Currently, you have to insert the following into .emacs in order for .R files to open in R-mode:

(autoload 'R-mode "ess-site.el" "" t)
(add-to-list 'auto-mode-alist '("\\.R\\'" . R-mode))

Could we please register .R file extensions with R-mode as part of ess installation?

The new SAS automatic mode conflicts with org-mode html export somehow. Attempting html export from org-mode results in the error:

ess-turn-on-SAS-listing-mode: Wrong type argument: arrayp, nil

I can't figure out why this mode is trying to turn on, perhaps a temporary file name that org is producing? Setting ess-automatic-sas-log-or-lst-mode to nil fixes this.

Backtrace shows the functon tree (I had to leave out the arguments with characters that won't display):

  replace-regexp-in-string("\\.lst$" ".log" nil)
  ess-turn-on-SAS-listing-mode()
  set-auto-mode(t)
  org-html-final-function

Now ESS puts all info files in the main directory of the elpa ESS folder. Could it be possible to put all info files in the "info" sub folder, so that it looks cleaner?

I get the following error when compiling ESS from package.el:

Compiling file /Users/dcurtis/.emacs.d/elpa/ess-20120924.1839/lisp/ess-bugs-d.el at Thu Oct  4 08:22:18 2012
ess-bugs-d.el:29:1:Error: Symbol's value as variable is void: ess-help-arg-regexp

It was reported in melpa/melpa#310

ess-comp.el:51:39:Error: Invalid read syntax: "?"
ess-debug.el:75:1:Error: Invalid read syntax: "?"
ess-r-d.el:186:1:Error: Invalid read syntax: "?"
ess-rutils.el:54:1:Error: Invalid read syntax: "?"
ess-site.el:397:1:Error: Invalid read syntax: "?"
ess-swv.el:89:1:Error: Invalid read syntax: "?"

With ess installed on Ubuntu raring, when start emacs24 as root, the site-wide startup file /etc/emacs/site-start.d/50ess.el hangs emacs startup.

As a workaround, I gzipped the file and emacs starts. But then emacs doesn't speak statistics.

uname -a

Linux larry 3.8.0-22-generic #33-Ubuntu SMP Thu May 16 15:17:14 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

root@larry:/etc/emacs/site-start.d# lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 13.04
Release: 13.04
Codename: raring

root@larry:/etc/emacs/site-start.d# apt-cache show ess
Package: ess
Status: install ok installed
Priority: optional
Section: gnu-r
Installed-Size: 5331
Maintainer: ESS Debian Maintainers [email protected]
Architecture: all
Version: 12.09-2-1quantal0
Depends: dpkg (>= 1.15.4) | install-info, emacs23 | emacs22 | emacs21 | emacsen
Suggests: xlispstat, r-base | r-base-core, pspp, jags
Conflicts: dhelp (<= 0.3.12)
Conffiles:
/etc/emacs/site-start.d/50ess.el 848ba5caab794872624a0c09f55c7a14
Description-en: Emacs mode for statistical programming and data analysis
ESS ("Emacs Speaks Statistics") is a GNU Emacs and XEmacs mode for
interactive statistical programming and data analysis. Languages
supported are the S family (S 3/4, S-PLUS 3/4/5/6/7, and R), SAS,
XLispStat, Stata, BUGS and Julia.
.
ESS grew out of the desire for bug fixes and extensions to S-mode and
SAS-mode as well as a consistent union of their features in one
package.
Homepage: http://ess.r-project.org

Package: ess
Priority: optional
Section: universe/math
Installed-Size: 5331
Maintainer: Ubuntu Developers [email protected]
Original-Maintainer: ESS Debian Maintainers [email protected]
Architecture: all
Version: 12.09-2-1
Depends: dpkg (>= 1.15.4) | install-info, emacs23 | emacs22 | emacs21 | emacsen
Suggests: xlispstat, r-base | r-base-core, pspp, jags
Conflicts: dhelp (<= 0.3.12)
Filename: pool/universe/e/ess/ess_12.09-2-1_all.deb
Size: 1639420
MD5sum: 30071b1812f752567f0493b5fb1bc47c
SHA1: 4e85ff2804f04555e994be75927e78c1d168fb4f
SHA256: 520aa6a01c3c5cf52c3d9909dfb95ebf392dbfd22263d217033468329d42acf8
Description-en: Emacs mode for statistical programming and data analysis
ESS ("Emacs Speaks Statistics") is a GNU Emacs and XEmacs mode for
interactive statistical programming and data analysis. Languages
supported are the S family (S 3/4, S-PLUS 3/4/5/6/7, and R), SAS,
XLispStat, Stata, BUGS and Julia.
.
ESS grew out of the desire for bug fixes and extensions to S-mode and
SAS-mode as well as a consistent union of their features in one
package.
Homepage: http://ess.r-project.org
Description-md5: f5e47a5777b8c61804112900078b7e16
Bugs: https://bugs.launchpad.net/ubuntu/+filebug
Origin: Ubuntu

It seems that when tracebug mode is enabled, commands like ess-eval-region save the text to be evaluated to a temp file and then source that file. However, this doesn't work when the R process in question is running on a remote host (via TRAMP), since the temp file is created on the local machine. I guess the temp file saving routine needs to be modified to save to the remote hosts's temporary directory when the R process is running on a remote host.

Hi, I step through code a lot, mainly with C-c C-n and C-c C-p

Whenever I hit the beginning of a loop I have to skip that line, setup the loop variables and step from within the block, example:

code.outside()

for( i in some.variable ){

code.inside()

}

I would happily step through the code outside, move passed the for line, and start stepping from the code in side.

How about creating an intelligent step, that would understand the loop and setup the variables in it, ie "i" in my case? And then allow me to step further inside.

T.

In the license section of the header of ess-bugs-d.el, it states the following:

;; In short: you may use this code any way you like, as long as you
;; don't charge money for it, remove this notice, or hold anyone liable
;; for its results.

This is not an accurate description of the GPL. There is nothing in the GPL which prohibits charging money for the software, and there are many other important aspects of it. I recommend deleting this paragraph altogether, since it does not appear in typical Elisp files.

My main problem is using el-get to build and use ESS on windows.

On windows I don't have tex installed, but the make clean, make all el-get recipe tries to build the documentation which fails, (which causes the package to not be included). Certainly having tex installed shouldn't be a requirement to use ESS.

While this is a bit of an el-get problem, it's a problem for el-get because there is no target that just builds the lisp. While this can be fixed by 'cd lisp && make' that doesn't feel right to me. I'd prefer the makefile have a target that just builds the lisp.

I'm more than willing to make a pull-request that does this if that'd be easier.

Hi,

I posted what I think is a bug to ESS-help but got no replies, so I'm posting it here in the hopes a developer sees it.

The problem I'm having is the following:

I'm using a combination of emacs + ESS + Auctex + knitr to edit .Rnw files.
It's working fairly well with the addition of a small function to call
my makefile and generate the final pdf.

A problem I'm encountering is with the syntax highlighting. I think
it's connected to the parsing of the latex input.

Here's an example:

\documentclass[a4paper]{article}

\begin{document}

\title{test}

\maketitle

<<chunk1>>=
a = list(x = 1, y = 2)
b = a$x
@

this is the test continued...

\end{document}

After the @ everything gets the same color. I think the parser is
getting confused because it expects another dollar sign after a$x, but
that's not latex math notation, it's an R command to select an item
from a list.

You can actually see that the Github highlighter has the same problem, but it's not supposed to know about weaving.

Here's the relevant portion of my .emacs and my package versions:

(require 'ess-site)
(ess-toggle-underscore nil) ; desabilite <- no modo ESS.
(setq inferior-R-args "--no-save --no-restore-data") ; dont restore workspace
(setq TeX-PDF-mode t) ; use pdflatex
(load "auctex.el" nil t t)
(load "preview-latex.el" nil t t)
(setq ess-ask-for-ess-directory nil) ;otherwise you are prompted each
time you start an interactive R session

Software versions:
emacs 24.1-2
emacs-ess 12.04.4-1
auctex 11.86-3

Any ideas how to solve this?

Thanks,
Cássio

Sometimes when I save my R script in ESS, an Adobe Acrobat window pops up asking me what language to use. I have not been able to determine the origin, but setting options(pdfviewer=NULL) did not fix it.

bla bla bla

The function ess-indent-or-complete (bound to TAB) generates an error message saying that looking-at-p does not have a function definition, when in R mode.

To reproduce edit a function in R-mode and hit tab in the middle of an incomplete expression.

For example, hitting TAB in the whitespace to the left of function in the fourth line of this expression generates the error.

emparams <- function(measure=1:5) {
as.vector(outer(c("a","b","itau"),measure,
function(x,y) paste(x,"[",y,"]",sep="")))
}

When looking for help for functions in multiple packages (like merge in base and data.table), R prompts the user to select which package they want help for:

> ?merge
Help on topic ‘merge’ was found in the following packages:

  Package               Library
  data.table            /Library/Frameworks/R.framework/Versions/2.15/Resources/library
  base                  /Library/Frameworks/R.framework/Resources/library

Choose one 

1: Merge Two Data Tables {data.table}
2: Merge Two Data Frames {base}

When using auto-complete, the quick-help feature (seemingly) is causing Emacs to hang indefinitely when it attempts to show documentation for such a function.

This is in version 12.09-2.

When I type

rtags(

in the console, the mini buffer shows only a few arguments as

rtags: path=".", pattern=".[RrSs]$", recrusive=F, src=list.files(path=pat{.})

but rtags has more arguments

 rtags(path = ".", pattern = "\\.[RrSs]$",
       recursive = FALSE,
       src = list.files(path = path, pattern = pattern,
                        full.names = TRUE,
                        recursive = recursive),
       keep.re = NULL,
       ofile = "", append = FALSE,
       verbose = getOption("verbose"))

I suspect this is a bug rather than an expected behavior because the plot() function has more arguments but all are displayed correctly.

I have emacs 24.0.95.1 and ess 12.4 installed for Debian Wheezy Linux 64bit.

Thanks a lot!

Feng

Did someone force push at some point to this repository? Because for some reason, even when i do a merge with master, I get a very weird history with commits that don't exist in the repository anymore?

That is, I am seeing commits that are not from me, but are not in the master branch...

1. Load up R file
2. start up R interpreter session in other buffer
3. try auto-complete (=> should work)
4. set ac-auto-start to nil
5. try auto-complete (=> should fail)

A colleague found this likely source of the problem in ess-r-d.el:

(if (< (length ac-prefix) ac-auto-start)

I think the latest commits caused this issue.
It is blocking me from using ESS.

I use GNU emacs 24.3 on ubuntu 13.04 64bit.

Using ESS-dev rev 5744, Emacs 24.3.50.1 (x86_64-pc-linux-gnu, GTK+ Version 3.8.2) of 2013-08-21 on dex, modified by Debian, with a minimal .emacs:

(setq load-path (cons "~/.emacs-site" load-path))
(load "~/.emacs-site/ess-dev/lisp/ess-site.el")

I have the following problem at R startup: After opening a .R file, and send the first command without previously opening R (C-c C-n directly), the following error is thrown in the mini-buffer: 'ess--inject-code-from-file: Opening input file: aucun fichier ou dossier de ce type, /home/mathieu/nilESSR.R' (in English: 'No such file or directory'). R starts normally, and I can do my stuff, although I have to resend the line first.

As far as I can see, I can evaluate line, paragraph and region. And as far as I can understand, a paragraph is delimited by empty lines.

the paragraph eval is nice, it means I can break up long lines on more lines and still easily run they at once in the inferior process.

How about introducing eval block? It would identify the block that the cursor is in and evaluate it. that means I could easily run a loop in a buffer:

for( foo in bar ){

  some()

  code()

  here()

  that.can.have()

  air.in.it()

}

I could easily evaluate that loop by standing in it and, instead of first selecting all of it and evaluate region.

Now instead I find myself adding ## on blank lines to be able to achieve this with eval paragraph:

for( foo in bar ){
  ##
  some()
  ##
  code()
  ##
  etc()
}

.. which is not desirable

The hs-minor-mode already has code that can detect blocks of code like this, so it should be doable.

Would it be possible, and implementable, to connect to an R session on another host, perhaps using ssh, perhaps with tramp.

X-forward will probably be tricky, but personally I'd be happy to run R with DISPLAY=:1 in for example vnc on that other host and just connect with a vnc viewer to see graphics.

When I do Alt-x julia emacs 24.2 shows this message and freezes. Installed via msys (windows 7 x86). The documentation phase of the makefile (makeinfo) also fails.

Any ideas?

github test

Emacs hangs with the above message on first use of ess (eg ess-eval-line)

I'll try to pinpoint where this changed exactly when I get some time. It worked fine as of 20130711.

emacs 24.3.1 on ubuntu.

testing notifications again

this is a minor error, however i thought i would still report it. so do this in an R window running in emacs : 1) type a double quote ("). 2) press Enter.. 3) press Ctrl-C twice. now it has exited from the previous command and is ready for a new one. 4) start pressing another command. it appears orange as if it were a comment and not a command. although the command works fine and the way to come out of this is that press a double quote (") again and press Enter and then Ctrl-C twice again, i still thought of reporting because sometimes small issues lead the developers to larger ones. thanks and thanks again for the ESS, its a boon for R users.

Before I worked with R from emacs, I had a function save.work("some tag") # that would save history and save image and create two files with timestamp and "some tag" in their filename in a dedicated projcet sub directory.

This allowed me to load both data and history from a compiled list of work sessions, and was a fast way to set my mental context back to where I were when I left off.

From emacs unfurtunately savehistory is not available, what other ways is there to achieve this?

It would be nice to have some simple functionality that would:

• save a valid R-history
• save.image
• save the R buffer with all its content
• label and timestamp it in a meaningful way

And then functionality that would restore it later, ie start a R process, load hitsory from last time, load the R buffer content from last time (maybe omitting parts where I by accident print big data objects?), load the data from last time, and essentially bring me back to where I were when I left off.

All of this should be doable, but I wouldn't quite know where to start.

T.

In a .Rnw file when moving from tex code to R chunk code, the line is immediately sent to R process. If I move the cursor from top to bottom the line <<>> is getting evaluated, if I move from bottom to top, then the last line of the chunk before @ is evaluated. This makes editing Sweave files quite annoying. I use ESS 12.09 with Aquamacs 2.2 on Mac OS X 10.6.8.

The current melpa package of ESS does not contain an info file. Can you please add it?

Example code:

for( foo in bar ){

while( some.thing() ){

if( this.or.that() ){

    a.function.call(  perhaps.an.array[1 #  <-- marker after the 1 here

The proposed lisp function if run now, would first close the bracket by adding a "]" , if I ran it again it it would add a ) to close the ( of the function call on that line.

Running it more times would then close { } blocks from if, while and for.

Seeing how ess can match blocks and parentheses as I typt them, it would probably be a quick fix to add this extra functionality (unless it exists! I searched and could not find it)

Hi, in lisp I can go to the end of a bracket (foo bar)<-here and hit C-x C-e and it will evaluate the list that ends with that bracket

Could something similar be implemented for R?

Im thinking along these lines:

foo <- sum( is.na( some.funcion() ) ) )

If I place the marker after a bracket, it would evaluate the nested code matching that level, ie (showing the marker with underscore):

foo <- sum( is.na( some.function()_ ) ) # would evaluate some.function()

foo <- sum (is.na( some.fun_ction() ) ) # would also evluate some.function()

foo <- sum( is.na( some.function() ) )_ # would evaluate sum( ... )

foo <-_ sum( is.na(some.function() ) ) # would evaluate the assignment (and the sum )

That way I could easily debug nested code on one code line (or paragraph) without having to re-select incremently larger potions of it before running ess-eval-region

ess-remote is defined in essd-els.el, but it is used in the function ess--R-load-ESSR in ess-r-d.el. So that if I used a personalized ess-site.el and don't load essd-els, it will fail to load ESSR.

once in a while I experience cases where the roxygen template generation will not work properly:

I just played around and concluded that if I have a function with parameter lines like

    ...
    col = '#ff0000',
    ignored = TRUE,
    ....

parameter ignored will not be included in the template (and if I put it there manually, it will be removed by the next update)

This seems to be caused by the '#' in the assigned default ...?

A fix or a suggestion for a workaround would be much appreciated

Benno

If I emacs -Q, I can manually M-: (require 'ess-site) M-x ess-mode successfully. However, If I try to delay loading ess-site until Emacs tries to open a .R file, then I get an error. Am I doing something wrong?

Trace:

Debugger entered--Lisp error: (error "Customise alist is not specified, nor  ess-local-customize-alist is set.")
  signal(error ("Customise alist is not specified, nor  ess-local-customize-alist is set."))
  error("Customise alist is not specified, nor  ess-local-customize-alist is set.")
  ess-mode()
  set-auto-mode-0(ess-mode nil)
  set-auto-mode()
  normal-mode(t)
  after-find-file(nil t)
  find-file-noselect-1(#<buffer comments.R> "~/Desktop/src/comments/R/comments.R" nil nil "~/Desktop/src/comments/R/comments.R" (910806962 16777217))
  find-file-noselect("/Users/andrew/Desktop/src/comments/R/comments.R" nil nil nil)
  find-file("/Users/andrew/Desktop/src/comments/R/comments.R")
  command-line-1(("comments.R"))
  command-line()
  normal-top-level()

System:

$ specs emacs os
Specs:

specs 0.3
https://github.com/mcandre/specs#readme

emacs --version
GNU Emacs 24.3.1
Copyright (C) 2013 Free Software Foundation, Inc.
GNU Emacs comes with ABSOLUTELY NO WARRANTY.
You may redistribute copies of Emacs
under the terms of the GNU General Public License.
For more information about these matters, see the file named COPYING.

system_profiler SPSoftwareDataType | grep 'System Version'
      System Version: OS X 10.8.3 (12D78)

I'm using ESS to evaluate a buffer of R in a process running remotely in Emacs. When I try to run ess-eval-buffer on a medium-sized .R file (just over 100 lines) I am seeing errors appear due to text not being sent to the R process correctly. My workflow is:

• I call M-x shell within Emacs
• ssh to the remote machine
• call M-x ess-remote on the shell buffer
• call ess-eval-buffer in the buffer of the R code

An example of the error I am seeing is:

R> R> + Error: unexpected symbol in:
"ggplot(uniquefull[Widgets.for.Session < 500]) + geom_freqpoly(aes(x=Widgets.for.Session, y=..count..,
group=TestCggplot(uniquefull[Widgets.for.Sessi
print"`

When the actual code I am trying to run is:

...some code here...
uniquefull[Test == "LooseTestpre50", TestConcat := "LooseTest"]

ggplot(uniquefull[Widgets.for.Session < 500]) + geom_freqpoly(aes(x=Widgets.for.Session, y=..count.., group=TestConcat, colour=TestConcat)) + xlab("Number of widgets people finish with") + ylab("Number of user-sessions") + ggtitle("Comparing number of people who finish with how many widgets\n split by what test they were in, or were likely to end in")

ggplot(uniquefull[Widgets.for.Session < 500]) + geom_freqpoly(aes(x=Widgets.for.Session, y=..density.., group=TestConcat, colour=TestConcat)) + xlab("Number of widgets people finish with") + ylab("Number of user-sessions") + ggtitle("Comparing number of people who finish with how many widgets\n split by what test they were in, or were likely to end in")

print("Instead of looking at test bands, look at strata of wotzits at exit")

...some more code...

These are the settings I have in my .emacs for ESS:

(require 'ess-rutils)
(require 'ess-eldoc)
(add-hook 'inferior-ess-mode-hook
      '(lambda()
         (setq ess-eval-visibly-p nil)
         (auto-complete-mode t)
         (define-key inferior-ess-mode-map "\C-ct" 'ess-R-object-tooltip)
         (setq comint-input-ring-size 99999)))

(setq comint-prompt-read-only t)
(setq comint-scroll-to-bottom-on-input t)
(setq comint-scroll-to-bottom-on-output t)
(setq comint-move-point-for-output t)
(setq ess-ask-for-ess-directory t)

(add-hook 'ess-mode-hook 
          '(lambda()
             (define-key ess-mode-map (kbd "C-\\") 'ess-eval-line)))

Does anyone know how to solve this error?

Hi, I have a great idea.

When I work with R, i typically code in the R-terminal, and when it works, I paste code to an R buffer and save R scripts. So I repeatedely select code and paste, again and again.

(I don't run R from emacs, just use the r-mode)

It would be neat if R could speak to emacs directly, something like this:

*) user starts a local listening service in emacs, something like M-x background-accept-r-code-into-buffer
*) user runs something like send.to.emacs( pattern="foo" )

• this command searches through history() and prompts user to select lines of code
  *) R then sends this code to the emacs listening service
  *) emacs accepts code and inserts it where cursor is at in a buffer.

This would require an R package to do this (I wouldn't know where to start), and extra functionality to ess (I wouldn't know where to start there either)

T.

Given the following R code, run under Windows 7 using the 32-bit version of R2.15 patched, ESS does not recognise the text in Thai script in the data frame. If I copy and paste the same code into RGui it works as expected.

thaidf <- data.frame(thdate = c("31 พฤษภาคม 2555","30 เมษายน 2555"),
value = c(100,110))
english_months <- seq(1:12)
thai_months <- c('มกราคม','กุมภาพันธ์','มีนาคม','เมษายน','พฤษภาคม','มิถุนายน',
'กรกฎาคม','สิงหาคม','กันยายน','ตุลาคม','พฤศจิกายน','ธันวาคม')

print(thaidf)
for (ii in seq_along(thai_months)) {
## convert months in Thai script to numerical
thaidf$thdate <- (sapply(thaidf$thdate,
function(x) {gsub(thai_months[ii],
english_months[ii], x, useBytes = TRUE)}))
}
print(thaidf)

In a freshly cloned copy (commit 43d10f1), the file ess-julia.jl seems to be missing, and when started with M-x julia, it gives the error:

julia> load("/Users/yavuz/Dropbox/.emacs.d/yavuz/ESS/etc/ess-julia.jl")

julia> Cannot find /Users/yavuz/Dropbox/.emacs.d/yavuz/ESS/etc/ess-julia.jl
 in realpath at file.jl:151
 in find_in_path at util.jl:197
 in require at util.jl:174

It would be nice to have ESS be more friendly to being installed with package.el which means using autoload in the proper places. Is there any insight into what would need to be done to make this happen?

My MacBook runs OS X Mountain Lion 10.8.2. I installed ESS version 12.09-2 from the tar file in the download area.

I can run a ESS session just fine, but when I exit Emacs and confirming that the R process should be killed, the screen immediately becomes grey and the computer reboots.

I have experienced this every time I have used ESS, but not since I realized that ESS was the reason for the reboot.

emacs version 24.3.50.1
ess version 13.05

In R, when load library 'lme4', and run ess-display-help-on-object (C-c C-v) will hung emacs.
with message:
Forming completions for .GlobalEnv...done
Forming completions for package:longpower...done
Forming completions for package:nlme...done
Forming completions for package:lme4...

I use ESS in conjunction with evil.

Sometimes I would like to evaluate a selected rectangular block. Eg. in

funkyFun(
  x = 5
 ,y = 10
 ,c = 25
)

I visually select block

x = 5
y = 10
c = 25

If I then do ess-eval-region, I get

R> x = 5
R> , y = 10
Error: unexpected ',' in " ,"
...

Using ESS-dev rev 5744, Emacs 24.3.50.1 (x86_64-pc-linux-gnu, GTK+ Version 3.8.2) of 2013-08-21 on dex, modified by Debian, with a minimal .emacs:

(setq load-path (cons "~/.emacs-site" load-path))
(load "~/.emacs-site/ess-dev/lisp/ess-site.el")

I have the following problem using the help system: If I call help on a function (C-c C-v), the following error is thrown in the mini-buffer: 'Error: could not find function ".ess_help"'. It can be fixed by creating the function in the R buffer:

.ess_help <- help

I know a similar problem used to affect ESS at some point, when the function was named .help (if I remember correctly).

ESS-mirror or just ESS?

Hi, if I type the following code in a buffer with r-mode:

foo <- bar[1]

I would then hit enter right after the "]" to start a new line.

As I input the second and closing "]", emacs pauses for parts of a second to hightlight the opening "[" after the r. The cursor jumps back to this point briefly before jumping to where I was typing.

If I hit enter during this pause, the "]" is removed.

So my keys would go like this:

f o o _ b a r [ 1 ]

(I use shift-underscore to input the assignment opreator)

The final instead of inputing a new line, deletes the "]" character unless I wait until emacs has jumped to the [ and back again to the ]

In a normal typing speed you can easily type 10-15 keys a second. Having to wait for some .5 or .8 or 1.0 second to wait for parens highlighting to complete, is an annoying interruption to the flow of fingers and mind.

Now I don't know if this parens highlighting is a feature of ESS or some other major or minor mode of emacs, but hopefully someone knows and can help me.

Yet another test.

I'm using R 2.14.2 under windows xp.
I found that the args can show for 'mean' but not for rnorm.

I checked that the output from ess--funargs-command:
For 'rnorm':

• + + + + + + + + + + + + + + + + + + + + Note: no visible global function definition for 'rnorm'
  Note: no visible global function definition for 'rnorm'
  [1] "stats" "n, mean=0, sd=1" "n=" "mean="
  [5] "sd="

For 'mean':

• + + + + + + + + + + + + + + + + + + + + [1] "base" "x, trim=0, na.rm=FALSE, ..."
  [3] "x=" "...="
  [5] "trim=" "na.rm="

So I think the function ess-get-words-from-vector needs to be improved.

I have the recent master branch of ESS with auto-complete enabled.

I like the popup thing that gives a function help when I try to complete a word. But it is also quite annoying that it always shows something like

No documentation for 'xxx' in specified packages and libraries: you cold try '??xxx'

when xxx is just a variable in the local buffer. Can this be a little bit smarter, i.e. only popup the help when the help is available, otherwise be quiet?

Thanks for the hard work!

Feng

Somebody working on gadfly support? Its weave function is similar to knitr.