bug#29737: 27.0.50; pixel-scroll-mode is laggy

[Valentin Ignatyev <valentjedi@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1977 bytes --]

Hi all.
I just tested this fix. While it doesn't solve the issue, I saw a little
improvement. And now I see the lag pattern. When I slowly start scrolling,
it scrolls for a while and then jumps about 10 rows at the time, then
scrolls a bit more and jumps again
If I swipe my trackpad too fast, emacs instantly eats 100% CPU and hangs
infinitely (it hangs not every time, but rather once in a while. It can
just go to the end of the file).
If I scroll with Ctrl-key as Eli suggested in the first thread mail, it
scrolls better than before the fix, but I still see this lag pattern
(scroll a bit and then jump), but in the smaller scale and it doesn't kill
emacs.
Setting mouse-wheel-progressive-speed to nil seem to make no difference.

Btw, while I do use macbook, my OS is Arch Linux here :)
I think I'll try to compile latest emacs-26 branch on macos itself and tell
there is defference. Maybe it'll shrink down the scope

Thanks for your work on this, I really appreciate it


On Sat, Dec 23, 2017 at 7:20 PM, Eli Zaretskii <eliz@gnu.org> wrote:

> > Date: Sat, 23 Dec 2017 12:18:56 +0900 (JST)
> > Cc: valentjedi@gmail.com, 29737@debbugs.gnu.org,
> tkk@misasa.okayama-u.ac.jp
> > From: Tak Kunihiro <tkk@misasa.okayama-u.ac.jp>
> >
> > > I'd like to fix this on the release branch, not on master.  Is the
> > > patch you sent good to go to the release branch?  Does it solve the
> > > display lags mentioned in the bug report?
> >
> > Yes, the patch I sent is good to go to the release branch.
>
> Thanks, pushed to the release branch.
>
> > I'm not 100% sure what 'the display lag' meant.
>
> It meant very slow scrolling, with redisplay falling far behind.
> After applying the patch, I see a definite improvement on my system.
>
> > After the patch, at least I tested with MacBook and do not recognize
> > 'the display lag'.
>
> Right.  Valentin, could you please see whether this change makes the
> problem go away, or at least makes it much less of a problem?  TIA.
>

[-- Attachment #2: Type: text/html, Size: 2818 bytes --]