If the fortran fails, we are raising an exception, and then the temporary files will be deleted, so the only way to see what happened would be to try to recreate the run somehow (although in principle we know the inputs). Can we keep the contents of the temp directory for inspection (which will contain amongst other things the input and output files used by the model)?
One possible option might be put top-level code in the handler that catches the exception, copies the files somewhere (maybe retaining only the last few such runs) and then re-raises it. But is there some tidier functionality built into pywps?
If the fortran fails, we are raising an exception, and then the temporary files will be deleted, so the only way to see what happened would be to try to recreate the run somehow (although in principle we know the inputs). Can we keep the contents of the temp directory for inspection (which will contain amongst other things the input and output files used by the model)?
One possible option might be put top-level code in the handler that catches the exception, copies the files somewhere (maybe retaining only the last few such runs) and then re-raises it. But is there some tidier functionality built into pywps?