R help - Nov 2005 - "Warning message: package '...' was built under R version 2.3.0"

Hi,

While installing precompiled packages I often get warnings like the one in
the subject. I usually ignore them, but I still do not understand why
windows packages are build with unreleased versions of R. 
Is there some way to get packages build under R-2.2.0?
What are potential problems that can result from that version mismatch?

My system: winXP
R version: 2.2.0

Jarek Tuszynski 

	[[alternative HTML version deleted]]

Tuszynski, Jaroslaw W. wrote:
> Hi,
> 
> While installing precompiled packages I often get warnings like the one in
> the subject. I usually ignore them, but I still do not understand why
> windows packages are build with unreleased versions of R. 
> Is there some way to get packages build under R-2.2.0?
Well, those in the corresponding "2.2" repository on CRAN should all
be
build with R-2.2.0. If not, please tell me which one is wrong.

Where did you get the packages for R-devel from?

Uwe Ligges

> What are potential problems that can result from that version mismatch?
> 
> My system: winXP
> R version: 2.2.0
> 
> Jarek Tuszynski 
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> 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

Uwe,

I think I used following repositories: PA1, PA2 and NC, which are the
closest to me and I think I got the same results in all. The package I
downloaded was "coin" which downloaded other packages. The consol
printout
follows:

> chooseCRANmirror()
> utils:::menuInstallPkgs()
trying URL
'http://www.ibiblio.org/pub/languages/R/CRAN/bin/windows/contrib/2.2/coin_0.
3-3.zip'
Content type 'application/zip' length 793299 bytes
opened URL
downloaded 774Kb

package 'coin' successfully unpacked and MD5 sums checked

The downloaded packages are in
        C:\Documents and Settings\tuszynskij\Local
Settings\Temp\Rtmp7594\downloaded_packages
updating HTML package descriptions
> library(coin)
Loading required package: survival
Loading required package: splines
Loading required package: mvtnorm
Loading required package: modeltools
Error in load(dataFile, ns) : ReadItem: unknown type 241
In addition: Warning messages:
1: package 'mvtnorm' was built under R version 2.3.0 
2: package 'modeltools' was built under R version 2.3.0 
Error: unable to load R code in package 'modeltools'
Error: package 'modeltools' could not be loaded

Jarek Tuszynski

-----Original Message-----
From: ligges at statistik.uni-dortmund.de
[mailto:ligges at statistik.uni-dortmund.de] 
Sent: Wednesday, November 16, 2005 11:27 AM
To: Tuszynski, Jaroslaw W.
Cc: r-help at stat.math.ethz.ch
Subject: Re: [R] "Warning message: package '...' was built under R
version
2.3.0"


Tuszynski, Jaroslaw W. wrote:
> Hi,
> 
> While installing precompiled packages I often get warnings like the 
> one in the subject. I usually ignore them, but I still do not 
> understand why windows packages are build with unreleased versions of 
> R. Is there some way to get packages build under R-2.2.0?
Well, those in the corresponding "2.2" repository on CRAN should all
be
build with R-2.2.0. If not, please tell me which one is wrong.

Where did you get the packages for R-devel from?

Uwe Ligges

> What are potential problems that can result from that version 
> mismatch?
> 
> My system: winXP
> R version: 2.2.0
> 
> Jarek Tuszynski
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> 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

I do have single version of R (2.2.0) and all packages I have were installed
by downloading precompiled packages from PA or NC repositories within last
month. 

I also just managed to download 'modeltools' from PA2 and its
DESCRIPTION
says:
Built: R 2.3.0; ; 2005-11-04 20:47:47; windows

R console:

	> utils:::menuInstallPkgs()
	--- Please select a CRAN mirror for use in this session ---
	trying URL
'http://lib.stat.cmu.edu/R/CRAN/bin/windows/contrib/2.2/modeltools_0.2-0.zip
'
	Content type 'application/zip' length 64426 bytes
	opened URL
	downloaded 62Kb

	package 'modeltools' successfully unpacked and MD5 sums checked

	The downloaded packages are in
	        C:\Documents and Settings\tuszynskij\Local
Settings\Temp\Rtmp4458\downloaded_packages
	updating HTML package descriptions
	> library(modeltools)
	Error in load(dataFile, ns) : ReadItem: unknown type 241
	In addition: Warning message:
	package 'modeltools' was built under R version 2.3.0 
	Error: unable to load R code in package 'modeltools'
	Error: package/namespace load failed for 'modeltools'

The same package downloaded from PA1 does not give me any warnings or
errors.

Jarek Tuszynski


-----Original Message-----
From: ligges at statistik.uni-dortmund.de
[mailto:ligges at statistik.uni-dortmund.de] 
Sent: Wednesday, November 16, 2005 1:04 PM
To: Tuszynski, Jaroslaw W.
Cc: r-help at stat.math.ethz.ch
Subject: Re: [R] "Warning message: package '...' was built under R
version
2.3.0"


Tuszynski, Jaroslaw W. wrote:
> Uwe,
> 
> I think I used following repositories: PA1, PA2 and NC, which are the 
> closest to me and I think I got the same results in all. The package I 
> downloaded was "coin" which downloaded other packages. The consol
> printout
> follows:
> 
> 
> 
>>chooseCRANmirror()
>>utils:::menuInstallPkgs()
> 
> trying URL 
> 'http://www.ibiblio.org/pub/languages/R/CRAN/bin/windows/contrib/2.2/c
> oin_0.
> 3-3.zip'
> Content type 'application/zip' length 793299 bytes
> opened URL
> downloaded 774Kb
> 
> package 'coin' successfully unpacked and MD5 sums checked
> 
> The downloaded packages are in
>         C:\Documents and Settings\tuszynskij\Local 
> Settings\Temp\Rtmp7594\downloaded_packages
> updating HTML package descriptions
> 
>>library(coin)
> 
> Loading required package: survival
> Loading required package: splines
> Loading required package: mvtnorm
> Loading required package: modeltools
> Error in load(dataFile, ns) : ReadItem: unknown type 241
> In addition: Warning messages:
> 1: package 'mvtnorm' was built under R version 2.3.0

I have just downloaded
http://www.ibiblio.org/pub/languages/R/CRAN/bin/windows/contrib/2.2/mvtnorm_
0.7-2.zip
and its DESCRIPTION file correctly has:
"Built: R 2.2.0; i386-pc-mingw32; 2005-10-14 14:44:12; windows"


My guess is that you are using one library for two different R versions 
and you have installed some R-devel package into the library you are now 
using with R-2.2.0 ...

You cannot mix binary packages for R < 2.3.0 with those for R-devel due 
to changes in environment handling.

Uwe Ligges

>> Yes. The Statlib mirror (which I suppose is what you call PA2) appears
to
>> be badly messed up.
>>
>> -- 
>>    O__  ---- Peter Dalgaard             ??ster Farimagsgade 5, Entr.B
>>   c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
>>  (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45)
35327918
>> ~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  FAX: (+45)
35327907

Peter,

Thanks for checking on this problem. 

My names for CRAN mirror sites (PA 1, PA 2, NC, etc.) came from that is
shown in windows RGui when you choose "Packages/Set CRAN mirror" menu.
All
mirrors are represented by "Country (city)" for most of the world
except US
which is shown as "USA (state abbreviation + number)". So what I call
"PA 1"
means "USA/Pennsylvania #1", etc. Names you are using ("Statlib
mirror") are
not visible in GUI. Just a clarification.

Jarek Tuszynski

Tuszynski, Jaroslaw W. wrote:
> Hi,
> 
> While installing precompiled packages I often get warnings like the one in
> the subject. I usually ignore them, but I still do not understand why
> windows packages are build with unreleased versions of R. 
> Is there some way to get packages build under R-2.2.0?
I think this part of your question was determined to be a problem with a 
certain mirror (but I didn't read all the messages on this...)
> What are potential problems that can result from that version mismatch?
Newer versions of R may store things differently.  We try hard to make 
sure that newer versions can read data from older ones, but it's hard to 
work in the opposite direction.

For one important change, an environment of NULL in 2.2.x has been 
replaced with an environment of baseenv() in 2.3.x.  2.3.x can usually 
translate the NULLs, but 2.2.x won't know what a baseenv() object is.

Duncan Murdoch