[Symphony] Symphony crashes in the preprocessor

[Menal Guzelsoy]

Bill,

We already have related fixes that are not in the repository yet. For
now, you can just comment out that part of the code. If you like, I
can also send you a quick patch.

Thanks,

Menal.

Menal Guzelsoy
1610 4620455



On Mon, Jun 21, 2010 at 6:18 PM, Bill Basden <wdblists at gmail.com> wrote:
> Hi Everyone,
>
> I've been getting some crashes due to segfaults in Symphony that are
> related to the pre-processor.
>
> The routine prep_basic (Master/master_prep_base.c) will call
> prep_improve_variable again if any rows have been deleted in case the
> column sizes have changed.  The strange thing is that in this call the
> row index (row_ind) is always passed as -1.
>
>   /* if we have deleted some rows, the column sizes might have changed,
>      so try to improve those variables one more time */
>   if (stats->rows_deleted > 0){
>      for (col_ind = 0; col_ind < n; col_ind++){
>         if (cols[col_ind].var_type != 'F' && cols[col_ind].col_size == 0){
>            termcode = prep_improve_variable(P, col_ind, -1, 0,
>                                             dive_level, TRUE, FALSE,
>                                             FALSE,
>                                             0.0,0.0, MAT_COL_ORDERED);
>            if (PREP_QUIT(termcode)){
>               return termcode;
>            }
>         }
>      }
>   }
>
> It looks to me that the prep_improve_variable function will always use
> that -1 as an index into mip->sense, mip->rhs or rows and  crash.  Am
> I missing something?
>
>
> Thanks,
> Bill
>
>
> _______________________________________________
> Symphony mailing list
> Symphony at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/symphony
>