RFC: change to XO sleep behavior

[Mikus Grinbergs]

>  > Please - make sure that it takes a multi-second duration of
>  > the power button to perform a *complete* shutdown.
> 
> no, definitely not.  training users to hold down power buttons
> is/was a huge mistake, since getting in the habit is a little
> like parking your car by letting it hit the front of your garage. 
> "but it never hurt the wall before?  what changed?"

Here, I disagree with you.  On all my desktop systems, I need to hold
down the power button in order for that system to drop power
*completely*.  I have no wish to learn a different "remove power"
procedure for a laptop (which here happens to be the XO).

More to the point, an "accidental" complete-power-down is more
disruptive than asking the user to hold down the power button (to
indicate an "intentional" complete-power-down).  The incident I
described consisted of the XO going into complete-power-down while I was
*working* with it (trying to set it up after a new build install) -- the
sooner-than-expected suspend confused me, and I was injudicious in what
I pressed to try to get it out of its suspend -- I *certainly* did not
expect the XO to end up in a complete-power-down.  [Once that happened,
I had to power-up, go into "My Settings" and disable suspend (to prevent
it from interrupting me again) -- only thus was I able to then get back
to where I had been before the catastrophic complete-power-down.]


--------
What I am pointing out is the difference in intent between pushing the
power button when the system is "awake" (then the user normally wants it
to go into sleep), versus pushing the power button when the system is
"not awake" (then the user normally wants it to come out of sleep).

Since "going beyond sleep into complete-power-down" is so drastic, I
myself prefer *that* intent to be indicated explicitly (such as by
holding down the power button).

mikus