[R-SIG-Mac] cannit get list of binary packages from the GUI

[Simon Urbanek]

David,

thanks, the mystery is finally solved. The list of package was overflowing one of the scripts that collects all packages so it produced an empty PACKAGE file due to "argument list too long". That should be now fixed and http://R.research.att.com/ mirror should be up to date now. I still don't know why it worked occasionally (like last time when I was investigating it) but the bottom line that it should be ok now.

Thanks,
Simon


On Mar 17, 2012, at 1:19 PM, David Winsemius wrote:

> 
> On Mar 17, 2012, at 10:32 AM, Prof Brian Ripley wrote:
> 
>> On 17/03/2012 11:27, Berend Hasselman wrote:
>>> 
>>> On 17-03-2012, at 12:06, David Winsemius wrote:
>>> 
>>>> 
>>>> On Mar 16, 2012, at 7:21 PM, Simon Urbanek wrote:
>>>> 
>>>>> 
>>>>> On Mar 16, 2012, at 8:55 AM, Prof Brian Ripley wrote:
>>>>> 
>>>>>> On 16/03/2012 12:31, Ken knoblauch wrote:
>>>>>>> Federico Calboli<f.calboli<at>   imperial.ac.uk>   writes:
>>>>>>>> Apologies, I forgot to say, I have R 2.14.2, latest 64-bit GUI
>>>>>>> from http://r.research.att.com/, OS 10.7.3
>>>>>>>>> I just noticed the following weird issue.  If I go to the menu
>>>>>>> 'Packages and Data',  select 'Package
>>>>>>>> Installer' the installer goes up, but when I tell it to give the
>>>>>>> list of binaries available (I like to keep up
>>>>>>>> to date, and check every couple of days) the list comes up empty.
>>>>>>> The list for CRAN sources, and for
>>>>>>>> Bioconductor binaries work fine.
>>>>>>>>> 
>>>>>>>>> I tried to change the mirror, but it does not make a difference.
>>>>>>>>> 
>>>>>>>>> Am I the only one with the problem? what's going on?
>>>>>>> 
>>>>>>> I second this observation.
>>>>>> 
>>>>>> What is going on is that http://cran.r-project.org/bin/macosx/leopard/contrib/2.14/ has an empty PACKAGES file.  Given that it hadn't been updated for a couple of weeks on Weds, you may just need to be patient.  But in any case, only Simon can fix this as only he has sufficient access to the package builder. So reporting to Simon is always going to be the most effective way forward for similar problems.
>>>>>> 
>>>>> 
>>>>> Hmm... it's not empty on the master machine -- use http://r.research.att.com as the mirror - it should work ...
>>>>> 
>>>> 
>>>> Could PACKAGES file be in some damaged state that R cannot process? When I change the repository to that value  in the R64 GUI and try to see either either the binary versions or the source versions of contributed packages in the Package Installer I get these warnings at the console:
>>>> 
>>>> Warning: unable to access index for repository  http://r.research.att.com/bin/macosx/leopard/contrib/2.14
>>>> 
>>>> Warning: unable to access index for repository  http://r.research.att.com/src/contrib
>>>> 
>>>> I am able to see the .tgz packages at http://r.research.att.com/bin/macosx/leopard/contrib/2.14/ with Firefox.
>>>> 
>>>> I am able to get a list of source packages (but not binary) to display in the Package Installer if I use a different repository. And I am able to install binary packages with ... for example:
>>>> 
>>>> install.packages("actuar", repo="http://r.research.att.com", type="mac.binary.leopard")
>>>> 
>>>> (Which updated an existing version to the current version.)
>>>> 
>>>> Admittedly I am running a somewhat behind the times version, but this is completely different behavior than it has exhibited up to a few days ago.
>>>> 
>>> 
>>> 
>>> I updated binary packages  this morning (17 March 2012  approx. 9:00 CET) without any problem at all from http://cran.r-project.org. Yesterday no update appeared possible (Get List couldn't retrieve the list of packages ).
>> 
>> Yes, I updated PACKAGES[.gz] on the CRAN master this morning.  There is clearly a problem mirroring to CRAN and I have alerted the relevant people.  Hopefully mirroring will not wipe out the files I copied over without putting later ones in their place.
> 
> In case anyone actually needs to use the Package Installer while this get investigated, I see that the CRAN repo at "the Hutch" still is in a state where it will populate the binaries list.
> 
> http://cran.fhcrc.org
> 
> Neither the AT&T, the CMU, nor the ETH CRAN repos will do so as of 1PM US Eastern time.
> 
> -- 
> David
>> 
>> 
>>> Using R 2.14.2.
>>> 
>> 
>> 
>>> Berend
>>> 
>>> R version 2.14.2 (2012-02-29)
>>> Platform: i386-apple-darwin9.8.0/i386 (32-bit)
>>> 
>>> locale:
>>> [1] en_GB.UTF-8/en_GB.UTF-8/en_GB.UTF-8/C/en_GB.UTF-8/en_GB.UTF-8
>>> 
>>> attached base packages:
>>> [1] stats     graphics  grDevices utils     datasets  methods   base
>>> _______________________________________________
>>> R-SIG-Mac mailing list
>>> R-SIG-Mac at r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>> 
>> 
>> -- 
>> Brian D. Ripley,                  ripley at 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
> 
> David Winsemius, MD
> West Hartford, CT
> 
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
> 
>