logilab-astng #2451 find a way to help inferences in some cases [open]

for instance we could take advantage for UserList and UserDict to know more things about list and dict. Would also need to give it some hints about particuliar classes / modules which are too dynamic to be properly understood (twisted...).

prioritynormal
typeenhancement
done in<not specified>
load5.000
load left5.000
closed by<not specified>