On Tue, Jun 20, 2006 at 11:31:47AM -0600, Warren Turkal wrote: > On Monday 19 June 2006 11:37, you wrote: > > Don't know, some script? > > Anyway, I contacted upstream and he agrees that the easiest explanation > > for the behaviour you see is something external intervening. > > I'm sorry to bother you in this direction but... > > Okay. I don't believe it and even an "rgrep 'synclient'" in my /etc turned up > nothing. I have also reinstalled the system about two days ago, so it is a > perfect time to do some more testing. I haven't even restored my home > directory yet. I just have an empty user with all of KDE and Xorg installed.
and still the same problem... right? > > I prepared a modified package that should completely discard the > > external-intervention-conspiracy :) > > Package is at http://people.debian.org/~malattia/packages/ and includes > > the below small patch to "hide" the SHMConfig struct in the driver and > > modifies synclient to ask for confirmation when trying to change a > > parameter value (so beware that X could hang waiting for an answer if > > the conspiracy is taking place). > > The .changes file is signed and you can check md5sums of files there. > > I am guessing that I need to run this with startx from console as starting > from the kdm init script won't give me the stdout and stdin that I want so I > can see the manual confirmation you added. Is this logic correct? well, yes, I suppose running from xdm will simply hang there or continue without touching synaptics settings as the added code in synclient is pretty dumb and doesn't check for errors (no tty should make the scanf return with an error and "yes" will be compared some random memory on the stack). > BTW, how does this explain the scenario where 'synclient -l' shows the > RightEdge set to 5980, but the edge scrolling doesn't work until I manually > execute 'synclient RightEdge=5980' to set the option? I'm speechless... :) > Thanks for your tireless efforts on this, sorry for being that obstinate on my side, but really, similar problems are always due to some header mismatch in [kqg]synaptics or similar, expecially when the shared memory segment layout changes as happened in 0.14.5. Thanks again for your patience, -- mattia :wq! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
