[ESS] Dumping functions with C-c C-d in R 2.0.0 has new "feat ure"/ RCS

Patrick Connolly p.connolly at hortresearch.co.nz
Wed Oct 27 22:25:23 CEST 2004

On Tue, 26-Oct-2004 at 11:46PM -0400, Richard M. Heiberger wrote:

|> What are you doing when you save a date to a file?

It's somewhat involved, and evidently not very interesting to other R
users, but since you ask....  

|> Are you doing something to the R object that is saved inside the
|> file .RData?  Are you talking about the myfile.r file that stores
|> object definitions?

Just before objects are saved to .RData, I have a little procedure
that adds the current date as a comment to any object that has none
already, then it saves the body of the functions that have been
updated or created to a text file.  I used the comment instead of a
general named attribute because it was an attribute that wasn't
displayed under "normal" use.

I could bore the list with ravings about how I use the date
information everyday, but I'll desist.  I've mentioned it before but
few were interested.

|> When is the behavior you don't like?  What is the appearance on the screen?
|> Are you talking about displaying a function in the *R* buffer?
|> Are your talking about what you see when you dump a function with C-c C-d
|> to /tmp/p.connolly.object.R?  What was the previous behavior?
|> Please send some selections from the *R* buffer that illustrate both
|> the behaviors you do like and the behaviors you don't like.

Here's a little function I made some years ago.  From the *R* buffer,
it looks like this:

> sub.col.names
function(x, patt = "D")
### Gets the names of x that fit a particular pattern patt
    col.names <- names(x)
    y <- col.names[grep(patt, col.names)]

In the /tmp/p.connolly.sub.col.names.R buffer, it looks like this:

"sub.col.names" <-
  structure(function(x, patt = "D")
### Gets the names of x that fit a particular pattern patt
              col.names <- names(x)
              y <- col.names[grep(patt, col.names)]
            , comment = "13/06/2000")

Until the new behaviour of dump(), the whole structure stuff didn't
happen.  Now, everything I want to edit is shunted 12 extra spaces to
the right.  In the small example, it's of no consequence, but when
there are panel functions in calls to lattice functions, I can really
do with that extra 12 spaces.  

It also means that I have to hand edit the comment -- something I had
happening automatically up to now.

Martin has given some information that will probably do what I need.


Patrick Connolly
Mt Albert
New Zealand 
Ph: +64-9 815 4200 x 7188
I have the world`s largest collection of seashells. I keep it on all
the beaches of the world ... Perhaps you`ve seen it.  ---Steven Wright 

More information about the ESS-help mailing list