chris warth | 17 Apr 02:16 2015

polymode/ess-mode interaction

I noticed that polymode behaves differently after visiting *.R files than
it does before visiting a *.R file.
The difference is caused by ess-mode defining the symbol 'r-mode' which
polymode tries to use.  If ESS hasn't been loaded yet, 'r-mode' is not
defined and polymode falls back to using Fundamental mode instead (after a

How can I ensure that r-mode from ESS is always defined and available
before loading a *.Rmd file with polymode?

I'm trying to do this in a hook to avoid loading ESS unless I really need
I tried this without success,

     (add-hook 'poly-markdown+r-mode-hook
   (lambda ()
 (message "in poly-markdown+r-mode-hook")
 (require 'ess)))

Also requiring the more specific 'ess-r-d in the poly-markdown+r-mode-hook
does not seem to help.

Ideas?  Thanks in advance!

Chris Warth
Fred Hutchinson Cancer Research Center
1100 Fairview Ave N.
Seattle, WA 98109-1924

	[[alternative HTML version deleted]]
(Continue reading)

Tamas Papp | 13 Apr 15:35 2015

process-send-string 500 char limitation


Working on ESS/Julia interaction, I ran into a problem with sending
large chunks of code to the inferior Julia process using

Looking into the source and the docs for process-send-string (used by
ESS-SEND-STRING), I noticed that

  "If STRING is more than 500 characters long, it is sent in several
  bunches.  This may happen even for shorter strings."

Is there a way to overcome this limitation and send potentially large
strings as a single piece? I need this because I worked out a way to
annotate pieces of code so that they show file and line number info,
using Base.include_string in Julia, but this breaks with
PROCESS-SEND-STRING because Julia has no + prompt like R.

Alternatives like writing code into a file exist, but would give the
wrong file & line number info.



ESS-help <at> mailing list

(Continue reading)

Tamas Papp | 10 Apr 13:36 2015

Julia in ESS


Now that ESS 15.03 is released, I was wondering what the plans of the
ESS team are regarding Julia support.

I have been an ESS user for a long time now (interacting with R), and
IMO it would be easier to make ESS fully support Julia than to duplicate
its functionality in another, new Emacs package, but I don't know if
ESS developers think this is worth supporting.

I would like to contribute to ESS to make it work better with Julia, but
I would appreciate some guidance from ESS developers on what their plans
are. In particular, are you guys planning to merge this PR 122, or is
there some reason that makes this approach non-viable and would require
another solution?

I have code piling up in
which I would be happy to merge eventually once these issues are
resolved, and continue working on ESS/Julia.



ESS-help <at> mailing list

Stephen Eglen | 7 Apr 12:20 2015

Re: Is this useful for discriminating Sweave vs knitr docs?

> I agree that the ESS should use the file ocal variables approach and not
> use the RStudio syntax, but I would suggest that ESS can offer to *add*
> the file local variables if the RStudio syntax is present and possibly
> vice versa as well, i.e. write, in *addition* to the file local
> variables, the RStudio syntax as well?

Thanks Martin, Rainer.

I'm a bit wary of having two syntaxes around (think about the poor end
user... what if two conflicting values are provided?).

Any of the Rstudio team on here and wish to comment?  I'm all for trying
to ensure we keep our conventions together a bit more closely.
(Sectioning comments is another that springs to mind.)


ESS-help <at> mailing list

Stephen Eglen | 7 Apr 00:27 2015

lintr and flyspell

Anyone been using lintr package for lint-checking of R code?  Its now
integrated into flycheck:

I've not tried it out yet - has anyone?


ESS-help <at> mailing list

Stephen Eglen | 6 Apr 23:50 2015

Is this useful for discriminating Sweave vs knitr docs?

Rstudio use a simple comment like:

% !Rnw weave = knitr

to ensure that Rstudio knows the current .Rnw file should be processed
with knitr (rather than Sweave).  Would that be useful for ESS to adopt?


ESS-help <at> mailing list

Paul Johnson | 4 Apr 05:00 2015

Melpa got me now. What you say about ess-smart-underscore ?

I finally got Melpa to work. I've tried a lot of packages in the last 2 days :)

I found ess-smart-underscore and I think I like it.  I think I wish
you'd put it into ESS itself.  Have you thought about it?

Without smart underscore, I still get a lot of accidental <- when I want _.

I watch the students in the lab here learning to use ESS/Emacs and the
two most frustrating keyboard things for them are 1) the arrow keys
don't work the way the expect and 2) variables with underscores in
them are very difficult to type.

To anticipate one counter argument, changing the behavior of _ will
inconvenience the people who have gotten used to the double _.
However, remember that the underscore was a forbidden character (was
assignment in S) and most of us who are really old still get a
creepy-crawley feeling when we see underscores in variable and
function names.  So, if you install smart underscore, you reduce the
discomfort by 1/2 by reducing the number of times they have to type


Paul E. Johnson
Professor, Political Science        Director
1541 Lilac Lane, Room 504      Center for Research Methods
University of Kansas                  University of Kansas     

(Continue reading)

Marius Hofert | 2 Apr 23:01 2015

Re: Multiple R versions not recognized on M-x R-[Tab]

Hoi Martin!

Works like a charm, thanks a lot.

[thanks also for the hint at 3.2.0]

Cheers to Zurich (and keep up the great work),


On Thu, Apr 2, 2015 at 4:07 PM, Martin Maechler
<maechler <at>> wrote:
> Dear Marius  (and all the other readers),
> All your R versions must be in your PATH, i.e.,  the  <rbuild>/bin/R
> must be in your PATH,
> but of course not "as R" but  as "R-3..." or "R-devel.."
> and "the way" is to use symbolic links :
> e.g.
> cd /usr/local/bin
> ln -s  /usr/local/R/R-3.2.0_build/bin/R  R-3.2.0
> Hint:  R 3.2.0 went from "alpha" to "beta" today ... release is in
> about two weeks.
> Best regards,
> Martin
(Continue reading)

Marius Hofert | 2 Apr 21:50 2015

Multiple R versions not recognized on M-x R-[Tab]


I work on Debian Testing and installed ESS (M-x ess-version gives: 15.03 [git:
6cbe608018deda6fa5608735c821094a553dd55c] (loaded from
/usr/share/emacs/site-lisp/ess/)) and have several R version installed. However,
I don't see them via M-x R-[Tab].

I saw this post:
but neither gives precise instructions how to get the 'automagically determine R
versions' feature.

Here are some more details:
- Emacs 24.4.1
- M-x ess-r-versions-created gives /usr/local/R/R
- M-x ess-rterm-version-paths gives nil
- M-x R-[Tab] gives:

Possible completions are:
R-fix-T-F R-initialize-on-start
R-mode R-site-search

Judging from the second post above, I tried this:

(setq ess-rterm-version-paths "/usr/local/R/R"); for finding all
installed R versions

... but M-x R-[Tab] still gives the same output as before
(Continue reading)

Feng Li | 2 Apr 15:28 2015

Inverse of many small matrices

Dear all,

I am working with a likelihood function that requires the inverse of 
many small covariance matrices for multivariate normal densities. When 
the sample size is large, this calculation is really heavy. Those 
matrices are independent but unfortunately I can hardly find a way to 
vectorize them.

Can anyone give me a hint to speed this up? Thanks in advance!



Feng Li, Ph.D.
School of Statistics and Mathematics
Central University of Finance and Economics
100081 Beijing, China

ESS-help <at> mailing list

David Scott | 2 Apr 13:34 2015


On Windows I have installed Emacs 24.4 and ESS 14.09 using Vince 
Goulet's distribution. I am having a lot of trouble with processing .Rnw 
and .Rmd files using knitr. I am aware of the M-n w and M-n e commands 
(and the related M-n W and M-n E commands). I am able to weave 
successfully, but exporting (which I understand should use pdfLaTeX for 
a .Rnw file) just seems to hang showing no output and I have no idea 
what is going on. This possibly happens when I have a problem in my file 
(such as a beamer slide which includes verbatim but is not specified as 
fragile), but I really don't know. For .Rmd files I resorted to a 
creating a .bat file using render from the rmarkdown package, which was 
reasonably successful, but was not the best workflow in my view. I would 
really like to be able to just process .Rnw files as previously with M-n 
s followed by M-n P, but I am far from that at present.

Another problem I have at the moment is with M-n w adding '[weaved]' to 
the name of the file. First of all that causes a proliferation of files, 
secondly clicking on the icon to view any resulting pdf causes a file 
not found error because the file name looked for doesn't have '[weaved]' 
added to it.

(To be totally unreasonable, the choice of the word 'weaved' to be added 
to the file name grates on me like finger nails on a blackboard. The 
appropriate adjective for an object which has been created by weaving is 

So. Can anyone explain to me how to process a .Rnw file and view the 
resulting pdf? Also, how to stop the addition of [weaved] to the file name?

Failing that, can anyone explain how to modify auto-mode-alist so that 
instead of polymode, it invokes the old noweb where there is a menu and 
(Continue reading)