[Clp] OsiClpSolverInterface::computeLargestAway sets LogLevel
Sebastian Niemann
sebastian.niemann at inform-software.com
Tue Oct 18 10:45:48 EDT 2011
Hi!
I think we are facing the same problem. From some point on our virtual
user print method is not called anymore. Debugging showed that this is
due to the call "temp.setLogLevel(0)" in
"OsiClpSolverInterface::computeLargestAway()".
Although temp itself is a copy, the CoinMessageHandler member handler_
is *not* copied in case a custom handler is used (see beginning of
method "ClpModel::gutsOfCopy"). The log level of "temp" (which - due to
the shallow copy - is the log level of our custom handler) is never
reset which then leads to silence...
Cheers,
Sebastian
On 07.10.2011 18:23, John Forrest wrote:
> I don't really understand what problem is.
>
> temp is a separate copy which will be deleted on exit - logLevel in
> modelPtr will not be changed (or have I misunderstood).
>
> John Forrest
>
> On Thu, 2011-10-06 at 13:59 +0300, Jan-Willem Goossens wrote:
>> It seems to me that the computeLargestAway() method in
>> OsiClpSolverInterface sets the LogLevel to a 'temporary' value, but
>> doesnt set it back:
>>
>> OsiClpSolverInterface::computeLargestAway()
>> {
>> // get largest scaled away from bound
>> ClpSimplex temp=*modelPtr_;
>> temp.setLogLevel(0);
>>
>>
>> I guess the original value should be saved, and then re-set?
>> Regards,
>>
>> Jan-Willem
>> _______________________________________________
>> Clp mailing list
>> Clp at list.coin-or.org
>> http://list.coin-or.org/mailman/listinfo/clp
>>
>
>
> _______________________________________________
> Clp mailing list
> Clp at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/clp
--
Sebastian Niemann
Airport Systems Division
INFORM GmbH, Pascalstr. 23, 52076 Aachen, Germany
Telefon: (+49) 2408 9456-3691
E-Mail: sebastian.niemann at inform-software.com
Web: http://www.inform-software.com
INFORM Institut fur Operations Research und Management GmbH
Registered AmtsG Aachen HRB1144 Gfhr. Adrian Weiler
More information about the Clp
mailing list