I don't know why the touchscreen is not attached in the same orientation as the internal screen. However I can provide a feature in the next release to workaround this problem. For example, to provide a configuration file to change the initial calibration.
-- $4 2013/8/26 下午9:31 於 "Sebastian Reichel" <s...@ring0.de> 寫道： > On Mon, Aug 26, 2013 at 07:59:46PM +0800, Shih-Yuan Lee (FourDollars) > wrote: > > > So your software keeps the touchscreen calibration working if the > > > display configuration is modified by xrandr etc., but depends on > > > an already calibrated default configuration? > > Yes, my program keeps touchscreen calibration working no matter what the > > display configuration is changed, but doesn't depend on an already > > calibrated default configuration. > > How is that supposed to work? AFAIK on some devices the > touchscreen is not attached in the same orientation as > the internal screen (IIRC I already saw an inverted one). > > So let's assume a testcase: > > * screen is _not_ inverted/rotated/scaled/... (= default) > * touchscreen is using default values and thus is inverted > > Thus the touchscreen input is broken out of the box. How > can you calibrate this without getting any user input? > How do you know, that the initial touchscreen input is > inverted/rotated/scaled/...? > > > It will calculate the Coordinate Transformation Matrix of > > Touchscreen everytime when xrandr changes the output > > configuration. > > Yeah, but that does not help for the initial calibration of > the touchscreen. It obviously is a nice way to keep the > touchscreen working when xrandr is used. > > -- Sebastian >
