[Cbc] Segfault dealing with infeasible lp

[John Forrest]

Haroldo,

Unable to reproduce error.  When I get to line 550, nonLinearCost_ has 
just been used successfully so if now NULL there must be something 
seriously wrong.

What configure options and compiler?

John
On 09/01/17 20:34, Haroldo Gambini Santos wrote:
>
> Hi Forrest,
>
> It seems that it still crashes in revision 2322:
>
>
> Welcome to the CBC MILP Solver
> Version: 2.9
> Build Date: Jan  9 2017
> Revision Number: 2322
>
> command line - cbc infeasible.lp (default strategy 1)
> cbc: ClpSimplex.cpp:550: int ClpSimplex::gutsOfSolution(double*, const 
> double*, bool): Assertion `nonLinearCost_' failed.
> Abortado (imagem do núcleo gravada)
>
>
>
>
> Em 09/01/2017 16:21, John Forrest escreveu:
>> Had already made a change in local copy of trunk (for a different 
>> reason) which fixes error.
>>
>> Hopefully fixed in stable and trunk
>>
>> John Forrest
>> On 09/01/17 16:25, Luís Borges de Oliveira wrote:
>>> Hello,
>>>
>>> We've come across an infeasible lp that causes CBC to segfault:
>>>
>>>     Welcome to the CBC MILP Solver
>>>     Version: 2.9
>>>     Build Date: Jan  5 2017
>>>     Revision Number: 2315
>>>
>>>     command line - Z:\siscog\sc-cbc\coin-or-cbc\build\bin\cbc.exe
>>>     infeasible.lp (default strategy 1)
>>>
>>>     Program received signal SIGSEGV, Segmentation fault.
>>>     0x0000000000737f80 in
>>>     ClpNonLinearCost::checkInfeasibilities(double) ()
>>>
>>> An attempt to minimise the lp size made the problem go away, so here 
>>> is the full lp: 
>>> https://drive.google.com/open?id=0B4aZo5-t5mnHQUlFWU1BRHRMRDA (4.5 MB).
>>>
>>> In case it helps, the infeasibility stems from following constraints 
>>> and binary variables:
>>>
>>>     cnd(R51,16): + a(R51,16,D141) + a(R51,17,D1093) + a(R51,19,D513)
>>>     <= 2
>>>     s_a(R51,16): + a(R51,16,D141) = 1
>>>     s_a(R51,17): + a(R51,17,D1093) = 1
>>>     s_a(R51,19): + a(R51,19,D513) = 1
>>>
>>>
>>> Cheers,
>>> Luís
>>>
>>>
>>> _______________________________________________
>>> Cbc mailing list
>>> Cbc at list.coin-or.org
>>> https://urldefense.proofpoint.com/v2/url?u=http-3A__list.coin-2Dor.org_mailman_listinfo_cbc&d=CwICAg&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=js2M0T-3OIMIVDvokcKjokJbk0F8QOCd0mT4FsVFE88&m=iozQ9OKrvb9wPclxXhYrd58gAiZa_F_xpncpIiISDj4&s=vZBatNpuyxBZlcMLfvt1kGxA3pWqA9xzUCfRs64JR6c&e=  
>>
>>
>>
>>
>> _______________________________________________
>> Cbc mailing list
>> Cbc at list.coin-or.org
>> https://urldefense.proofpoint.com/v2/url?u=http-3A__list.coin-2Dor.org_mailman_listinfo_cbc&d=CwICAg&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=pLOfVNEEHf-xhIqn1-uzYcZ6Q7UefG6Bg6rXCKTMiAA&m=tijMMVTOV6quP22Hf_4jMxAjTsI-J-jW7VS-TLad0kI&s=EVxvUBuZRiQQVQRUmeEBSbDtyF-tf_IGuBQEv3brA-I&e=  
>
> -- 
> ==================================================
> Haroldo Gambini Santos
> D.Sc, Computer Science
> Universidade Federal de Ouro Preto
> http://www.decom.ufop.br/haroldo/


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20170110/3eb87c2e/attachment.html>