From: Alex P. <pes...@ya...> - 2008-02-28 09:19:41
|
On Wednesday 27 February 2008 15:29, Adriano dos Santos Fernandes wrote: > Adriano dos Santos Fernandes escreveu: > > * Not the dumb concept of Vulcan provider requiring another C interface, > > but the official API going direct to the engine. I'm not sure if you're > > understand me, but for me is clear that the same API available to > > clients should be available to external modules (engines) and there is > > no reason for engine providers entry-points be different than y-valve > > entry-points. > > To make things more clear: having a unified API (now ISC/GDS) we can > encapsulate this API in C++ classes and use it in our utilities and in > internal dynamic queries. Once we establish these classes as a public > API it will be direct usable for users in external engines and the client. With new public API we may define that handles are pointers, and forget this part of problem. Current converter from 32 bit handles will remain to support legacy ISC API, and that legacy ISC API suuport library can be designed to be able to talk to any library (yValve, engine, remote). |