Tim-Philipp Müller [Tue, 6 Jul 2010 09:48:27 +0000 (10:48 +0100)]
checks: rewrite gsttagsetter test to use fewer g_usleep()
Something about that seems to interact badly with some schedulers,
so do things differently.
Fixes #623469.
Tim-Philipp Müller [Tue, 6 Jul 2010 09:19:09 +0000 (10:19 +0100)]
docs: add new add gst_clock_id_wait_async_full() to docs
API: add gst_clock_id_wait_async_full
See #623589.
Edward Hervey [Tue, 6 Jul 2010 08:28:14 +0000 (10:28 +0200)]
gstvalue: Improve gst_value_{set|get}_caps docs
Alessandro Decina [Tue, 6 Jul 2010 08:35:09 +0000 (10:35 +0200)]
win32: export gst_clock_id_wait_async_full
Alessandro Decina [Tue, 6 Jul 2010 08:31:25 +0000 (10:31 +0200)]
tests: remove ABI checks for GstClockEntry.
Alessandro Decina [Mon, 5 Jul 2010 16:45:55 +0000 (18:45 +0200)]
clock: document that GstClockEntry should be treated as ana opaque structure.
Alessandro Decina [Mon, 5 Jul 2010 11:10:09 +0000 (13:10 +0200)]
clock: use the new gst_clock_id_wait_async_full.
Use the new gst_clock_id_wait_async_full in gst_clock_set_master.
Also add some tests.
Alessandro Decina [Mon, 5 Jul 2010 11:01:53 +0000 (13:01 +0200)]
clock: fix refcounting bug in gst_clock_set_
Make sure clock->clockid is unreffed before clock->master.
gst_clock_id_unschedule (clock->clockid) tries to access clock->master. If
clock->master is unreffed before and it's deallocated, _unschedule could access
free'd memory.
Alessandro Decina [Mon, 5 Jul 2010 10:56:40 +0000 (12:56 +0200)]
clock: add gst_clock_id_wait_async_full.
Add gst_clock_id_wait_async_full. It's the same as gst_clock_id_wait_async but
allows passing a GDestroyNotify to destroy user_data.
Stefan Kost [Mon, 5 Jul 2010 14:50:33 +0000 (17:50 +0300)]
docs: improve the api docs for new GstPadLinkChecks and its use
Stefan Kost [Mon, 5 Jul 2010 09:21:51 +0000 (12:21 +0300)]
README: update after removal of "old" dir.
Remove "old" and add a line about "examples".
Tim-Philipp Müller [Sun, 4 Jul 2010 16:34:30 +0000 (17:34 +0100)]
Remove old 0.8 tests and examples from git tree
Doesn't really look like anything's worth keeping.
Tim-Philipp Müller [Sat, 3 Jul 2010 15:39:40 +0000 (16:39 +0100)]
check: skip silly test that segfaults when in a CK_FORK=no environment
See #623469.
Tim-Philipp Müller [Sat, 3 Jul 2010 14:13:14 +0000 (15:13 +0100)]
checks: make fakesrc check work in a CK_FORK=no environment
Reset have_eos at the beginning of each test.
See #623469.
Tim-Philipp Müller [Sat, 3 Jul 2010 13:09:36 +0000 (14:09 +0100)]
checks: run tests calling gst_deinit() last so things work with CK_FORK=no
Because gst_init() will fail once gst_deinit() has been called.
See #623469.
Tim-Philipp Müller [Sat, 3 Jul 2010 13:04:32 +0000 (14:04 +0100)]
checks: don't assume element factory is not loaded yet
It may already be loaded if check is being run with CK_FORK=no.
See #623469.
Tim-Philipp Müller [Thu, 1 Jul 2010 18:58:09 +0000 (19:58 +0100)]
docs: fix a couple of typos in the manual
Spotted by Alexander Saprykin.
Fixes #622379.
Edward Hervey [Thu, 1 Jul 2010 15:56:33 +0000 (17:56 +0200)]
gstcaps: Make sure _normalize() is applied on all structures.
We need to use gst_caps_get_size() in the loop counter since some
structures could be added while iterating.
Fixes #623301
Tim-Philipp Müller [Wed, 30 Jun 2010 12:16:35 +0000 (13:16 +0100)]
docs: update 'XML in GStreamer' section in application developer's manual
Tim-Philipp Müller [Tue, 29 Jun 2010 17:48:05 +0000 (18:48 +0100)]
0.10.29.3 pre-release
Wim Taymans [Fri, 25 Jun 2010 17:03:27 +0000 (19:03 +0200)]
multiqueue: implement acceptcaps function
Our acceptcaps function can simply forward the query.
Tim-Philipp Müller [Mon, 28 Jun 2010 14:28:59 +0000 (15:28 +0100)]
Bump automake requirement to 1.10
For maintainability reasons and $(builddir).
Fixes #622944.
Tim-Philipp Müller [Mon, 28 Jun 2010 12:56:00 +0000 (13:56 +0100)]
tools: mention --eos-on-shutdown on gst-launch man page
Sebastian Dröge [Mon, 28 Jun 2010 08:20:39 +0000 (10:20 +0200)]
utils: Don't use G_GNUC_CONST for the uint64 scaling functions
They are actually *not* const functions because on architectures
without int128 instructions the parameters were changed.
gcc re-used the parameters on the stack for multiple calls though
and the changed parameters were used for the second call then.
Fixes bug #623003.
Tim-Philipp Müller [Sat, 26 Jun 2010 16:48:31 +0000 (17:48 +0100)]
0.10.29.2 pre-release
Tim-Philipp Müller [Sat, 26 Jun 2010 16:47:55 +0000 (17:47 +0100)]
po: update translations
Tim-Philipp Müller [Sat, 26 Jun 2010 09:16:36 +0000 (10:16 +0100)]
examples: remove xml example build system bits and purge from tree
Fixes make distcheck.
Tim-Philipp Müller [Sat, 26 Jun 2010 08:59:31 +0000 (09:59 +0100)]
xml: keep dummy gst_xml_get_type() function for g-i even if rest of GstXML is removed
Add a minimal gst_xml_get_type() function, so that gobject-introspection doesn't
break the compilation if we're compiling with GST_REMOVE_DEPRECATED defined or
--disable-loadsave having been passed to configure. Until someone figures out
a better way at least.
Tim-Philipp Müller [Sat, 26 Jun 2010 00:01:49 +0000 (01:01 +0100)]
Don't include <libxml/parser.h> from public headers if GST_DISABLE_DEPRECATED is defined
Since everything GstXML related has been deprecated, we can now skip the
libxml includes from the public headers when GST_DISABLE_DEPRECATED is
defined.
See #463435.
Tim-Philipp Müller [Fri, 25 Jun 2010 23:18:10 +0000 (00:18 +0100)]
examples: add missing stdlib.h include in typefind example
Stefan Kost [Fri, 25 Jun 2010 18:23:22 +0000 (21:23 +0300)]
dot-dump: terminate truncated strings and escape special chars
Fixes syntax errors in generated dot files for caps with strings.
Edward Hervey [Fri, 25 Jun 2010 16:52:02 +0000 (18:52 +0200)]
pad: more documentation regarding the new flags
Wim Taymans [Fri, 25 Jun 2010 16:18:55 +0000 (18:18 +0200)]
pad: make the NOTHING link check flag be 0
Make the pad link check of NOTHING be 0. This way we have a flag for each
feature and 0 when no flags are set.
Sebastian Dröge [Fri, 25 Jun 2010 16:24:47 +0000 (18:24 +0200)]
tests: Remove GstXML tests
Sebastian Dröge [Fri, 25 Jun 2010 16:13:57 +0000 (18:13 +0200)]
gstxml: Deprecate GstXml and related functions
Pipeline serialisation to and from XML is horribly broken for all
but the most simple use cases, and will likely never be fixed.
Make sure everyone playing around with these tools is aware of
this, to avoid frustration. See countless bug reports in bugzilla.
Fixes bug #622685.
Sebastian Dröge [Fri, 25 Jun 2010 16:11:30 +0000 (18:11 +0200)]
controller: Fix build with GST_REMOVE_DEPRECATED
Edward Hervey [Thu, 24 Jun 2010 10:19:20 +0000 (12:19 +0200)]
benchmarks: Use gst_element_link_pads_full
We're testing caps negotiation, not pad linking. Brings the startup
time down 100 fold.
Tim-Philipp Müller [Thu, 24 Jun 2010 16:53:36 +0000 (17:53 +0100)]
tools: deprecate gst-xmllaunch and print fat warning if someone tries to use it
Pipeline serialisation to and from XML is horribly broken for all
but the most simple use cases, and will likely never be fixed.
Make sure everyone playing around with these tools is aware of
this, to avoid frustration. See countless bug reports in bugzilla.
Tim-Philipp Müller [Thu, 24 Jun 2010 16:22:13 +0000 (17:22 +0100)]
gst-launch: rename new --no-play command line option to --no-sigusr-handler
--no-play seems a bit confusing.
Fixes #621867.
Stefan Kost [Thu, 24 Jun 2010 12:07:11 +0000 (15:07 +0300)]
Automatic update of common submodule
From 73ff93a to a519571
Tim-Philipp Müller [Wed, 23 Jun 2010 10:02:16 +0000 (11:02 +0100)]
binaryregistry: ignore the plugin cache if the filter environment has changed
Make sure that we properly update the registry and the cache file whenever
the filter environment changes or there's no more filter set.
Tim-Philipp Müller [Thu, 27 May 2010 11:36:10 +0000 (12:36 +0100)]
pluginloading: add support for whitelisting based on plugin or source module name and path
This feature is primarily intended for use in plugin modules' unit tests.
Consider the following situation: gst-plugins-good is built against an
installed GStreamer core. An older version of gst-plugins-good is also
installed in that prefix, along with random other plugin modules. Now,
when doing 'make check' in the just-built gst-plugins-good tree, we
want to only load plugins from GStreamer core, gst-plugins-base, and
gst-plugins-good, but not random other modules (we don't want any unit
tests to fail just because some module in gst-plugins-bad has a broken
plugin_init, for example). Also, we want to only load gst-plugins-good
modules from the locally-built source tree, but not any of the older
gst-plugins-good modules installed. This is usually assured by loading
the ones in the source tree first (by adding that path first to the
right environment variables), but it gets tricky when plugins are
moved, removed, merged, or renamed, or the plugin filename changes.
Note that 'make check' should really work right without doing
'make install' or uninstalling the old gst-plugins-good package (or
any other gst-plugins-foo package) first.
Enter GST_PLUGIN_LOADING_WHITELIST. This environment variable may
contain source-package@path-prefix pairs separated by the platform
search path separator (G_SEARCHPATH_SEPARATOR_S). The source package
and path prefix are separated by the '@' character. The path prefix is
entirely optional, as is the '@' separator if no path is given.
It is also possible to filter based on plugin names instead of the name
of the source-package by specifying one or more plugin names separated
by commas before the optional path prefix.
In short, the following match patterns are possible:
plugin1,plugin2@pathprefix or
plugin1,plugin2@* or just
plugin1,plugin2 or
source-package@pathprefix or
source-package@* or just
source-package
So for our gst-plugins-good unit test example above, we would set the
environment variable on *nix to something like this (will likely be a
relative path in practice):
gstreamer:gst-plugins-base:gst-plugins-good@/path/to/src/gst-plugins-good
Fixes #619815 and #619717.
Edward Hervey [Wed, 23 Jun 2010 15:24:07 +0000 (17:24 +0200)]
gstghostpad: We don't need any checks when linking target pad
https://bugzilla.gnome.org/show_bug.cgi?id=622504
Edward Hervey [Wed, 23 Jun 2010 15:00:17 +0000 (17:00 +0200)]
gstutils: New gst_element_link_pads_full method
Links the elements with the specified pad linking checks.
API:gst_element_link_pads_full
https://bugzilla.gnome.org/show_bug.cgi?id=622504
Edward Hervey [Wed, 23 Jun 2010 14:45:19 +0000 (16:45 +0200)]
GstPad: Add new pad linking method with configurable checks.
To be used for cases where we don't need all checks to be validated.
API: gst_pad_link_full
API: GstPadLinkCheck
https://bugzilla.gnome.org/show_bug.cgi?id=622504
Thiago Santos [Tue, 15 Jun 2010 21:26:01 +0000 (18:26 -0300)]
tag: Adds GST_TAG_IMAGE_ORIENTATION tag
Adds a new tag to inform about the image orientation and how
to rotate and flip it before display.
Note that this tag is a string with a predefined set of
possible values.
API: GST_TAG_IMAGE_ORIENTATION
Fixes #619508
Tim-Philipp Müller [Tue, 22 Jun 2010 17:53:15 +0000 (18:53 +0100)]
gstobject: deprecate gst_object_{set|get}_name_prefix()
The name prefix stuff has never been used for anything and it doesn't
look like we'll ever want to use it for anything.
Fixes #621006.
Johan Dahlin [Tue, 22 Jun 2010 13:20:53 +0000 (10:20 -0300)]
Add gobject-introspection annotations for GstPadIntLinkFunction
Fixes build with latest gobject-introspection from git.
https://bugzilla.gnome.org/show_bug.cgi?id=622025
Wim Taymans [Mon, 21 Jun 2010 09:41:46 +0000 (11:41 +0200)]
queue2: fix merging of ranges
When we merge two ranges, don't updata the current range writing_pos with
whereever we were writing earlier in the merged range. Spotted by bilboed.
Tim-Philipp Müller [Sat, 19 Jun 2010 10:19:37 +0000 (11:19 +0100)]
registry: also skip .deps dirs when scanning for plugins
No need to descend into .deps dirs in uninstalled setups, we know
these don't contain any plugins.
Thiago Santos [Thu, 17 Jun 2010 14:39:04 +0000 (11:39 -0300)]
taglist: add gst_tag_list_peek_string_index to avoid a copy
Adds a variation of the _get_string_index function that doesn't copy
the string.
API: gst_tag_list_peek_string_index
https://bugzilla.gnome.org/show_bug.cgi?id=621896
Tim-Philipp Müller [Fri, 18 Jun 2010 11:00:53 +0000 (12:00 +0100)]
typefind: make sure buffers' metadata is writable before setting caps on them
Fixes warnings when using playbin2 with dvb:// streams, where typefind
comes after mpegtsparse.
Tim-Philipp Müller [Thu, 17 Jun 2010 14:52:57 +0000 (15:52 +0100)]
docs: update introspected plugin docs for gstdoc-scanobj changes
Update common for latest gstdoc-scanobj and inspect xml files for
escaping and pad template order changes.
Sebastian Dröge [Thu, 17 Jun 2010 11:19:24 +0000 (13:19 +0200)]
basesink: Initialize jitter to prevent printing an uninitialized variable if waiting for the clock failed
Tim-Philipp Müller [Thu, 17 Jun 2010 09:34:51 +0000 (10:34 +0100)]
po: update for new strings
Tim-Philipp Müller [Thu, 17 Jun 2010 08:33:43 +0000 (09:33 +0100)]
manual: fix another typo and some inaccuracies
Fix a wrong statement and flesh out section on messages and queries
a bit.
Alexander Saprykin [Thu, 17 Jun 2010 07:05:58 +0000 (09:05 +0200)]
manual: Fix another typo
Alexander Saprykin [Thu, 17 Jun 2010 07:05:28 +0000 (09:05 +0200)]
manual: Fix typo
Johan Dahlin [Wed, 16 Jun 2010 16:11:06 +0000 (13:11 -0300)]
elementfactory: Add an allow-none annotation
https://bugzilla.gnome.org/show_bug.cgi?id=621773
Johan Dahlin [Wed, 16 Jun 2010 16:10:26 +0000 (13:10 -0300)]
miniobject: Add introspection annotations
These are required to know how to unref/ref and
convert to/from a GValue.
https://bugzilla.gnome.org/show_bug.cgi?id=621773
Johan Dahlin [Wed, 16 Jun 2010 16:10:13 +0000 (13:10 -0300)]
event: Add out annotations
https://bugzilla.gnome.org/show_bug.cgi?id=621773
Johan Dahlin [Wed, 16 Jun 2010 16:10:06 +0000 (13:10 -0300)]
query: Add out annotations
https://bugzilla.gnome.org/show_bug.cgi?id=621773
Johan Dahlin [Wed, 16 Jun 2010 16:09:57 +0000 (13:09 -0300)]
message: Add out annotations
https://bugzilla.gnome.org/show_bug.cgi?id=621773
Wim Taymans [Wed, 16 Jun 2010 11:00:30 +0000 (13:00 +0200)]
fdsink: make sync property work correctly
Don't override the default get_times vmethod so that we can use the sync
property.
Set the default sync property to FALSE. It used to be set to TRUE but because
the get_times was NULL, it always behaved like FALSE.
Fixes #621530
Benjamin Gaignard [Tue, 15 Jun 2010 16:48:53 +0000 (18:48 +0200)]
element: Improve gst_element_get_name() docs
Fixes bug #621660.
Edward Hervey [Tue, 15 Jun 2010 14:49:04 +0000 (16:49 +0200)]
Automatic update of common submodule
From 9339ccc to 35617c2
Stefan Kost [Tue, 15 Jun 2010 13:53:35 +0000 (16:53 +0300)]
Automatic update of common submodule
From 5adb1ca to 9339ccc
Stefan Kost [Tue, 15 Jun 2010 13:34:37 +0000 (16:34 +0300)]
Automatic update of common submodule
From 57c89b7 to 5adb1ca
Stefan Kost [Tue, 15 Jun 2010 12:31:12 +0000 (15:31 +0300)]
Automatic update of common submodule
From c804988 to 57c89b7
Edward Hervey [Tue, 15 Jun 2010 09:48:26 +0000 (11:48 +0200)]
Revert "GstPad: Add GST_PAD_NEGOTIABLE GstPadFlag"
This reverts commit
dc38e75d88bd8921895821f7afed01cab30e46c9.
boom
Edward Hervey [Tue, 15 Jun 2010 09:48:17 +0000 (11:48 +0200)]
Revert "gstpad: Return pad template in get_caps if pad is not negotiable"
This reverts commit
7460321a600438966d7152ab2b4318be48eadce0.
crack
Edward Hervey [Tue, 15 Jun 2010 09:48:07 +0000 (11:48 +0200)]
Revert "pad: fix comment"
This reverts commit
8e92cb4a7d56cdfa4674315c64b58c1b1b9d8208.
whatever...
Edward Hervey [Tue, 15 Jun 2010 09:47:57 +0000 (11:47 +0200)]
Revert "element: only clear negotiable when going to NULL"
This reverts commit
8f5ec1f737c3b37538b2307aef160d9d21f1c422.
bleeeeh
Tim-Philipp Müller [Tue, 15 Jun 2010 09:46:08 +0000 (10:46 +0100)]
info: add dummy TRACE log level macros for when debugging is disabled
Forgot those when adding the original API, just like the API markers
in the commit message:
API: GST_TRACE
API: GST_TRACE_OBJECT
API: GST_CAT_TRACE
API: GST_CAT_TRACE_OBJECT
API: GST_LEVEL_TRACE
Fixes compilation with --disable-gst-debug
Tim-Philipp Müller [Tue, 15 Jun 2010 00:15:54 +0000 (01:15 +0100)]
gst-launch: print more errors to stderr and suppress more output in quiet mode
If --quiet is given, don't print anything but errors. Also, make
sure errors are always printed to stderr and not to stdout.
Fixes #621595.
Wim Taymans [Mon, 14 Jun 2010 16:07:23 +0000 (18:07 +0200)]
element: only clear negotiable when going to NULL
Don't clear the negotiable flag when going to READY because then it will never
be set to TRUE again.
Wim Taymans [Mon, 14 Jun 2010 15:33:16 +0000 (17:33 +0200)]
pad: fix comment
Edward Hervey [Mon, 17 May 2010 13:06:37 +0000 (15:06 +0200)]
gstpad: Return pad template in get_caps if pad is not negotiable
https://bugzilla.gnome.org/show_bug.cgi?id=618644
Edward Hervey [Mon, 17 May 2010 13:04:48 +0000 (15:04 +0200)]
GstPad: Add GST_PAD_NEGOTIABLE GstPadFlag
A pad is 'negotiable' when its container element is in a state greater
than GST_STATE_READY
API:gst_pad_is_negotiable
API:gst_pad_set_negotiable
API:GST_PAD_NEGOTIABLE
https://bugzilla.gnome.org/show_bug.cgi?id=618644
Wim Taymans [Mon, 14 Jun 2010 14:51:39 +0000 (16:51 +0200)]
fakesink: use method to set sync property
Use the basesink method to configure the sync property instead of poking the
parent structure.
Wim Taymans [Mon, 14 Jun 2010 14:50:41 +0000 (16:50 +0200)]
filesink: use the default get_times function
Use the default get_times function of basesink so that we honour the sync
property instead of never synchronizing to the clock.
Fixes #621530
Wim Taymans [Mon, 14 Jun 2010 14:20:18 +0000 (16:20 +0200)]
basetransform: reevaluate proxy_alloc when reconfigured
When we reconfigure the transform element, make sure we reevaluate the proxying
of buffer_alloc the next time around.
Fixes #621332
Sebastian Dröge [Mon, 14 Jun 2010 13:39:57 +0000 (15:39 +0200)]
caps: Don't use invalid fraction range in the unit test
Sebastian Dröge [Mon, 14 Jun 2010 13:30:08 +0000 (15:30 +0200)]
gstvalue: Add some more assertions and checks for valid input parameters
Mark Nauwelaerts [Thu, 27 May 2010 13:13:18 +0000 (15:13 +0200)]
adapter: optimize progressive masked_scan
Retain the last scanned buffer entry and offset, so we can resume buffer
scanning there in case of a typical progressive scan.
Also potentially optimize _copy subsequently occurring in that area.
Mark Nauwelaerts [Thu, 27 May 2010 10:15:22 +0000 (12:15 +0200)]
adapter: add extended masked_scan_uint32_peek that also provides matching value
Also add to .def and docs.
Fixes #619828.
API: gst_adapter_masked_scan_uint32_peek
Edward Hervey [Mon, 14 Jun 2010 11:38:41 +0000 (13:38 +0200)]
win32: fix .def file
Tim-Philipp Müller [Mon, 14 Jun 2010 11:25:33 +0000 (12:25 +0100)]
docs: add Since: tag for new gst_caps_steal_structure
Tim-Philipp Müller [Fri, 11 Jun 2010 14:36:10 +0000 (15:36 +0100)]
docs: fix example to use a category name that actually exists
Edward Hervey [Mon, 14 Jun 2010 09:39:40 +0000 (11:39 +0200)]
gstcaps: New gst_caps_steal_structure() method
This allows removing structures from caps without them being freed. Helpful when
plugins need to move around structures without having to do an expensive structure
copy.
API:gst_caps_steal_structure
https://bugzilla.gnome.org/show_bug.cgi?id=621527
Sebastian Dröge [Mon, 14 Jun 2010 11:10:07 +0000 (13:10 +0200)]
configure: Don't add G_THREADS_MANDATORY to GST_ALL_CFLAGS
It's already included in GLIB_EXTRA_CFLAGS
Sebastian Dröge [Mon, 14 Jun 2010 11:07:33 +0000 (13:07 +0200)]
configure: use GLIB_EXTRA_CFLAGS
Sebastian Dröge [Mon, 14 Jun 2010 11:02:08 +0000 (13:02 +0200)]
Automatic update of common submodule
From 7a0fdf5 to c804988
Wim Taymans [Mon, 14 Jun 2010 11:01:55 +0000 (13:01 +0200)]
capsfilter: fix printf format
Wim Taymans [Mon, 14 Jun 2010 10:39:47 +0000 (12:39 +0200)]
capsfilter: implement custom accept_caps method
Implement a custom acceptcaps function. We can simply check if there is an
intersection with the new caps. This makes the accept caps function much faster.
See #621190
Wim Taymans [Mon, 14 Jun 2010 10:36:54 +0000 (12:36 +0200)]
basetransform: add accept_caps vmethod
Allow subclasses to override the acceptcaps function because in some cases a
custom implementation can be much much faster than the default one.
See #621190
Sebastian Dröge [Mon, 14 Jun 2010 09:30:20 +0000 (11:30 +0200)]
Automatic update of common submodule
From 6da3bab to 7a0fdf5
Olivier Crête [Fri, 11 Jun 2010 22:49:02 +0000 (18:49 -0400)]
capsfilter: Remove transform_size
GstBaseTransform now assumes that the size is the same if there is not
transform_size.
https://bugzilla.gnome.org/show_bug.cgi?id=621334
Olivier Crête [Fri, 11 Jun 2010 22:46:30 +0000 (18:46 -0400)]
basetransform: Assume size is the same if no transform_size/get_unit_size
Subclasses that don't implemen transform_size should be assumed to produce output
buffers of the same size.
https://bugzilla.gnome.org/show_bug.cgi?id=621334
Sebastian Dröge [Mon, 14 Jun 2010 06:18:00 +0000 (08:18 +0200)]
gstvalue: Don't initialize arrays from variables