On 7 February 2013 11:46, Michael Hanselmann <[email protected]> wrote: > 2013/2/7 Bernardo Dal Seno <[email protected]>: >> On 7 February 2013 09:01, Iustin Pop <[email protected]> wrote: >>> I'm not sure I like the deepcopy there… since we don't care about the >>> actual contents, only if it's different, what about serialising it back >>> as Michael said? Then you don't need to keep an intermediate >>> representation in memory, just serialise it back and compare with >>> raw_data, which is still available. >> >> The problem with serialization is that there is no guarantee the >> dictionaries are serialized in a stable way. So I didn't choose that >> way. > > You don't need to serialize to a string, but only to a dict > (….ToDict()). That way you can easily compare them.
If you remove deepcopy and leave only ToDict(), unit tests would fail (that was my first idea). Bernardo
