GitLab

Fixes bug #630437.

Clock IDs were leaked if the clock got disposed before the worker thread
got a chance to reap unscheduled entries.

Fixes bug #630439.

Because of the awkward refcounting in prepare_output_buffer, we might end up
with writable buffers that point to the same data. Check for those cases so that
we avoid a useless memcpy and keep valgrind quiet.

Fixes #628176

Only try to build (pseudo-)C++ unit test if a working C++ compiler has been
found, otherwise the build will fail. (We do this to make sure our headers
are 'C++ clean').

Adds a new tag to indicate the error in horizontal positioning
in meters. This is one of the available 'gps error' fields in
exif, for example.

API: GST_TAG_GEO_LOCATION_HORIZONTAL_ERROR

This reverts commit 9b424b15.

Fixes #503592

This aims to catch cpp issues in core. Add c++ boilerplate to configure.

From aa0d1d0 to 5e3c9bf

Fixes bug #629494.

Add a function to add a list of buffers to the bufferlist.

Add a function to retrieve a list of buffers containing the first N bytes from
the adapter. This can be done without a memcpy and should make it possible to
transfer the list to a GstBufferList later.

Optimize _peek() some more by reusing already assembled data when we can.

When we have already assembled some data before, reuse this data and only copy
the part that is new.

Move some common code into one place

Add support for 0 sized buffers. This is interesting in combination with the
timestamp functions.

Fixes #629553

Add gst_is_initialized() guard to gst_element_factory_make(), so
people who forgot to call gst_init() get a useful warning for what
seems to be a common enough mistake.

Sprinkle some G_UNLIKELY(), return TRUE/FALSE constants, avoid an
unnecessary g_value_unset(), move g_value_init()+set_int64_range()
closer to where they're needed.

Don't create a new GValueArray copy for every single _add_buffering_range()
call, but append to the existing value array owned by the structure instead.

Split out functions that do the actual work, so we avoid doing
the same g_return_if_fail() checks multiple times for each call.

Add _set_value() variants that take ownership of the value passed
instead of making a copy of the value. This is useful for setting
values to things that aren't refcounted (e.g. GValueArrays or
strings or string arrays, etc.).

API: gst_structure_take_value()
API: gst_structure_id_take_value()

https://bugzilla.gnome.org/show_bug.cgi?id=629831

Add an option to forward all the internal messages that would otherwise be
filtered such as EOS, SEGMENT and ASYNC messages.
This allows the application to, for example, detect that a partial pipeline is
prerolled or reached eos.
The original messages are wrapped inside an element message because the parent
bins are not supposed to see those internal messages escape.

Saves us a strlen() call.

Update the buffering status even when we are not using a queue so that EOS can
properly finish the buffering.

A flush-stop event would make a pad unflushing, causing it
to start acting as an activated pad. This, for example,
could lead to the chain function being called when stuff
isn't initialized.

This could happend when setting qtdemux to NULL while a seek
was being handled in the upstream filesrc (in push mode).

This patch makes it check if it is activated before setting
it to unflushing.

Hide a compatibility typedef.

https://bugzilla.gnome.org/show_bug.cgi?id=629241
https://bugzilla.gnome.org/show_bug.cgi?id=550616

Now one can use GST_NO_STATE_IGNORE_ELEMENTS=1 make generic/states.check
to try elements that would normaly be skipped.