Anyone fixed the bug I reported 'segfault in leastsq'? I see a lot of progress was made to track it down - but I don't think I know enough about the code to try to fix it.