syndaemon lies about its functionality, -R doesn't use XRecord
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
xserver-xorg-input-synaptics (Ubuntu) |
Confirmed
|
Undecided
|
Unassigned |
Bug Description
With a friend we checked the processor wakeups of my laptop with powertop. We spotted syndaemon. Investigating it, I found out that it polls, although an option was set to block and wait using some XRecord extension of X11. The option -R is also set in the current configuration of syndaemon in 12.10. Also the man page doesn't tell one about the deactivated -R option. I compiled it myself, with the option compiled in (need to install libxtst-dev) and it blocked and waited like expected. When I read about the lengthy bugreports of ancient times about polling in syndaemon.
https:/
As I understand them, X11 crashes randomly when XRecord is used. And the chosen solution was to deactivate the usage of XRecord in syndaemon.
I am not satisfied with the situation.
* man page shouldn't list -R if it is not compiled in; it was misleading to me
* syndaemon shouldnt be started with -R, when it has no affect; it was misleading to me
* there should be a power measurement documented somewhere which defends polling against power drainage; the decision is weird on battery powered devices
* XRecord should be deactivated in X11; who guarantees that another app than syndaemon comes by using XRecord?
* XRecord should be fixed in X11 (the bug description and tried fixes arent exhaustive by far!)
Status changed to 'Confirmed' because the bug affects multiple users.