Bug 93923

Summary:	Kinetic scrolling not stopped by leaving finger on Synaptics touchpad
Product:	Wayland	Reporter:	Milan Bouchet-Valat <nalimilan>
Component:	libinput	Assignee:	Peter Hutterer <peter.hutterer>
Status:	RESOLVED NOTOURBUG	QA Contact:
Severity:	normal
Priority:	medium	CC:	bordjukov, carlosg, peter.hutterer
Version:	unspecified
Hardware:	Other
OS:	All
Whiteboard:
i915 platform:		i915 features:
Attachments:	evemu-record output for 5 sequences of kinetic scrolling + leaving finger evemu-record output for upward vs downward kinetic scrolling

Description Milan Bouchet-Valat 2016-01-29 09:55:20 UTC

When using vertical kinetic scrolling on my Synaptics touchpad (with one finger, on the side), leaving the finger on the touchpad without moving does not stop the scrolling at all, which it does with the Synaptics driver. Also, doing a finger move in the reverse direction to stop the scrolling has a small delay. Or at least a larger part of the finger move has no effect on the pointer.


This is on Fedora 23 with libinput 1.1.5. I have an hp Pavilion dm4 with this touchpad:
# Input device name: "SynPS/2 Synaptics TouchPad"
# Input device ID: bus 0x11 vendor 0x02 product 0x07 version 0x1b1

(Offspring of Bug 91615)

Comment 1 Milan Bouchet-Valat 2016-01-29 10:00:35 UTC

Created attachment 121384 [details]
evemu-record output for 5 sequences of kinetic scrolling + leaving finger

In each sequence, I scrolled downward enough to have a long list continue scrolling on its own, and then I put the finger on the side of the touchpad without moving it. The list continued to scroll.

Comment 2 Milan Bouchet-Valat 2016-01-29 10:02:20 UTC

Created attachment 121385 [details]
evemu-record output for upward vs downward kinetic scrolling

In this recording, I'm scrolling upward and downward alternatively, with strong moves.

Comment 3 Peter Hutterer 2016-01-31 23:26:53 UTC

verified that libinput sends the 0/0 terminating scroll event, and the 1.1.5 libinput driver passes it on. so this is a bug in GTK somewhere, Carlos, any ideas?

Comment 4 Milan Bouchet-Valat 2016-02-04 22:07:07 UTC

FWIW, I've just checked that the problem does not appear with the Synaptics driver. Anyway, I'll file a bug against GTK.

Comment 5 Peter Hutterer 2016-02-04 22:09:24 UTC

synaptics does kinetic scrolling in the driver and keeps sending scroll event (indistinguishable from normal scroll events). libinput only sends a notification that the finger was lifted, the actual scroll events are generated by the client. That has some advantages and disadvantages, but it also means that the two cannot be easily compared, they happen in different layers of the stack.

Comment 6 Milan Bouchet-Valat 2016-02-05 08:13:07 UTC

I see, thanks.

Comment 7 Milan Bouchet-Valat 2016-02-05 14:18:39 UTC

Looks like Carlos has been working on this recently at https://bugzilla.gnome.org/show_bug.cgi?id=756729. I'm not commenting there as you seem to be subscribed already. Do you think I should file a new bug anyway?

Comment 8 Peter Hutterer 2016-02-08 00:04:04 UTC

Note: that bug was related to the Wayland bits of GTK, so if you're running an X session then you should file a new bug. If you're running Wayland, it's also worth filing a new bug because that one is fixed, so any regression/bug should be handled separately to avoid confusion.

Comment 9 Milan Bouchet-Valat 2016-02-27 11:27:17 UTC

OK, filed a bug in GTK at https://bugzilla.gnome.org/show_bug.cgi?id=762767

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.