Can you please open a ticket about this. We'll need to think about this and we do not want to forget.
On Tuesday, 11 June 2013 16:47:08 UTC-5, Niphlod wrote: > > @derek : request.now is just a field that represents the row that needs to > go to the db. is then either inserted or updated accordingly to the > record_id found (more precisely, it tries to update and then if no updates > are detected it inserts it). at the beginning of the try_store_whatever > there's the self._unchanged() check, that is the one "producing" the issue. > > @massimo and devs: apparently this is not solvable empirically basing our > unchanged check on the hash of the pickled string (even if deepcopying > seems to return consistent behaviour). Given that the pickle docs doesn't > assure that consistent strings are generated, I think the only way is to > create a slightly different implementation of Storage for session (or just > change the Storage implementation) to keep a record of the Storage actually > being changed. If I'm not mistaken, it's a dict, so every call to > __setitem__ > __delitem__ > pop (if key exists) > setdefault (if key does not exists) > update > clear ? > to the underlying dict needs to flip a "something" (boolean) that gets > checked in the _unchanged() function. > > On Tuesday, June 11, 2013 11:14:06 PM UTC+2, Niphlod wrote: >> >> apparently "the issue" (i.e. different strings can be generated for the >> same object) is documented at >> http://docs.python.org/2/library/pickle.html#id10 >>> >>> >>>>> -- --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
