GitLab

Less cracktastic in the end.

Handle corner case where we try to read beyond the end of the
last file part, in which case we want to return a short read.
If we get fewer bytes than expected for any other file part,
we should just error out, since something fishy's going on
then.

Looks like some parsers (in some versions at least) expect the
offsets to be set, and behave weird if that's not the case
(e.g. off-by-one in h264parse).

Add new splitfilesrc element that presents multiple files
(selectable via a location regex) as one single contiguous
file.

This reverts commit d6a9de9e.

setcaps functions aren't supposed to take ownership of the caps passed

https://bugzilla.gnome.org/show_bug.cgi?id=662664

The element does not modify the data anyway.

Using 'iso' conflicts with mp4mux variant type, ismlmux now
uses iso-fragmented

Fixes #656823

PulseAudio 1.0 supports per-source-output volumes, and this exposes the
functionality via the GstStreamVolume interface.

When compiled against pre-1.0 PulseAudio, the interface is not
implemented, and the "volume" or "mute" properties are not available.
This bit of ugliness will go away when we can depend on PulseAudio 1.0
or greater.

https://bugzilla.gnome.org/show_bug.cgi?id=595055

Makes sure we don't leak a refcount if the object is disposed before a
NEWSEGMENT turns up.

Which I accidentally broke when fixing flv videos breaking on
spurious timestamp discontinuities in broken files.

https://bugzilla.gnome.org/show_bug.cgi?id=631430

It also work in this format. Avoids the need for conversion.

The plugins were using _fixed_caps_ and thus not adjusting to new upstream
sizes. Spotted by Tim Müller.

This is corrrupts the memoy when resizing. Add a FIXME to make it resizeable
once that is solved.

The plugins were using _fixed_caps_ and thus not adjusting to new upstream
sizes. Spotted by Tim Müller.

Do not forget to free the filename strings when deleting
the list of files.

Tests fail to build because g_mkdtemp is available from glib since
2.26.

This patch adds a condition around the redefinition of
g_mkdtemp on the tests to only build it if glib is older than
2.26.

This allows decoding at least one sample where something has
stuffed some ID3 tag before the (supposedly initial) FMT\ .

https://bugzilla.gnome.org/show_bug.cgi?id=660249

https://bugzilla.gnome.org/show_bug.cgi?id=661841

Let's initialize them to zero. It looks alright, but then it
also looks alright with v3, or with the corresponding pixels
from the source. I don't know what the original intent would
be, and the original effectv source also has this bug/feature.

https://bugzilla.gnome.org/show_bug.cgi?id=661841

https://bugzilla.gnome.org/show_bug.cgi?id=664548