GRBgeterrormsg() vs GRBgetmerrormsg()

進行中

Mike Heinen

• 2025年01月03日 20:59
• 編集済み

We have a fair number of places in code that look like the following 

int ret = GRBaddvar(model, numnz, vind, vval, obj, lb, ub, vtype, name);
if (ret)
{
  printf("GRBaddvar() failed. Failed to add Gurobi variable. Error code %d: %s", ret, GRBgeterrormsg(GRBgetenv(model)));
}

I recently noticed in the gurobi_c.h header file that there is a GRBgetmerrormsg() function defined which takes a GRBmodel instead of a GRBenv struct. Does this mean we could change all these to just do:

int ret = GRBaddvar(model, numnz, vind, vval, obj, lb, ub, vtype, name);
if (ret)
{
printf("GRBaddvar() failed. Failed to add Gurobi variable. Error code %d: %s", ret, GRBgetmerrormsg(model));
}

関連するエンティティ

• C

0

• 

  Maliheh Aramon

  Gurobi Staff

  • 2025年01月03日 21:33

  Hi Mike, 

  As you know, only GRBgeterrormsg() function is included in our official documentation. Therefore, we recommend using this officially documented function.

  You are correct that the Gurobi C API includes a function named GRBgetmerrormsg(GRBmodel *model). However, this function internally calls GRBgeterrormsg() after verifying the validity of the model pointer. Is there a specific reason you prefer to use GRBgetmerrormsg instead of GRBgeterrormsg?

  Best regards,

  Maliheh

  0
• 

  Mike Heinen

  • 2025年01月06日 14:11

  

  

  

  No reason other than making the code a little cleaner on our end.

  0

サインインしてコメントを残してください。