[Bioc-devel] build/check reports and the upcoming release: windows question

Dan Tenenbaum dtenenba at fhcrc.org
Wed Mar 28 18:29:23 CEST 2012


Hi Leo,

On Wed, Mar 28, 2012 at 7:32 AM, Dan Tenenbaum <dtenenba at fhcrc.org> wrote:
> On Wed, Mar 28, 2012 at 6:38 AM, Leo Lahti <leo.lahti at iki.fi> wrote:
>> Dear BioC-devel,
>>
>> The NetResponse package builds in Linux & Mac with no problems.
>>
>> For Windows the build/check report mentions that the platform is "not
>> supported" (http://bioconductor.org/checkResults/2.10/bioc-LATEST/).
>>
>> This is due to inclusion of certain graphics tools that are apparently
>> not available in default Windows systems.
>>
>
> I think it's because multicore and doMC are not supported on Windows.
> Most of the functionality of these packages can be found in the
> 'parallel' package, however, which is part of R and is cross-platform.
>
>

Looks like you have done some work towards this. I removed 'doMC' from
DESCRIPTION and the package builds and checks fine on Windows. I do
see the following in check, which  is probably something you want to
investigate:

* checking R code for possible problems ... NOTE
detect.responses: no visible global function definition for '%dopar%'
detect.responses: no visible global function definition for 'foreach'
plot.data: no visible global function definition for 'melt'

I'm sorry I didn't notify you of this problem when it arose and
instead marked the package as unsupported on windows. I should have
contacted you; that's what I normally do in cases like this.

In any case, I have removed the "unsupported on windows" designation
so the package will make it into release on all three platforms.
Thanks,
Dan


>> As no errors/warnings are produced, however, is the package still
>> going to be accepted to the next BioC release, or should I try to
>> modify the package dependencies so that Windows will become supported
>> for the tests?
>>
>
> The package will make it into release, it just won't have a windows
> binary available unless multicore and doMC are dropped.
>
> Dan
>
>> best regards.
>> Leo
>>
>> _______________________________________________
>> Bioc-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel



More information about the Bioc-devel mailing list