When set to 'true', the node will remain alive without any error produced if
it cannot be linked to its specified target; in all other cases, an error will
be produced and the node will be destroyed.
Two new properties are available to change the behavior of the linking policy:
- target.dont-fallback: when set to true, the node shouldn't be allowed to
fallback to another available target other than the one specified in the
target.object property or metadata.
- target.dont-move: when set to true, wireplumber should ignore the
target.object metadata, so that it isn't possible to dynamically move the node
to another target using metadata.
See #524
This patch improves the smart filters unlinking logic by only unlinking smart
filters that are disabled, instead of any kind of filters. The patch also
removes redundant filter hooks.
This patch updates the deprecated policy-bluetooth.lua script so that it works
with the current version. The script has been moved into the device sub-folder,
and renamed to autoswitch-blueooth-profile.lua. The settings-manager is also
used for the configuration, and the actual configuration has been moved from
linkind.conf to bluetooth.conf.
Handle client-requested and flatpak access on the session manager side.
Pipewire daemon doesn't know about the intended permission hierarchy, so
it is better done on the session manager side, where all the rules are.
The "default" access is used for normal clients, in the use case where
Pipewire server will not assign permissions itself but leaves it to the
session manager. In this use case only session manager has
"unrestricted".
Make "default" equal to "unrestricted" in the default access
configuration.
This allows users to enable/disable smart filter policy per filter. The property
is considered false by default, meaning that smart filter policy is disabled for
all filters by default.
I don't fully remember where this originates from, but it looks like
some sort of hack to workaround a race condition where the event handler
would try to iterate over items that were just created and were not yet
fully ready to be used.
Since all scripts run in a sandbox with their own global environment,
it means that they don't interfere with each other's global variables.
Therefore, all file-wide variables can be declared global without
any change in behavior. In my understanding, it is better to do so
because this means that any code accessing those variables is going
to access them directly from the global environment table with a simple
lookup rather than having each variable referenced in the local closure
of each function separately.
The filter's destination target was not being found properly due to iterating
the matching rules table with ipairs instead of pairs... the later is meant to
be used with JSON objects. In addition, the filters were not being re-evaluated
and linked properly when a device node was removed, this is because of a typo in
the find-best-target.lua script.
See #501
After a timeout, it's possible that something else may have
destroyed the node already. This can happen for example with virtual
sinks that were created from another process.
Fixes: #467
