[Pm-utils] Re: pm-utils/pm/hooks 05led, 1.6, NONE 95ibm_led, NONE, 1.1 95led,1.5,NONE

[Peter Jones]

On Fri, 2006-04-28 at 16:59 +0100, Richard Hughes wrote:
> On Fri, 2006-04-28 at 11:45 -0400, Peter Jones wrote:
> > On Fri, 2006-04-28 at 15:26 +0000, Richard Hughes wrote:
> > 
> > > Log Message:
> > > 2006-04-28  Richard Hughes  <richard at hughsie.com>
> > > * pm/hooks/05led: Remove this file as we want the led to turn on at
> > > the *end* of the hooks stuff. We shouldn't have this and 95led.
> > > * pm/hooks/95led: rename to 95ibm_led as this is ibm specific. Make
> > > sure the led blinks at suspend, and turns off when we resume.
> > Huh?  No, this was behaving right -- we put it on blink while we're
> > running scripts, and then solid when we're done and it is totally
> > suspending.
> 
> Sorry, I should have cc'd the list first. (adding as cc)
> 
> I don't think we should be blinking during the hooks stuff as we should
> be doing it for a *split* second.

Huh?  No, it's 3 or 4 seconds that it blinks...

>  This has to be really simple, after
> all it is just a suspend LED, and I don't think we should be encouraging
> people to change stuff mid-executing-hooks just by choosing a really
> small number and a big number and hoping it all works okay.

It's not encouraging them to *randomly* pick numbers.  The people you're
talking about are distro packagers.  They are always going to have to
consider the order that hooks run in, whenever they add hooks.  The same
is true if, for some reason, a user wants to add his own hooks.

One has really got to be mindful of the hook order, or else one is going
to get results that weren't as expected.  That's _why_ the ordering
mechanism is so prominent.

That being said, there are really very few places where one needs to do
something at two different spots in the process.  So I don't think it's
worth worrying about either way.
-- 
  Peter