In the cases you describe, OS power management doesn't work properly.
Normally listening to the radio will cause network activity, and the
machine is not idle in this case. Same goes for drive access.
As I recall, my machines PPC, Intel, 10.4, 10.5 never slept to
interrupt a network flow they were serving. I've seen the case you
describe on linux machines, for sure. Actually, I doubt I've seen a
linux machine sleep over drive activity, but everything is possible with
linux :)

Are you streaming from the server, or directly from the players ? Did
you try resetting the firmware on the Mac, or evaluating how it behaves
with a fresh OS (installed on an external drive, for example) ?
Anyway, the newest "busy signal" of srvPowerControl provides a hook to
a script that could "move the mouse" or something. I'm not sure
srvPowerControl tries to read back the output of the script.

The current situation for srvPowerControl as I read it is as such:
- on a well behaved machine, that doesn't sleep when its system is
active somehow, srvPowerControl gives the right order at the right time.
With the desktop login/logout hooks, it will even work appropriately for
Desktop machines. The 3 macs I've had over the last few years all had
"smart enough" activity detection.
- on a machine that is poorly managed by the OS, srvPowerControl works
well for single-purpose servers: remove all PM independent of
srvPowerControl and let it decide alone when to sleep/wake the machine.


For multi-purpose machines (Desktops, audio + file + video server...)
with dubious activity detection at the OS level, either the OS PM or
srvPowerControl will manage to ruin the show.
Personally, I had developed, before srvPowerControl was out, a custom
PM daemon that watches different things on a multi-purpose linux server.
I integrate srvPowerControl mostly as a front-end for user-defined PM
actions in the web GUI and players menus. 
That daemon is a lot of awkward and platform/application specific code,
it performs adequately but is none too smart.

I think the silver bullet is in developing dummy a device driver for
SqueezeBox Server. Receiving power events, the driver could block or let
go OS sleep according to what the SBS wants. I had hopes the new control
panels for SBS would include this on OS X and windows but apparently
that is not the case.
I'm not sure a dummy driver with some sort of  bidirectional link to
SBS is really complex and hard to maintain. But it is platform-dependent
and has to be written in the language of the OS' device toolkit,
probably C or a derivative. 
Volunteers with knowledge of IOKit, D-bus and their Windows equivalent
are needed, first to evaluate this idea. 

Anyway, I invite you to check out the srvPowerControl thread regularly.
I am sure that, sometime, somehow, you'll get your machine to operate as
intended.


-- 
epoch1970
------------------------------------------------------------------------
epoch1970's Profile: http://forums.slimdevices.com/member.php?userid=16711
View this thread: http://forums.slimdevices.com/showthread.php?t=69121

_______________________________________________
plugins mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/plugins

Reply via email to