I believe the correct fix is to disable the fallback in hald.
Various parties are interested in acpi events - not just hald, but X as mentioned above, and probably others. But if any of those parties grab /proc/acpi/events then no other party can. To me, that implies that the only party that should be grabbing it is one that is going to pass those events on to everyone else, and that's acpid. If you don't the acpi event broadcaster, acpi events won't be broadcast, and that's fair enough. It's not worth a special kludge to ensure that one party can get them.
Of course, hald is also a kind of event broadcaster, but it's a large, sophisticated, high-level event broadcaster. It seems better to have it depend on acpid for receiving acpi events than to rewrite acpid to depend on it.
On my machine, acpid is only 180k; the actual daemon is only 18k. Let's make hald depend on or recommend it, and take out the fallback code altogether.
-- __ \/ o\ Paul Crowley, [EMAIL PROTECTED] /\__/ http://www.ciphergoth.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
