[BioC] error with get(getOption("device")) in maanova with R 2.8.0

[Wolfgang Huber]

Dear Rob

the package "maanova" passed "R CMD check" with R 2.8, and this is all 
that Bioconductor can and will test on community-contributed packages. 
(And believe me that for 294 packages, that is already lot to ask!)

Obviously, not all the functionality of a package is exhaustively tested 
in the tests that are invoked by "R CMD check" (man page examples, 
vignette, explicit tests). You might raise this with the package 
authors. But I would remark that it is unrealistic to expect that "R CMD 
check" will find all, possibly subtle, errors.

Best wishes
      Wolfgang

----------------------------------------------------
Wolfgang Huber, EMBL-EBI, http://www.ebi.ac.uk/huber


Rob Tempelman ha scritto:
> Dear Wolfgang,
> 
>   Indeed, there needs to be a fix.  What I don't understand is how come this
> sort of thing doesn't get diagnosed when maanova is included as part of 
> 
> "2008-10-22 BioC 2.3 consisting of 294 packages and designed to work with R
> 2.8.z, was released today"
> 
> as indicated from the main page of bioconductor.org?
> 
> I suppose I would need to go through all of functions myself and make my own
> fix.
> 
> 
> Rob
> 
> -----Original Message-----
> From: Wolfgang Huber [mailto:huber at ebi.ac.uk] 
> Sent: Wednesday, November 12, 2008 4:53 AM
> To: Rob Tempelman
> Cc: bioconductor at stat.math.ethz.ch
> Subject: Re: [BioC] error with get(getOption("device")) in maanova with R
> 2.8.0
> 
> Rob Tempelman ha scritto:
>> I downloaded the latest version of maanova (version 1.12.0) from
> bioconductor that is presumably designed to work with R 2.8.0.  However,
> upon using some of the functions (say gridcheck on the dataset kidney
> provided with the package), I get the following error message:
>>> data(kidney)
>>> gridcheck(kidney.raw)
>> Error in get(getOption("device")) : invalid first argument
>>
>>
>> It is my understanding that all of the get(getOption("device")) commands
> now need to be replaced by dev.new() in R 2.8.0.  If so, could this be done
> with this particular package?
>> Thanks,
>>
>> Rob Tempelman
> 
> 
> 
> Hi Rob,
> 
> just a guess, but you or the authors of maanova might find it 
> instructive to check the R 2.8 release notes "CHANGES IN R VERSION 
> 2.8.0": http://cran.at.r-project.org/src/base/NEW
> 
> 
> The 'factory-fresh' defaults for options("device") have been
> 	changed to refer to the devices as functions in the grDevices
> 	namespace and not as names.  This makes it more likely that
> 	the incorrect (since R 2.5.0) assumption in packages that
> 	get(getOption("device"))() will work will catch users of those
> 	packages.
> 
> 
> 
> Best wishes
>       Wolfgang
> 
> ----------------------------------------------------
> Wolfgang Huber, EMBL-EBI, http://www.ebi.ac.uk/huber


--