GNU bug report logs -
#32932
27.0.50; render bugs on macOS Mojave
Previous Next
Full log
View this message in rfc822 format
On Wed, Jan 29, 2020 at 12:04 PM Alan Third <alan <at> idiocy.org> wrote:
>
> On Wed, Jan 29, 2020 at 08:32:25AM -0800, Aaron Jensen wrote:
> > On Wed, Jan 29, 2020 at 2:08 AM Alan Third <athird <at> googlemail.com> wrote:
> > >
> > > Hi Aaron, just as a test can you try disabling powerline?
> >
> > Hi Alan, I don't use powerline, I use doom-modeline and disabling it
> > has no effect.
> >
> > I actually have worse news. My original test was actually with an
> > unoptimized 27, as I forgot to actually run make (sigh). With 28, on
> > my 4k display, Emacs is unusable even with the modeline disabled. My
> > frame is only on about 3/5s of the screen, so it's not even taking up
> > the entire 4k. The latency on every paint is probably in the 200-400ms
> > range. I haven't measured it, but that's what it feels like.
>
> Really it shouldn’t be slow for updating relatively small areas, so if
> every paint is that slow then we’ve got real trouble.
>
> Do you use a transparent background in Emacs?
No
>
> Can you try applying these three changes, and see if they make any
> difference. I’d like you to try the first two both with and without
> the third.
With regard to scrolling (full repaints) they make no difference,
which you likely suspected. These are my scroll-up-benchmark times for
one of my .org files:
Emacs 27: 2.6s
Emacs 28
No patches: 13.0s
First 2: 13.2s
All 3: 13.8s
In terms of just repainting when I move the point, applying the first
2 patches made a noticeable difference. It is still sluggish compared
to 27. I'm not sure I can tell a difference with the 3rd patch applied
as well. It seems similar or maybe slightly faster.
Aaron
This bug report was last modified 5 years and 93 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.