wrong retention time alignement computation in the 2.2.0 branch

a severe problem was introduced in e7c05996 version 2.2.0-7.

The retention time target that is computed for each peptide from observed MS2 retention times was not corrected using alignment informations. This is a cause of wrong retention time references and a source of error in match between run. This problem has been partly compensated by the post matching mecanism that can retreive good peaks if they were not found, but this certainly introduce a lot of noise since 21 06 2016.

This problem should have been fixed in 42f3bae8.

After a severe refactoring, that simplifies the code and make it more efficient, this bug was found and fixed.

We have to check it.