4 ms·

My only issue with the touch pad is that ctrl switches scrolling into ctrl+scrolling mid-scroll-throw. To be clearer, if you 2-finger scroll quickly and lift o

My only issue with the touch pad is that ctrl switches scrolling into ctrl+scrolling mid-scroll-throw. To be clearer, if you 2-finger scroll quickly and lift off before stopping, the scroll continues as if it had been thrown. This is great. If you then press ctrl (in anticipation of making a new tab, for instance), suddenly you are zooming in/out instead of scrolling up/down. This is not great. I don't know enough about user-input to know if this is unavoidable or if it could be fixed driver or browser side.

muro 8y ago
Doh, that was the spurious zooming back when I had a xps15, I never figured out what the cause was. Thank you!
- exikyut 8y ago
  My reply beside your comment may be interesting
exikyut 8y ago
This is because Synaptics is silently handling capacitative multitouch events from the trackpad and generating a synthetic stream of mouse button 4/5 (scroll wheel) events (with haptic decay and everything). As soon as CTRL accompanies the scroll wheel events... woops. This is primarily an input layer design failure: - scroll up/down is not its own event, built out in such a way that the "CTRL+" could be selectively ignored - Toolkits don't separate input sources sufficiently enough that '"CTRL+" from one device and mouse button 4/5 from another should be ignored' could be implemented. (X11 does expose everything distinctly/discretely, but pretty much nothing on top (Qt, especially GTK) really takes advantage of it. For example you can have unlimited mouse cursors on the screen, but beyond moving and resizing windows at the same time, nothing really does anything novel with this capability, and the input models (single window focused at one time, etc) don't really permit innovation either.) Hmm...... although... maybe if you took evdev apart at a very low level, and/or maybe the synaptics driv--wait. Hack synaptics to bits so it also _watches_ the keyboard, specifically CTRL keypresses, and stops sending its synthesized stream of buttonpress/releases. Maybe.

muro 8y ago

Doh, that was the spurious zooming back when I had a xps15, I never figured out what the cause was. Thank you!

exikyut 8y ago
My reply beside your comment may be interesting

exikyut 8y ago

This is because Synaptics is silently handling capacitative multitouch events from the trackpad and generating a synthetic stream of mouse button 4/5 (scroll wheel) events (with haptic decay and everything). As soon as CTRL accompanies the scroll wheel events... woops. This is primarily an input layer design failure: - scroll up/down is not its own event, built out in such a way that the "CTRL+" could be selectively ignored - Toolkits don't separate input sources sufficiently enough that '"CTRL+" from one device and mouse button 4/5 from another should be ignored' could be implemented. (X11 does expose everything distinctly/discretely, but pretty much nothing on top (Qt, especially GTK) really takes advantage of it. For example you can have unlimited mouse cursors on the screen, but beyond moving and resizing windows at the same time, nothing really does anything novel with this capability, and the input models (single window focused at one time, etc) don't really permit innovation either.) Hmm...... although... maybe if you took evdev apart at a very low level, and/or maybe the synaptics driv--wait. Hack synaptics to bits so it also _watches_ the keyboard, specifically CTRL keypresses, and stops sending its synthesized stream of buttonpress/releases. Maybe.