[Bioc-devel] Biobase 1.15.33 depends on ALL problem

[Martin Morgan]

Thanks Keith for the report.

You've entered a bit of a twilight zone, and the sailing will be a bit
rough.

The underlying problem involves changes in R, not Biobase. I don't
know the details, but a series of changes between R r42739 and 42758
use a different way of representing package versions. It looks like
your R lies in the middle of this. I am speculating that the unusual
errors about package version mismatches relate to this; these problems
do not occur on more recent R on linux; there might be
windows-specific issues, which I will investigate later today.

Usually, the solution would be 'update R'. You'll want to be careful,
though, to update R to R-alpha, rather than R-devel, as R-devel has
branched in preparation for the release in October. The problem is
that R-alpha is available as *source*; if you are used to downloading
the binary version of R-devel, then you'll end up with R-2.7.0dev
(which we do not intend to be able to support with biocLite for at
least a couple of weeks).

Martin

Keith Satterley <keith at wehi.EDU.AU> writes:

> In addition to what I wrote previously, my immediate problem is this error:
>
>  > library("affy")
> Error: package 'Biobase' 1.15.33 was found, but >= c(1, 13, 41) is required by 
> 'affy'
>
> affy has a Depends...Biobase (>= 1.13.41) in its DESCRIPTION file
>
> I'm assuming this is a problem not unique to my installation of R
>
> cheers,
> Keith
>
>
> Keith Satterley wrote:
>> When I used biocLite("Biobase") I couldn't get version 1.15.33, so I downloaded 
>> the source and I was doing and "R CMD check Biobase" on the source code. I 
>> installed the latest Rtools.
>> 
>> Initially I got the warnings:
>> 
>> Found the following sources/headers with CR or CRLF line endings:
>>    src/envir.c
>>    src/Rinit.c
>> 
>> which I fixed to avoid these warnings.
>> 
>> There were also 6 warnings under:
>>   Package vignettes without corresponding PDF:
>> relating to .Rnw files in Biobase/inst/doc.
>> 
>> However, what are these errors reported in 00check.log:
>> 
>> *** Weave Errors ***
>> File 
>> W:/aaa-R/dev/R-2.6.0dev/Biobase/compile_version/ver-1.15.31/Biobase.Rcheck/inst/doc/Qviews.Rnw 
>> :
>> Error :  chunk 1 (label=doa)
>> Error : package 'Biobase' 1.15.33 is loaded, but >= c(1, 12, 0) is required by 'ALL'
>> *** Source Errors ***
>> File 
>> W:/aaa-R/dev/R-2.6.0dev/Biobase/compile_version/ver-1.15.31/Biobase.Rcheck/inst/doc/Qviews.R 
>> :
>> Error : package 'Biobase' 1.15.33 is loaded, but >= c(1, 12, 0) is required by 'ALL'
>> 
>> 
>> I'm not very familiar with this sort of code, but on looking at the ALL package 
>> which has "Depends: Biobase (>= 1.12.0)" in its DESCRIPTION file and on looking 
>> at the svn log of Biobase I'm wondering if it is related to a change on Revision 
>> 26853, which says:
>> 
>> Accomodate changes in base::package_version
>> 
>> - use numeric_version rather than package_version
>> 
>> - internal .as.numeric_version
>> 
>> - stronger checks on valid version comparisons
>> 
>> Need I worry about this? It installs OK and loads OK when I run R.
>>  > sessionInfo()
>> R version 2.6.0 Under development (unstable) (2007-09-02 r42744)
>> i386-pc-mingw32
>> 
>> locale:
>> LC_COLLATE=English_Australia.1252;LC_CTYPE=English_Australia.1252;LC_MONETARY=English_Australia.1252;LC_NUMERIC=C;LC_TIME=English_Australia.1252
>> 
>> attached base packages:
>> [1] tools     stats     graphics  grDevices datasets  utils     methods   base
>> 
>> other attached packages:
>> [1] Biobase_1.15.33
>> 
>> cheers,
>> 
>> Keith
>> 
>> ========================
>> Keith Satterley
>> Bioinformatics Division
>> The Walter and Eliza Hall Institute of Medical Research
>> Parkville, Melbourne,
>> Victoria, Australia
>> 
>> _______________________________________________
>> Bioc-devel at stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>
> _______________________________________________
> Bioc-devel at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel

-- 
Martin Morgan
Bioconductor / Computational Biology
http://bioconductor.org