|
Packit |
a94d48 |
Future Enhancements
|
|
Packit |
a94d48 |
-------------------
|
|
Packit |
a94d48 |
|
|
Packit |
a94d48 |
* systemd support
|
|
Packit |
a94d48 |
- Support receiving unix sockets from systemd.
|
|
Packit |
a94d48 |
- Make daemonize() do what's right for systemd.
|
|
Packit |
a94d48 |
- Can we auto-detect systemd? Maybe the presence of the environment vars
|
|
Packit |
a94d48 |
for the unix sockets?
|
|
Packit |
a94d48 |
- Maintain compatibility with upstart, etc...
|
|
Packit |
a94d48 |
|
|
Packit |
a94d48 |
* Look into using libnl for netlink
|
|
Packit |
a94d48 |
- Can this be done? Is it a good idea?
|
|
Packit |
a94d48 |
|
|
Packit |
a94d48 |
OLD IDEAS
|
|
Packit |
a94d48 |
|
|
Packit |
a94d48 |
* DBUS support
|
|
Packit |
a94d48 |
- Do we really need DBUS support? Probably not.
|
|
Packit |
a94d48 |
- Tim says this would be easy to implement. He mentioned "mced" which
|
|
Packit |
a94d48 |
is similar to acpid and handles DBUS.
|
|
Packit |
a94d48 |
http://code.google.com/p/mcedaemon/
|
|
Packit |
a94d48 |
- Everyone seems to agree that the whole point of acpid is
|
|
Packit |
a94d48 |
to avoid the big desktop stuff, and therefore, this is probably
|
|
Packit |
a94d48 |
not a good idea.
|
|
Packit |
a94d48 |
- However, there are still pieces of those big desktop environments that
|
|
Packit |
a94d48 |
connect to acpid via sockets. DBUS should be a nicer interface.
|
|
Packit |
a94d48 |
- Implement with a build-time option to turn it on/off? This way users
|
|
Packit |
a94d48 |
can reduce the dependencies if they want. Or can we detect the presence
|
|
Packit |
a94d48 |
of DBUS at runtime?
|
|
Packit |
a94d48 |
|
|
Packit |
a94d48 |
* Allow socket-connected clients to filter incoming events
|
|
Packit |
a94d48 |
- This might be useful on older laptops that used to generate a slew
|
|
Packit |
a94d48 |
of events. I think that on modern hardware, this is not needed.
|