GitLab

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

When the player notifies that the current media has changed, it must be
considered "selected" by the randomizer.

Otherwise, when the player requests the next media to play, the playlist
will always "peek" the same next item from the randomizer, which will
never actually "select" it.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Test that additions and removals behave correctly in loop mode.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

randomizer_RemoveAt() did not manage the 'history' cursor correctly: it
explicitly tested "r->history == 0", which meant "no history" in
an earlier implementation, but not anymore.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

This will avoid regression of bug #22305.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

New items must always be added _before_ the 'history' cursor, so they
will always move it to the right.

The confusion came from an earlier implementation, where "history == 0"
was synonymous with "no history" (and it was not possible to consider
the whole array as history).

Fixes #22305Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Do not dereference new_media when it is NULL.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

refs BCDisassembly1RightSideTxt2.mov

ever seen one ?

Signed-off-by: Konstantin Pavlov <thresh@videolan.org>

Signed-off-by: Konstantin Pavlov <thresh@videolan.org>

The aarch64 assembly in gnutls still needs to be disabled on windows
in general though; the issue is that it unconditionally uses ELF
specific directives (.type, .size, .section), so the disabling isn't
related to any specific issue in clang, but only that gnutls' assembly
for this architecture isn't portable.

Add a comment to clarify the issue.

Not all systems have aclocal-1.15.

In the distributed tar.gz ltmain.sh is read only so we can't patch it on UNIX
systems.

TARGET_OS_MAC is actually true when building for iPhone as well,
therefore not usable in this case.

ex: https://livestream.peer5.com/video/kite/index.m3u8

This will only be enabled when building windows targets and will only work when
compiling with Clang.