Re: [Bioc-devel] GSL on Windows

2016-11-18 Thread Dan Tenenbaum
You need to make sure both i386 and x64 versions of GSL are installed under c:\GnuWin32, for a directory structure like this: c:\GnuWin32\i386 c:\GnuWin32\x64 (Also I would recommend using a more descriptive name than c:\GnuWin32, perhaps call it c:\GSL?) And then notice how the

[Bioc-devel] GSL on Windows

2016-11-18 Thread Dong Li
Hi Sorry about previous post on support.bioconductor.org. I am trying to use GSL on Windows to build my package. And I use src/Makevars.win as Martin's package https://github.com/Bioconductor-mirror/DirichletMultinomial. PKG_LIBS += -L$(LIB_GSL)/lib -lgsl -lgslcblas PKG_CPPFLAGS +=

Re: [Bioc-devel] vmatchPattern Returns Out of Bounds Indices

2016-11-18 Thread Hervé Pagès
On 11/18/2016 02:00 AM, Dario Strbenac wrote: Good day, These questions really belong to the support site. I suppose, although it seemed like an unexpected issue at first because it's not documented within ?lowlevel-matching so users don't know what to expect. You'll get that behaviour by

Re: [Bioc-devel] Coverage test badge not updating

2016-11-18 Thread Leonardo Collado Torres
Hi Valerie, Thanks! I do see the badges now but the percent reported seems off. Particularly: * http://bioconductor.org/packages/release/bioc/html/recount.html shows a 65% badge with a link to https://codecov.io/github/Bioconductor-mirror/recount?branch=release-3.4 which has a 97% coverage *

Re: [Bioc-devel] Coverage test badge not updating

2016-11-18 Thread Obenchain, Valerie
Coverage is now up and running for the release packages. I think all looks well... let me know if you see any problems. Valerie On 11/01/2016 03:14 PM, Hervé Pagès wrote: > Hi Leo, > > I can confirm that while Bioc 3.4 was in devel, package coverage was > computed only for the devel version of

Re: [Bioc-devel] vmatchPattern Returns Out of Bounds Indices

2016-11-18 Thread Dario Strbenac
Good day, > These questions really belong to the support site. I suppose, although it seemed like an unexpected issue at first because it's not documented within ?lowlevel-matching so users don't know what to expect. > You'll get that behaviour by allowing indels. This reveals a discrepancy