commit | 37516c968c1bc6b6910316f21dfd9d30b2eb66fd | [log] [tgz] |
---|---|---|
author | Sameer Agarwal <sameeragarwal@google.com> | Sun Aug 07 12:54:15 2022 -0700 |
committer | Sameer Agarwal <sameeragarwal@google.com> | Sun Aug 07 12:56:42 2022 -0700 |
tree | 352c288b0a8a9d64efc2e4d3f27416d8b6496e75 | |
parent | d8dad14eed1cd162526277b08012a2f82024098b [diff] |
Fix a bug in InnerProductComputer. InnerProductComputer was assuming that the input matrix had at least one structurally non-zero block. As a result sometimes when InnerProductComputer.SubMatrix test generated matrices where the submatrix was structurally zero it would cause a segmentation fault. This changes fixes this issue and reduces the threshold for the minimum block density so that there is a much higher chance of such matrices occuring as part of testing. Fixes https://github.com/ceres-solver/ceres-solver/issues/820 Change-Id: Iec3a2431e646f0c2aac9e0b994531fa72323f329
Ceres Solver is an open source C++ library for modeling and solving large, complicated optimization problems. It is a feature rich, mature and performant library which has been used in production at Google since 2010. Ceres Solver can solve two kinds of problems.
Please see ceres-solver.org for more information.