Subject | Re: XSQLDA/XSQLVAR issues |
---|---|
Author | Aleksey Karyakin |
Post date | 2005-02-21T17:28:54Z |
--- In Firebird-Architect@yahoogroups.com, "Aleksey Karyakin"
<karyakin@c...> wrote:
is local-oriented. Literally mapping it to remote objects and calls
would result in keeping the diagnosic object on the server. However,
it's inacceptable for client-server performance, where diagnostic
object should be always local, implemented by the client proxy, and
received by value from server along with response packet.
This actually means API/remote protocol split, which is fine for me
but anyone who believe in location transparency and uniform layering
would disagree :)
<karyakin@c...> wrote:
>exception
> The diagnostic object is implemented by the underlying API level.
> This is a pure C API, completely portable, doesn't rely on
> handling or thread-local data and is object-oriented.In reality, there's a conceptual problem. The proposed diagnostic API
is local-oriented. Literally mapping it to remote objects and calls
would result in keeping the diagnosic object on the server. However,
it's inacceptable for client-server performance, where diagnostic
object should be always local, implemented by the client proxy, and
received by value from server along with response packet.
This actually means API/remote protocol split, which is fine for me
but anyone who believe in location transparency and uniform layering
would disagree :)
> Regards,
> Aleksey Karyakin