[Bioc-devel] Build failing for MutationalPatterns

[Martin Morgan]

On 11/30/2017 10:01 AM, Janssen-10, R.R.E. wrote:
> Dear Martin,
> 
> Martin writes:
> 
>> On 11/29/2017 11:14 AM, Janssen-10, R.R.E. wrote:
>>> Hi Martin,
>>>
>>> Thanks for your reply!
>>>
>>> Martin Morgan writes:
>>>
>>>> On 11/29/2017 09:58 AM, Janssen-10, R.R.E. wrote:
>>>>> Dear Bioconductor,
>>>>>
>>>>> The build for MutationalPatterns has been failing for a week.  The exact same code built fine, because the last commit ups the version number to 1.4.1, which is the version that is available on Bioconductor.
>>>>>
>>>>> I also cannot reproduce the error with the code from upstream on my computer. What can we do about this to fix the build on your infrastructure?
>>>>
>>>> I reproduced this on my linux laptop, installing the package and then
>>>> running
>>>>
>>>>     MutationalPatterns/vignettes master$ Rdev CMD Sweave --pdf
>>>> Introduction_to_MutationalPatterns.Rnw
>>>>
>>>> The key to reproducibility is using the same software as the build
>>>> system -- currently, R-devel and Bioc-devel with current packages
>>>>
>>>>     $ Rdev -e "BiocInstaller::biocVersion()" # 3.7
>>>>     $ Rdev -e "BiocInstaller::isDevel()"   # TRUE
>>>>     $ Rdev -e "BiocInstaller::biocValid()" # TRUE
>>>>
>>>> and to be sure that you're running in a clean checkout of your package,
>>>> e.g., by cloning into tmp
>>>>
>>>>     $ cd /tmp
>>>>     $ git clone https://git.bioconductor.org/packages/MutationalPatterns
>>>
>>>
>>> Apparently, I need to use a version of R that hasn't been released yet:
>>>
>>> $ R
>>> R version 3.4.2 (2017-09-28) -- "Short Summer"
>>> ...
>>>> library(BiocInstaller)
>>> Bioconductor version 3.6 (BiocInstaller 1.28.0), ?biocLite for help
>>>> useDevel()
>>> Error: 'devel' version requires a more recent R
>>>
>>> It seems that R version 3.4.3 is going to be released tomorrow, so I'll wait for that.
>>
>>> Thanks for verifying that this error can be reproduced.
>>
>> For the current Bioconductor cycle, we are using the 'devel' version of
>> R, see
>>
>>    http://bioconductor.org/developers/how-to/useDevel/
>>
>> The 'devel' version of R is always available, e.g., after clicking
>> through to your operating system of choice at
>> https://cran.r-project.org/bin/ and looking for something along the
>> lines of r-devel snapshot.
> 
> Thanks.
> 
> I have Ubuntu "xenial", for which I installed "r-base-dev" following the instructions here:
> https://cran.r-project.org/bin/linux/ubuntu/
> 
> sudo apt-get install r-base-dev


This is 'base R, but with development headers', what you want is 'r-devel'.

This

   https://cran.r-project.org/bin/linux/debian/#r-devel

seems to suggest that the way to get this is via svn checkout and local 
build

 
https://cran.r-project.org/bin/linux/debian/#installing-r-devel-or-a-release-branch-from-svn

Martin

> 
> This provides R 3.4.2, from which I cannot use the devel stuff.
> 
> I then upgraded GNU Guix's R to 3.4.3, but from there I also cannot use the devel stuff from Bioconductor.
> I attached the dependency graph for that R environment, including the dependencies for MutationalPatterns.
> 
> So, I tested this in a container with the attached dependency graph, and I still receive the same error:
> Error: 'devel' version requires a more recent R
> 
> On R 3.4.3 I cannot reproduce the issue, so I suspect it's how our package integrates with Bioconductor 3.7.
> Is there a complete guide to get the development version up and running?
> 
> Also, should packages in Bioconductor release 3.6 be built with the development version of R?
> 
> Thank you for your time.
> 
> Kind regards,
> Roel Janssen
> 
> ------------------------------------------------------------------------------
> 
> De informatie opgenomen in dit bericht kan vertrouwelijk zijn en is
> uitsluitend bestemd voor de geadresseerde. Indien u dit bericht onterecht
> ontvangt, wordt u verzocht de inhoud niet te gebruiken en de afzender direct
> te informeren door het bericht te retourneren. Het Universitair Medisch
> Centrum Utrecht is een publiekrechtelijke rechtspersoon in de zin van de W.H.W.
> (Wet Hoger Onderwijs en Wetenschappelijk Onderzoek) en staat geregistreerd bij
> de Kamer van Koophandel voor Midden-Nederland onder nr. 30244197.
> 
> Denk s.v.p aan het milieu voor u deze e-mail afdrukt.
> 
> ------------------------------------------------------------------------------
> 
> This message may contain confidential information and ...{{dropped:12}}