From: | Shachar Shemesh <psql(at)shemesh(dot)biz> |
---|---|
To: | William ZHANG <uniware_at_zedware_dot_org(at)antispam(dot)com> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: OLE DB driver |
Date: | 2004-01-11 20:21:00 |
Message-ID: | 4001B02C.9080008@shemesh.biz |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
William ZHANG wrote:
>"Shachar Shemesh" <psql(at)shemesh(dot)biz> wrote:3FFF1607(dot)9090806(at)shemesh(dot)biz(dot)(dot)(dot)
>
>
>>Andreas Pflug wrote:
>>
>>
>>
>>>I wonder if this could be implemented as a wrapper around libpq. This
>>>way, the OLEDB driver would benefit from a proven piece of software.
>>>
>>>Regards,
>>>Andreas
>>>
>>>
>>That's what I'm doing. I'm not sure why other drivers didn't do that as
>>well (maybe there was no native Windows port of libpq at the time?).
>>
>>
>
>AFAIK, rely on libpq is a fast way but not always convenient.
>
>
Seems like more convinent than hacking the client/server protocol
myself. I may come up to a wall later on, and then I may reconsider.
>OLEDB has many features borrowed from ODBC, and it is a
>substitute for ODBC. Microsoft is now making ODBC obsolete and
>advocating OLEDB instead. You'd better read pgodbc's source
>code, too.
>
>
>
I did some of that, and did some ODBC hacking too (very little). The
fact that ODBC uses it's own protocol implementation, I think, is a
mistake. There are places where some intervention is going to be
unavoidable, I'm afraid, but that has nothing to do with what protocol
to use for communicating with the back end. If Postgres doesn't support
writeable cursors, then it doesn't support writeable cursors. One has to
emulate them. I will defenitely use some borrowing from pgsqlODBC (the
main reason my OLE DB is LGPL was so I can do that).
--
Shachar Shemesh
Lingnu Open Systems Consulting
http://www.lingnu.com/
From | Date | Subject | |
---|---|---|---|
Next Message | Bruce Momjian | 2004-01-12 00:25:50 | My travels |
Previous Message | Tom Lane | 2004-01-11 19:32:07 | Re: libpq thread safety |