[Scip] 3.1.1 build problem

[Ambros Gleixner]

Hi Victor,

another solution would be to compile with the old SoPlex 1.x interface, 
which works by adding

    LEGACY=true SPX_LEGACY=true

to the make options (of SoPlex and SCIP).  This deactivates rational 
arithmetic in SoPlex which is needed for the iterative refinement 
procedure to work effectively.  If you only use SCIP, however, it does 
not make a difference.

Kind regards,

ambros



On 05.03.2014 21:52, Matthias Miltenberger wrote:
> Hi Victor,
>
> in the new SoPlex version we are using a quite new feature of c++
> dealing with explicit conversions between double and rational types.
> With the latest gcc you should have no problems compiling.
> Sorry about the trouble.
>
> Best
> Matthias
>
> On March 5, 2014 9:00:11 PM CET, Victor Miller <victorsmiller at gmail.com>
> wrote:
>
>     I have gotten things to work by doing make COMP=intel .  I don't
>     know why gcc was so finicky.
>
>     Victor
>
>
>     On Wed, Mar 5, 2014 at 2:45 PM, Victor Miller
>     <victorsmiller at gmail.com <mailto:victorsmiller at gmail.com>> wrote:
>
>         I just tried to build scipoptsuite 3.1.1 using gcc 4.4.7 on a
>         redhat system.    It terminated with the errors:
>
>         -> compiling obj/O.linux.x86_64.gnu.opt/lib/changesoplex.o
>         In file included from src/basevectors.h:24,
>                                      from src/spxlpbase.h:28,
>                                      from src/spxlp.h.:24,
>                                      from src/spxsolver.h:29,
>                                      from src/changesoplex.cpp:22:
>         src/rational.h:103: error: only declarations of constructors can
>         be 'explicit'
>         src/rational.h:106: error: only declarations of constructors can
>         be 'explicit'
>
>
>         Suggestions?
>
>         Victor
>
>
>     ------------------------------------------------------------------------
>
>     Scip mailing list
>     Scip at zib.de
>     http://listserv.zib.de/mailman/listinfo/scip
>
>
> --
> Sent from mobile device - please excuse brevity
> Matthias Miltenberger
> www.zib.de/miltenberger <http://www.zib.de/miltenberger>
>
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip
>

-- 
____________________________________________________________
Ambros M. Gleixner
Zuse Institute Berlin - Matheon - Berlin Mathematical School
http://www.zib.de/gleixner