I just wondered: Why is it sensible to have the dispatch switches in each inset?
As far as I can tell, the same functionality can achieved by virtual functions for individual LFUNs and a single big switch in the inset base class. The only disadvantage I see is that we'll have 100+ additional member functions in the base class (which is certainly not Nice), but the total effort should be about the same. I am asking not because I want to change this but because I just met a nuisance of the current scheme: If D derives from B and B implements lfuns l1,l2,...l10 and D wants to override just l1, it has to explicitly forward calls to l2...l10 to B. Which is Not Nice, and moreover error-prone when lfun "handlers" get added to B. Our current scheme is conceptually not very different from extending the insets' vtable "manually" by the jump table in the dispatch switch - with the need to do some things manually the compiler would take care of if left to it... Do I miss something crucial? Andre' -- Those who desire to give up Freedom in order to gain Security, will not have, nor do they deserve, either one. (T. Jefferson)
