But not the DecklLinkAPIDispatch.cpp files.

The structure passed through the prog-map can now contain a
PCR_<prog_id>=sink_<PID> key-value pair.

Alex Ashley authored Feb 05, 2019 and Tim-Philipp Müller committed Feb 19, 2019

To make curlhttpsrc behave more like souphttpsrc, set the
BUFFER_OFFSET in its output buffers to match the segment
start. This means that in a HTTP RANGE request, the BUFFER_OFFSET
will match the value in the RANGE request.

Alex Ashley authored Feb 05, 2019 and Tim-Philipp Müller committed Feb 19, 2019

To make it closer to a drop-in replacement for souphttpsrc,
expose the same gst_error_message_with_details as souphttpsrc,
so that applications can received the HTTP status code and reason
when an error occurs.

Alex Ashley authored Feb 05, 2019 and Tim-Philipp Müller committed Feb 19, 2019

curlhttpsrc uses a single thread running the
gst_curl_http_src_curl_multi_loop() function to handle receiving
data and messages from libcurl. Each instance of curlhttpsrc adds
an entry into a queue in GstCurlHttpSrcMultiTaskContext and waits
for the multi_loop to perform the HTTP request.

Valgrind has shown up race conditions and memory leaks:
1. gst_curl_http_src_change_state() does not wait for the multi_loop
   to complete before going to the NULL state, which means that
   an instance of GstCurlHttpSrc can be released while
   gst_curl_http_src_curl_multi_loop() still has a reference to it.
2. if multiple elements try to be removed from the queue at once,
   only the last one is deleted.
3. source->caps is leaked
4. curl multi_handle is leaked
5. leak of curl_handle if URI not set
6. leak of http_headers when reusing element
7. null pointer dereference in negotiate caps
8. double-free of the default user-agent string
9. leak of multi_task_context.task

This commit changes the logic so that each element has a connection
status, which is used by the multi_loop to decide when to remove an
element from its queue. An instance of curlhttpsrc will not enter
the NULL state until its reference has been removed from the queue.

When shutting down the curl multi loop, the memory allocated from the
call to curl_multi_init() is now released.

When gstadaptivedemux uses a URI source element, it will re-use
it for multiple requests, moving it between READY and PLAYING
between each request. curlhttpsrc was leaking the http_headers
structure in this use case.

The gst_curl_http_src_negotiate_caps() function extracts the
"response-headers" field from the http_headers, but did not check
that this field might be NULL.

If the user-agent property is set, the global user-agent string
was freed. This caused a double-free error if the user-agent is
ever set a second time during the execution of the process.

There are situations within curlhttpsrc where the code needs
both the global multi_task_context mutex and the per-element
buffer_mutex. To avoid deadlocks, it is vital that the order in
which these are requested is always the same. This commit modifies
the locking order to always be in the order:
 1. multi_task_context.task_rec_mutex
 2. buffer_mutex

Fixes #876

Alex Ashley authored Feb 05, 2019 and Tim-Philipp Müller committed Feb 19, 2019

Based upon the souphttpsrc tests, add unit tests for the curlhttpsrc
element. The souphttpsrc tests are able to use an HTTP server that
is provided as part of the soup library. This does not exist in the
curl library, therefore these tests provide a very simple HTTP server
using the GIO library.

These curlhttpsrc tests contain one new test that does not come from
the souphttpsrc tests. The test_multiple_http_requests test tries to
reproduce the way in which GstAdaptiveDemux makes use of URI source
elements. GstAdaptiveDemux creates a bin with the httpsrc element
and a queue element and sets the locked state of that bin to TRUE,
so that it does not follow the state transitions of its parent. It
then moves this bin to the PLAYING state to start each download and
back to READY when the download completes.

Haihao Xiang authored Feb 14, 2019 and Víctor Manuel Jáquez Leal committed Feb 19, 2019

Tested with the pipeline below:

gst-launch-1.0 videotestsrc num-buffers=1 ! msdkvpp ! \
video/x-raw\(memory:DMABuf\),format=UYVY ! msdkvpp ! video/x-raw, \
format=YUY2 ! filesink location=a.yuv

Haihao Xiang authored Feb 14, 2019 and Víctor Manuel Jáquez Leal committed Feb 19, 2019

This make the pipeline below works:

gst-launch-1.0 videotestsrc num-buffers=1 ! msdkvpp ! \
video/x-raw,format=UYVY ! filesink location=a.yuv

Once https://github.com/intel/media-driver/pull/526 in the media-driver
is merged, the pipeline below also works:

gst-launch-1.0 videotestsrc num-buffers=1 ! msdkvpp ! \
video/x-raw\(memory:DMABuf\),format=UYVY ! filesink location=a.yuv

Haihao Xiang authored Feb 14, 2019 and Víctor Manuel Jáquez Leal committed Feb 19, 2019

Tested with the pipeline below:

gst-launch-1.0 videotestsrc ! msdkvpp ! \
video/x-raw\(memory:DMABuf\),format=BGRx ! glimagesink

The VCD source was ported in 2014 (commit 89eb1e9a), but the necessary
"cdxaparse" plugin, which is used to "Parse a .dat file (VCD) into
raw mpeg1" was never ported.

This means that the probable main user for the feature, totem, hasn't
actually been able to play back VCDs, since 2012, when it switched to
using GStreamer 1.0.

Note that even if cdxaparse was finally ported, a lot of work would
still be necessary before it is considered usable. Notably, it is
missing disc image support [1] and some VCDs just cannot be opened for
reading [2].

[1]: gstreamer/gst-plugins-bad#898
[2]: gstreamer/gst-plugins-bad#899

args... is not understood by MSVC:

error C2010: '.': unexpected in macro parameter list

pthread is not portable, so we can't use a pthread mutex use GMutex
instead.

Instead of creating a region, adding nothing to it, setting that as the
input region and destroying the region, you can instead just pass NULL
to wl_surface_set_input_region for the same effect.

Fixes #702

If bundle was used in combination with rtx, only the bundled transport
stream would have correctly configured rtx parameters.

Iterate over the payloads upfront in the bundled case to ensure the
correct payload mapping is set for the RTX elements.

Edward Hervey authored Feb 13, 2019 and Tim-Philipp Müller committed Feb 13, 2019

OpenSSL will take care of returning valid context if there are
only non-fatal issues. Don't abort in those cases and instead just
print out the issues

Fixes gstreamer/gst-plugins-bad#811

This release requires all buffer descriptor pointers to be valid, even
when we provide zero input buffers.

The MPEG-TS packetiser should use the upstream DTS for
skew correction when running in that mode, as the DTS
carries the upstream arrival time. The PTS (if it's
set at all) is less useful, and can be invalid.