[ESS] Aquamacs ESS slow scroll (buffering?)

Steve Lianoglou lianoglou.steve at gene.com
Tue Nov 26 18:43:47 CET 2013

Hi Paul,

On Mon, Nov 25, 2013 at 8:45 PM, Paul Johnson <pauljohn32 at gmail.com> wrote:
> Here's what I found out last night.
> Steve suggested using "real" Emacs. Hey, Steve: What's your time output?  I
> downloaded Emacs-universal, It has no ESS built in. Bummer. I took
> Emacs-Modified for Mac on Prof Goulet's website, which appears to start
> from same Emacs and add the useful stuff.
> I want to force full output, so I add the options line in this test
> function:
> set.seed(23456)
> y <-  lapply(1:50000, function(x) paste0(letters[1:(rpois(1, 10))],
> collapse=""))
> options("max.print" = 50000)
> system.time(print(y))
> I started the test, watched Emacs scroll for a while, went and took a long
> shower, watched a quarter of the Chiefs-Chargers football game, then
> watched Emacs.app some more.

Indeed -- I only suggested using a more vanilla emacs to alleviate the
annoyance you alluded to when having to support your students with
Aquamacs, since aquamacs tweaks emacs to make it more osx like (which,
you find annoying).

As I mentioned, the time output on any version of emacs still
exhibited the same problem you were posting about.

I did ask, however, if you were sure this was a osx specific issue, as
I think was your initial suggestion.

Is it true that you do not see this degraded performance on you linux setup?

> I got an email suggesting "How about RStudio?"
> [49999]]
> [1] "abcdef"
> [[50000]]
> [1] "abcdefghijklmnopq"
> user  system elapsed
> 2.540   0.103   2.529
> I was surprised how fast that is. There's no colored syntax. But I can
> scroll up and down the whole output frame without any fight. However, ...

This is also (pleasantly) surprising, as RStudio previously had some
issues w/ speed (with its coding buffer at least), which was further
exacerbated when running RStudio on Mavericks.

They provided a new pre-release of RStudio which apparently makes it
even faster, so if you haven't tried that, it might go even faster:



Steve Lianoglou
Computational Biologist

More information about the ESS-help mailing list