Hello,

I am calling mrqmin 2 or more times within the same program to compare different fits. In some cases, the ma value is changed and my program seg fault. As I understant it, oneda and some others vectors are initialized on the first call to mrqmin (say with ma=3) because of static modifier. If I try a second fit with say ma=4, then oneda and others won't be resize and seg fault occurs for oneda[3][0]? Am I missing something? Is the static modifier necessary?

Thank you,

Philippe

If you just want to compare different fits, there are no nee to call mrqmin at all. Instead you can call mrqcof directly to calculate chi-square (which I assume you use to compare fits) and hence you does not have to bother about the size of ma.

See also
http://www.nr.com/forum/showthread.php?s=&threadid=134