NJOY2016.56

This version mainly addressed issue #130.

This relates to changes made by the consistency checks when a checker acer run is requested. NJOY only rarely modifies data. It does so for secondary particle distributions that use LAW=4 (isotropic angular distribution and continuous tabulated energy distributions) or LAW=44 (Kalbach-Mann).

When the consis routine determines the maximum outgoing energy is too high compared to epmax, it sets the last value of the outgoing energy array to this value, and all values before it to monotonically increasing values that are lower than epmax. It then adjusts the pdf values to maintain the associated cdf. This often leads to the introduction of artificial spikes in the plots produced by acer.

This behaviour has now been modified. Whenever this problem is detected, the consis routine will now reset the first energy larger to epmax to a value slightly below it and adjust the pdf value so that the cdf value for this energy point is 1.0. All other energy points after this energy point along with their associated data is set to zero. The locators in the xss array are left unchanged.

Due to this new behaviour, and because the routine that writes the xss array does not check locators, the change routine was modified as well to verify consistency of the locators with the actual position in the xss array. Whenever the locator is inconsistent, it will either just advance to the locator (thus taking into account the gaps created by the consis routine while writing a message to the user) or issue an error when the locator points to a position lower than the current one.

Test case 55 was added to properly test this for a LAW=44. Up to now, no case for LAW=4 has been found.

The change routine was also simplified to improve its readability by adding subroutines to write a list of integers, a list of real numbers, a single integer and a single real number.

A large number of compiler warnings were also fixed in this pull request (mostly unused or uninitialised variables).