Re: [Rd] R CMD BATCH vs R CMD batch

[Rainer M Krug]

Dirk Eddelbuettel  writes:

> On 28 October 2015 at 21:39, Marius Hofert wrote:
> | Out of laziness I just used "R CMD batch" instead of "R CMD BATCH". I
> | didn't get an error so didn't think about the consequences... One
> | consequence is (at least on Mac OS X 10.11 but probably in more
> | generality) that R_BATCH_OPTIONS are ignored, which was kind of fatal
> | in my case... I am thus wondering whether it makes sense to either a)
> | have R_BATCH_OPTIONS also be respected for "R CMD batch" or b) simply
> | not allow "R CMD batch" as a valid command (so requiring to use "R CMD
> | BATCH"). Both approaches might be delicate... just wanted to point
> | this issue out...
>
> Same reason we have 'R CMD INSTALL' as there often is /usr/bin/install with
> different options ...
>
> In general 'R CMD foo' will run for any 'foo' in the path:
>
>edd@max:~$ R CMD date
>Wed Oct 28 21:05:01 CDT 2015
>edd@max:~$ 

So what is R CMD exactly doing in this example? The output is the same if
I say only the command (using pwd as otherwise the time has changed...): 

,
| 09:35:03 ~$ R CMD pwd
| /Users/rainerkrug
| 09:35:37 ~$ R CMD pwd
| /Users/rainerkrug
| 09:37:44 ~$ pwd
| /Users/rainerkrug
| 09:37:49 ~$
`

And this happens, except in cases where the foo is defined as a CMD in R 
(build, ...):

,
| Commands:
|   BATCH   Run R in batch mode
|   COMPILE Compile files for use with R
|   SHLIB   Build shared library for dynamic loading
|   INSTALL Install add-on packages
|   REMOVE  Remove add-on packages
|   build   Build add-on packages
|   check   Check add-on packages
|   LINKFront-end for creating executable programs
|   Rprof   Post-process R profiling files
|   Rdconv  Convert Rd format to various other formats
|   Rd2pdf  Convert Rd format to PDF
|   Rd2txt  Convert Rd format to pretty text
|   Stangle Extract S/R code from Sweave documentation
|   Sweave  Process Sweave documentation
|   Rdiff   Diff R output ignoring headers etc
|   config  Obtain configuration information about R
|   javareconf  Update the Java configuration variables
|   rtags Create Emacs-style tag files from C, R, and Rd files
`

Unless I miss something, is this is an inconsistency in R?

Rainer


>
> Dirk

-- 
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation Biology, 
UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Stellenbosch University
South Africa

Tel :   +33 - (0)9 53 10 27 44
Cell:   +33 - (0)6 85 62 59 98
Fax :   +33 - (0)9 58 10 27 44

Fax (D):+49 - (0)3 21 21 25 22 44

email:  rai...@krugs.de

Skype:  RMkrug

PGP: 0x0F52F982


signature.asc
Description: PGP signature
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Rd] R CMD BATCH vs R CMD batch

[Rainer M Krug]

Deepayan Sarkar  writes:

> On Thu, Oct 29, 2015 at 2:09 PM, Rainer M Krug  wrote:
>> Dirk Eddelbuettel  writes:
>>
>>> On 28 October 2015 at 21:39, Marius Hofert wrote:
>>> | Out of laziness I just used "R CMD batch" instead of "R CMD BATCH". I
>>> | didn't get an error so didn't think about the consequences... One
>>> | consequence is (at least on Mac OS X 10.11 but probably in more
>>> | generality) that R_BATCH_OPTIONS are ignored, which was kind of fatal
>>> | in my case... I am thus wondering whether it makes sense to either a)
>>> | have R_BATCH_OPTIONS also be respected for "R CMD batch" or b) simply
>>> | not allow "R CMD batch" as a valid command (so requiring to use "R CMD
>>> | BATCH"). Both approaches might be delicate... just wanted to point
>>> | this issue out...
>>>
>>> Same reason we have 'R CMD INSTALL' as there often is /usr/bin/install with
>>> different options ...
>>>
>>> In general 'R CMD foo' will run for any 'foo' in the path:
>>>
>>>edd@max:~$ R CMD date
>>>Wed Oct 28 21:05:01 CDT 2015
>>>edd@max:~$
>>
>> So what is R CMD exactly doing in this example? The output is the same if
>> I say only the command (using pwd as otherwise the time has changed...):
>>
>> ,
>> | 09:35:03 ~$ R CMD pwd
>> | /Users/rainerkrug
>> | 09:35:37 ~$ R CMD pwd
>> | /Users/rainerkrug
>> | 09:37:44 ~$ pwd
>> | /Users/rainerkrug
>> | 09:37:49 ~$
>> `
>>
>> And this happens, except in cases where the foo is defined as a CMD in R 
>> (build, ...):
>>
>> ,
>> | Commands:
>> |   BATCH   Run R in batch mode
>> |   COMPILE Compile files for use with R
>> |   SHLIB   Build shared library for dynamic loading
>> |   INSTALL Install add-on packages
>> |   REMOVE  Remove add-on packages
>> |   build   Build add-on packages
>> |   check   Check add-on packages
>> |   LINKFront-end for creating executable programs
>> |   Rprof   Post-process R profiling files
>> |   Rdconv  Convert Rd format to various other formats
>> |   Rd2pdf  Convert Rd format to PDF
>> |   Rd2txt  Convert Rd format to pretty text
>> |   Stangle Extract S/R code from Sweave documentation
>> |   Sweave  Process Sweave documentation
>> |   Rdiff   Diff R output ignoring headers etc
>> |   config  Obtain configuration information about R
>> |   javareconf  Update the Java configuration variables
>> |   rtags Create Emacs-style tag files from C, R, and Rd 
>> files
>> `
>>
>> Unless I miss something, is this is an inconsistency in R?
>
> One important difference (not sure if the only one) is that R CMD
> defines more environment variables. Compare
>
> env | grep -i tex
> R CMD env | grep -i tex
>
> So for example
>
> R CMD pdflatex
>
> will behave differently from plain pdflatex.

Reading Dirk's email again, I think I get the picture of what ~R CMD foo~ is
doing: running ~foo~ after doing some initial magic.

Out of interest: What is the magic ~R CMD~ is doing? Is it documented
anywhere?

Rainer

>
> -Deepayan

-- 
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation Biology, 
UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Stellenbosch University
South Africa

Tel :   +33 - (0)9 53 10 27 44
Cell:   +33 - (0)6 85 62 59 98
Fax :   +33 - (0)9 58 10 27 44

Fax (D):+49 - (0)3 21 21 25 22 44

email:  rai...@krugs.de

Skype:  RMkrug

PGP: 0x0F52F982


signature.asc
Description: PGP signature
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Rd] R CMD BATCH vs R CMD batch

[Deepayan Sarkar]

On Thu, Oct 29, 2015 at 2:09 PM, Rainer M Krug  wrote:
> Dirk Eddelbuettel  writes:
>
>> On 28 October 2015 at 21:39, Marius Hofert wrote:
>> | Out of laziness I just used "R CMD batch" instead of "R CMD BATCH". I
>> | didn't get an error so didn't think about the consequences... One
>> | consequence is (at least on Mac OS X 10.11 but probably in more
>> | generality) that R_BATCH_OPTIONS are ignored, which was kind of fatal
>> | in my case... I am thus wondering whether it makes sense to either a)
>> | have R_BATCH_OPTIONS also be respected for "R CMD batch" or b) simply
>> | not allow "R CMD batch" as a valid command (so requiring to use "R CMD
>> | BATCH"). Both approaches might be delicate... just wanted to point
>> | this issue out...
>>
>> Same reason we have 'R CMD INSTALL' as there often is /usr/bin/install with
>> different options ...
>>
>> In general 'R CMD foo' will run for any 'foo' in the path:
>>
>>edd@max:~$ R CMD date
>>Wed Oct 28 21:05:01 CDT 2015
>>edd@max:~$
>
> So what is R CMD exactly doing in this example? The output is the same if
> I say only the command (using pwd as otherwise the time has changed...):
>
> ,
> | 09:35:03 ~$ R CMD pwd
> | /Users/rainerkrug
> | 09:35:37 ~$ R CMD pwd
> | /Users/rainerkrug
> | 09:37:44 ~$ pwd
> | /Users/rainerkrug
> | 09:37:49 ~$
> `
>
> And this happens, except in cases where the foo is defined as a CMD in R 
> (build, ...):
>
> ,
> | Commands:
> |   BATCH   Run R in batch mode
> |   COMPILE Compile files for use with R
> |   SHLIB   Build shared library for dynamic loading
> |   INSTALL Install add-on packages
> |   REMOVE  Remove add-on packages
> |   build   Build add-on packages
> |   check   Check add-on packages
> |   LINKFront-end for creating executable programs
> |   Rprof   Post-process R profiling files
> |   Rdconv  Convert Rd format to various other formats
> |   Rd2pdf  Convert Rd format to PDF
> |   Rd2txt  Convert Rd format to pretty text
> |   Stangle Extract S/R code from Sweave documentation
> |   Sweave  Process Sweave documentation
> |   Rdiff   Diff R output ignoring headers etc
> |   config  Obtain configuration information about R
> |   javareconf  Update the Java configuration variables
> |   rtags Create Emacs-style tag files from C, R, and Rd files
> `
>
> Unless I miss something, is this is an inconsistency in R?

One important difference (not sure if the only one) is that R CMD
defines more environment variables. Compare

env | grep -i tex
R CMD env | grep -i tex

So for example

R CMD pdflatex

will behave differently from plain pdflatex.

-Deepayan

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Rd] R CMD BATCH vs R CMD batch

[peter dalgaard]

On 29 Oct 2015, at 10:44 , Rainer M Krug  wrote:

> Out of interest: What is the magic ~R CMD~ is doing? Is it documented
> anywhere?


R is open source (and shell scripts are considered self-documenting by some)

On Unix-alikes, R is a shell script which, if called with 1st argument CMD, 
calls ${R_HOME}/bin/Rcmd, which is another shell script that ends with 

case "${1}" in
## this was a separate command prior to 2.10.0
  Rd2txt)
cmd="${R_HOME}/bin/Rdconv"
extra="-t txt"
;;
## removed in 2.15.0
  Rd2dvi)
echo "R CMD Rd2dvi is defunct: use Rd2pdf instead"
exit 1
;;
  *)
if test -x "${R_HOME}/bin/${1}"; then
  cmd="${R_HOME}/bin/${1}"
else
  cmd="${1}"
fi
;;
esac
shift

exec "${cmd}" ${extra} "${@}"

I.e., except for setting up variables and such, R CMD ${foo} essentially looks 
for ${R_HOME}/bin/${foo} and executes it if it exists and is executable, 
otherwise it just executes ${foo}.

Notice that, somewhat contrary to what Dirk said, this logic works at the mercy 
of the file system when it comes to case sensitivity. On OSX with the default 
case-insensitive setup we get

Peters-iMac:IBP pd$ R CMD install
Error: ERROR: no packages specified

which comes from R's package installer, whereas a case-sensitive FS would pick 
up /usr/bin/install instead. That is just how the "test" shell built-in works: 

Peters-iMac:~ pd$ test -x `R RHOME`/bin/INSTALL ; echo $?
0
Peters-iMac:~ pd$ test -x `R RHOME`/bin/UNSTALL ; echo $?
1
Peters-iMac:~ pd$ test -x `R RHOME`/bin/install ; echo $?
0


-- 
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Office: A 4.23
Email: pd@cbs.dk  Priv: pda...@gmail.com

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Rd] R CMD BATCH vs R CMD batch

[Rainer M Krug]

peter dalgaard  writes:

> On 29 Oct 2015, at 10:44 , Rainer M Krug  wrote:
>
>> Out of interest: What is the magic ~R CMD~ is doing? Is it documented
>> anywhere?
>
>
> R is open source (and shell scripts are considered self-documenting by 
> some)

True - should have looked in the sources...

>
> On Unix-alikes, R is a shell script which, if called with 1st argument
> CMD, calls ${R_HOME}/bin/Rcmd, which is another shell script that ends
> with
>
> case "${1}" in
> ## this was a separate command prior to 2.10.0
>   Rd2txt)
> cmd="${R_HOME}/bin/Rdconv"
> extra="-t txt"
> ;;
> ## removed in 2.15.0
>   Rd2dvi)
> echo "R CMD Rd2dvi is defunct: use Rd2pdf instead"
> exit 1
> ;;
>   *)
> if test -x "${R_HOME}/bin/${1}"; then
>   cmd="${R_HOME}/bin/${1}"
> else
>   cmd="${1}"
> fi
> ;;
> esac
> shift
>
> exec "${cmd}" ${extra} "${@}"
>
> I.e., except for setting up variables and such, R CMD ${foo}
> essentially looks for ${R_HOME}/bin/${foo} and executes it if it
> exists and is executable, otherwise it just executes ${foo}.

Thanks for that detailed explanation of the mechanism - it is much
clearer now. And as usual, if one knows how something is happening, it's
not "magical" anymore. 

>
> Notice that, somewhat contrary to what Dirk said, this logic works at
> the mercy of the file system when it comes to case sensitivity. On OSX
> with the default case-insensitive setup we get
>
> Peters-iMac:IBP pd$ R CMD install
> Error: ERROR: no packages specified
>
> which comes from R's package installer, whereas a case-sensitive FS
> would pick up /usr/bin/install instead. That is just how the "test"
> shell built-in works:
>
> Peters-iMac:~ pd$ test -x `R RHOME`/bin/INSTALL ; echo $?
> 0
> Peters-iMac:~ pd$ test -x `R RHOME`/bin/UNSTALL ; echo $?
> 1
> Peters-iMac:~ pd$ test -x `R RHOME`/bin/install ; echo $?
> 0

Yeah - true. I was already starting to wonder as I always was using ~R
CMD install~ and it always worked on my Mac.

Thanks again,

Rainer

-- 
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation Biology, 
UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Stellenbosch University
South Africa

Tel :   +33 - (0)9 53 10 27 44
Cell:   +33 - (0)6 85 62 59 98
Fax :   +33 - (0)9 58 10 27 44

Fax (D):+49 - (0)3 21 21 25 22 44

email:  rai...@krugs.de

Skype:  RMkrug

PGP: 0x0F52F982


signature.asc
Description: PGP signature
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

[Rd] R CMD BATCH vs R CMD batch

[Marius Hofert]

Hi,

Out of laziness I just used "R CMD batch" instead of "R CMD BATCH". I
didn't get an error so didn't think about the consequences... One
consequence is (at least on Mac OS X 10.11 but probably in more
generality) that R_BATCH_OPTIONS are ignored, which was kind of fatal
in my case... I am thus wondering whether it makes sense to either a)
have R_BATCH_OPTIONS also be respected for "R CMD batch" or b) simply
not allow "R CMD batch" as a valid command (so requiring to use "R CMD
BATCH"). Both approaches might be delicate... just wanted to point
this issue out...

Cheers,
M

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Rd] R CMD BATCH vs R CMD batch

[Marius Hofert]

Great, thanks, Dirk. Nice example :-)


On Wed, Oct 28, 2015 at 10:05 PM, Dirk Eddelbuettel  wrote:
>
> On 28 October 2015 at 21:39, Marius Hofert wrote:
> | Out of laziness I just used "R CMD batch" instead of "R CMD BATCH". I
> | didn't get an error so didn't think about the consequences... One
> | consequence is (at least on Mac OS X 10.11 but probably in more
> | generality) that R_BATCH_OPTIONS are ignored, which was kind of fatal
> | in my case... I am thus wondering whether it makes sense to either a)
> | have R_BATCH_OPTIONS also be respected for "R CMD batch" or b) simply
> | not allow "R CMD batch" as a valid command (so requiring to use "R CMD
> | BATCH"). Both approaches might be delicate... just wanted to point
> | this issue out...
>
> Same reason we have 'R CMD INSTALL' as there often is /usr/bin/install with
> different options ...
>
> In general 'R CMD foo' will run for any 'foo' in the path:
>
>edd@max:~$ R CMD date
>Wed Oct 28 21:05:01 CDT 2015
>edd@max:~$
>
> Dirk
>
> --
> http://dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Rd] R CMD BATCH vs R CMD batch

[Dirk Eddelbuettel]

On 28 October 2015 at 21:39, Marius Hofert wrote:
| Out of laziness I just used "R CMD batch" instead of "R CMD BATCH". I
| didn't get an error so didn't think about the consequences... One
| consequence is (at least on Mac OS X 10.11 but probably in more
| generality) that R_BATCH_OPTIONS are ignored, which was kind of fatal
| in my case... I am thus wondering whether it makes sense to either a)
| have R_BATCH_OPTIONS also be respected for "R CMD batch" or b) simply
| not allow "R CMD batch" as a valid command (so requiring to use "R CMD
| BATCH"). Both approaches might be delicate... just wanted to point
| this issue out...

Same reason we have 'R CMD INSTALL' as there often is /usr/bin/install with
different options ...

In general 'R CMD foo' will run for any 'foo' in the path:

   edd@max:~$ R CMD date
   Wed Oct 28 21:05:01 CDT 2015
   edd@max:~$ 

Dirk

-- 
http://dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel