my taurus build of sage 4.5.1 has liblapack.so just fine... On Aug 2, 10:11 pm, Willem Jan Palenstijn <w...@usecode.org> wrote: > On Tue, Aug 03, 2010 at 05:23:17AM +1000, Peter Jeremy wrote: > > On 2010-Aug-02 06:58:10 -0700, Dima Pasechnik <dimp...@gmail.com> wrote: > > >install the cvxopt-1.1.2.spkg from the link I posted on #6456, > > >run spkg-check, first without the 1-line change below, then with) > > > >diff -r 117baef5ef34 patches/setup.py > > >--- a/patches/setup.py Sun Aug 01 11:48:42 2010 -0700 > > >+++ b/patches/setup.py Mon Aug 02 09:53:17 2010 -0400 > > >@@ -13,7 +13,7 @@ > > > libraries = > > >['m','lapack','gsl','gslcblas','blas','cblas','atlas'] > > > else: > > > # libraries = ['m','lapack','blas','cblas','atlas','gfortran'] > > >- libraries = > > >['m','lapack','gsl','gslcblas','blas','cblas','gfortran','atlas'] > > >+ libraries = > > >['m','lapack','gsl','gslcblas','blas','cblas','gfortran','atlas','gcc_s'] > > > I would not recommend this fix as it will be needed wherever liblapack > > is used (and will not work if there are any cases where liblapack is > > dlopen()d. I believe the better fix is to add a dependency on libgcc_s > > to liblapack.so as per my fix in #9600. > > I'm trying to reproduce this, but on skynet's taurus I don't even get a > liblapack.so, but only liblapack.a. (The build isn't done yet, but the lapack > package is done.) > > Am I missing something? I checked, but none of my recent sage builds on > multiple machines seem to have liblapack.so. > > -Willem Jan
-- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
