Re: [gpm] little chance in 50-no-critical-on-ac.patch (ubuntu)

[Richard Hughes <hughsient gmail com>]

On Mon, 2006-02-20 at 16:35 -0500, David Zeuthen wrote:
> On Mon, 2006-02-20 at 20:52 +0100, Micha�Arnauts wrote:
> > This is what is in CVS:
> > (in battery_status_changed_primary)
> > 
> > 	if (warning_type == GPM_WARNING_ACTION) {
> > 		gpm_tray_icon_notify (GPM_TRAY_ICON (manager->priv->tray_icon),
> > 				      5000,
> > 				      _("Critical action"),
> > 				      NULL,
> > 				      _("The battery is below the critical level and "
> > 					"this computer is about to shutdown."));
> > 		/* wait 10 seconds for user-panic */
> > 		g_usleep (1000 * 1000 * 10);
> 
> So you're block the the UI of g-p-m here? That's bad, isn't it?
> Shouldn't you be using g_timeout_add() instead and have some kind of
> state machine?

Why a state machine? What's wrong with doing:

static gboolean
manager_critical_action_do (GpmManager *manager)
{
	manager_policy_do (manager, GPM_PREF_BATTERY_CRITICAL);
	return FALSE;
}

...
	/* wait 10 seconds for user-panic */
	g_timeout_add (1000*10, (GSourceFunc)manager_critical_action_do, manager);
...

That would solve the UI issue wouldn't it?

Richard.