[Bioc-devel] r+w permissions in release branches

[Alejandro Reyes]

Hi Andrej,

Yes, that would help, that would be also a solution for my case, 
installing an old version of the cran package (stamod in my case)

However, I don't know if this could be a "general solution for all 
users" since when installing a package via
biocLite,  the latest version of the cran package is installed 
regardless the R/BiocInstaller version you are using.  Users
would need to download the versions of the dependencies that they need 
and install them manually.


Alejandro



> Dear Kasper,
>
> regarding your issue with R-2.15: I was wondering whether using an
> older version of Rcpp from
> http://cran.r-project.org/src/contrib/Archive/Rcpp/ would help?
>
> Cheers,
> Andrzej
>
> On Tue, Apr 22, 2014 at 2:46 PM, Kasper Daniel Hansen
> <kasperdanielhansen at gmail.com> wrote:
>> This is because commits to this branch of Bioconductor has been disabled
>> and it is intentional.
>>
>> But it raises the larger question, recently touched upon in a lengthy
>> thread on R-devel, on whether this is a good state of affairs for
>> Bioconductor.  Specifically the issue has to do with dependency of a
>> Bioconductor package on a CRAN package and what happens when CRAN packages
>> gets updated in a way that breaks backwards compability.  Right now, we
>> (Bioconductor) might get hosed.
>>
>> For example, we recently deployed a new computing cluster here at Hopkins.
>>   I maintain our R installation and some users have asked for an install of
>> Bioconductor using the latest version of R-2.15, for reproducibility
>> reasons.  I have a number of scripts which installs a standard suite of
>> packages we use here.  The issue I am facing is that Rcpp has been updated
>> and does not seem to be available for this version of R.  This indirectly
>> breaks crlmm, lumi, minfi, charm, methylumi, bead array,
>> arrayQualityMetrics to mention but a few we use on our end.  This seems
>> somewhat undesirable from a reproducibility perspective - I cannot even
>> install the packages!
>>
>> Best,
>> Kasper
>>
>>
>>
>>
>>
>> On Tue, Apr 22, 2014 at 2:19 PM, Alejandro Reyes <alejandro.reyes at embl.de>wrote:
>>
>>> Dear Dan, Dear developers list,
>>>
>>> Due a recent change in one cran package, DEXSeq 1.8.0 (for the R version
>>> 3.0.*) stop working. I fixed this conflict in the release branch of
>>> bioconductor and tried to commit my changes. But I don't seem to have write
>>> access, e.g:
>>>
>>> $ svn ci --username a.reyes -m "fixed conflicts with newest version of
>>> cran package"
>>> Sending        DESCRIPTION
>>> svn: Commit failed (details follow):
>>> svn: access to '/bioconductor/!svn/ver/81643/branches/RELEASE_2_13/madman/Rpacks/DEXSeq/DESCRIPTION'
>>> forbidden
>>>
>>> I  also noticed that I also don't have read access...
>>>
>>> svn co --username a.reyes https://hedgehog.fhcrc.org/
>>> bioconductor/branches/RELEASE_2_13/madman/Rpacks/DEXSeq
>>> svn: access to 'https://hedgehog.fhcrc.org/bioconductor/branches/RELEASE_
>>> 2_13/madman/Rpacks/DEXSeq' forbidden
>>>
>>> I was wondering if this intentional? If so, what would be the way to solve
>>> this kind of problems (e.g. a dependency changing outside bioconductor that
>>> breaks previous versions of a bioconductor package)?
>>>
>>> Best regards,
>>> Alejandro
>>>
>>> _______________________________________________
>>> Bioc-devel at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>
>>          [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> Bioc-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel