Fixes build for upnp >= 1.6.23 and less than 1.8.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

When an Android device is plugged with MTP disabled, the tracks may not
be retrieved.

In that branch of code, pp_items was never initialized, so the behavior
was undefined in CloseDevice() (on device unplug), often causing a
segfault.

Initialize it to NULL to fix the problem.
Signed-off-by: Romain Vimont <rom@rom1v.com>
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

This is an Objective C module, therefore it should use OBJCFLAGS instead
of CFLAGS.

Callbacks now take const void* as second argument and some members can
only be accessed via getter functions.
Signed-off-by: Sebastian Ramacher <sramacher@debian.org>
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

CID #1398416

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

Nearly all messages had level Info, which is unnecessary.
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

This is merely a workaround until we have proper pagination
refs #18558

Extract the Chromecast name and model from the TXT records ("md", "fn")
and use them as name, if available, to make it easier to identify the
Chromecast.

This makes no functional differences, but it dramatically reduces the
number of events.

No need to lock here. The item is not visible to any other threads
until services_discovery_Add(Sub)Item().

Add a separate callback wrapper for item with category, which is long
since deprecated. Remove the category parameter from the normal
callback wrapper.

A single category containing all SD items was just a layer of
inconvenience. Remove it, and match the udev SD behaviour.

A single category containing all SD items was just a layer of
inconvenience. Remove it, and match the udev SD behaviour.

libupnp cannot handle more than one interface/IP at a time, so we need to make
sure we use an appropriate one.
Original patch by: Steve Lhomme <robux4@videolabs.io>

Fix #14526

Following the changes introduced by 7ecd9271, the affected lines would
result in a warning due to signed vs unsigned integer comparision.
Signed-off-by: Thomas Guillem <thomas@gllm.fr>

Now this works not only on GCC.

CID #1402708

(I would think that the bug is contextually obvious but...)