<html> <head> <meta http-equiv="content-type" content="text/html; charset=GB2312"> </head> <body bgcolor="#FFFFFF" text="#000000"> Hi xf86-input-evdev developers, <div class="moz-forward-container"> I use mtdev "read_mtdev.cpp" in test.tar.bz2 to get ABS_MT_POSITION_X and ABS_MT_POSITION_Y sample "2s_zoom_in_sample.txt" in the same tarball for each ABS_MT_SLOT when multi-touching. But ABS_MT_POSITION_X and Y counted more than once for each slot "sample_plot.cpp", the plot chart shown as "2slots_zoom_in.png". When ./sample_plot 2s_zoom_in_sample.txt 0 to only show slot[0] as "slot0_zoom_in.png", there are three blue lines, but instead should be only one. When ./sample_plot 2s_zoom_in_sample.txt 1 to only show slot[1] as "slot1_zoom_in.png", there are also three red lines, but instead should be only one too. My question is whether or not I process the ABS_MT_POSITION_X and Y for EV_ABS motion event type in "m_process_absolute_motion_event" uncorrectly? I read the source code "xf86-input-evdev" <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://cgit.freedesktop.org/xorg/driver/xf86-input-evdev/tree/src/evdev.c">http://cgit.freedesktop.org/xorg/driver/xf86-input-evdev/tree/src/evdev.c</a> It mentioned that <pre><code>MT axes are counted twice - once as ABS_X (which the kernel keeps for backwards compatibility), once as ABS_MT_POSITION_X. So we need to keep a mapping of those axes to make sure we only count them once</code></pre> It means I have to filter or remove some _X and _Y from the sample ? HOWTO : other issue; You can use scons <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://www.scons.org">http://www.scons.org</a> to build the test case. "sample_slot.cpp" use Vtk <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://www.vtk.org">http://www.vtk.org</a> to render plot chart, you might need to install libvtk. Many thanks! </div> </body> </html>