Re: broken coordinate_matrix::sort with gcc 4.7 (was: patches for #7297, #7296, #6514, #6511 on trunk - please verify)
Gunter Winkler <[email protected]>
| Newsgroups | gmane.comp.lib.boost.ublas |
|---|---|
| Message-ID | <[email protected]> |
Hello, Am Tuesday 28 August 2012 schrieb sguazt: > > However, the compilation of test_inplace_solve (together with > test_coordinate_matrix_sort and test_assignment) still fails. > Here is the output of compiler that I get when I try to compile > test_inplace_solve (sorry it's very long): > It looks like the new gcc 4.7 breaks the index_triple_array hack (see storage.hpp). I introduced this class to sort the three parallel arrays (index1, index2, value) from coordinate matrix without the need to create a temporary permutation vector. Unfortunately the index triple (which essentially holds the index and references of the three arrays) must behave like a reference but is not allowed to be referenced. see storage.hpp:1844 Looks like this hack does not work with gcc 4.7 any more. Maybe we can improve coordinate_matrix::sort() somehow to use the zip_iterator. (see storage_sparse.hpp:4396) What happens if you remove USE_COORDINATE_MATRIX from the sparse test set in libs/numeric/ublas/test/Jamfile.v2 ? Then all tests should succeed. mfg Gunter _______________________________________________ ublas mailing list [email protected] http://lists.boost.org/mailman/listinfo.cgi/ublas Sent to: [email protected]
signature.asc
(application/pgp-signature, 198 B)
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) iEYEABECAAYFAlA9GdMACgkQghsXb/J0PcF0xACfV4y8aESExfhjm0mSHLeLaNRU rt0An39kY8uGEidEjgIsZPKtZmYQ9QPu =TRFf -----END PGP SIGNATURE-----