platform/upstream/gstreamer.git
15 years agogst/playback/gstqueue2.c: Fix compile error from ignored return value.
David Schleef [Sat, 16 Jun 2007 03:42:14 +0000 (03:42 +0000)]
gst/playback/gstqueue2.c: Fix compile error from ignored return value.

Original commit message from CVS:
* gst/playback/gstqueue2.c:
Fix compile error from ignored return value.

15 years agogst/playback/gstqueue2.c: Fix build on MacOSX.
Edward Hervey [Wed, 13 Jun 2007 18:20:57 +0000 (18:20 +0000)]
gst/playback/gstqueue2.c: Fix build on MacOSX.

Original commit message from CVS:
* gst/playback/gstqueue2.c: (gst_queue_create_read):
Fix build on MacOSX.

15 years agogst/playback/gstqueue2.c: Fix a division by zero when the max percent is <= 0. Fixes...
Wim Taymans [Tue, 12 Jun 2007 08:38:06 +0000 (08:38 +0000)]
gst/playback/gstqueue2.c: Fix a division by zero when the max percent is <= 0. Fixes #446572. also update the bufferi...

Original commit message from CVS:
Patches by: Thiago Sousa Santos <thiagossantos at gmail dot com>
* gst/playback/gstqueue2.c: (update_buffering),
(gst_queue_locked_enqueue):
Fix a division by zero when the max percent is <= 0. Fixes #446572.
also update the buffering status when receiving events. Fixes #446551.

15 years agogst/playback/gstqueue2.c: Wait for preroll before attempting to forward a duration...
Thiago Sousa Santos [Mon, 11 Jun 2007 11:32:26 +0000 (11:32 +0000)]
gst/playback/gstqueue2.c: Wait for preroll before attempting to forward a duration query upstream.

Original commit message from CVS:
Based on patch by: Thiago Sousa Santos <thiagossantos at gmail dot com>
* gst/playback/gstqueue2.c: (gst_queue_peer_query),
(gst_queue_handle_src_query):
Wait for preroll before attempting to forward a duration query upstream.
Fixes #445505.

15 years agogst/playback/gstqueue2.c: Fix compilation.
Wim Taymans [Thu, 7 Jun 2007 09:11:27 +0000 (09:11 +0000)]
gst/playback/gstqueue2.c: Fix compilation.

Original commit message from CVS:
* gst/playback/gstqueue2.c: (gst_queue_get_range):
Fix compilation.

15 years agogst/playback/gstqueue2.c: Add pull based scheduling and fix some deadlocks. Fixes...
Thiago Sousa Santos [Wed, 6 Jun 2007 13:36:26 +0000 (13:36 +0000)]
gst/playback/gstqueue2.c: Add pull based scheduling and fix some deadlocks. Fixes #444523.

Original commit message from CVS:
Patch by: Thiago Sousa Santos <thiagossantos at gmail dot com>
* gst/playback/gstqueue2.c: (gst_queue_init),
(gst_queue_handle_sink_event), (gst_queue_chain),
(gst_queue_get_range), (gst_queue_src_checkgetrange_function),
(gst_queue_sink_activate_push), (gst_queue_src_activate_push),
(gst_queue_src_activate_pull):
Add pull based scheduling and fix some deadlocks. Fixes #444523.
Does not yet completely work because duration queries upstream won't
block yet.

15 years agoSome more fseeko checks.
Wim Taymans [Wed, 6 Jun 2007 09:08:50 +0000 (09:08 +0000)]
Some more fseeko checks.

Original commit message from CVS:
* configure.ac:
* gst/playback/gstqueue2.c: (gst_queue_create_read):
Some more fseeko checks.

15 years agogst/playback/gstqueue2.c: Include stdio to define fseeko.
Wim Taymans [Tue, 5 Jun 2007 17:02:13 +0000 (17:02 +0000)]
gst/playback/gstqueue2.c: Include stdio to define fseeko.

Original commit message from CVS:
* gst/playback/gstqueue2.c: (gst_queue_have_data),
(gst_queue_create_read), (gst_queue_read_item_from_file),
(gst_queue_open_temp_location_file), (gst_queue_locked_enqueue):
Include stdio to define fseeko.

15 years agogst/playback/gstqueue2.c: Add support for filebased buffering. Fixes #441264.
Thiago Sousa Santos [Tue, 5 Jun 2007 16:14:23 +0000 (16:14 +0000)]
gst/playback/gstqueue2.c: Add support for filebased buffering. Fixes #441264.

Original commit message from CVS:
Based on patch by: Thiago Sousa Santos <thiagossantos at gmail dot com>
* gst/playback/gstqueue2.c: (gst_queue_class_init),
(gst_queue_init), (gst_queue_finalize),
(gst_queue_write_buffer_to_file), (gst_queue_have_data),
(gst_queue_create_read), (gst_queue_read_item_from_file),
(gst_queue_open_temp_location_file),
(gst_queue_close_temp_location_file), (gst_queue_locked_flush),
(gst_queue_locked_enqueue), (gst_queue_locked_dequeue),
(gst_queue_is_empty), (gst_queue_is_filled),
(gst_queue_change_state), (gst_queue_set_temp_location),
(gst_queue_set_property):
Add support for filebased buffering. Fixes #441264.

15 years agogst/playback/gstqueue2.c: Tweak the buffering thresholds a little.
Wim Taymans [Thu, 17 May 2007 15:22:44 +0000 (15:22 +0000)]
gst/playback/gstqueue2.c: Tweak the buffering thresholds a little.

Original commit message from CVS:
* gst/playback/gstqueue2.c: (update_rates):
Tweak the buffering thresholds a little.
Update the buffer size with the previously calculate rate instead of
only when we calculate a new rate so that we get smoother buffering
updates.
* gst/playback/Makefile.am:
* gst/playback/gsturidecodebin.c: (gst_uri_decode_bin_base_init),
(gst_uri_decode_bin_class_init), (gst_uri_decode_bin_init),
(gst_uri_decode_bin_finalize), (gst_uri_decode_bin_set_property),
(gst_uri_decode_bin_get_property), (unknown_type),
(add_element_stream), (no_more_pads_full), (no_more_pads),
(source_no_more_pads), (new_decoded_pad), (array_has_value),
(gen_source_element), (has_all_raw_caps), (analyse_source),
(remove_decoders), (make_decoder), (remove_source),
(source_new_pad), (setup_source), (decoder_query_init),
(decoder_query_duration_fold), (decoder_query_duration_done),
(decoder_query_position_fold), (decoder_query_position_done),
(decoder_query_latency_fold), (decoder_query_latency_done),
(decoder_query_seeking_fold), (decoder_query_seeking_done),
(decoder_query_generic_fold), (gst_uri_decode_bin_query),
(gst_uri_decode_bin_change_state), (plugin_init):
New element that intergrates a source, optional buffering element and
decodebin.

15 years agogst/playback/gstqueue2.c: fix build.
Wim Taymans [Thu, 17 May 2007 13:36:11 +0000 (13:36 +0000)]
gst/playback/gstqueue2.c: fix build.

Original commit message from CVS:
* gst/playback/gstqueue2.c: (gst_queue_get_type),
(gst_queue_class_init), (gst_queue_finalize), (update_time_level),
(apply_segment), (apply_buffer), (update_buffering),
(reset_rate_timer), (update_rates), (gst_queue_locked_flush),
(gst_queue_locked_enqueue), (gst_queue_locked_dequeue),
(gst_queue_handle_sink_event), (gst_queue_is_filled),
(gst_queue_chain), (gst_queue_push_one), (gst_queue_loop),
(plugin_init):
fix build.

15 years agogst/playback/: On our way to playbin2 this is the new network queue that does bufferi...
Wim Taymans [Thu, 17 May 2007 11:57:44 +0000 (11:57 +0000)]
gst/playback/: On our way to playbin2 this is the new network queue that does buffering all by itself using high and ...

Original commit message from CVS:
* gst/playback/Makefile.am:
* gst/playback/gstqueue2.c: (gst_queue_get_type),
(gst_queue_class_init), (gst_queue_init), (gst_queue_finalize),
(gst_queue_getcaps), (gst_queue_bufferalloc),
(gst_queue_acceptcaps), (update_time_level), (apply_segment),
(apply_buffer), (update_buffering), (reset_rate_timer),
(update_rates), (gst_queue_locked_flush),
(gst_queue_locked_enqueue), (gst_queue_locked_dequeue),
(gst_queue_handle_sink_event), (gst_queue_is_empty),
(gst_queue_is_filled), (gst_queue_chain), (gst_queue_push_one),
(gst_queue_loop), (gst_queue_handle_src_event),
(gst_queue_handle_src_query), (gst_queue_sink_activate_push),
(gst_queue_src_activate_push), (gst_queue_change_state),
(gst_queue_set_property), (gst_queue_get_property), (plugin_init):
On our way to playbin2 this is the new network queue that does buffering
all by itself using high and low watermarks. It can also measure up and
downstream bandwidth to optimally size the queue.

15 years agoparse: Fix memleak of unused delayed links
David Schleef [Thu, 29 Oct 2009 05:03:44 +0000 (22:03 -0700)]
parse: Fix memleak of unused delayed links

Attach the DelayedLink structure to the element, so that when
the element is disposed, the DelayedLink is freed.

15 years agopad: make _fixate_caps() also truncate when needed
Rob Clark [Wed, 9 Sep 2009 20:37:11 +0000 (15:37 -0500)]
pad: make _fixate_caps() also truncate when needed

The default gst_pad_fixate_caps() previously would only fixate each individual
struct. In case there are multiple structs, the resulting caps would still not
be fixed. In the spirit of how individual structs are fixated, this patch
changes gst_pad_fixate_caps() to remove all but the first struct.

Fixes #595886

15 years agoinfo: fix docs
Wim Taymans [Mon, 21 Sep 2009 09:44:13 +0000 (11:44 +0200)]
info: fix docs

15 years agooptimisation : Use g_object_newv where possible.
Edward Hervey [Wed, 28 Oct 2009 08:26:32 +0000 (09:26 +0100)]
optimisation : Use g_object_newv where possible.

This avoids:
* triple-checking for the GType when type-checking is enabled (see #597260)
* Avoids going through an expensive no-argument checking which landed in
  glib-2.22
* Avoids going through 2 extrac functions (g_object_new -> g_object_new_valist)

15 years agodocs: include annotation glossary to have working links.
Stefan Kost [Wed, 28 Oct 2009 08:15:12 +0000 (10:15 +0200)]
docs: include annotation glossary to have working links.

15 years agoannotations: add annotations to gst_init_check too
Stefan Kost [Wed, 28 Oct 2009 08:14:36 +0000 (10:14 +0200)]
annotations: add annotations to gst_init_check too

15 years agodocs: tell more about what happens in gst_init.
Stefan Kost [Wed, 28 Oct 2009 07:58:52 +0000 (09:58 +0200)]
docs: tell more about what happens in gst_init.

Add links to gst_update_registry and the env-vars.

15 years agodocs: remove reference to OGI and rephrase sections docs
Stefan Kost [Wed, 28 Oct 2009 07:21:01 +0000 (09:21 +0200)]
docs: remove reference to OGI and rephrase sections docs

The OGI links are dead, so remove them. Also remove the paragraph that pointed
to OGI and DS. Only mentioning DS there made it a but pointless. Add a generic
paragraph instead that tells a bit about the usecases gstreamer covers.

15 years agoRemove GST_DEBUG_FUNCPTR where they're pointless
Tim-Philipp Müller [Wed, 28 Oct 2009 00:29:30 +0000 (00:29 +0000)]
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 tenth of a polar bear.

15 years agocapsfilter: sprinkle some GST_DEBUG_FUNCPTR
Tim-Philipp Müller [Wed, 28 Oct 2009 00:07:48 +0000 (00:07 +0000)]
capsfilter: sprinkle some GST_DEBUG_FUNCPTR

15 years agodesign: add some ideas for SKIP mode trickmodes
Wim Taymans [Tue, 27 Oct 2009 14:47:39 +0000 (15:47 +0100)]
design: add some ideas for SKIP mode trickmodes

15 years agomultiqueue: buffering is implemented now
Wim Taymans [Fri, 23 Oct 2009 14:20:02 +0000 (10:20 -0400)]
multiqueue: buffering is implemented now

15 years agomultiqueue: make sure percent increases
Wim Taymans [Fri, 23 Oct 2009 01:24:24 +0000 (21:24 -0400)]
multiqueue: make sure percent increases

Keep track of the last posted percent message and make sure the next percent
messages are strictly increasing.

15 years agomultiqueue: refactor buffering code
Wim Taymans [Thu, 22 Oct 2009 20:38:12 +0000 (16:38 -0400)]
multiqueue: refactor buffering code

Move the buffering update code to a separate function so that we can call it
when the buffering state changes due to EOS.
Avoid dividing by 0.

15 years agomultiqueue: add buffering support
Wim Taymans [Thu, 22 Oct 2009 18:09:01 +0000 (14:09 -0400)]
multiqueue: add buffering support

Add support for buffering mode where we post BUFFERING messages based on the
level of the queues. It currently operates on the first queue that goes over or
under the high/low thresholds.

15 years agomultiqueue: don't check visible items in buffering
Wim Taymans [Thu, 22 Oct 2009 18:07:31 +0000 (14:07 -0400)]
multiqueue: don't check visible items in buffering

In buffering mode we want to ignore the max visible items to decide when the
queue is filled. Instead, we only look at the number of bytes and/or time in the
queue.

15 years agomultiqueue: hook up low/high percent
Wim Taymans [Wed, 21 Oct 2009 15:30:40 +0000 (11:30 -0400)]
multiqueue: hook up low/high percent

Hook up the low/high percent properties for the buffering mode.

15 years agomultiqueue: hook up property for buffering
Wim Taymans [Wed, 21 Oct 2009 15:24:47 +0000 (11:24 -0400)]
multiqueue: hook up property for buffering

15 years agomultiqueue: small cleanups
Wim Taymans [Thu, 22 Oct 2009 14:21:39 +0000 (15:21 +0100)]
multiqueue: small cleanups

Remove unused variable to avoid confusion
Fix some typo

15 years agomultiqueue: add FIXME for wrong code
Wim Taymans [Thu, 22 Oct 2009 13:41:52 +0000 (09:41 -0400)]
multiqueue: add FIXME for wrong code

Needs further investigation

15 years agomultiqueue: fix debug output
Wim Taymans [Wed, 21 Oct 2009 18:20:29 +0000 (14:20 -0400)]
multiqueue: fix debug output

15 years agomultiqueue: avoid shadowing function argument
Wim Taymans [Wed, 21 Oct 2009 18:15:05 +0000 (14:15 -0400)]
multiqueue: avoid shadowing function argument

Don't shadow the sq argument in the underrun_cb function but use
a different variable name to iterate the other queues.
Use the same variable name in the overrun_cb function.

15 years agomultiqueue: make queue arg explicit
Wim Taymans [Wed, 21 Oct 2009 18:12:12 +0000 (14:12 -0400)]
multiqueue: make queue arg explicit

Make the queue argument to IS_FILLED explicit

15 years agomultiqueue: fix properties
Wim Taymans [Wed, 21 Oct 2009 15:17:08 +0000 (11:17 -0400)]
multiqueue: fix properties

Fix properties, make the extra-size properties as not implemented.

15 years agodebugutils: allow to hide/show pad status with graphdetails flag
Stefan Kost [Sun, 25 Oct 2009 19:35:09 +0000 (21:35 +0200)]
debugutils: allow to hide/show pad status with graphdetails flag

15 years agotypefindhelper: Remove obsolete FIXME
Sebastian Dröge [Sat, 24 Oct 2009 11:14:25 +0000 (13:14 +0200)]
typefindhelper: Remove obsolete FIXME

It's not necessary (and not a good idea) to cache the typefind factory
list anymore.

15 years agoregistry: private is a C++ keyword, don't use it
Sebastian Dröge [Sat, 24 Oct 2009 09:58:25 +0000 (11:58 +0200)]
registry: private is a C++ keyword, don't use it

...otherwise C++ compilers will complain when including gstregistry.h

15 years agodocs: add Since tag to docs for new API
Tim-Philipp Müller [Sat, 24 Oct 2009 09:21:30 +0000 (10:21 +0100)]
docs: add Since tag to docs for new API

And tell gtk-doc that GstRegistryPrivate is private.

15 years agotypefind: Keep typefind factories sorted in the registry. Fixes #599147
Edward Hervey [Wed, 21 Oct 2009 07:48:41 +0000 (09:48 +0200)]
typefind: Keep typefind factories sorted in the registry. Fixes #599147

This avoids having to do the sorting everytime we use typefind

The behaviour of gst_type_find_factory_get_list has subtlely changed
in the sense that the order was previously undefined, whereas now
it returns them sorted by rank and then by name.

15 years agoregistry: Cache element and typefind factories. Fixes 598896
Edward Hervey [Wed, 21 Oct 2009 07:45:47 +0000 (09:45 +0200)]
registry: Cache element and typefind factories. Fixes 598896

This avoids unneeded list/filtering if the registry hasn't changed

15 years agogstpluginfeature: API : new gst_plugin_feature_list_copy() method
Edward Hervey [Wed, 21 Oct 2009 07:40:49 +0000 (09:40 +0200)]
gstpluginfeature: API : new gst_plugin_feature_list_copy() method

This allows copying AND incrementing the refcount at the same time,
avoiding a double iteratio of the GList

15 years agogstregistry: Add a cookie for detecting feature list changes
Edward Hervey [Sat, 24 Oct 2009 08:05:59 +0000 (10:05 +0200)]
gstregistry: Add a cookie for detecting feature list changes

We also create a private structure, since we will need to add more
data there in following patches.

15 years agomessage: don't use typechecking cast macros
Wim Taymans [Fri, 23 Oct 2009 17:19:04 +0000 (13:19 -0400)]
message: don't use typechecking cast macros

Simply use casting macros for accessing the message fields like we do for
buffers and events. Avoids some costly typechecking that does not really buy us
much.

15 years agoRevert "gstmessage: Avoid expensive src/type/timestamp fetch."
Wim Taymans [Fri, 23 Oct 2009 17:13:52 +0000 (13:13 -0400)]
Revert "gstmessage: Avoid expensive src/type/timestamp fetch."

This reverts commit 61cf93a334b79a2d8493e531cc44ba45a4209805.

15 years agogstmessage: Avoid expensive src/type/timestamp fetch.
Edward Hervey [Fri, 23 Oct 2009 15:51:27 +0000 (17:51 +0200)]
gstmessage: Avoid expensive src/type/timestamp fetch.

If we've already checked that we have a valid message, use the entries
directly.

15 years agogstcaps: Use inlined version of _is_any()/_is_empty()
Edward Hervey [Fri, 23 Oct 2009 15:47:43 +0000 (17:47 +0200)]
gstcaps: Use inlined version of _is_any()/_is_empty()

CAPS_IS_ANY and CAPS_IS_EMPTY are the equivalent of their gst_caps_*
counterpart except that they avoid the typechecking and are inlined.

CAPS_IS_EMPTY_SIMPLE only checks for empty caps (without checking if
the caps is ANY).

15 years agoTODO: add item to TODO list
Wim Taymans [Thu, 22 Oct 2009 20:42:13 +0000 (16:42 -0400)]
TODO: add item to TODO list

We currently do a little too much work when we push the first buffer around
resulting in excessive caps checking. We can probably make this a little less
expensive.

15 years agogstpipeline: Simplify base time checking slightly
Jan Schmidt [Thu, 22 Oct 2009 11:52:46 +0000 (12:52 +0100)]
gstpipeline: Simplify base time checking slightly

Simplify checking and distribution of the base time - don't re-check
the value of a local variable that was set 3 lines earlier.

15 years agogsturi: Optimisation: Avoid type-checking in sorting method.
Edward Hervey [Thu, 22 Oct 2009 11:15:15 +0000 (13:15 +0200)]
gsturi: Optimisation: Avoid type-checking in sorting method.

We already know the list only contains plugin features

15 years agogsturi: Optimisation : Cast when we're sure of the type.
Edward Hervey [Thu, 22 Oct 2009 11:13:56 +0000 (13:13 +0200)]
gsturi: Optimisation : Cast when we're sure of the type.

Also directly access GstElementFactory->uri_type instead of going
through a function that will (once again) check whether it's a
GstElementFactory

15 years agotests/fakesink: Add some debugging
Edward Hervey [Thu, 22 Oct 2009 10:33:37 +0000 (12:33 +0200)]
tests/fakesink: Add some debugging

15 years agotests/faeksink: Lower the number of threads to avoid timeouts
Edward Hervey [Thu, 22 Oct 2009 10:33:01 +0000 (12:33 +0200)]
tests/faeksink: Lower the number of threads to avoid timeouts

We just end up with way too much contention in glib otherwise.

15 years agogstregistry: Use hash table when finding a feature. Fixes #598895
Edward Hervey [Mon, 19 Oct 2009 07:06:16 +0000 (09:06 +0200)]
gstregistry: Use hash table when finding a feature. Fixes #598895

15 years agotee: implement custom acceptcaps function
Wim Taymans [Wed, 21 Oct 2009 20:26:01 +0000 (16:26 -0400)]
tee: implement custom acceptcaps function

Implement a custom acceptcaps function on the sinkpad. We can accept any caps as
long as it is accepted by all downstream peer elements.

15 years agomultiqueue: avoid lock for taking the counter
Wim Taymans [Wed, 21 Oct 2009 17:38:57 +0000 (13:38 -0400)]
multiqueue: avoid lock for taking the counter

The counter for incomming data is already protected with the STREAM_LOCK so we
don't need to add another lock around it.

15 years agoregistry: hash the plugin basename
Wim Taymans [Wed, 21 Oct 2009 03:28:54 +0000 (23:28 -0400)]
registry: hash the plugin basename

Maintain a hashtable of the plugin basename. We can then use this
hashtable to speedup the search for an existing plugin and avoid
a whole lot of strcmp calls.

15 years agoregistry: speed up _strlen
Wim Taymans [Wed, 21 Oct 2009 03:27:41 +0000 (23:27 -0400)]
registry: speed up _strlen

Make the _strlen function a little tighter

15 years agoregistry: name is never NULL
Wim Taymans [Wed, 21 Oct 2009 01:43:58 +0000 (21:43 -0400)]
registry: name is never NULL

When looking up a feature by name, we never call this internal
function with NULL so we don't have to check for it.

15 years agoregistry: refactor plugin lookup
Wim Taymans [Wed, 21 Oct 2009 01:39:11 +0000 (21:39 -0400)]
registry: refactor plugin lookup

We keep lookup plugins by their basename. Avoid creating a basename
from a filename if we can.

15 years agoregistry: do quick check for . files
Wim Taymans [Wed, 21 Oct 2009 01:01:55 +0000 (21:01 -0400)]
registry: do quick check for . files

Do a quick check for . files before calling the strcmp functions

15 years agoRemove executable bits on xml unit test.
Michael Smith [Tue, 20 Oct 2009 19:21:09 +0000 (12:21 -0700)]
Remove executable bits on xml unit test.

15 years agodocs: Fix docs for gst_caps_set_simple()
Benjamin Otte [Mon, 19 Oct 2009 14:47:10 +0000 (16:47 +0200)]
docs: Fix docs for gst_caps_set_simple()

15 years agodocs: Modify docs string slightly.
Jan Schmidt [Mon, 19 Oct 2009 12:02:30 +0000 (13:02 +0100)]
docs: Modify docs string slightly.

Apparently starting the last line of a docs string with 'returns' both
confuses and enrages gtk-doc. Use a slightly different wording instead.

15 years agopluginloader: When a plugin is blacklisted, output a GST_ERROR line.
Jan Schmidt [Mon, 19 Oct 2009 11:29:35 +0000 (12:29 +0100)]
pluginloader: When a plugin is blacklisted, output a GST_ERROR line.

15 years agotools/gst-inspect: Check we're not handling NULL pointers.
Edward Hervey [Mon, 19 Oct 2009 11:30:10 +0000 (13:30 +0200)]
tools/gst-inspect: Check we're not handling NULL pointers.

15 years agotools/gst-inspect: Remove dead assignment
Edward Hervey [Mon, 19 Oct 2009 11:29:40 +0000 (13:29 +0200)]
tools/gst-inspect: Remove dead assignment

15 years agotrace: Do not poison gst_trace_add_entry()
Peter Kjellerstedt [Wed, 14 Oct 2009 08:54:32 +0000 (10:54 +0200)]
trace: Do not poison gst_trace_add_entry()

Since gst_trace_add_entry() is a macro, gcc will barf when it is
defined in case it has been poisoned due to trace support being
disabled.

15 years agocontroller: just cast in internal API where we have checked parameters already
Stefan Kost [Sun, 18 Oct 2009 20:18:58 +0000 (23:18 +0300)]
controller: just cast in internal API where we have checked parameters already

15 years agocontroller: use g_slice for controlled property structures
Stefan Kost [Sun, 18 Oct 2009 20:15:07 +0000 (23:15 +0300)]
controller: use g_slice for controlled property structures

Use g_slide instead of nomal g_new, Also don't init struct with 0 as we need to
init it anyway with the real values.
Also join the 3 flags checks into one.

15 years agolibs/base/typefindhelper: Remove useless typechecking in tight loop
Edward Hervey [Sun, 18 Oct 2009 15:17:17 +0000 (17:17 +0200)]
libs/base/typefindhelper: Remove useless typechecking in tight loop

The list against which we run the comparefunc will only contain
GstPluginFeature, therefore remove the 6 expensive type checks we do
for every single comparision.

15 years agocaps: fix typo in docs
Wim Taymans [Fri, 16 Oct 2009 10:39:54 +0000 (12:39 +0200)]
caps: fix typo in docs

15 years agowin32: Add new API symbol
Edward Hervey [Fri, 16 Oct 2009 07:43:08 +0000 (09:43 +0200)]
win32: Add new API symbol

15 years agoAutomatic update of common submodule
Stefan Kost [Fri, 16 Oct 2009 07:13:53 +0000 (10:13 +0300)]
Automatic update of common submodule

From 85d1530 to 0702fe1

15 years agoImprove caps setters API
Benjamin Otte [Wed, 7 Oct 2009 13:32:18 +0000 (15:32 +0200)]
Improve caps setters API

This patch adds gst_caps_set_value() and allows gst_caps_set_simple() to
work on non-simple caps. See the API documentation for the functions
about what they do.

The intention of these changes is to ease working with caps in caps
transform functions. An example for this would be ffmpegcolorspace,
where the caps transform function could be changed to look roughly like
this (pseudocode ahead):
  result = gst_caps_copy (template_caps);
  value = gst_structure_get_value (gst_caps_get_structure (caps, 0),
    "widh");
  gst_caps_set_value (result, value);
  /* same for height, framerate and par */
  return caps;
which is much cleaner and easier to understand than the current code.

https://bugzilla.gnome.org/show_bug.cgi?id=597690

15 years agoAdd XML_LIBS when building tests that use xml-specific functions
Benjamin Otte [Fri, 2 Oct 2009 08:15:55 +0000 (10:15 +0200)]
Add XML_LIBS when building tests that use xml-specific functions

15 years agopo: Don't create backup .po files
Jan Schmidt [Thu, 15 Oct 2009 15:35:59 +0000 (16:35 +0100)]
po: Don't create backup .po files

As well as preventing creation of useless backup files, it works
around a bug in gettext 0.17 on OS/X

15 years agobasesrc: fix race in PLAYING->PAUSED->PLAYING
Wim Taymans [Thu, 15 Oct 2009 14:30:36 +0000 (16:30 +0200)]
basesrc: fix race in PLAYING->PAUSED->PLAYING

When we quickly switch from PLAYING to PAUSED and back to PLAYING it's possible
in some cases that the task refuses to start, This is because when we go to
PAUSED, we unschedule the clock timeout, which could return UNSCHEDULED when
we're back to PLAYING, causing the task to PAUSE again with a wrong-state.

This patch checks if we are running when we return with an UNSCHEDULED return
value and if we are, try to create a new buffer.

Fixes #597550

15 years agodocs: clarify preset api docs
Stefan Kost [Thu, 15 Oct 2009 09:16:05 +0000 (12:16 +0300)]
docs: clarify preset api docs

15 years agodocs: fix controller sections docs
Stefan Kost [Wed, 14 Oct 2009 14:57:40 +0000 (17:57 +0300)]
docs: fix controller sections docs

15 years agoAutomatic update of common submodule
Edward Hervey [Wed, 14 Oct 2009 08:40:50 +0000 (10:40 +0200)]
Automatic update of common submodule

From 6380d4b to 85d1530

15 years agoiterator: Fix a documentation typo
Peter Kjellerstedt [Wed, 14 Oct 2009 08:16:31 +0000 (10:16 +0200)]
iterator: Fix a documentation typo

15 years agobuild: ...and add missing endif
Sebastian Dröge [Wed, 14 Oct 2009 06:57:52 +0000 (08:57 +0200)]
build: ...and add missing endif

15 years agobuild: Also don't run make check-exports if debugging is disabled
Sebastian Dröge [Wed, 14 Oct 2009 06:57:04 +0000 (08:57 +0200)]
build: Also don't run make check-exports if debugging is disabled

15 years agobuild: Only run make check-exports if no public API was disabled
Peter Kjellerstedt [Wed, 14 Oct 2009 06:50:31 +0000 (08:50 +0200)]
build: Only run make check-exports if no public API was disabled

Fixes bug #598297.

15 years agogstobject: Replace recursive gst_object_has_ancestor() with an iterative version
Sebastian Dröge [Wed, 14 Oct 2009 06:30:07 +0000 (08:30 +0200)]
gstobject: Replace recursive gst_object_has_ancestor() with an iterative version

This is slightly more efficient because the compiler can't do tail
recursion here and has to keep all stack frames.

Not that efficiency is that important here but I already had
the iterative version somewhere else and both are easy to read.

15 years agogstobject: Add simple unit test for gst_object_has_ancestor()
Sebastian Dröge [Wed, 14 Oct 2009 06:29:32 +0000 (08:29 +0200)]
gstobject: Add simple unit test for gst_object_has_ancestor()

15 years agonetclientclock: fix timestamp comparission, Fixes #597407
Tommi Myöhänen [Tue, 13 Oct 2009 16:12:50 +0000 (19:12 +0300)]
netclientclock: fix timestamp comparission, Fixes #597407

15 years agocheck: Disable the test_fail_abstract_new() test entirely on OS/X
Jan Schmidt [Mon, 12 Oct 2009 20:51:55 +0000 (21:51 +0100)]
check: Disable the test_fail_abstract_new() test entirely on OS/X

Fixes a compiler warning from the function being compiled but not
used.

15 years agodebug: Mark the GST_POLL symbol as extern to avoid multiply-defined error
Jan Schmidt [Mon, 12 Oct 2009 13:57:35 +0000 (14:57 +0100)]
debug: Mark the GST_POLL symbol as extern to avoid multiply-defined error

15 years agoUpdate common to 6380d4b370f078f0cca7240428ea9f6639571ff5
Jan Schmidt [Mon, 12 Oct 2009 13:47:30 +0000 (14:47 +0100)]
Update common to 6380d4b370f078f0cca7240428ea9f6639571ff5

15 years agogstpoll: Make the new GST_POLL debug completely private
Jan Schmidt [Mon, 12 Oct 2009 13:24:04 +0000 (14:24 +0100)]
gstpoll: Make the new GST_POLL debug completely private

Make the GST_POLL debug category symbol private to libgstreamer, as
there should be no external users of it.

15 years agochecks: Disable a fairly silly gstobject test on OS/X
Jan Schmidt [Mon, 12 Oct 2009 13:22:34 +0000 (14:22 +0100)]
checks: Disable a fairly silly gstobject test on OS/X

This test used to SIGBUS on OS/X but now SIGSEGV's instead on
Snow Leopard. It's not worth the effort to figure out which platform
should produce which error for what is fundamentally a pretty silly
test, so just disable it on OS/X

15 years agolibs/gst/check: Make writing threadsafe. Backported from libcheck trunk
Edward Hervey [Mon, 12 Oct 2009 11:50:51 +0000 (13:50 +0200)]
libs/gst/check: Make writing threadsafe. Backported from libcheck trunk

15 years agolibs/gst/check: Run gst-indent on libcheck.
Edward Hervey [Mon, 12 Oct 2009 11:49:35 +0000 (13:49 +0200)]
libs/gst/check: Run gst-indent on libcheck.

15 years agogstpluginloader: Don't wait forever on gst_poll_wait.
Edward Hervey [Mon, 12 Oct 2009 10:02:34 +0000 (12:02 +0200)]
gstpluginloader: Don't wait forever on gst_poll_wait.

This allows the macosx versions to properly error out when fds are closed.

This is only a temporary fix until the pluginloader is switched to not
use GstPoll but GIOChannels.

15 years agogstpoll: Only take into account active fds
Edward Hervey [Mon, 12 Oct 2009 10:01:59 +0000 (12:01 +0200)]
gstpoll: Only take into account active fds

This is needed so that select properly errors out on macosx (sigh)

15 years agogstpoll: Add some debugging statements
Edward Hervey [Mon, 12 Oct 2009 08:07:03 +0000 (10:07 +0200)]
gstpoll: Add some debugging statements

15 years agogstpoll: Use the error fdset when using select/pselect.
Edward Hervey [Mon, 12 Oct 2009 08:01:01 +0000 (10:01 +0200)]
gstpoll: Use the error fdset when using select/pselect.

This is needed to properly detect fds that are closed or that got
an error