Re: yet another Java oddity

[Mike Miller]

On Fri, Feb 8, 2013 at 5:37 PM, Michael D. Godfrey wrote:
> It may make sense to put code in Octave initialization which checks the FPU
> state,
> and modifies it if necessary.  This will prevent this problem from coming
> back again.

I'm not sure I follow you, can you explain?

The FPU state is reset for every new process, so there's no need to
check it or set it when Octave starts up, if that's what you mean. In
theory any library we link into Octave could set it to whatever value
it wants at any time. So far both Java and ATLAS do this and do not
reset it when they return control to Octave.

-- 
mike