Re: [eclipse-clp-users] 'Problem is infeasible!' in eplex goes to stdout

From: Sergii Dymchenko <kit1980_at_...6...>
Date: Wed, 20 Aug 2014 00:58:29 -0700
Hi,

I've investigated the issue further and found that the message appears not
for all infeasible instances, but only for some of them. The issue present
in the version 6.1 #191 and not present in the version 6.1 #164 - so it was
introduced sometime in between.

Please see the file attached for the demo.

Sergii.


On Sat, Aug 16, 2014 at 7:20 PM, Sergii Dymchenko <kit1980_at_...6...> wrote:

> Hi Kish,
>
> Yes, I use CLP/CBC bundled with ECLiPSe. I'll provide code that
> demonstrates the issue on Monday.
>
> I agree that suppressing all warning_output is a bad idea. I used to
> do that, but now I usually just redirect warning_output and log_output
> to stderr because I use ECLiPSe in batch mode and so it matters what
> goes to stdout. BTW, shouldn't warning_output and log_output by
> defaulr go to stderr, not stdout?
>
> Sergii.
>
> On 8/16/14, Kish Shen <kisshen_at_...5...> wrote:
> > Hi Sergii,
> >
> > On 16/08/2014 00:54, Sergii Dymchenko wrote:> Hi,
> >  >
> >  > I've downloaded ECLiPSe version 6.1 #191 and now my programs that
> > uses eplex
> >  > often output 'Problem is infeasible!' to stdout.
> >  >
> >
> > You did not say which solver you are using with eplex, but I assume you
> > are using CLP/CBC. Is this correct?
> >
> >> Could you make  'Problem is infeasible!' message go to warning stream so
> >> it
> >> can be silenced with `:- set_stream(warning_output, null).`?
> >
> > No, ECLiPSe does not control where the solver send its messages, but the
> > solver should not be sending anything directly to stdout, as its
> > messages should be sent to the correct channel depending on the message
> > type (error, warning, result, log). Could you please provide a simple
> > example that produce this message (and could you also confirm the solver
> > you are using), so that I can report the issue to the correct people
> > (the developers of CLP/CBC, if this is the solver you are using).
> >
> > By the way, it is a bad idea to suppress all warning_output output, as
> > this is an ECLiPSe stream, and all ECLiPSe warnings are sent there, and
> > ignoring warnings sent by ECLiPSe is generally a bad idea. If you want
> > to suppress warnings from the solver, you should redirect the eplex
> > warning_channel instead --
> >
> > :- lp_set(warning_channel, -warning_output).
> >
> > (see the documentation for lp_set/2 for more detail).
> >
> > Cheers,
> >
> > Kish
> >
> >
>



Received on Wed Aug 20 2014 - 07:58:36 CEST

This archive was generated by hypermail 2.3.0 : Tue Apr 16 2024 - 09:13:20 CEST