Add the minimum-update-interval property to the clock, with a default
of 50ms and don't send polling requests faster than that. That helps to
ensure we spread the initial observations out a little - startup takes
a little longer, but tracking is more stable.
Move the discont skew limiting code inside an if statement, so that
it's only done when the linear regression succeeds and the clock
parameters might actually change.
Allow setting a GstBus on the network clock client
via a new 'bus' object property. If a bus is set, the
clock will output an element message containing statistics
about new clock observations and the clock correlation.
When the local clock is synchronised with the remote, limit the
maximum jump in the clock at any point to be one average RTT to
the server. Also, publish in the bus message whether we are
synched with the remote or not.
* GstGlobalDeviceMonitor was renamed to GstDeviceMonitor
* Expand GST_MESSAGE_DEVICE to the full enum value names
* Correct the incorrect references to the GstDeviceProvider interfaces
* Describe caps arguments for gstcheck interface
* Add missing docs for GstNetAddressMeta and its add function
* Add docs for toc helper macros
* Avoid refering to GstValueList type as done elsewhere
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732786
They are very confusing for people, and more often than not
also just not very accurate. Seeing 'last reviewed: 2005' in
your docs is not very confidence-inspiring. Let's just remove
those comments.
Fix a typo in a doc string - the property is round-trip-limit, not
roundtrip-limit.
Remove a bogus GST_WARNING that can print an uninitialised variable
and is redundant anyway.
Sometimes, packets might take a very long time to return. Such packets
usually are way too late and destabilize the regression with their
obsolete data. On Wi-Fi, round-trips of over 7 seconds have been observed.
If the limit is set to a nonzero value, packets with a round-trip period
larger than the limit are ignored.
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
https://bugzilla.gnome.org/show_bug.cgi?id=712385
Keep a rolling average of the round trip time for network clock
observations, favouring shorter round trips as being more accurate.
Don't pass any clock observation to the clock slaving if it has a
round-trip time greater than 2 times the average.
Actual shifts in the network topology will be noticed after some
time, as the rolling average incorporates the new round trip times.
Even though this parameter is not used, it should be const to fit in with the
coding standards for other similar parameters. Client code already passes in
const strings under the expectation that they won’t be modified.
https://bugzilla.gnome.org/show_bug.cgi?id=710442
The _1_0 suffixed environment variables override the
non-suffixed ones, so if we're in an environment that
sets the _1_0 suffixed ones, such as jhbuild, we need
to set those to make sure ours actually always get
used.
Using multiple libraries causes problems for the C# bindings and
will for similiar languages such as Java when there are bindings
for them.
Also change --library=libgstfoo-X.la to --library=gstfoo-X as
the man page suggests it should be done.
https://bugzilla.gnome.org/show_bug.cgi?id=679315
No need to use a custom main context and custom timeout sources,
just use g_socket_condition_timed_wait() instead, which was added
for exactly this case.
Also seems to help with the unit test deadlocking with glib 2.33.x
https://bugzilla.gnome.org/show_bug.cgi?id=681575
Remove GST_MAJORMINOR and replace it by GST_API_VERSION
Also set GST_VERSION_{MAJOR,MINOR,MICRO,NANO} explicitely
now.
All versions are at 1.0.0 now for the release soon but
API/ABI can still change until the 1.0.0 release.
Next release versions until 1.0.0 will be 0.10.9X and
these will be release candidates. GST_VERSION_* will
nonetheless stay at 1.0.0.0.
Split out the registration of the metadata API and its implementation. Make a
GType for each metadata API. This allows us to store extra information with the
API type such as the tags.
Change the buffer API so that we can get the metadata using the API GType.
Change the query API so that we use the metadata API GType in the allocation
query instead of a string.
Update netaddress and unit tests
Add support for adding tags to the metadata. with some standard keys, this
should make it possible to describe what the metadata refers to. We should be
able to use this information to decide if a transformation destroys the metadata
or not.
Flesh out the transform method. Add a type and extra info to the transform
function so that implementation can transform the metadata.
Remove the copy function and replace with the more generic transform.
Remove the android/ top dir
Fixe the Makefile.am to be androgenized
To build gstreamer for android we are now using androgenizer which generates the needed Android.mk files.
Androgenizer can be found here: http://git.collabora.co.uk/?p=user/derek/androgenizer.git
Apply fix from libgstbase to all core libs now that we know that it
works. Should fix problems with g-ir-scanner using the wrong
(ie. system) libgstreamer, leading to linking errors such as
undefined reference to `gst_clock_single_shot_id_reinit'.
https://bugzilla.gnome.org/show_bug.cgi?id=637549
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
Sets up a GST_PKG_CONFIG_PATH variable for use in Makefile.am
(avoids trailing ':' in PKG_CONFIG_PATH used). A useful side
effect of this is also that the PKG_CONFIG_PATH environment
is now logged in the configure output.
Point g-ir-scanner to the .la file of our library, which hopefully
makes it find the right dependencies in all cases (ie. our locally
built libgstreamer and not the system-installed one). This is also
how it's done in Gtk+ and how it's documented in the wiki, see
http://live.gnome.org/GObjectIntrospection/AutotoolsIntegration
Based on patches by Vincent Untz and Alan Knowles.
Fixes#603710.
Our own pkgconfig directory should come first, so that pkg-config uses
the in-tree libgstreamer and not some external one when --pkg=gstreamer-0.10
is passed to g-ir-scanner.
See #603710.
Original commit message from CVS:
Patch by: David Schleef <ds@schleef.org>
* libs/gst/net/Makefile.am:
Add WIN32_LIBS to libgstnet LIBADD. Fixes#557300.
Original commit message from CVS:
* gst/gstutils.c:
* gst/gstutils.h:
API: Deprecate gst_atomic_int_set(), g_atomic_int_set() should be used
now that we depend on new enough GLib.
* gst/gstcaps.c: (gst_static_caps_get):
* gst/gstclock.c: (gst_clock_entry_new):
* gst/gstinfo.c: (_gst_debug_init), (gst_debug_set_colored),
(gst_debug_set_default_threshold), (_gst_debug_category_new),
(gst_debug_category_set_threshold):
* libs/gst/base/gstbasesink.c: (gst_base_sink_init),
(gst_base_sink_set_qos_enabled):
* libs/gst/net/gstnettimeprovider.c:
(gst_net_time_provider_set_property):
Use g_atomic_int_set() instead of gst_atomic_int_set().
Original commit message from CVS:
Patch by: Kwang Yul Seo <kwangyul.seo gmail com>
* libs/gst/net/gstnettimepacket.h:
* docs/libs/gstreamer-libs-sections.txt:
Define socklen_t as int if it's not defined yet. Fixes compilation
with MSVC6 and other versions where socklen_t is not defined in
the windows headers (#518022).
Original commit message from CVS:
Patch by: Ole André Vadla Ravnås
<ole dot andre dot ravnas at tandberg dot com>
* docs/gst/gstreamer-sections.txt:
* gst/gstpoll.c: (find_index), (gst_poll_free_winsock_event),
(gst_poll_update_winsock_event_mask), (gst_poll_new),
(gst_poll_free), (gst_poll_fd_init), (gst_poll_add_fd_unlocked),
(gst_poll_remove_fd), (gst_poll_fd_ctl_write),
(gst_poll_fd_ctl_read_unlocked), (gst_poll_fd_has_closed),
(gst_poll_fd_has_error), (gst_poll_fd_can_read_unlocked),
(gst_poll_fd_can_write), (gst_poll_wait),
(gst_poll_set_controllable), (gst_poll_restart),
(gst_poll_set_flushing):
* gst/gstpoll.h:
* libs/gst/net/gstnetclientclock.c: (gst_net_client_clock_new):
* libs/gst/net/gstnettimeprovider.c: (gst_net_time_provider_start),
(gst_net_time_provider_new):
* plugins/elements/gstfdsink.c: (gst_fd_sink_start):
* plugins/elements/gstfdsrc.c: (gst_fd_src_start):
* tests/benchmarks/gstpollstress.c: (main):
* tests/check/gst/gstpoll.c: (GST_START_TEST), (gst_poll_suite):
Remove GstPollMode from the API, it does not make sense to let the
application control this.
Add support for Win32.
Fix the testsuite. Fixes#520671.
Original commit message from CVS:
Patch by: Ole André Vadla Ravnås <ole.andre.ravnas@tandberg.com>
* gst/gsttrace.c:
* libs/gst/net/gstnetclientclock.c: (gst_net_client_clock_new):
* libs/gst/net/gstnettimepacket.c: (gst_net_time_packet_send):
* libs/gst/net/gstnettimeprovider.c: (gst_net_time_provider_new):
Fix a couple of missing includes for MSVC2005 and a C99 issue. Also,
starting with 2.14.0, GLib won't provide a pipe() macro any longer,
so use _pipe() directly (#492077).
* win32/common/dirent.c: (_treaddir):
Add a couple of casts to make it build without warnings with MSVC.
* win32/common/libgstreamer.def:
Add some more symbols that need to be exported.
Original commit message from CVS:
2006-10-09 Zaheer Abbas Merali <zaheerabbas at merali dot org>
* libs/gst/net/gstnetclientclock.c: (gst_net_client_clock_thread):
Remove local debugging system and use Gstreamer's instead.
Original commit message from CVS:
2006-10-09 Zaheer Abbas Merali <zaheerabbas at merali dot org>
Patch by: Josep Torre Valles <josep@fluendo.com>
* common/m4/gst-error.m4:
Disable warning of statement not reached on Forte.
* gst/gstmessage.h:
Fix warning on Forte (value doesn't fit on enumeration).
* libs/gst/base/gstbasesink.c: (gst_base_sink_chain_unlocked):
Fix warning on Forte (value doesn't fit on enumeration).
* libs/gst/net/gstnetclientclock.c: (gst_net_client_clock_thread):
DEBUG macro says it takes minimum of 2 args and so Forte
complains about the use with just 1 arg.
* plugins/elements/gstfdsink.c:
* plugins/elements/gstfdsrc.c:
* plugins/elements/gstfilesink.c:
* plugins/elements/gstfilesrc.c:
Use correct return type for the uri handler implementations.
All these fix warnings in Forte. Fixes bug #360860.
Original commit message from CVS:
2006-10-05 Zaheer Abbas Merali <zaheerabbas at merali dot org>
Patch by: Josep Torre Valles <josep@fluendo.com>
* common/m4/gst-error.m4:
* configure.ac:
* libs/gst/net/Makefile.am:
Fix a compilation issue with Forte on Solaris. inet_aton is in
libresolv.
Original commit message from CVS:
* configure.ac:
set CFLAGS and friends to -O0 if gcov is being used
add GCOV LIBS
* gst/Makefile.am:
* libs/gst/base/Makefile.am:
* libs/gst/check/Makefile.am:
* libs/gst/controller/Makefile.am:
* libs/gst/dataprotocol/Makefile.am:
* libs/gst/net/Makefile.am:
* plugins/elements/Makefile.am:
* plugins/indexers/Makefile.am:
add makefile rules to generate gcov data and clean up
* tests/check/Makefile.am:
add a coverage target that generates an html overview
of coverage data
Original commit message from CVS:
* gst/gstinfo.c:
Promote GST_DEBUG_CATEGORY_STATIC in example in docs.
* libs/gst/net/gstnetclientclock.c:
* libs/gst/net/gstnettimeprovider.c:
Use GST_DEBUG_CATEGORY_STATIC here too (#342503).
Original commit message from CVS:
* libs/gst/net/gstnettimepacket.c: (gst_net_time_packet_send):
MSG_DONTWAIT is not defined on Cygwin, so work
around that (fixes#317048).
Original commit message from CVS:
2005-12-12 Andy Wingo <wingo@pobox.com>
* libs/gst/net/gstnettimeprovider.c
(gst_net_time_provider_class_init, gst_net_time_provider_init)
(gst_net_time_provider_set_property)
(gst_net_time_provider_get_property): Export "active" as a GObject
property.
(gst_net_time_provider_thread): Only respond to time queries if
the time provider is active.
* libs/gst/net/gstnettimeprovider.h: Add an "active" boolean to
NetTimeProvider, preserving binary compat.
Original commit message from CVS:
* configure.ac:
* gst/Makefile.am:
* gst/base/Makefile.am:
* gst/check/Makefile.am:
* gst/elements/Makefile.am:
* gst/net/Makefile.am:
update LDFLAGS use some more
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
Update for new API and API changes.
* gst/gstobject.h:
Documentation fix: GST_TRYLOCK -> GST_OBJECT_TRYLOCK
* gst/gstvalue.c:
Documentation typo fix.
* gst/net/gstnettimepacket.c:
Documentation fixes for arguments.
Original commit message from CVS:
* docs/design/part-TODO.txt:
Update TODO.
* gst/net/gstnetclientclock.c: (gst_net_client_clock_thread):
* gst/net/gstnetclientclock.h:
Use parent fields for timeout and window_size.
Original commit message from CVS:
2005-11-23 Andy Wingo <wingo@pobox.com>
* check/net/gstnetclientclock.c (test_functioning): Adjust to
rate_num/rate_denom change.
* gst/net/gstnetclientclock.c
(gst_net_client_clock_observe_times): Take the SLAVE_LOCK not the
OBJECT_LOCK. Don't call add_observation with the lock.
* gst/gstclock.c (gst_clock_init): Initialize the rate as a
fraction.
(gst_clock_adjust_unlocked): Adjust using uint64_scale and the
rate fraction.
(gst_clock_set_calibration, gst_clock_get_calibration): Change to
deal with rate as a fraction whose numerator and denominator are
GstClockTime values.
(gst_clock_set_master): Only use the OBJECT_LOCK to set the
master; the other fields are protected by the SLAVE_LOCK.
(do_linear_regression): Note that this must be called with the
SLAVE_LOCK.
(gst_clock_add_observation): Take the SLAVE_LOCK, not the
OBJECT_LOCK. Call set_calibration instead of touching the
variables directly.
(gst_clock_set_property, gst_clock_get_property): Protect
master/slave parameters with the SLAVE_LOCK.
* gst/gstclock.h (GstClock): Remove rate, add rate_numerator and
rate_denominator. PR3C1S3. Add a new lock, the SLAVE_LOCK, and
note that all of the instance variables that add_observation and
the set_master functions use are protected by that lock and not
the OBJECT_LOCK.
(GST_CLOCK_SLAVE_LOCK, GST_CLOCK_SLAVE_UNLOCK): New macros.
* gst/gstclock.c (gst_clock_add_observation): No longer requires
the caller to take the object lock.
Original commit message from CVS:
* gst/base/gstbasesink.c: (gst_base_sink_class_init),
(gst_base_sink_wait), (gst_base_sink_do_sync),
(gst_base_sink_handle_event):
* gst/base/gstbasesink.h:
No need to store the clock, the parent element class already
has it.
* gst/gstbin.c: (gst_bin_set_clock_func), (gst_bin_add_func):
Updates for clock_set returning a gboolean
* gst/gstclock.c: (gst_clock_entry_new), (gst_clock_id_wait),
(gst_clock_id_wait_async), (gst_clock_class_init),
(gst_clock_init), (gst_clock_finalize),
(gst_clock_get_internal_time), (gst_clock_get_time),
(gst_clock_slave_callback), (gst_clock_set_master),
(gst_clock_get_master), (do_linear_regression),
(gst_clock_add_observation), (gst_clock_set_property),
(gst_clock_get_property):
* gst/gstclock.h:
Implement master/slave. When setting a clock as a slave, a
periodic timeout is scheduled to sample master and slave times.
Then the slave clock is recalibrated to match offset and rate
of the master clock.
Update logging a bit.
Add flag so that a clock can state that is cannot be slaved to
another clock.
* gst/gstelement.c: (gst_element_set_clock):
* gst/gstelement.h:
The set_clock returns a gboolean for when an element cannot
deal with the selected clock in the pipeline.
* gst/gstpipeline.c: (gst_pipeline_change_state),
(gst_pipeline_set_clock):
* gst/gstpipeline.h:
Handle the case where the selected clock cannot be set on
the pipeline.
* gst/net/gstnetclientclock.c: (gst_net_client_clock_class_init),
(gst_net_client_clock_init), (gst_net_client_clock_finalize),
(gst_net_client_clock_set_property),
(gst_net_client_clock_get_property),
(gst_net_client_clock_observe_times):
* gst/net/gstnetclientclock.h:
Use regression code in GstClock parent, remove duplicated
functionality.
Original commit message from CVS:
2005-11-18 Andy Wingo <wingo@pobox.com>
* gst/net/gstnetclientclock.c: Turn off debugging.
* check/net/gstnetclientclock.c (test_functioning): Assert that the
times connverge somewhat. Can't make a real test.
Original commit message from CVS:
2005-11-18 Andy Wingo <wingo@pobox.com>
* gst/net/gstnetclientclock.c (do_linear_regression): Use all
integer arithmetic. Return the minimum of the domain, which can be
set as "internal" for gst_clock_set_calibration.
(gst_net_client_clock_observe_times): Call _set_calibration.
(gst_net_client_clock_new): Call _set_calibration instead of
rate_offset.
* check/net/gstnetclientclock.c (test_functioning): Use the right
adjustment api.
* gst/gstclock.h:
* gst/gstclock.c (gst_clock_get_calibration)
(gst_clock_set_calibration): New functions, obsolete the ones I
added yesterday. Doh. Precision issues mean we have to extrapolate
from a point in the more recent past than 1970.
(gst_clock_get_rate_offset, gst_clock_set_rate_offset): Mark as
obsolete.
(gst_clock_adjust_unlocked): Use the right calibration data.
Original commit message from CVS:
2005-11-18 Andy Wingo <wingo@pobox.com>
* gst/net/gstnetclientclock.c (gst_net_client_clock_thread):
Whoops, check the right fd. Also add some debugging.
(gst_net_client_clock_observe_times): Adjust for int64 offset.
(do_linear_regression): Add a crapload of debugging. Subtract off
the minimum values from the input series to discard unneeded bits.
Use only int arithmetic. There is still double arithmetic when
calculating the intercept that needs fixing. Return boolean to
indicate success; FALSE would mean the domain or range is too
great. Still needs fixes.
Original commit message from CVS:
2005-11-18 Andy Wingo <wingo@pobox.com>
* gst/net/gstnetclientclock.c (gst_net_client_clock_class_init):
Port goes all the way to MAXUINT16.
* gst/net/gstnettimeprovider.c: Make the port range the same as
for the kernel: 0 assigns, otherwise ports are less than
MAXUINT16.
* check/net/gstnettimeprovider.c: Adapt for 0 == kernel assigns
port change.
* check/net/gstnetclientclock.c (test_functioning): Add the start
of another test.
Original commit message from CVS:
2005-11-18 Andy Wingo <wingo@pobox.com>
* gst/net/gstnetclientclock.c (gst_net_client_clock_new): Adjust
the clock initially so it produces values around the base time.
(gst_net_client_clock_class_init): Typo fix.
(gst_net_client_clock_thread): Add note on when the socket gets
closed.
Original commit message from CVS:
* gst/net/gstnetclientclock.c: (do_linear_regression),
(gst_net_client_clock_do_select), (gst_net_client_clock_thread):
Fix compilation, uninitialized vars and a forgotten continue.
Original commit message from CVS:
2005-11-17 Andy Wingo <wingo@pobox.com>
* check/Makefile.am (check_PROGRAMS):
* check/net/gstnetclientclock.c: Add a most minimal test for the
net client clock. More to come later.
* gst/net/gstnet.h:
* gst/net/Makefile.am: Add netclientclock.
* gst/net/gstnetclientclock.h:
* gst/net/gstnetclientclock.c: New files, implement an untested
GstClock that takes its time from a network time provider.
Implements the algorithm in network-clock.scm.
* tests/network-clock.scm (*window-size*): Rename from
*queue-length*.
* tests/network-clock.scm (network-time):
* tests/network-clock-utils.scm (q-push): Update callers.
Original commit message from CVS:
2005-11-16 Andy Wingo <wingo@pobox.com>
* gst/net/Makefile.am:
* gst/net/gstnet.h:
* gst/net/gstnettimeprovider.c:
* gst/net/gstnettimeprovider.h: Use the timepacket stuff in the
provider, include it from gstnet.h, and add it to the build.
* gst/net/gstnettimepacket.h:
* gst/net/gstnettimepacket.c: New files, abstracts out the packet
sending and receiving.
Original commit message from CVS:
2005-11-16 Andy Wingo <wingo@pobox.com>
* gst/net/Makefile.am:
* gst/net/gstnet.h: New part of core to hold network elements and
objects. Put in core because it exposes API that applications want
to use. The library is named libgstnet-tempname right now because
of the existing libgstnet in gst-plugins-base. Solution is
probably to rename the one in plugins-base; will file a bug for
the freeze break.
* gst/net/gstnettimeprovider.c:
* gst/net/gstnettimeprovider.h: New object to export a GstClock's
get_time call over the network.
* configure.ac:
* gst/Makefile.am (lib_LTLIBRARIES): Add gstnet to the build.
* check/Makefile.am:
* check/net/gstnettimeprovider.c: A most minimal test suite. Will
get additions shortly.
