R help - Jul 2006 - --no-save and --save toggle from inside R? + BATCH stderr

dear R wizards:  is it possible to instruct R to save or no-save from
inside R?  or does this have to be given at invokation on the
command-line?    The same question applies to "--no-restore-data",
although this presumably would have to be decided in a .First()
function or something like it.

on a similar note, I would love a CMD BATCH invokation to output just
one line to stderr at the end of its run which tells me whether the
batch job ended prematurely because of an error.  I know this can be
undesirable if run in another program that refuses output (e.g., a web
browser, where this can cause havoc if issued before the
content-type:), so it would require that some switch like --silent
would suppress this one line summary/error.  This is not urgent---I
can use the R return code to write a wrapper for this facility, but
this might be nice default behavior.  My guess is that most of us
won't mind to see a flag in writing if something went badly wrong.

sincerely,

/iaw

ivo welch wrote:
> dear R wizards:  is it possible to instruct R to save or no-save from
> inside R?  
You can specify the action when you call q() or quit(), but you can't 
change the default action.
> or does this have to be given at invokation on the
> command-line?    The same question applies to
"--no-restore-data",
> although this presumably would have to be decided in a .First()
> function or something like it.
>   
This one can't currently be changed.

Duncan Murdoch
> on a similar note, I would love a CMD BATCH invokation to output just
> one line to stderr at the end of its run which tells me whether the
> batch job ended prematurely because of an error.  I know this can be
> undesirable if run in another program that refuses output (e.g., a web
> browser, where this can cause havoc if issued before the
> content-type:), so it would require that some switch like --silent
> would suppress this one line summary/error.  This is not urgent---I
> can use the R return code to write a wrapper for this facility, but
> this might be nice default behavior.  My guess is that most of us
> won't mind to see a flag in writing if something went badly wrong.
>
> sincerely,
>
> /iaw
>
> ______________________________________________
> R-help at 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
>

thank you.  exactly what I needed.

can I add it as a suggestion for the next R version to add
an option to quit that prints only one line to stderr (or stdout, but
the real one, not the batch output one) that gives a 1-line job
summary---"no error" (or nothing), "5 warnings", "5
warnings, fatal
error: blah".  except in rare circumstances, this is likely to be the
user-preferred behavior.

(taking a hand where I have gotten a finger---can I redefine a quit
function that can invoke the original quit() with an option?)

regards,

/iaw

the following diff to the shell script invoking R prints an error
message if R terminates with an error code:

112c112,119
<       exec sh "${R_HOME}/bin/Rcmd" "${@}" ;;
---
>       sh "${R_HOME}/bin/Rcmd" "${@}"
>       rc="$?"
>       if [ "$rc" -ne 0 ]; then
>          shift ;
>          echo "${@}: Error Return Code: $rc"
>       fi
>       exit $?
>       ;;
regards,

/iaw