MOOG remembers bad times #35
Description
If the compiled version of MOOG gets into some really bad territory and calls stop, Python does recover from this OK (see #23). However, some of these instances are when MOOG is trying to work out which step size to start with (e.g., step sizes in synthesis points or along the curve-of-growth). When this happens (and MOOG dies), the variables within MOOG remain.
That means in the next star you analyse, MOOG will still complain that it cannot decide on a step size and fall over for all stars subsequently analysed in that Python session. Obviously, that's a problem.
The workaround at the moment is simply to analyse one star per Python session. Debugging this will mean ensuring the appropriate variables are reinitialised in the MyAbfind.f file.