Raise Gurobi's own error message when it fails before finding a solution - #3574

Description

When gurobipy raises a GurobiError inside optimize(), both GUROBI interfaces (gurobi_qpif.py and gurobi_conif.py) catch it together with every other exception and return SOLVER_ERROR. The user then only sees the generic message:

cvxpy.error.SolverError: Solver 'GUROBI' failed. Try another solver, or solve with verbose=True for more information.

The most common case is a size-limited license (the pip gurobipy one). Gurobi's own message says exactly what is wrong, but it gets lost:

import cvxpy as cp
x = cp.Variable(2001)
cp.Problem(cp.Minimize(cp.sum(x)), [x >= 0]).solve(solver=cp.GUROBI)

With this PR, both interfaces re-raise the GurobiError as a SolverError with Gurobi's message, chained with from e, when no solution was found:

cvxpy.error.SolverError: Solver 'GUROBI' failed: Model too large for size-limited license; visit https://gurobi.com/unrestricted for more information

If Gurobi found a solution before the error (SolCount > 0), the current behaviour is kept. The HiGHS and OSQP interfaces already follow the same raise SolverError(e) pattern.

Tests: test_gurobi_error_message in test_qp_solvers.py (LP path) and in TestGUROBI in test_conic_solvers.py (SOCP path). Both patch gurobipy.Model.optimize to raise the license error, so they don't depend on the installed license. Both fail on master and pass with this change. I also checked the snippet above against a real restricted license, through both the QP and the conic path.

Issue link (if applicable): addresses part of #1548, the Gurobi case reported in this comment. The issue is broader than that: unpacking solver metadata before raising. CPLEX and COPT swallow errors around the solve call in the same way, and SCIP only logs a warning. These could be follow-ups.

Type of change

• New feature (backwards compatible)
• New feature (breaking API changes)
• Bug fix
• Other (Documentation, CI, ...)

Contribution checklist

• Add our license to new files.
• Check that your code adheres to our coding style.
• Write unittests.
• Run the unittests and check that they’re passing.
• Run the benchmarks to make sure your change doesn’t introduce a regression.