[Bioc-devel] Build error on Moscato2 for mzR

[Laurent Gatto]

On 16 April 2016 09:42, Blattmann Peter wrote:

> Dear all, 
>
> I'm maintaining the SWATH2stats package and we just use the MSstats
> package in the vignette for showing an example workflow. As MSstats
> depends on MSnbase and this depends on mzR we receive an error to
> build our SWATH2stats package now.
>
> In light of the release schedule (where packages should pass R cmd
> build and check by April 15) how shall we proceed?  Is it no problem
> that our package is included in the current release if we have install
> and build errors only on moscato2 but it builds correctly on the other
> operating systems? Or should we at one point take out the example
> workflow from the vignette? Or will the problem in mzR be solved in
> time and we don't have to worry?

The problem has been fixed by Dan with help from Kasper and Jim and the
Rcpp team. Confirmation on Bioc build servers on Sunday.

Details here

 https://github.com/sneumann/mzR/issues/36

Best wishes,

Laurent

> Best
>
> Peter
>
> -----Original Message-----
> From: Bioc-devel [mailto:bioc-devel-bounces at r-project.org] On Behalf Of Steffen Neumann
> Sent: Thursday, April 07, 2016 11:03 AM
> To: Samuel Wieczorek; mailman, bioc-devel
> Subject: Re: [Bioc-devel] Build error on Moscato2 for mzR
>
> Hi Samuel,
>
> On Do, 2016-04-07 at 10:28 +0200, Samuel Wieczorek wrote:
> [...]
>> 
>> which needs the mzR package.
>> The latter package has build error since a few days and this 
>> propagates errors on the other packages.
> We're painfully aware of the build failure of mzR. I opened an issue for it onhttps://github.com/sneumann/mzR/issues/36
>
> I know one person who was able to build mzR on the new toolchainhttps://github.com/sneumann/mzR/issues/34#issuecomment-190231
> 559
>
> Unfortunately none of us is a Windows wizard, and I even failed to install a working development environment on a virtual machine. If you or someone else would help out with a patch for issue #36, that would be awesome. I am currently unsure if we need to modify the CFLAGS so that only one of the files contains the std symbols, or the LDFLAGS so that the linker realises that it can chose one of the symbols.
>
> Please head over to the github issue and continue discussion for fixes there, I can also test and upload any suggested patches to the SinglePackageBuilder.
>
> Thanks for any help,
> yours,
> Steffen


-- 
Laurent Gatto | @lgatt0
http://cpu.sysbiol.cam.ac.uk/
http://lgatto.github.io/