Carl Hetherington wrote:

Hi,

Another error that valgrind has spotted is the unitialised indices 4 and 5 in the NR::Matrix tp at the top of matrix_is_isometry() (nr-arena-shape.cpp:362).

Should I fix this here, or would it be better to make the NR::Matrix default constructor zero-init its members?

I zeroed the members (actually, it doesn't matter unless 4 or 5 is a special value, which they shouldn't be :)

njh, the zeroator.