Re: [R] Rconsole wishlist

From: Prof Brian Ripley <ripley_at_stats.ox.ac.uk>
Date: Fri 04 Mar 2005 - 17:35:04 EST

The problem is that the commands history is GUI-specific (there are four mechanisms, readline on Unix, one for the GUI MacOS X (I believe), one for RGui and one for rterm. And we will have to cope with both GNU and BSD readline shortly). We try to keep them consistent, and that means a lowest-common denominator approach.

History is *not* intended to be a `command log' but a way to retrieve recentish commands, and the console buffer is just that, a buffer. They are implemented for fast access backwards and so not really appropriate for large logs. If you want a session log, look at 'split' option to sink().

With the advent of multibyte characters (supported in R as from 2.1.0) we need to redesign some of this (rterm's command line will not support such locales for 2.1.0). It should be possible to make the default R_HISTSIZE unlimited.

The console buffer is less clear: how do you want to be warned? This could happen in the middle of a single print() command: it could even happen that a single line of output exceeds the buffer size.

Finally: R-devel is for R development issues, so if people want to pursue this, please move there. (I am trying to be non-technical in this reply, and probably not succeeding.)

On Fri, 4 Mar 2005, Kjetil Brinchmann Halvorsen wrote:

> Duncan Mackay wrote:
>
>> Sorry, yes, Rgui under WinXP (SP2). But while Windows date stamps the
>> history file
>>
>>
>>> file.info(".Rhistory")
>>>
>> size isdir mode mtime ctime
>> .Rhistory 5377 FALSE 666 2005-03-04 10:37:52 2005-03-04 10:37:52
>> atime
>> .Rhistory 2005-03-04 13:54:11
>>
>> the problem is that there can be multiple sessions stored in .Rhistory and
>> the session dates aren't stored there. Moreover, it seems to me that the
>> history buffer can also overflow without warning after long sessions or
>> many
>> repeated sessions and so that you can inadvertently lose parts of your
>> command log.
>>
> Yes. But you can define an environment variable R_HISTSIZE (or some similar
> name, do a
> R site search to find. Not defined on my machine now. That should really be
> in the help file for
> savehistory()) to avoid the problem.
>
>
>> (Is this right, anyone?) Perhaps it would be preferable for R
>> to save each session's command history in a separate history file, along
>> the
>> lines of
> NO. it is better to have just one file as now.
>
> Kjetil
>
>
>> .Last <- function() {
>> savefilename <- paste("Rhistory",date())
>> savefilename <- gsub(" ","_",savefilename)
>> savefilename <- gsub(":",".",savefilename)
>> savefilename <- paste(savefilename,".txt",sep="")
>> if(interactive()) try(savehistory(savefilename))
>> cat("Current history saved in file: ",savefilename,"\n")
>> }
>>
>> but this doesn't address any overflow issues.
>>
>> Duncan
>>
>>
>> -----Original Message-----
>> From: Liaw, Andy [mailto:andy_liaw@merck.com] Sent: Friday, 4 March 2005
>> 11:14 AM
>> To: 'Duncan Mackay'; R-news
>> Subject: RE: [R] Rconsole wishlist
>>
>>
>> I'm guessing you're talking about Rgui on Windows, but please don't leave
>> us
>> guessing.
>>
>> If you run R under Ess/(X)Emacs, you have the entire session that can be
>> saved in a (transcript) file.
>>
>> Does your OS not put date stamps on file?
>>
>>
>>> file.info(".Rhistory")
>>>
>> size isdir mode mtime ctime
>> .Rhistory 1025 FALSE 666 2005-03-03 19:27:31 2004-08-13 10:45:09
>> atime
>> .Rhistory 2005-03-03 19:27:31
>>
>> Andy
>>
>>
>>
>>> From: Duncan Mackay
>>>
>>> Hi all,
>>> Wouldn't it be nice (??!!) if R automatically issued a
>>> warning message when
>>> the R console buffer was about to fill so that you could save all your
>>> output into a text file? (I know about sink(), but I think it would be
>>> good
>>> to have an easier mechanism to save a complete record of messages and
>>> function output). And on a similar vein, wouldn't it also be nice if R
>>> automatically entered a date stamp into the history file??
>>>
>>> Duncan
>>>
>>>
>>> *****************************************
>>> Dr. Duncan Mackay
>>> School of Biological Sciences
>>> Flinders University
>>> GPO Box 2100
>>> Adelaide
>>> S.A. 5001
>>> AUSTRALIA
>>>
>>> Ph (08) 8201 2627 FAX (08) 8201 3015
>>>
>>>
>> http://www.scieng.flinders.edu.au/biology/people/mackay_d/index.html
>>
>>
>> [[alternative HTML version deleted]]
>>
>> ______________________________________________
>> R-help@stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide!
>> http://www.R-project.org/posting-guide.html
>>
>>
>>
>>
>>
>> ----------------------------------------------------------------------------
>> --
>> Notice: This e-mail message, together with any attachments,...{{dropped}}
>>
>> ______________________________________________
>> R-help@stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide!
>> http://www.R-project.org/posting-guide.html
>>
>>
>>
>>
>
>
> --
>
> Kjetil Halvorsen.
>
> Peace is the most effective weapon of mass construction.
> -- Mahdi Elmandjra
>
>
>
>
>
> --
> No virus found in this outgoing message.
> Checked by AVG Anti-Virus.
>
> ______________________________________________
> R-help@stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
>

-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

______________________________________________
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Received on Fri Mar 04 17:52:32 2005

This archive was generated by hypermail 2.1.8 : Fri 03 Mar 2006 - 03:30:40 EST