From: Olivier M. <om...@ti...> - 2003-10-27 17:31:52
|
Dear, On Tue, 28 Oct 2003 02:43:21 +1100, Helen Borrie wrote : > As it stands, the OPO parameter hack is only the best *stop-gap* > solution. It *is* a hack. IBO can't tell whether it is connecting to an > OPO server or not, until *after* it has connected. Currently, it has to > know that, because the IBO parser can *only* handle the bad ordering for > now. For the long haul, there has to be a better solution that is handled > on a connection basis.. May I then suggest we make the server-side setting visible to applications which care about it through, for instance, a call to isc_database_info with some new token like : 'isc_info_fixed_param_ordering' ? IBO could connect, request that info (it probably request already some other informations right after connection already) and adjust its behaviour based on it. A FB >= 1.5 would support that info item and would return a quad with value 0 meaning old-buggy behaviour and 1 meaning new fixed behaviour. When requesting this info item from a non FB server, IBO would get an error returned, meaning old-parameter ordering. Oh by the way, what about recent InterBase versions ? Do they still implement the bug or did they corrected it ? If corrected what did they do to support older and newer applications ? There might be ideas to get from there too. And that info would be of interest for IBO I suppose. -- Best Regards, Olivier Mascia |