Compiler error with GCC 15 in Fedora Rawhide

[musicinmybrain]

The system compiler was just updated to a GCC 15 pre-release in Fedora Rawhide, the development version of the distribution, and I see a regression in the form of a compiler error.

I already looked at #319 and applied 4a1e139 as a patch.

Now I see:

In file included from ../../linbox/matrix/sparse-matrix.h:87:
../../linbox/matrix/sparsematrix/sparse-tpl-matrix-omp.h: In member function 'void LinBox::SparseMatrix<Field_, LinBox::SparseMatrixFormat::TPL_omp>::rebind<Tp1_>::operator()(other&, const LinBox::SparseMatrix<Field_, LinBox::SparseMatrixFormat::TPL_omp>::Self_t&)':
../../linbox/matrix/sparsematrix/sparse-tpl-matrix-omp.h:321:43: error: 'const LinBox::SparseMatrix<Field_, LinBox::SparseMatrixFormat::TPL_omp>::Self_t' has no member named 'data'; did you mean 'data_'? [-Wtemplate-body]
  321 |                         Ap.data_.resize(A.data.size());  
      |                                           ^~~~
      |                                           data_
../../linbox/matrix/sparsematrix/sparse-tpl-matrix-omp.h:323:39: error: 'const LinBox::SparseMatrix<Field_, LinBox::SparseMatrixFormat::TPL_omp>::Self_t' has no member named 'data'; did you mean 'data_'? [-Wtemplate-body]
  323 |                              v_p != A.data.end(); ++ v_p, ++ vp_p)
      |                                       ^~~~
      |                                       data_

Looking at

linbox/linbox/matrix/sparsematrix/sparse-tpl-matrix-omp.h

Lines 321 to 323 in 4a1e139

	Ap.data_.resize(A.data.size());
	for (v_p = A.data_.begin(), vp_p = Ap.data_.begin();
	v_p != A.data.end(); ++ v_p, ++ vp_p)

and noting that there is already a mixture of .data and .data_, I tried

diff --git a/linbox/matrix/sparsematrix/sparse-tpl-matrix-omp.h b/linbox/matrix/sparsematrix/sparse-tpl-matrix-omp.h
index feca4cf35..c97b1c8c1 100644
--- a/linbox/matrix/sparsematrix/sparse-tpl-matrix-omp.h
+++ b/linbox/matrix/sparsematrix/sparse-tpl-matrix-omp.h
@@ -318,9 +318,9 @@ class SparseMatrix<Field_, SparseMatrixFormat::TPL_omp> : public BlackboxInterfa
                        typedef typename selfvec::const_iterator selfiter;
                        otheriter vp_p; selfiter v_p;
 
-                       Ap.data_.resize(A.data.size());
+                       Ap.data_.resize(A.data_.size());
                        for (v_p = A.data_.begin(), vp_p = Ap.data_.begin();
-                            v_p != A.data.end(); ++ v_p, ++ vp_p)
+                            v_p != A.data_.end(); ++ v_p, ++ vp_p)
                                hom.image (vp_p->elt, v_p->elt);
                }
        };

and this seemed to work, so I’m going to offer that as a PR.