Open Bug 1883286 Opened 1 year ago Updated 9 months ago

Touchpad scroll speed is way too fast on Linux

Categories

(Core :: Widget: Gtk, defect, P3)

Firefox 123
x86_64
Linux
defect

Tracking

()

UNCONFIRMED

People

(Reporter: luca.swainson, Unassigned)

References

Details

Attachments

(1 file)

24.30 KB, text/plain
Details

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:123.0) Gecko/20100101 Firefox/123.0

Steps to reproduce:

Install firefox 123.0 on arch linux with wayland gnome. Scroll on any webpage with two fingers using a laptop touchpad.

Actual results:

Two finger scrolling is way too fast! Web pages easily fling from top to bottom

Expected results:

Two finger scrolling moves the content at a reasonable pace.

The Bugbug bot thinks this bug should belong to the 'Core::Widget: Gtk' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Component: Untriaged → Widget: Gtk
Product: Firefox → Core
OS: Unspecified → Linux
Hardware: Unspecified → x86_64

Can you run Firefox with MOZ_ENABLE_WAYLAND=0 env variable and check if X11 scroll is better?
Thanks.

Flags: needinfo?(luca.swainson)
Priority: -- → P3

(In reply to Martin Stránský [:stransky] (ni? me) from comment #2)

Can you run Firefox with MOZ_ENABLE_WAYLAND=0 env variable and check if X11 scroll is better?
Thanks.

Hi, I just checked and the scroll sensitivity is still too high, but there is no longer any momentum, meaning it stops when I lift my fingers. Either way, the sensitivity is normal when scrolling with a peripheral mouse or with a touchscreen.

Flags: needinfo?(luca.swainson)

Botond, any idea here?
Thanks.

Flags: needinfo?(botond)

(In reply to Martin Stránský [:stransky] (ni? me) from comment #4)

Botond, any idea here?

I think the findings in comment 3 suggest that the issue is at a higher level in the stack than our scrolling code:

  • If the sensitivity was a consequence of our pan-gesture processing code, it would not be observed with MOZ_ENABLE_WAYLAND=0 which disables that code.
    • That said, decreasing the value of the pref apz.gtk.pangesture.page_delta_mode_multiplier from its default 1.0 to a lower value is likely to help as a workaround.
  • If the sensitivity was a consequence of Firefox scrolling code common to scroll-wheel and pan-gesture, it would be observed with a peripheral mouse as well.

I think a diagnostic step that would give us some more info is to test the behaviour of another GTK3 application. Do you perhaps know of something similar to xev, but which works with Wayland, and which operates at the GTK level so that it would e.g. print GdkEventScroll deltas, so we can see how they compare between the touchpad and the peripheral mouse?

Flags: needinfo?(botond) → needinfo?(stransky)

(In reply to Botond Ballo [:botond] from comment #5)

  • That said, decreasing the value of the pref apz.gtk.pangesture.page_delta_mode_multiplier from its default 1.0 to a lower value is likely to help as a workaround.

Based on the discussion in bug 1752862 and related bugs, another option worth trying is setting apz.gtk.pangesture.delta_mode to 2 (that's "pixel" mode; with that in place, fine-tuning can be done by varying apz.gtk.pangesture.pixel_delta_mode_multiplier rather than apz.gtk.pangesture.page_delta_mode_multiplier).

(In reply to Botond Ballo [:botond] from comment #5)

I think a diagnostic step that would give us some more info is to test the behaviour of another GTK3 application. Do you perhaps know of something similar to xev, but which works with Wayland, and which operates at the GTK level so that it would e.g. print GdkEventScroll deltas, so we can see how they compare between the touchpad and the peripheral mouse?

We may use gedit and open large text file there and then try the scroll. If Wayland is involved we can print deltas by WAYLAND_DEBUG=1 and filter out the increments, like:

WAYLAND_DEBUG=1 gedit 2>&1 | grep "axis"

this one prints scroll deltas from Wayland compositor. Luca, can you try that and attach output here?
Thanks.

Flags: needinfo?(stransky) → needinfo?(luca.swainson)
Attached file output
(In reply to Martin Stránský [:stransky] (ni? me) from comment #7) > (In reply to Botond Ballo [:botond] from comment #5) > > I think a diagnostic step that would give us some more info is to test the behaviour of another GTK3 application. Do you perhaps know of something similar to `xev`, but which works with Wayland, and which operates at the GTK level so that it would e.g. print `GdkEventScroll` deltas, so we can see how they compare between the touchpad and the peripheral mouse? > > We may use gedit and open large text file there and then try the scroll. If Wayland is involved we can print deltas by WAYLAND_DEBUG=1 and filter out the increments, like: > > ``` > WAYLAND_DEBUG=1 gedit 2>&1 | grep "axis" > ``` > > this one prints scroll deltas from Wayland compositor. Luca, can you try that and attach output here? > Thanks. Hi, I did as you said and the scroll did seem faster than gtk4 software. It generated quite a bit of output, which I will paste below:

Some input that could be useful on this, I think this could be related to Firefox being on GTK3 instead of 4. In GTK4 they changed the mutliplier value for the touchpad scrolling speed to be more sane, making GTK4 apps scroll at a reasonnable speed, where GTK3 apps scroll too fast, similar to firefox does by default under wayland. If you want to check this out, try a GTK4 app, such as any libadwaita app, and a GTK3 app, such as Geany. This fast scroll issue can also be seen in Chrome (and other chromium based apps, including Electron) when using the Wayland mode.

https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/4508#note_1416720

A needinfo is requested from the reporter, however, the reporter is inactive on Bugzilla. Given that the bug is still UNCONFIRMED, closing the bug as incomplete.

For more information, please visit BugBot documentation.

Status: UNCONFIRMED → RESOLVED
Closed: 9 months ago
Flags: needinfo?(luca.swainson)
Resolution: --- → INCOMPLETE

Not sure why this was marked as resolved? There was no needinfo request on my end.

Status: RESOLVED → UNCONFIRMED
Resolution: INCOMPLETE → ---

Luca, did tweaking the prefs discussed in comment 6 help?

Botond: I'm not Luca, but I agree with this bug. FWIW I tried comment 6 prefs and I do agree they are better than the defaults.

(In reply to Botond Ballo [:botond] from comment #12)

Luca, did tweaking the prefs discussed in comment 6 help?

Yes, setting apz.gtk.pangesture.delta_mode to 2 and apz.gtk.pangesture.pixel_delta_mode_multiplier to 20 made it usable. These seem like they would be better defaults than the current settings.

You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: