[Rd] Problem with creating default NAMESPACE on FreeBSD

[Rainer Hurling]

On 22.07.2011 14:50 (UTC+1), Prof Brian Ripley wrote:
> One possibility is that you are building from a tarball snapshot without
> mentioning it, and the (unspecified) tarball is incomplete.
>
> For R-devel using tarballs is not a wise idea: they are really only
> checked carefully in pre-release periods. (One of the checks is that
> AFAIR CRAN builds from the tarballs, but CRAN is not manned at present.)
>
> But as far as I can the 'make dist' procedure does work now and did at
> some point in the last week.
>
> On Fri, 22 Jul 2011, Prof Brian Ripley wrote:
>
>> It does not seem to happen to anyone else, including me on FreeBSD. Is
>> this a completely clean install? src/library/datasets/Makefile
>> definitely installs a NAMESPACE file as part of the mkR1 target, so I
>> would check that you have a src/library/datasets/NAMESPACE file in
>> your sources.
>>
>> (There was a short window on the 16th when it happened as a commit for
>> src/library/datasets/NAMESPACE was missed at first.)

Many thanks for answering and sorry for the late reaction. We had some 
trouble in business yesterday.

It was not a totally clean installation before. I only deinstalled the 
old R version before installing the new one in the same place (make 
deinstall / make reinstall with a selfmade port math/R-devel for testing 
purposes). Other packages etc. had not been deleted before.

I forgot to mention that there were warnings in the build for several 
packages, like that:

installing to /usr/ports/math/R-devel/work/R-devel/library/mgcv/libs
** R
** data
** inst
** byte-compile and prepare package for lazy loading
** help
*** installing help indices
** building package indices ...
** testing if installed package can be loaded
During startup - Warning message:
package 'datasets' in options("defaultPackages") was not found

BUT:
Today I tried again with newest version r56478 and surprisingly, all 
that problems disappeared. Even no more warnings at build time.

For sureness I checked again older versions (between 15.07. and 21.07) 
but the build warnings there exist, with version r56478 they went away.

Rainer Hurling


P.S.: Thanks for the hint with the problems of the tarball. I had not 
been aware of it. In this case here it (mostly) also works with the 
tarball version.



>> On Fri, 22 Jul 2011, Rainer Hurling wrote:
>>
>>> After building and installing R-devel from 21/07/2011 on FreeBSD
>>> 9.0-CURRENT (amd64), starting R gives the following warning:
>>>
>>> [..snip..]
>>> Beim Start - Warnmeldung:
>>> package 'datasets' in options("defaultPackages") was not found
>>>
>>> This happens with all versions after 14/07/2011. I think it is
>>> related to
>>>
>>> NEWS Sat, 16 Jul 2011
>>> CHANGES IN R-devel PACKAGE INSTALLATION
>>> Packages without explicit ‘NAMESPACE’ files will have a default one
>>> created at build or INSTALL time, so all packages will have
>>> namespaces. A consequence of this is that ‘.First.lib’ functions need
>>> to be renamed, usually as ‘.onLoad’ but sometimes as ‘.onAttach’.
>>>
>>>
>>> Package 'datasets' seems to be installed correctly under
>>> /usr/local/lib/R/library/:
>>>
>>> ls -l datasets/
>>> total 18
>>> -rw-r--r-- 1 root wheel - 289 22 Jul 08:12:04 2011 DESCRIPTION
>>> -rw-r--r-- 1 root wheel - 5547 22 Jul 08:12:04 2011 INDEX
>>> drwxr-xr-x 2 root wheel - 512 22 Jul 08:12:04 2011 Meta
>>> drwxr-xr-x 2 root wheel - 512 22 Jul 08:12:04 2011 R
>>> drwxr-xr-x 2 root wheel - 512 22 Jul 08:12:04 2011 data
>>> drwxr-xr-x 2 root wheel - 512 22 Jul 08:12:04 2011 help
>>> drwxr-xr-x 2 root wheel - 512 22 Jul 08:12:04 2011 html
>>>
>>>
>>> Calling options("defaultPackages") shows all relevant packages:
>>>
>>> $defaultPackages
>>> [1] "datasets" "utils" "grDevices" "graphics" "stats" "methods"
>>>
>>>
>>> But trying to load library 'datasets' fails with the following message:
>>>
>>> library(datasets)
>>> Fehler in library(datasets) :
>>> package 'datasets' does not have a NAMESPACE and should be re-installed
>>>
>>>
>>> Googling around for some days now it looks like this does not happen
>>> to anyone else (at least there are no infos). Does anybody knows what
>>> is going on here?
>>>
>>> Any help would be appreciated,
>>> Rainer Hurling
>>>
>>>
>>>
>>> sessionInfo()
>>> R Under development (unstable) (2011-07-21 r56467)
>>> Platform: amd64-portbld-freebsd9.0 (64-bit)
>>>
>>> locale:
>>> [1]
>>> de_DE.ISO8859-15/de_DE.ISO8859-15/de_DE.ISO8859-15/C/de_DE.ISO8859-15/de_DE.ISO8859-15
>>>
>>>
>>> attached base packages:
>>> [1] stats graphics grDevices utils methods base