[Coin-discuss] seg fault using Cbc/Clp with OsiCbcSolverInterface

[Kish Shen]

Hi,

I am running some tests using the OsiCbcSolverInterface, and I got a seg
fault trying to solve this problem (fiber):

Coin0510I Presolve is modifying 2 integer bounds and re-presolving
Coin0506I Presolve 20 (-398) rows, 20 (-1278) columns and 70 (-4201) elements
Clp0006I 0  Obj 407431 Primal inf 5.97197 (6) Dual inf 80791.2 (1)
Clp0006I 5  Obj 409451
Clp0001I Primal infeasible - objective value 409451
Coin0505I Presolved problem not optimal, resolve after postsolve
Segmentation fault

In this particular case, I am solving the MIP problem using Osi, rather
than directly using Cbc -- i.e. I just call Osi's branchAndBound(), with 
presolve turned on for initial solve and resolve.

I have generated the C++ code that will reproduce this problem. If anyone
(I guess probably John Forrest) is interested in looking at this, please
let me know and I can send it.

Thanks and cheers,

Kish Shen
IC-Parc
Imperial College London
U.K.