[Debichem-devel] Bug#649381: mpqc: Please enable mp2r12

[Michael Banck]

Hi,

On Mon, Nov 21, 2011 at 08:38:30AM +0100, Daniel Leidert wrote:
> Am Montag, den 21.11.2011, 01:46 +0100 schrieb Michael Banck:
> > On Sun, Nov 20, 2011 at 03:26:08PM +0100, Michael Banck wrote:
> > > This is because libSCcints and (at least) libSCmbptr12 are not shipped
> > > in libsc7.
> > 
> > Turns out those are only enabled when libint is present:
> > 
> > |AC_CHECK_HEADERS(libint/libint.h,
> > |  AC_CHECK_LIB(int,init_libint_base,
> > |    HAVE_LIBINT=yes
> > |    AC_DEFINE(HAVE_LIBINT)
> > |    AC_SUBST(HAVE_LIBINT)
> > |    LIBS="-lint $LIBS"
> > |    )
> > |  )
> > [...]
> > |if test ! X$HAVE_LIBR12 = Xyes -o ! X$HAVE_LIBINT = Xyes; then
> > |  EXCLUDED_DIRS="-x SRC_LIB_CHEMISTRY_QC_CINTS -x SRC_LIB_CHEMISTRY_QC_MBPTR12 $EXCLUDED_DIRS"
> > |fi
> > 
> > We should really look into getting libint into Debian.
> 
> Well libint 1.x is not a big problem (I already started the basic
> packaging work some time ago). The configuration stuff is pretty ugly,
> but I guess, we can handle it. But the 1.x series is not longer
> supported. libint2 is the library to go, but it is (AFAIK) developed
> together with/for newer mpqc. I have doubts, if it is really useful to
> package libint1. IMO we should really check out, what is necessary to
> package current mpqc code and if this mpqc works with all the Debian
> packages currently relying on the last mpqc release.

Well, libint1 also enhances cp2k and (I think) at least abinit.  As I
said, I will contact the libint/mpqc author, but I doubt that a release
will happen in time for wheezy, and I am even less confident about a stable
libint2 library release.

So I think having libint1 as a stop-gap measure in case mpqc3 is way off
would be sensible.


Michael