Hi Chaskiel, Chaskiel Grundman wrote: > Between 0.109-3 and 0.109-5, the "is gnome-power-manager running" check > moved into /etc/acpi/sleepbtn.sh from > /usr/share/acpi-support/suspendorhibernate. > > Unfortunately, the /etc/acpi/events/sleepbtn config file dispatches > sleep button events to /etc/acpi/sleep.sh, not /etc/acpi/sleepbtn.sh, > bypassing the check entirely. This results in double suspends when > gnome-power-manager is configured to handle suspend button events.
> Also, why do you think you need to use acpi_fakekey to submit an event > to gnome-power-manager? g-p-m seems to get the original event itself > (not sure if this is via hal or direct from acpid). While mostly > harmless, this can result in g-p-m reporting that suspends failed or > were disallowed, because it gets two suspend requests in a short period > of time (currently, it says 'Policy timeout is not valid. Please wait > a few seconds') Aargh. Thanks for pointing this out to me. There are TWO functions for acpi-support, which are mixed up here: 1. Suspend support. This is what /etc/acpi/sleepbtn does: it handles the NORMAL suspend button and calls its own suspend code -- but only if it's not handled otherwise. 2. Translate buttons that are not correctly translated by the kernel. If you check /etc/acpi/events/* for "sleepbtn.sh", you will find that there are four scripts that call sleepbtn.sh, for four brands of laptops. These two functions were confused when the suspend support was refactored recently. Originally, sleep.sh, which is meant to implement (1), checked for gnome-power-manager and did nothing if it was called when gnome-power-manager was active. This check should be re-added. It will be in the next upload. Thanks very much for reporting! Cheers, Bart -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
