Closed mlubin closed 9 years ago
The duals of NonPos rows should be <= 0, so we need to add this post-processing since we transform NonPos to NonNeg for ECOS.
Coverage increased (+0.12%) when pulling dd36e4421b6432936096453d7213231e7a70c257 on duals into ef1b05a134aaa9b3a484e3db8aba5d8590062bd2 on master.
Test?
Working on a QP -> conic wrapper, and I've got tests passing for LPs in JuMP with ECOS now.
The duals of NonPos rows should be <= 0, so we need to add this post-processing since we transform NonPos to NonNeg for ECOS.