You essentially have the right idea.
If you have, for instance (R) and (S) mandelic acid, you can react the mixture with, for instance, enantiomerically pure (R)-alpha-methylbenzylamine to form diastereomeric salts (or further, amides) that you may be able to recrystallise or elute separately. If you had a mixture of (R) and (S)alpha-methylbenzylamines you would form the enantiomeric pair of diastereomers (R),(R) and (S),(S), which are identical in a achiral environment, and the (R),(S) and (S),(R) forms which too are identical, but diastereomeric to the first pair.
Remember, enantiomers react identically, unless in an chiral enviroment.