Add support RTP buffers with multiple memory blocks. We allow one block for the
header, one for the extension data, N for data and one memory block for the
padding.
Remove the validate function, we validate now when we map because we need to
parse things in order to map multiple memory blocks.
RTCP header can be (2^16 + 1) * 4 bytes long, so when validating a bogus
packet it was possible to get a 16bit overflow resulting in a length of 0.
This would put the gst_rtcp_buffer_validate_data function in a endless loop.
https://bugzilla.gnome.org/show_bug.cgi?id=667313
gst_basertppayload -> gst_base_rtp_payload
Add pts/dts support in the depayloader
Remove old timestamp code
Add a default getcaps function so subclasses can chain up to it instead of
relying on the return value of the getcaps function.
Without the perfect timestamp machinery, the RTP timestamp can be
computed directly from the running time of a buffer, but the perfect
timestamp patch broke that assumption. This patch restores it by
having the first perfect timestamp be the running time of that buffer
and counting from there.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=654434
... which allows adding additional packets and may be needed to counteract
the shrink that implicitly occurred during a map/unmap cycle when adding
a previous packet.
Use the caps event instead of the setcaps function to configure caps.
Use a default event handler for the base rtp payloader instead of the awkward
way of handling the return value.
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
1) We need to lock and get a strong ref to the parent, if still there.
2) If it has gone away, we need to handle that gracefully.
This is necessary in order to safely modify a running pipeline. Has been
observed when a streaming thread is doing a buffer_alloc() while an
application thread sends an event on a pad further downstream, and from
within a pad probe (holding STREAM_LOCK) carries out the pipeline plumbing
while the streaming thread has its buffer_alloc() in progress.
Fixes introspection failures caused by type assertions/warnings.
Since we now moved from _get_type() functions to external GType
variables in a couple of places, we actually have to call gst_init()
to make sure these are set when we use GST_TYPE_FOO.
Make sure to use the PKG_CONFIG_PATH set at configure time instead of
just relying on an env-var set one. This makes sure both g-ir-compiler
and g-ir-scanner use the same PKG_CONFIG_PATH for determining include
paths etc.
When we have an invalid running-time (because we clipped, for example) use the
RTP base time for timestamping instead of generating wrong RTP timestamps.
Add a new function called gst_rtp_buffer_list_from_buffer() that takes
a GstBuffer containing a RTP packets and spits out a GstBufferList
containing two buffers, one with the header and the other with the payload.
RFC 5285 describes a generic method to add multiple header extensions to RTP packets.
These functions parse these headers and return them, both for the one-byte header and the
two bytes headers.
This is pretty much an FAQ, so try to make the error message a bit
more helpful. Also, don't tell people to file a bug in bugzilla
about this (which is what happens if the default error message for
CORE_NEGOTIATION is used).
When calling gobject-introspection scanner, make sure our own
freshly-built libs within the source tree (well, build dir) come
first in the PKG_CONFIG_PATH. May or may not help to make sure
that it doesn't pick up older external plugins-base libs (or
.gir files) from outside the source tree / build directory as
dependencies of the introspected lib instead of using the
stuff we just built in a sibling directory.
https://bugzilla.gnome.org/show_bug.cgi?id=623698
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/AutotoolsIntegrationFixes#603710.
Use new girdir and typlibdir from core .pc files, so we can figure
out the right includes to pass to the gobject-introspection tools,
whether core is installed in the same prefix as gobject-introspection
or in a different prefix or uninstalled. This also keeps us from adding
bogus paths to the includes that only work if core is uninstalled.
Also add some missing includes/pkgs where needed.
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.
Add various conversion functions between time<->bytes<->rtptime that will be
used later on.
Refactor the min/max packet length code so that it can be used for both
sample/frame based payloaders. Cache the returned values.
code cleanups.
When we discover a DISCONT buffer, make the outgoing RTP timestamps have the
same gap as the GStreamer timestamps gap.
Have a custom sample/frame function to generate an offset that the base class
will use for generating RTP timestamps. This results in perfect RTP timestamps
on the output buffers.
Refactor setting metadata on output buffers.
Add some more functionality to _flush().
Handle DISCONT on the input buffers and set the marker bit and DISCONT flag on
the next outgoing buffer.
Flush the pending data on EOS.
Always use the adapter when we need to fragment the incomming buffer. Use more
modern adapter functions to avoid malloc and memcpy. The overall result is that
the code looks cleaner while it should be equally fast and in some case avoid a
memcpy and malloc.
Use the adapter timestamping functions for more precise timestamps in case of
weird disconts.
Cache some values instead of recalculating them.
Add gst_base_rtp_audio_payload_flush() to flush a certain amount of bytes from
the internal adapter.
API: GstBaseRTPAudioPayload::gst_base_rtp_audio_payload_flush()
Allow subclasses to use the OFFSET field on RTP buffers to influence the way in
which RTP timestamps are generated. Usually timestamps are created from the
GStreamer timestamps on the buffer, which could result in imperfect RTP
timestamps.
... which is the default seed when creating a new GRand. Because
GLib in older versions used buffered IO this would take a lot of time.
Instead use the global GRand for getting random numbers and keep the
three instance GRand for backward compatibility with a simple seed.
Fixes bug #593284.
Original commit message from CVS:
Patch by: Olivier Crete <tester at tester ca>
* gst-libs/gst/rtp/gstrtcpbuffer.c: (gst_rtcp_packet_remove):
* gst-libs/gst/rtp/gstrtcpbuffer.h:
Implement gst_rtcp_packet_remove(). Fixes#563174.
* tests/check/libs/rtp.c: (GST_START_TEST), (rtp_suite):
Add unit test for some RTCP functions.
Original commit message from CVS:
* gst-libs/gst/rtp/gstrtpbuffer.c: (gst_rtp_buffer_validate_data),
(gst_rtp_buffer_set_packet_len), (gst_rtp_buffer_get_packet_len),
(gst_rtp_buffer_get_header_len), (gst_rtp_buffer_get_version),
(gst_rtp_buffer_set_version), (gst_rtp_buffer_get_padding),
(gst_rtp_buffer_set_padding), (gst_rtp_buffer_pad_to),
(gst_rtp_buffer_get_extension), (gst_rtp_buffer_set_extension),
(gst_rtp_buffer_get_extension_data),
(gst_rtp_buffer_set_extension_data), (gst_rtp_buffer_get_ssrc),
(gst_rtp_buffer_set_ssrc), (gst_rtp_buffer_get_csrc_count),
(gst_rtp_buffer_get_csrc), (gst_rtp_buffer_set_csrc),
(gst_rtp_buffer_get_marker), (gst_rtp_buffer_set_marker),
(gst_rtp_buffer_get_payload_type),
(gst_rtp_buffer_set_payload_type), (gst_rtp_buffer_get_seq),
(gst_rtp_buffer_set_seq), (gst_rtp_buffer_get_timestamp),
(gst_rtp_buffer_set_timestamp),
(gst_rtp_buffer_get_payload_subbuffer),
(gst_rtp_buffer_get_payload_len), (gst_rtp_buffer_get_payload):
Avoid expensive type checks we already did as part of the
_validate() function that should be called first.
Original commit message from CVS:
* gst-libs/gst/rtp/gstbasertpdepayload.c: (create_segment_event),
(gst_base_rtp_depayload_push_full),
(gst_base_rtp_depayload_set_gst_timestamp):
Fix some cases where a newsegment event was not sent.
Original commit message from CVS:
* gst-libs/gst/rtp/gstbasertpdepayload.c:
(gst_base_rtp_depayload_setcaps), (gst_base_rtp_depayload_chain),
(gst_base_rtp_depayload_set_gst_timestamp),
(gst_base_rtp_depayload_change_state):
* gst-libs/gst/rtp/gstbasertpdepayload.h:
Add some more G_LIKELY
Fail when the setcaps function was not called.
* gst-libs/gst/rtp/gstbasertppayload.c:
(gst_basertppayload_set_outcaps):
Propagate return value of setcaps.
Original commit message from CVS:
* gst-libs/gst/rtp/gstbasertppayload.c:
(gst_basertppayload_change_state):
Configure the next seqnum and timestamp in the state change so that they
can be queried soon after.
Original commit message from CVS:
* gst-libs/gst/rtp/gstbasertppayload.c: (gst_basertppayload_push),
(gst_basertppayload_change_state):
Simply converting the running time into an RTP timestamp by scaling it
based on the clock-rate is good enough for making an RTP timestamp. This
has the added benefit that we can later on expose a property with the
RTP timestamp of running time 0, as is needed for RTSP servers to
generate the response of the PLAY request.
Original commit message from CVS:
* gst-libs/gst/rtp/gstbasertpdepayload.c:
(gst_base_rtp_depayload_chain),
(gst_base_rtp_depayload_handle_sink_event),
(gst_base_rtp_depayload_push_full),
(gst_base_rtp_depayload_change_state):
Check sequence numbers, mark input buffers with a discont flag for the
subclass when we detected a gap, drop duplicate buffers. We do this
because one can use the element without a jitterbuffer in front and we
don't want to feed the subclasses invalid or reordered data.
Do an error when the subclass did not provide a process function instead
of crashing.
Some other small cleanups.
Original commit message from CVS:
* gst-libs/gst/rtp/gstbasertpdepayload.c:
(gst_base_rtp_depayload_chain):
Validate the RTP packet before further processing it. It's just too
dangerous to accept random packets and people are not forced to use a
jitterbuffer or session manager to filter out the bad packets.
* gst-libs/gst/rtp/gstrtpbuffer.c:
(gst_rtp_buffer_set_extension_data),
(gst_rtp_buffer_get_payload_subbuffer):
Small cleanups.
When setting extension data in a buffer that is too small, we fail and
we should not set the extension bit.
Change GST_WARNINGS into g_warning because they really are
programming errors.
* tests/check/libs/rtp.c: (GST_START_TEST):
Catch the g_warnings now in the unit tests and that fact that failing to
set extension data left the extension bit untouched.
Original commit message from CVS:
Patch by: Bernard B <b-gnome at largestprime dot net>
* gst-libs/gst/rtp/gstrtpbuffer.c: (gst_rtp_buffer_compare_seqnum):
Fix seqnum compare function for bordercase values and fix the docs
again. Fixes#533075.
* tests/check/libs/rtp.c: (GST_START_TEST), (rtp_suite):
Add a testcase for seqnum compare function.
Original commit message from CVS:
* gst-libs/gst/rtp/gstbasertppayload.c: (gst_basertppayload_init),
(gst_basertppayload_sink_setcaps),
(gst_basertppayload_sink_getcaps):
Rename the setcaps/getcaps function internally to make it clear that
they are called for the sink pad.
Original commit message from CVS:
* gst-libs/gst/rtp/gstbasertpdepayload.c:
(gst_base_rtp_depayload_class_init),
(gst_base_rtp_depayload_handle_sink_event), (create_segment_event),
(gst_base_rtp_depayload_packet_lost),
(gst_base_rtp_depayload_set_gst_timestamp):
* gst-libs/gst/rtp/gstbasertpdepayload.h:
Catch packet-lost events from the jitterbuffer and convert them into a
vmethod call (lost-packet) so that depayloaders can do something smart.
Also add a default packet-lost function that sends out a segment update
to the decoders.
