years agoinputselector: Notify when the tags property of the selector sinkpads changes
Sebastian Dröge [Thu, 4 Jun 2009 06:56:29 +0000 (08:56 +0200)]
inputselector: Notify when the tags property of the selector sinkpads changes

First part of bug #584686.

13 years agoselector: remove not needed instance var (previous commit).
Stefan Kost [Thu, 28 May 2009 07:38:50 +0000 (10:38 +0300)]
selector: remove not needed instance var (previous commit).

13 years agooutputselector: implement pad_alloc on active pad.
Stefan Kost [Thu, 28 May 2009 07:12:58 +0000 (10:12 +0300)]
outputselector: implement pad_alloc on active pad.

13 years agoinput-selector: Forward segment events for the active pad immediately.
Jan Schmidt [Thu, 23 Apr 2009 10:04:46 +0000 (11:04 +0100)]
input-selector: Forward segment events for the active pad immediately.

When a segment event is received on the active pad, forward it downstream
immediately instead of deferring it until the next data buffer arrives. This
fixes problems with segment updates never being sent downstream, like those
needed for sparse streams, or for closing previously opened segments.

This fixes playback of DVD menus with a still video frame and an audio track,
for example.

Fixes: #577843

13 years agooutputselector: reset state when going to READY
Wim Taymans [Fri, 27 Mar 2009 10:20:02 +0000 (11:20 +0100)]
outputselector: reset state when going to READY

Reset the last-buffer, the pending pad and the segment when going to the READY
state.
Fixes #576712.

13 years agoselector: merge the tags
Wim Taymans [Tue, 24 Mar 2009 14:23:03 +0000 (15:23 +0100)]
selector: merge the tags

Merge the tags received on the input-selector sinkpads instead of only keeping
the last one we saw.

13 years agodocs: various doc fixes
Stefan Kost [Wed, 25 Feb 2009 09:45:05 +0000 (11:45 +0200)]
docs: various doc fixes

No short-desc as we have them in the element details.
Also keep things (Makefile.am and sections.txt) sorted.
Reword ambigous returns. No text after since please.

13 years agoinputselector: Fix compilation, activate_sinkpad() has no notify parameter
Sebastian Dröge [Wed, 11 Feb 2009 16:21:20 +0000 (17:21 +0100)]
inputselector: Fix compilation, activate_sinkpad() has no notify parameter

13 years agoinput-selector: Activate and notify pad before processing events.
Michael Smith [Wed, 11 Feb 2009 00:22:54 +0000 (16:22 -0800)]
input-selector: Activate and notify pad before processing events.

Events should trigger pad selection if we don't already have an
explicitly selected pad, so that events prior to first buffer don't get
lost.

13 years agoUnref event if we don't forward it, unref pads when done with them.
Michael Smith [Sat, 31 Jan 2009 02:27:03 +0000 (18:27 -0800)]
Unref event if we don't forward it, unref pads when done with them.

13 years agoplugins/elements/gstinputselector.c: Ensure we emit notify::active-pad when auto...
Michael Smith [Thu, 4 Dec 2008 17:51:37 +0000 (17:51 +0000)]
plugins/elements/gstinputselector.c: Ensure we emit notify::active-pad when auto-selecting a pad due to it having activit...

Original commit message from CVS:
* plugins/elements/gstinputselector.c:
Ensure we emit notify::active-pad when auto-selecting a pad
due to it having activity and us not having an existing active
pad. Fixes #563147