When in error, attempt to execute a step called "error", if it exists.
This is rarely needed, but it might be useful in some cases to cleanup
some state, rolling back changes, when something fails.
This allows for Lua test scripts to be grouped into a label. So that they can be
treated different in terms of the init infrastructure that the script-tester
lays out for those tests.
Call the event "rescan-for-<context>", where <context> is either
"linking" or "default-nodes" and can be expanded in the future.
Add an additional hook in default-nodes to trigger the
rescan-for-default-nodes event when sources and sinks come and go.
Use the directory name and file name to construct the hook's name,
like a path. This way, when you see a hook name, it is clear where
to find that hook in the source code.
* client: Logic that deals with configuring clients (basically, permissions)
* device: Anyhing that that deals with configuring devices (profiles, routes, ...)
* node: Anything that deals with node objects: configuring nodes, changing
their state, their properties and also creating new nodes (but NOT linking them)
* linking: All the logic for creating links between nodes (and obviously,
deciding which links to create)
* monitors: Scripts that deal with hardware subsystems, mainly monitoring
hardware changes and reflecting them on pipewire
* default-nodes: All the logic for selecting the default sinks and sources
When the core is shutting down, the weak reference to it is cleared before
we have a change to use it, but it's ok because everything is getting destroyed
at this point, including the hook that we want to unregister.
But keep the additional comments and debug statement improvements.
Also, merge the stuff from common.h, which was previously shared
with m-default-nodes.
* Use more hooks and no custom object managers
* Use the settings manager for the config values
* Allow fully disabling the hooks when both restore-props and restore-target
are disabled in the settings
* Change the format AND the name of the state file; use json directly
in the values now that we can
* Remove entirely the hook priority numbers and use before/after dependencies
* Split the WpEvent code out of WpEventDispatcher
* Add methods on WpEvent to interface with it from the WpEventDispatcher.
As a bonus, we can now also implement tooling to inspect which hooks would
in theory run for an event and write tests around that
* Removed some internal debugging facilities and log calls, will redo it later.
* Using spa_list now for the list of hooks, to reduce the number of allocations
happening in the "hook collection" algorithm
* Switched some internal data to use g_new0 instead of g_slice_new0
* Added g_free to free WpEvent structures... surprisingly, we were leaking them
before
